Reduce number of iteration in test_physical_replication to reduce test time

.config/hakari.toml

@@ -23,30 +23,10 @@ platforms = [
 ]
 
 [final-excludes]
-workspace-members = [
-    # vm_monitor benefits from the same Cargo.lock as the rest of our artifacts, but
-    # it is built primarly in separate repo neondatabase/autoscaling and thus is excluded
-    # from depending on workspace-hack because most of the dependencies are not used.
-    "vm_monitor",
-    # All of these exist in libs and are not usually built independently.
-    # Putting workspace hack there adds a bottleneck for cargo builds.
-    "compute_api",
-    "consumption_metrics",
-    "desim",
-    "metrics",
-    "pageserver_api",
-    "postgres_backend",
-    "postgres_connection",
-    "postgres_ffi",
-    "pq_proto",
-    "remote_storage",
-    "safekeeper_api",
-    "tenant_size_model",
-    "tracing-utils",
-    "utils",
-    "wal_craft",
-    "walproposer",
-]
+# vm_monitor benefits from the same Cargo.lock as the rest of our artifacts, but
+# it is built primarly in separate repo neondatabase/autoscaling and thus is excluded
+# from depending on workspace-hack because most of the dependencies are not used.
+workspace-members = ["vm_monitor"]
 
 # Write out exact versions rather than a semver range. (Defaults to false.)
 # exact-versions = true

.config/nextest.toml

@@ -1,2 +1,2 @@
 [profile.default]
-slow-timeout = { period = "60s", terminate-after = 3 }
+slow-timeout = { period = "20s", terminate-after = 3 }

.dockerignore

@@ -8,7 +8,6 @@
 !scripts/combine_control_files.py
 !scripts/ninstall.sh
 !vm-cgconfig.conf
-!docker-compose/run-tests.sh
 
 # Directories
 !.cargo/
@@ -18,10 +17,9 @@
 !libs/
 !neon_local/
 !pageserver/
-!patches/
 !pgxn/
 !proxy/
-!storage_scrubber/
+!s3_scrubber/
 !safekeeper/
 !storage_broker/
 !storage_controller/

.gitattributes

@@ -1,2 +0,0 @@
-# allows for nicer hunk headers with git show
-*.rs diff=rust

.github/ISSUE_TEMPLATE/config.yml

@@ -1,6 +0,0 @@
-
-blank_issues_enabled: true
-contact_links:
-  - name: Feature request
-    url: https://console.neon.tech/app/projects?modal=feedback
-    about: For feature requests in the Neon product, please submit via the feedback form on `https://console.neon.tech`

.github/actionlint.yml

@@ -1,15 +1,14 @@
 self-hosted-runner:
   labels:
     - arm64
+    - dev
+    - gen3
     - large
-    - large-arm64
+    # Remove `macos-14` from the list after https://github.com/rhysd/actionlint/pull/392 is merged.
+    - macos-14
     - small
-    - small-arm64
     - us-east-2
 config-variables:
-  - BENCHMARK_PROJECT_ID_PUB
-  - BENCHMARK_PROJECT_ID_SUB
   - REMOTE_STORAGE_AZURE_CONTAINER
   - REMOTE_STORAGE_AZURE_REGION
   - SLACK_UPCOMING_RELEASE_CHANNEL_ID
-  - DEV_AWS_OIDC_ROLE_ARN

.github/actions/allure-report-generate/action.yml

@@ -150,7 +150,7 @@ runs:
 
         # Use aws s3 cp (instead of aws s3 sync) to keep files from previous runs to make old URLs work,
         # and to keep files on the host to upload them to the database
-        time s5cmd --log error cp "${WORKDIR}/report/*" "s3://${BUCKET}/${REPORT_PREFIX}/${GITHUB_RUN_ID}/"
+        time aws s3 cp --recursive --only-show-errors "${WORKDIR}/report" "s3://${BUCKET}/${REPORT_PREFIX}/${GITHUB_RUN_ID}"
 
         # Generate redirect
         cat <<EOF > ${WORKDIR}/index.html
@@ -183,7 +183,7 @@ runs:
       uses: actions/cache@v4
       with:
         path: ~/.cache/pypoetry/virtualenvs
-        key: v2-${{ runner.os }}-${{ runner.arch }}-python-deps-${{ hashFiles('poetry.lock') }}
+        key: v2-${{ runner.os }}-python-deps-${{ hashFiles('poetry.lock') }}
 
     - name: Store Allure test stat in the DB (new)
       if: ${{ !cancelled() && inputs.store-test-results-into-db == 'true' }}

.github/actions/download/action.yml

@@ -26,7 +26,7 @@ runs:
         TARGET: ${{ inputs.path }}
         ARCHIVE: /tmp/downloads/${{ inputs.name }}.tar.zst
         SKIP_IF_DOES_NOT_EXIST: ${{ inputs.skip-if-does-not-exist }}
-        PREFIX: artifacts/${{ inputs.prefix || format('{0}/{1}/{2}', github.event.pull_request.head.sha || github.sha, github.run_id, github.run_attempt) }}
+        PREFIX: artifacts/${{ inputs.prefix || format('{0}/{1}', github.run_id, github.run_attempt) }}
       run: |
         BUCKET=neon-github-public-dev
         FILENAME=$(basename $ARCHIVE)

.github/actions/neon-branch-create/action.yml

@@ -3,14 +3,14 @@ description: 'Create Branch using API'
 
 inputs:
   api_key:
-    description: 'Neon API key'
+    desctiption: 'Neon API key'
     required: true
   project_id:
-    description: 'ID of the Project to create Branch in'
+    desctiption: 'ID of the Project to create Branch in'
     required: true
   api_host:
-    description: 'Neon API host'
-    default: console-stage.neon.build
+    desctiption: 'Neon API host'
+    default: console.stage.neon.tech
 outputs:
   dsn:
     description: 'Created Branch DSN (for main database)'

.github/actions/neon-branch-delete/action.yml

@@ -3,17 +3,17 @@ description: 'Delete Branch using API'
 
 inputs:
   api_key:
-    description: 'Neon API key'
+    desctiption: 'Neon API key'
     required: true
   project_id:
-    description: 'ID of the Project which should be deleted'
+    desctiption: 'ID of the Project which should be deleted'
     required: true
   branch_id:
-    description: 'ID of the branch to delete'
+    desctiption: 'ID of the branch to delete'
     required: true
   api_host:
-    description: 'Neon API host'
-    default: console-stage.neon.build
+    desctiption: 'Neon API host'
+    default: console.stage.neon.tech
 
 runs:
   using: "composite"

.github/actions/neon-project-create/action.yml

@@ -3,19 +3,22 @@ description: 'Create Neon Project using API'
 
 inputs:
   api_key:
-    description: 'Neon API key'
+    desctiption: 'Neon API key'
     required: true
   region_id:
-    description: 'Region ID, if not set the project will be created in the default region'
+    desctiption: 'Region ID, if not set the project will be created in the default region'
     default: aws-us-east-2
   postgres_version:
-    description: 'Postgres version; default is 16'
-    default: '16'
+    desctiption: 'Postgres version; default is 15'
+    default: 15
   api_host:
-    description: 'Neon API host'
-    default: console-stage.neon.build
+    desctiption: 'Neon API host'
+    default: console.stage.neon.tech
+  provisioner:
+    desctiption: 'k8s-pod or k8s-neonvm'
+    default: 'k8s-pod'
   compute_units:
-    description: '[Min, Max] compute units'
+    desctiption: '[Min, Max] compute units; Min and Max are used for k8s-neonvm with autoscaling, for k8s-pod values Min and Max should be equal'
     default: '[1, 1]'
 
 outputs:
@@ -34,6 +37,10 @@ runs:
       # A shell without `set -x` to not to expose password/dsn in logs
       shell: bash -euo pipefail {0}
       run: |
+        if [ "${PROVISIONER}" == "k8s-pod" ] && [ "${MIN_CU}" != "${MAX_CU}" ]; then
+          echo >&2 "For k8s-pod provisioner MIN_CU should be equal to MAX_CU"
+        fi
+
         project=$(curl \
           "https://${API_HOST}/api/v2/projects" \
           --fail \
@@ -45,7 +52,7 @@ runs:
               \"name\": \"Created by actions/neon-project-create; GITHUB_RUN_ID=${GITHUB_RUN_ID}\",
               \"pg_version\": ${POSTGRES_VERSION},
               \"region_id\": \"${REGION_ID}\",
-              \"provisioner\": \"k8s-neonvm\",
+              \"provisioner\": \"${PROVISIONER}\",
               \"autoscaling_limit_min_cu\": ${MIN_CU},
               \"autoscaling_limit_max_cu\": ${MAX_CU},
               \"settings\": { }
@@ -68,5 +75,6 @@ runs:
         API_KEY: ${{ inputs.api_key }}
         REGION_ID: ${{ inputs.region_id }}
         POSTGRES_VERSION: ${{ inputs.postgres_version }}
+        PROVISIONER: ${{ inputs.provisioner }}
         MIN_CU: ${{ fromJSON(inputs.compute_units)[0] }}
         MAX_CU: ${{ fromJSON(inputs.compute_units)[1] }}

.github/actions/neon-project-delete/action.yml

@@ -3,14 +3,14 @@ description: 'Delete Neon Project using API'
 
 inputs:
   api_key:
-    description: 'Neon API key'
+    desctiption: 'Neon API key'
     required: true
   project_id:
-    description: 'ID of the Project to delete'
+    desctiption: 'ID of the Project to delete'
     required: true
   api_host:
-    description: 'Neon API host'
-    default: console-stage.neon.build
+    desctiption: 'Neon API host'
+    default: console.stage.neon.tech
 
 runs:
   using: "composite"

.github/actions/run-python-test-set/action.yml

@@ -43,7 +43,7 @@ inputs:
   pg_version:
     description: 'Postgres version to use for tests'
     required: false
-    default: 'v16'
+    default: 'v14'
   benchmark_durations:
     description: 'benchmark durations JSON'
     required: false
@@ -56,14 +56,14 @@ runs:
       if: inputs.build_type != 'remote'
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build_type }}-artifact
+        name: neon-${{ runner.os }}-${{ inputs.build_type }}-artifact
         path: /tmp/neon
 
     - name: Download Neon binaries for the previous release
       if: inputs.build_type != 'remote'
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build_type }}-artifact
+        name: neon-${{ runner.os }}-${{ inputs.build_type }}-artifact
         path: /tmp/neon-previous
         prefix: latest
 
@@ -71,7 +71,7 @@ runs:
       if: inputs.build_type != 'remote'
       uses: ./.github/actions/download
       with:
-        name: compatibility-snapshot-${{ runner.arch }}-${{ inputs.build_type }}-pg${{ inputs.pg_version }}
+        name: compatibility-snapshot-${{ inputs.build_type }}-pg${{ inputs.pg_version }}
         path: /tmp/compatibility_snapshot_pg${{ inputs.pg_version }}
         prefix: latest
         # The lack of compatibility snapshot (for example, for the new Postgres version)
@@ -83,12 +83,13 @@ runs:
       uses: actions/checkout@v4
       with:
         submodules: true
+        fetch-depth: 1
 
     - name: Cache poetry deps
       uses: actions/cache@v4
       with:
         path: ~/.cache/pypoetry/virtualenvs
-        key: v2-${{ runner.os }}-${{ runner.arch }}-python-deps-${{ hashFiles('poetry.lock') }}
+        key: v2-${{ runner.os }}-python-deps-${{ hashFiles('poetry.lock') }}
 
     - name: Install Python deps
       shell: bash -euxo pipefail {0}
@@ -113,8 +114,6 @@ runs:
         export PLATFORM=${PLATFORM:-github-actions-selfhosted}
         export POSTGRES_DISTRIB_DIR=${POSTGRES_DISTRIB_DIR:-/tmp/neon/pg_install}
         export DEFAULT_PG_VERSION=${PG_VERSION#v}
-        export LD_LIBRARY_PATH=${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/lib
-        export BENCHMARK_CONNSTR=${BENCHMARK_CONNSTR:-}
 
         if [ "${BUILD_TYPE}" = "remote" ]; then
           export REMOTE_ENV=1
@@ -130,8 +129,8 @@ runs:
           exit 1
         fi
         if [[ "${{ inputs.run_in_parallel }}" == "true" ]]; then
-          # -n sets the number of parallel processes that pytest-xdist will run
-          EXTRA_PARAMS="-n12 $EXTRA_PARAMS"
+          # -n16 uses sixteen processes to run tests via pytest-xdist
+          EXTRA_PARAMS="-n16 $EXTRA_PARAMS"
 
           # --dist=loadgroup points tests marked with @pytest.mark.xdist_group
           # to the same worker to make @pytest.mark.order work with xdist
@@ -169,28 +168,23 @@ runs:
           EXTRA_PARAMS="--durations-path $TEST_OUTPUT/benchmark_durations.json $EXTRA_PARAMS"
         fi
 
-        if [[ $BUILD_TYPE == "debug" && $RUNNER_ARCH == 'X64' ]]; then
+        if [[ "${{ inputs.build_type }}" == "debug" ]]; then
           cov_prefix=(scripts/coverage "--profraw-prefix=$GITHUB_JOB" --dir=/tmp/coverage run)
+        elif [[ "${{ inputs.build_type }}" == "release" ]]; then
+          cov_prefix=()
         else
           cov_prefix=()
         fi
 
         # Wake up the cluster if we use remote neon instance
         if [ "${{ inputs.build_type }}" = "remote" ] && [ -n "${BENCHMARK_CONNSTR}" ]; then
-          QUERIES=("SELECT version()")
-          if [[ "${PLATFORM}" = "neon"* ]]; then
-            QUERIES+=("SHOW neon.tenant_id")
-            QUERIES+=("SHOW neon.timeline_id")
-          fi
-
-          for q in "${QUERIES[@]}"; do
-            ${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/psql ${BENCHMARK_CONNSTR} -c "${q}"
-          done
+          ${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/psql ${BENCHMARK_CONNSTR} -c "SELECT version();"
         fi
 
         # Run the tests.
         #
-        # --alluredir saves test results in Allure format (in a specified directory)
+        # The junit.xml file allows CI tools to display more fine-grained test information
+        # in its "Tests" tab in the results page.
         # --verbose prints name of each test (helpful when there are
         # multiple tests in one file)
         # -rA prints summary in the end
@@ -199,6 +193,7 @@ runs:
         #
         mkdir -p $TEST_OUTPUT/allure/results
         "${cov_prefix[@]}" ./scripts/pytest \
+          --junitxml=$TEST_OUTPUT/junit.xml \
           --alluredir=$TEST_OUTPUT/allure/results \
           --tb=short \
           --verbose \
@@ -211,13 +206,13 @@ runs:
         fi
 
     - name: Upload compatibility snapshot
-      # Note, that we use `github.base_ref` which is a target branch for a PR
-      if: github.event_name == 'pull_request' && github.base_ref == 'release'
+      if: github.ref_name == 'release'
       uses: ./.github/actions/upload
       with:
-        name: compatibility-snapshot-${{ runner.arch }}-${{ inputs.build_type }}-pg${{ inputs.pg_version }}
+        name: compatibility-snapshot-${{ inputs.build_type }}-pg${{ inputs.pg_version }}-${{ github.run_id }}
         # Directory is created by test_compatibility.py::test_create_snapshot, keep the path in sync with the test
         path: /tmp/test_output/compatibility_snapshot_pg${{ inputs.pg_version }}/
+        prefix: latest
 
     - name: Upload test results
       if: ${{ !cancelled() }}

.github/actions/set-docker-config-dir/action.yml

@@ -1,36 +0,0 @@
-name: "Set custom docker config directory"
-description: "Create a directory for docker config and set DOCKER_CONFIG"
-
-# Use custom DOCKER_CONFIG directory to avoid conflicts with default settings
-runs:
-  using: "composite"
-  steps:
-  - name: Show warning on GitHub-hosted runners
-    if: runner.environment == 'github-hosted'
-    shell: bash -euo pipefail {0}
-    run: |
-      # Using the following environment variables to find a path to the workflow file
-      # ${GITHUB_WORKFLOW_REF} - octocat/hello-world/.github/workflows/my-workflow.yml@refs/heads/my_branch
-      # ${GITHUB_REPOSITORY}   - octocat/hello-world
-      # ${GITHUB_REF}          - refs/heads/my_branch
-      # From https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/variables
-
-      filename_with_ref=${GITHUB_WORKFLOW_REF#"$GITHUB_REPOSITORY/"}
-      filename=${filename_with_ref%"@$GITHUB_REF"}
-
-      # https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/workflow-commands-for-github-actions#setting-a-warning-message
-      title='Unnecessary usage of `.github/actions/set-docker-config-dir`'
-      message='No need to use `.github/actions/set-docker-config-dir` action on GitHub-hosted runners'
-      echo "::warning file=${filename},title=${title}::${message}"
-
-  - uses: pyTooling/Actions/with-post-step@74afc5a42a17a046c90c68cb5cfa627e5c6c5b6b # v1.0.7
-    env:
-      DOCKER_CONFIG: .docker-custom-${{ github.run_id }}-${{ github.run_attempt }}
-    with:
-      main: |
-        mkdir -p "${DOCKER_CONFIG}"
-        echo DOCKER_CONFIG=${DOCKER_CONFIG} | tee -a $GITHUB_ENV
-      post: |
-        if [ -d "${DOCKER_CONFIG}" ]; then
-          rm -r "${DOCKER_CONFIG}"
-        fi

.github/actions/upload/action.yml

@@ -8,7 +8,7 @@ inputs:
     description: "A directory or file to upload"
     required: true
   prefix:
-    description: "S3 prefix. Default is '${GITHUB_SHA}/${GITHUB_RUN_ID}/${GITHUB_RUN_ATTEMPT}'"
+    description: "S3 prefix. Default is '${GITHUB_RUN_ID}/${GITHUB_RUN_ATTEMPT}'"
     required: false
 
 runs:
@@ -45,7 +45,7 @@ runs:
       env:
         SOURCE: ${{ inputs.path }}
         ARCHIVE: /tmp/uploads/${{ inputs.name }}.tar.zst
-        PREFIX: artifacts/${{ inputs.prefix || format('{0}/{1}/{2}', github.event.pull_request.head.sha || github.sha, github.run_id , github.run_attempt) }}
+        PREFIX: artifacts/${{ inputs.prefix || format('{0}/{1}', github.run_id, github.run_attempt) }}
       run: |
         BUCKET=neon-github-public-dev
         FILENAME=$(basename $ARCHIVE)

.github/workflows/_benchmarking_preparation.yml

@@ -1,154 +0,0 @@
-name: Prepare benchmarking databases by restoring dumps
-
-on:
-  workflow_call:
-    # no inputs needed
-    
-defaults:
-  run:
-    shell: bash -euxo pipefail {0}
-
-jobs:
-  setup-databases:
-    strategy:
-      fail-fast: false
-      matrix:
-        platform: [ aws-rds-postgres, aws-aurora-serverless-v2-postgres, neon ] 
-        database: [ clickbench, tpch, userexample ]
-  
-    env:
-      LD_LIBRARY_PATH: /tmp/neon/pg_install/v16/lib
-      PLATFORM: ${{ matrix.platform }}
-      PG_BINARIES: /tmp/neon/pg_install/v16/bin
-
-    runs-on: [ self-hosted, us-east-2, x64 ]
-    container:
-      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:pinned
-      options: --init
-
-    steps:
-    - name: Set up Connection String
-      id: set-up-prep-connstr
-      run: |
-        case "${PLATFORM}" in
-          neon)
-            CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR }} 
-            ;;
-          aws-rds-postgres)
-            CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_CONNSTR }} 
-            ;;
-          aws-aurora-serverless-v2-postgres)
-            CONNSTR=${{ secrets.BENCHMARK_RDS_AURORA_CONNSTR }} 
-            ;;
-          *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}"
-            exit 1
-            ;;
-        esac
-
-        echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT  
-
-    - uses: actions/checkout@v4
-
-    - name: Download Neon artifact
-      uses: ./.github/actions/download
-      with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
-        path: /tmp/neon/
-        prefix: latest
-
-    # we create a table that has one row for each database that we want to restore with the status whether the restore is done    
-    - name: Create benchmark_restore_status table if it does not exist
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.set-up-prep-connstr.outputs.connstr }}
-        DATABASE_NAME: ${{ matrix.database }}
-      # to avoid a race condition of multiple jobs trying to create the table at the same time, 
-      # we use an advisory lock
-      run: |
-        ${PG_BINARIES}/psql "${{ env.BENCHMARK_CONNSTR }}" -c "
-        SELECT pg_advisory_lock(4711);  
-        CREATE TABLE IF NOT EXISTS benchmark_restore_status (
-        databasename text primary key,
-        restore_done boolean
-        );
-        SELECT pg_advisory_unlock(4711);
-        "
-    
-    - name: Check if restore is already done
-      id: check-restore-done
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.set-up-prep-connstr.outputs.connstr }}
-        DATABASE_NAME: ${{ matrix.database }}
-      run: |
-        skip=false
-        if ${PG_BINARIES}/psql "${{ env.BENCHMARK_CONNSTR }}" -tAc "SELECT 1 FROM benchmark_restore_status WHERE databasename='${{ env.DATABASE_NAME }}' AND restore_done=true;" | grep -q 1; then
-          echo "Restore already done for database ${{ env.DATABASE_NAME }} on platform ${{ env.PLATFORM }}. Skipping this database."
-          skip=true
-        fi
-        echo "skip=${skip}" | tee -a $GITHUB_OUTPUT
-
-    - name: Check and create database if it does not exist
-      if: steps.check-restore-done.outputs.skip != 'true'
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.set-up-prep-connstr.outputs.connstr }}
-        DATABASE_NAME: ${{ matrix.database }}
-      run: |
-        DB_EXISTS=$(${PG_BINARIES}/psql "${{ env.BENCHMARK_CONNSTR }}" -tAc "SELECT 1 FROM pg_database WHERE datname='${{ env.DATABASE_NAME }}'")
-        if [ "$DB_EXISTS" != "1" ]; then
-          echo "Database ${{ env.DATABASE_NAME }} does not exist. Creating it..."
-          ${PG_BINARIES}/psql "${{ env.BENCHMARK_CONNSTR }}" -c "CREATE DATABASE \"${{ env.DATABASE_NAME }}\";"
-        else
-          echo "Database ${{ env.DATABASE_NAME }} already exists."
-        fi
-
-    - name: Download dump from S3 to /tmp/dumps
-      if: steps.check-restore-done.outputs.skip != 'true'
-      env:
-        DATABASE_NAME: ${{ matrix.database }}
-      run: |
-        mkdir -p /tmp/dumps
-        aws s3 cp s3://neon-github-dev/performance/pgdumps/$DATABASE_NAME/$DATABASE_NAME.pg_dump /tmp/dumps/ 
-
-    - name: Replace database name in connection string
-      if: steps.check-restore-done.outputs.skip != 'true'
-      id: replace-dbname
-      env:
-        DATABASE_NAME: ${{ matrix.database }}
-        BENCHMARK_CONNSTR: ${{ steps.set-up-prep-connstr.outputs.connstr }}
-      run: |
-        # Extract the part before the database name
-        base_connstr="${BENCHMARK_CONNSTR%/*}"
-        # Extract the query parameters (if any) after the database name
-        query_params="${BENCHMARK_CONNSTR#*\?}"
-        # Reconstruct the new connection string
-        if [ "$query_params" != "$BENCHMARK_CONNSTR" ]; then
-          new_connstr="${base_connstr}/${DATABASE_NAME}?${query_params}"
-        else
-          new_connstr="${base_connstr}/${DATABASE_NAME}"
-        fi
-        echo "database_connstr=${new_connstr}" >> $GITHUB_OUTPUT  
-
-    - name: Restore dump
-      if: steps.check-restore-done.outputs.skip != 'true'
-      env:
-        DATABASE_NAME: ${{ matrix.database }}
-        DATABASE_CONNSTR: ${{ steps.replace-dbname.outputs.database_connstr }}
-        # the following works only with larger computes: 
-        # PGOPTIONS: "-c maintenance_work_mem=8388608 -c max_parallel_maintenance_workers=7"
-        # we add the || true because:
-        # the dumps were created with Neon and contain neon extensions that are not 
-        # available in RDS, so we will always report an error, but we can ignore it
-      run: |
-        ${PG_BINARIES}/pg_restore --clean --if-exists --no-owner --jobs=4 \
-        -d "${DATABASE_CONNSTR}" /tmp/dumps/${DATABASE_NAME}.pg_dump || true
-
-    - name: Update benchmark_restore_status table
-      if: steps.check-restore-done.outputs.skip != 'true'
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.set-up-prep-connstr.outputs.connstr }}
-        DATABASE_NAME: ${{ matrix.database }}
-      run: |
-        ${PG_BINARIES}/psql "${{ env.BENCHMARK_CONNSTR }}" -c "
-        INSERT INTO benchmark_restore_status (databasename, restore_done) VALUES ('${{ env.DATABASE_NAME }}', true)
-        ON CONFLICT (databasename) DO UPDATE SET restore_done = true;
-        "

.github/workflows/_build-and-test-locally.yml

@@ -1,303 +0,0 @@
-name: Build and Test Locally
-
-on:
-  workflow_call:
-    inputs:
-      arch:
-        description: 'x64 or arm64'
-        required: true
-        type: string
-      build-tag:
-        description: 'build tag'
-        required: true
-        type: string
-      build-tools-image:
-        description: 'build-tools image'
-        required: true
-        type: string
-      build-type:
-        description: 'debug or release'
-        required: true
-        type: string
-      pg-versions:
-        description: 'a json array of postgres versions to run regression tests on'
-        required: true
-        type: string
-
-defaults:
-  run:
-    shell: bash -euxo pipefail {0}
-
-env:
-  RUST_BACKTRACE: 1
-  COPT: '-Werror'
-  AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
-  AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
-
-jobs:
-  build-neon:
-    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', inputs.arch == 'arm64' && 'large-arm64' || 'large')) }}
-    container:
-      image: ${{ inputs.build-tools-image }}
-      credentials:
-        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
-        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
-      # Raise locked memory limit for tokio-epoll-uring.
-      # On 5.10 LTS kernels < 5.10.162 (and generally mainline kernels < 5.12),
-      # io_uring will account the memory of the CQ and SQ as locked.
-      # More details: https://github.com/neondatabase/neon/issues/6373#issuecomment-1905814391
-      options: --init --shm-size=512mb --ulimit memlock=67108864:67108864
-    env:
-      BUILD_TYPE: ${{ inputs.build-type }}
-      GIT_VERSION: ${{ github.event.pull_request.head.sha || github.sha }}
-      BUILD_TAG: ${{ inputs.build-tag }}
-
-    steps:
-      - name: Fix git ownership
-        run: |
-          # Workaround for `fatal: detected dubious ownership in repository at ...`
-          #
-          # Use both ${{ github.workspace }} and ${GITHUB_WORKSPACE} because they're different on host and in containers
-          #   Ref https://github.com/actions/checkout/issues/785
-          #
-          git config --global --add safe.directory ${{ github.workspace }}
-          git config --global --add safe.directory ${GITHUB_WORKSPACE}
-          for r in 14 15 16; do
-            git config --global --add safe.directory "${{ github.workspace }}/vendor/postgres-v$r"
-            git config --global --add safe.directory "${GITHUB_WORKSPACE}/vendor/postgres-v$r"
-          done
-
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Set pg 14 revision for caching
-        id: pg_v14_rev
-        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v14) >> $GITHUB_OUTPUT
-
-      - name: Set pg 15 revision for caching
-        id: pg_v15_rev
-        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v15) >> $GITHUB_OUTPUT
-
-      - name: Set pg 16 revision for caching
-        id: pg_v16_rev
-        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v16) >> $GITHUB_OUTPUT
-
-      # Set some environment variables used by all the steps.
-      #
-      # CARGO_FLAGS is extra options to pass to "cargo build", "cargo test" etc.
-      #   It also includes --features, if any
-      #
-      # CARGO_FEATURES is passed to "cargo metadata". It is separate from CARGO_FLAGS,
-      #   because "cargo metadata" doesn't accept --release or --debug options
-      #
-      # We run tests with addtional features, that are turned off by default (e.g. in release builds), see
-      # corresponding Cargo.toml files for their descriptions.
-      - name: Set env variables
-        env:
-          ARCH: ${{ inputs.arch }}
-        run: |
-          CARGO_FEATURES="--features testing"
-          if [[ $BUILD_TYPE == "debug" && $ARCH == 'x64' ]]; then
-            cov_prefix="scripts/coverage --profraw-prefix=$GITHUB_JOB --dir=/tmp/coverage run"
-            CARGO_FLAGS="--locked"
-          elif [[ $BUILD_TYPE == "debug" ]]; then
-            cov_prefix=""
-            CARGO_FLAGS="--locked"
-          elif [[ $BUILD_TYPE == "release" ]]; then
-            cov_prefix=""
-            CARGO_FLAGS="--locked --release"
-          fi
-          {
-            echo "cov_prefix=${cov_prefix}"
-            echo "CARGO_FEATURES=${CARGO_FEATURES}"
-            echo "CARGO_FLAGS=${CARGO_FLAGS}"
-            echo "CARGO_HOME=${GITHUB_WORKSPACE}/.cargo"
-          } >> $GITHUB_ENV
-
-      - name: Cache postgres v14 build
-        id: cache_pg_14
-        uses: actions/cache@v4
-        with:
-          path: pg_install/v14
-          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-pg-${{ steps.pg_v14_rev.outputs.pg_rev }}-${{ hashFiles('Makefile', 'Dockerfile.build-tools') }}
-
-      - name: Cache postgres v15 build
-        id: cache_pg_15
-        uses: actions/cache@v4
-        with:
-          path: pg_install/v15
-          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-pg-${{ steps.pg_v15_rev.outputs.pg_rev }}-${{ hashFiles('Makefile', 'Dockerfile.build-tools') }}
-
-      - name: Cache postgres v16 build
-        id: cache_pg_16
-        uses: actions/cache@v4
-        with:
-          path: pg_install/v16
-          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-pg-${{ steps.pg_v16_rev.outputs.pg_rev }}-${{ hashFiles('Makefile', 'Dockerfile.build-tools') }}
-
-      - name: Build postgres v14
-        if: steps.cache_pg_14.outputs.cache-hit != 'true'
-        run: mold -run make postgres-v14 -j$(nproc)
-
-      - name: Build postgres v15
-        if: steps.cache_pg_15.outputs.cache-hit != 'true'
-        run: mold -run make postgres-v15 -j$(nproc)
-
-      - name: Build postgres v16
-        if: steps.cache_pg_16.outputs.cache-hit != 'true'
-        run: mold -run make postgres-v16 -j$(nproc)
-
-      - name: Build neon extensions
-        run: mold -run make neon-pg-ext -j$(nproc)
-
-      - name: Build walproposer-lib
-        run: mold -run make walproposer-lib -j$(nproc)
-
-      - name: Run cargo build
-        run: |
-          PQ_LIB_DIR=$(pwd)/pg_install/v16/lib
-          export PQ_LIB_DIR
-          ${cov_prefix} mold -run cargo build $CARGO_FLAGS $CARGO_FEATURES --bins --tests
-
-      # Do install *before* running rust tests because they might recompile the
-      # binaries with different features/flags.
-      - name: Install rust binaries
-        env:
-          ARCH: ${{ inputs.arch }}
-        run: |
-          # Install target binaries
-          mkdir -p /tmp/neon/bin/
-          binaries=$(
-            ${cov_prefix} cargo metadata $CARGO_FEATURES --format-version=1 --no-deps |
-            jq -r '.packages[].targets[] | select(.kind | index("bin")) | .name'
-          )
-          for bin in $binaries; do
-            SRC=target/$BUILD_TYPE/$bin
-            DST=/tmp/neon/bin/$bin
-            cp "$SRC" "$DST"
-          done
-
-          # Install test executables and write list of all binaries (for code coverage)
-          if [[ $BUILD_TYPE == "debug" && $ARCH == 'x64' ]]; then
-            # Keep bloated coverage data files away from the rest of the artifact
-            mkdir -p /tmp/coverage/
-
-            mkdir -p /tmp/neon/test_bin/
-
-            test_exe_paths=$(
-              ${cov_prefix} cargo test $CARGO_FLAGS $CARGO_FEATURES --message-format=json --no-run |
-              jq -r '.executable | select(. != null)'
-            )
-            for bin in $test_exe_paths; do
-              SRC=$bin
-              DST=/tmp/neon/test_bin/$(basename $bin)
-
-              # We don't need debug symbols for code coverage, so strip them out to make
-              # the artifact smaller.
-              strip "$SRC" -o "$DST"
-              echo "$DST" >> /tmp/coverage/binaries.list
-            done
-
-            for bin in $binaries; do
-              echo "/tmp/neon/bin/$bin" >> /tmp/coverage/binaries.list
-            done
-          fi
-
-      - name: Run rust tests
-        env:
-          NEXTEST_RETRIES: 3
-        run: |
-          PQ_LIB_DIR=$(pwd)/pg_install/v16/lib
-          export PQ_LIB_DIR
-          LD_LIBRARY_PATH=$(pwd)/pg_install/v16/lib
-          export LD_LIBRARY_PATH
-
-          #nextest does not yet support running doctests
-          ${cov_prefix} cargo test --doc $CARGO_FLAGS $CARGO_FEATURES
-
-          # run all non-pageserver tests
-          ${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E '!package(pageserver)'
-
-          # run pageserver tests with different settings
-          for io_engine in std-fs tokio-epoll-uring ; do
-            for io_buffer_alignment in 0 1 512 ; do
-              NEON_PAGESERVER_UNIT_TEST_VIRTUAL_FILE_IOENGINE=$io_engine NEON_PAGESERVER_UNIT_TEST_IO_BUFFER_ALIGNMENT=$io_buffer_alignment ${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES  -E 'package(pageserver)'
-            done
-          done
-
-          # Run separate tests for real S3
-          export ENABLE_REAL_S3_REMOTE_STORAGE=nonempty
-          export REMOTE_STORAGE_S3_BUCKET=neon-github-ci-tests
-          export REMOTE_STORAGE_S3_REGION=eu-central-1
-          ${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E 'package(remote_storage)' -E 'test(test_real_s3)'
-
-          # Run separate tests for real Azure Blob Storage
-          # XXX: replace region with `eu-central-1`-like region
-          export ENABLE_REAL_AZURE_REMOTE_STORAGE=y
-          export AZURE_STORAGE_ACCOUNT="${{ secrets.AZURE_STORAGE_ACCOUNT_DEV }}"
-          export AZURE_STORAGE_ACCESS_KEY="${{ secrets.AZURE_STORAGE_ACCESS_KEY_DEV }}"
-          export REMOTE_STORAGE_AZURE_CONTAINER="${{ vars.REMOTE_STORAGE_AZURE_CONTAINER }}"
-          export REMOTE_STORAGE_AZURE_REGION="${{ vars.REMOTE_STORAGE_AZURE_REGION }}"
-          ${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E 'package(remote_storage)' -E 'test(test_real_azure)'
-
-      - name: Install postgres binaries
-        run: cp -a pg_install /tmp/neon/pg_install
-
-      - name: Upload Neon artifact
-        uses: ./.github/actions/upload
-        with:
-          name: neon-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-artifact
-          path: /tmp/neon
-
-      # XXX: keep this after the binaries.list is formed, so the coverage can properly work later
-      - name: Merge and upload coverage data
-        if: inputs.build-type == 'debug'
-        uses: ./.github/actions/save-coverage-data
-
-  regress-tests:
-    # Don't run regression tests on debug arm64 builds
-    if: inputs.build-type != 'debug' || inputs.arch != 'arm64'
-    needs: [ build-neon ]
-    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', inputs.arch == 'arm64' && 'large-arm64' || 'large')) }}
-    container:
-      image: ${{ inputs.build-tools-image }}
-      credentials:
-        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
-        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
-      # for changed limits, see comments on `options:` earlier in this file
-      options: --init --shm-size=512mb --ulimit memlock=67108864:67108864
-    strategy:
-      fail-fast: false
-      matrix:
-        pg_version: ${{ fromJson(inputs.pg-versions) }}
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Pytest regression tests
-        uses: ./.github/actions/run-python-test-set
-        timeout-minutes: 60
-        with:
-          build_type: ${{ inputs.build-type }}
-          test_selection: regress
-          needs_postgres_source: true
-          run_with_real_s3: true
-          real_s3_bucket: neon-github-ci-tests
-          real_s3_region: eu-central-1
-          rerun_flaky: true
-          pg_version: ${{ matrix.pg_version }}
-        env:
-          TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}
-          CHECK_ONDISK_DATA_COMPATIBILITY: nonempty
-          BUILD_TAG: ${{ inputs.build-tag }}
-          PAGESERVER_VIRTUAL_FILE_IO_ENGINE: tokio-epoll-uring
-
-      # Temporary disable this step until we figure out why it's so flaky
-      # Ref https://github.com/neondatabase/neon/issues/4540
-      - name: Merge and upload coverage data
-        if: |
-          false &&
-          inputs.build-type == 'debug' && matrix.pg_version == 'v16'
-        uses: ./.github/actions/save-coverage-data

.github/workflows/actionlint.yml

@@ -24,7 +24,7 @@ jobs:
 
   actionlint:
     needs: [ check-permissions ]
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
       - uses: reviewdog/action-actionlint@v1
@@ -36,16 +36,3 @@ jobs:
           fail_on_error: true
           filter_mode: nofilter
           level: error
-
-      - name: Disallow 'ubuntu-latest' runners
-        run: |
-          PAT='^\s*runs-on:.*-latest'
-          if grep -ERq $PAT .github/workflows; then
-            grep -ERl $PAT .github/workflows |\
-            while read -r f
-            do
-              l=$(grep -nE $PAT $f | awk -F: '{print $1}' | head -1)
-              echo "::error file=$f,line=$l::Please use 'ubuntu-22.04' instead of 'ubuntu-latest'"
-            done
-            exit 1
-          fi

.github/workflows/approved-for-ci-run.yml

@@ -18,7 +18,6 @@ on:
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.event.pull_request.number }}
-  cancel-in-progress: false
 
 env:
   GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
@@ -44,7 +43,7 @@ jobs:
       contains(fromJSON('["opened", "synchronize", "reopened", "closed"]'), github.event.action) &&
       contains(github.event.pull_request.labels.*.name, 'approved-for-ci-run')
 
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     steps:
       - run: gh pr --repo "${GITHUB_REPOSITORY}" edit "${PR_NUMBER}" --remove-label "approved-for-ci-run"
@@ -60,7 +59,7 @@ jobs:
       github.event.action == 'labeled' &&
       contains(github.event.pull_request.labels.*.name, 'approved-for-ci-run')
 
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     steps:
       - run: gh pr --repo "${GITHUB_REPOSITORY}" edit "${PR_NUMBER}" --remove-label "approved-for-ci-run"
@@ -69,41 +68,15 @@ jobs:
         with:
           ref: main
           token: ${{ secrets.CI_ACCESS_TOKEN }}
-      
-      - name: Look for existing PR
-        id: get-pr
-        env:
-          GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
-        run: |
-          ALREADY_CREATED="$(gh pr --repo ${GITHUB_REPOSITORY} list --head ${BRANCH} --base main --json number --jq '.[].number')"
-          echo "ALREADY_CREATED=${ALREADY_CREATED}" >> ${GITHUB_OUTPUT}
-      
-      - name: Get changed labels
-        id: get-labels
-        if: steps.get-pr.outputs.ALREADY_CREATED != ''
-        env:
-          ALREADY_CREATED: ${{ steps.get-pr.outputs.ALREADY_CREATED }}
-          GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
-        run: |
-          LABELS_TO_REMOVE=$(comm -23 <(gh pr --repo ${GITHUB_REPOSITORY} view ${ALREADY_CREATED} --json labels --jq '.labels.[].name'| ( grep -E '^run' || true ) | sort) \
-          <(gh pr --repo ${GITHUB_REPOSITORY} view ${PR_NUMBER} --json labels --jq '.labels.[].name' | ( grep -E '^run' || true ) | sort ) |\
-          ( grep -v run-e2e-tests-in-draft || true ) | paste -sd , -)
-          LABELS_TO_ADD=$(comm -13 <(gh pr --repo ${GITHUB_REPOSITORY} view ${ALREADY_CREATED} --json labels --jq '.labels.[].name'| ( grep -E '^run' || true ) |sort) \
-          <(gh pr --repo ${GITHUB_REPOSITORY} view ${PR_NUMBER} --json labels --jq '.labels.[].name' |  ( grep -E '^run' || true ) | sort ) |\
-          paste -sd , -)
-          echo "LABELS_TO_ADD=${LABELS_TO_ADD}" >> ${GITHUB_OUTPUT}
-          echo "LABELS_TO_REMOVE=${LABELS_TO_REMOVE}" >> ${GITHUB_OUTPUT}
 
       - run: gh pr checkout "${PR_NUMBER}"
 
       - run: git checkout -b "${BRANCH}"
 
       - run: git push --force origin "${BRANCH}"
-        if: steps.get-pr.outputs.ALREADY_CREATED == ''
 
       - name: Create a Pull Request for CI run (if required)
-        if: steps.get-pr.outputs.ALREADY_CREATED == ''
-        env: 
+        env:
           GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
         run: |
           cat << EOF > body.md
@@ -114,33 +87,16 @@ jobs:
             Feel free to review/comment/discuss the original PR #${PR_NUMBER}.
           EOF
 
-          LABELS=$( (gh pr --repo "${GITHUB_REPOSITORY}" view ${PR_NUMBER}  --json labels --jq '.labels.[].name'; echo run-e2e-tests-in-draft  )| \
-          grep -E '^run' | paste -sd , -)
-          gh pr --repo "${GITHUB_REPOSITORY}" create --title "CI run for PR #${PR_NUMBER}" \
+          ALREADY_CREATED="$(gh pr --repo ${GITHUB_REPOSITORY} list --head ${BRANCH} --base main --json number --jq '.[].number')"
+          if [ -z "${ALREADY_CREATED}" ]; then
+            gh pr --repo "${GITHUB_REPOSITORY}" create --title "CI run for PR #${PR_NUMBER}" \
                                                        --body-file "body.md" \
                                                        --head "${BRANCH}" \
                                                        --base "main" \
-                                                       --label ${LABELS} \
+                                                       --label "run-e2e-tests-in-draft" \
                                                        --draft
-      - name: Modify the existing pull request (if required)
-        if: steps.get-pr.outputs.ALREADY_CREATED != ''
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          LABELS_TO_ADD: ${{ steps.get-labels.outputs.LABELS_TO_ADD }}
-          LABELS_TO_REMOVE: ${{ steps.get-labels.outputs.LABELS_TO_REMOVE }}
-          ALREADY_CREATED: ${{ steps.get-pr.outputs.ALREADY_CREATED }}
-        run: |
-          ADD_CMD=
-          REMOVE_CMD=
-          [ -z "${LABELS_TO_ADD}" ] || ADD_CMD="--add-label ${LABELS_TO_ADD}"
-          [ -z "${LABELS_TO_REMOVE}" ] || REMOVE_CMD="--remove-label ${LABELS_TO_REMOVE}"
-          if [ -n "${ADD_CMD}" ] || [ -n "${REMOVE_CMD}" ]; then
-            gh pr --repo "${GITHUB_REPOSITORY}" edit ${ALREADY_CREATED} ${ADD_CMD} ${REMOVE_CMD}
           fi
 
-      - run: git push --force origin "${BRANCH}"
-        if: steps.get-pr.outputs.ALREADY_CREATED != ''
-             
   cleanup:
     # Close PRs and delete branchs if the original PR is closed.
 
@@ -152,7 +108,7 @@ jobs:
       github.event.action == 'closed' &&
       github.event.pull_request.head.repo.full_name != github.repository
 
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     steps:
       - name: Close PR and delete `ci-run/pr-${{ env.PR_NUMBER }}` branch

.github/workflows/benchmarking.yml

@@ -38,11 +38,6 @@ on:
         description: 'AWS-RDS and AWS-AURORA normally only run on Saturday. Set this to true to run them on every workflow_dispatch'
         required: false
         default: false
-      run_only_pgvector_tests:
-        type: boolean
-        description: 'Run pgvector tests but no other tests. If not set, all tests including pgvector tests will be run'
-        required: false
-        default: false
 
 defaults:
   run:
@@ -55,118 +50,11 @@ concurrency:
 
 jobs:
   bench:
-    if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
-    permissions:
-      contents: write
-      statuses: write
-      id-token: write # Required for OIDC authentication in azure runners
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - DEFAULT_PG_VERSION: 16
-            PLATFORM: "neon-staging"
-            region_id: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
-            RUNNER: [ self-hosted, us-east-2, x64 ]
-            IMAGE: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:pinned
-          - DEFAULT_PG_VERSION: 16
-            PLATFORM: "azure-staging"
-            region_id: 'azure-eastus2'
-            RUNNER: [ self-hosted, eastus2, x64 ]
-            IMAGE: neondatabase/build-tools:pinned
     env:
       TEST_PG_BENCH_DURATIONS_MATRIX: "300"
       TEST_PG_BENCH_SCALES_MATRIX: "10,100"
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: ${{ matrix.DEFAULT_PG_VERSION }}
-      TEST_OUTPUT: /tmp/test_output
-      BUILD_TYPE: remote
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
-      PLATFORM: ${{ matrix.PLATFORM }}
-
-    runs-on: ${{ matrix.RUNNER }}
-    container:
-      image: ${{ matrix.IMAGE }}
-      options: --init
-
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: Configure AWS credentials # necessary on Azure runners
-      uses: aws-actions/configure-aws-credentials@v4
-      with:
-        aws-region: eu-central-1
-        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
-        role-duration-seconds: 18000 # 5 hours
-
-    - name: Download Neon artifact
-      uses: ./.github/actions/download
-      with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
-        path: /tmp/neon/
-        prefix: latest
-
-    - name: Create Neon Project
-      id: create-neon-project
-      uses: ./.github/actions/neon-project-create
-      with:
-        region_id: ${{ matrix.region_id }}
-        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
-        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-
-    - name: Run benchmark
-      uses: ./.github/actions/run-python-test-set
-      with:
-        build_type: ${{ env.BUILD_TYPE }}
-        test_selection: performance
-        run_in_parallel: false
-        save_perf_report: ${{ env.SAVE_PERF_REPORT }}
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
-        # Set --sparse-ordering option of pytest-order plugin
-        # to ensure tests are running in order of appears in the file.
-        # It's important for test_perf_pgbench.py::test_pgbench_remote_* tests
-        extra_params:
-          -m remote_cluster
-          --sparse-ordering
-          --timeout 14400
-          --ignore test_runner/performance/test_perf_olap.py
-          --ignore test_runner/performance/test_perf_pgvector_queries.py
-          --ignore test_runner/performance/test_logical_replication.py
-          --ignore test_runner/performance/test_physical_replication.py
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
-        VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
-        PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
-
-    - name: Delete Neon Project
-      if: ${{ always() }}
-      uses: ./.github/actions/neon-project-delete
-      with:
-        project_id: ${{ steps.create-neon-project.outputs.project_id }}
-        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-
-    - name: Create Allure report
-      id: create-allure-report
-      if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
-
-    - name: Post to a Slack channel
-      if: ${{ github.event.schedule && failure() }}
-      uses: slackapi/slack-github-action@v1
-      with:
-        channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: |
-          Periodic perf testing: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
-      env:
-        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
-
-  replication-tests:
-    if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
-    env:
-      POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      DEFAULT_PG_VERSION: 14
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -180,81 +68,69 @@ jobs:
     steps:
     - uses: actions/checkout@v4
 
-
     - name: Download Neon artifact
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
+        name: neon-${{ runner.os }}-release-artifact
         path: /tmp/neon/
         prefix: latest
 
-    - name: Run Logical Replication benchmarks
-      uses: ./.github/actions/run-python-test-set
+    - name: Create Neon Project
+      id: create-neon-project
+      uses: ./.github/actions/neon-project-create
       with:
-        build_type: ${{ env.BUILD_TYPE }}
-        test_selection: performance/test_logical_replication.py
-        run_in_parallel: false
-        save_perf_report: ${{ env.SAVE_PERF_REPORT }}
-        extra_params: -m remote_cluster --timeout 5400
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
-      env:
-        VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
-        PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
-        NEON_API_KEY: ${{ secrets.NEON_STAGING_API_KEY }}
-        BENCHMARK_PROJECT_ID_PUB: ${{ vars.BENCHMARK_PROJECT_ID_PUB }}
-        BENCHMARK_PROJECT_ID_SUB: ${{ vars.BENCHMARK_PROJECT_ID_SUB }}
+        region_id: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
+        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
+        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
 
-    - name: Run Physical Replication benchmarks
+    - name: Run benchmark
       uses: ./.github/actions/run-python-test-set
       with:
         build_type: ${{ env.BUILD_TYPE }}
-        test_selection: performance/test_physical_replication.py
+        test_selection: performance
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
-        extra_params: -m remote_cluster --timeout 5400
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        # Set --sparse-ordering option of pytest-order plugin
+        # to ensure tests are running in order of appears in the file.
+        # It's important for test_perf_pgbench.py::test_pgbench_remote_* tests
+        extra_params: -m remote_cluster --sparse-ordering --timeout 5400 --ignore test_runner/performance/test_perf_olap.py
       env:
+        BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
         PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
-        NEON_API_KEY: ${{ secrets.NEON_STAGING_API_KEY }}
+
+    - name: Delete Neon Project
+      if: ${{ always() }}
+      uses: ./.github/actions/neon-project-delete
+      with:
+        project_id: ${{ steps.create-neon-project.outputs.project_id }}
+        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
 
     - name: Create Allure report
-      id: create-allure-report
       if: ${{ !cancelled() }}
       uses: ./.github/actions/allure-report-generate
-      with:
-        store-test-results-into-db: true
-      env:
-        REGRESS_TEST_RESULT_CONNSTR_NEW: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}
 
     - name: Post to a Slack channel
       if: ${{ github.event.schedule && failure() }}
       uses: slackapi/slack-github-action@v1
       with:
-        channel-id: "C06T9AMNDQQ" # on-call-compute-staging-stream
-        slack-message: |
-          Periodic replication testing: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
+        channel-id: "C033QLM5P7D" # dev-staging-stream
+        slack-message: "Periodic perf testing: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
       env:
         SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
 
   generate-matrices:
-    if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
     # Create matrices for the benchmarking jobs, so we run benchmarks on rds only once a week (on Saturday)
     #
     # Available platforms:
-    # - neonvm-captest-new: Freshly created project (1 CU)
-    # - neonvm-captest-freetier: Use freetier-sized compute (0.25 CU)
-    # - neonvm-captest-azure-new: Freshly created project (1 CU) in azure region
-    # - neonvm-captest-azure-freetier: Use freetier-sized compute (0.25 CU) in azure region
-    # - neonvm-captest-reuse: Reusing existing project
+    # - neon-captest-new: Freshly created project (1 CU)
+    # - neon-captest-freetier: Use freetier-sized compute (0.25 CU)
+    # - neon-captest-reuse: Reusing existing project
     # - rds-aurora: Aurora Postgres Serverless v2 with autoscaling from 0.5 to 2 ACUs
     # - rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
     env:
       RUN_AWS_RDS_AND_AURORA: ${{ github.event.inputs.run_AWS_RDS_AND_AURORA || 'false' }}
-      DEFAULT_REGION_ID: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     outputs:
       pgbench-compare-matrix: ${{ steps.pgbench-compare-matrix.outputs.matrix }}
       olap-compare-matrix: ${{ steps.olap-compare-matrix.outputs.matrix }}
@@ -264,37 +140,23 @@ jobs:
     - name: Generate matrix for pgbench benchmark
       id: pgbench-compare-matrix
       run: |
-        region_id_default=${{ env.DEFAULT_REGION_ID }}
-        runner_default='["self-hosted", "us-east-2", "x64"]'
-        runner_azure='["self-hosted", "eastus2", "x64"]'
-        image_default="369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:pinned"
         matrix='{
-          "pg_version" : [
-            16
-          ],
-          "region_id" : [
-            "'"$region_id_default"'"
-            ],
           "platform": [
-            "neonvm-captest-new",
-            "neonvm-captest-reuse",
+            "neon-captest-new",
+            "neon-captest-reuse",
             "neonvm-captest-new"
           ],
           "db_size": [ "10gb" ],
-          "runner": ['"$runner_default"'],
-          "image": [ "'"$image_default"'" ],
-          "include": [{ "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-freetier",       "db_size": "3gb" ,"runner": '"$runner_default"', "image": "'"$image_default"'" },
-                      { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-new",            "db_size": "10gb","runner": '"$runner_default"', "image": "'"$image_default"'" },
-                      { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-new",            "db_size": "50gb","runner": '"$runner_default"', "image": "'"$image_default"'" },
-                      { "pg_version": 16, "region_id": "azure-eastus2",          "platform": "neonvm-azure-captest-freetier", "db_size": "3gb" ,"runner": '"$runner_azure"',   "image": "neondatabase/build-tools:pinned" },
-                      { "pg_version": 16, "region_id": "azure-eastus2",          "platform": "neonvm-azure-captest-new",      "db_size": "10gb","runner": '"$runner_azure"',   "image": "neondatabase/build-tools:pinned" },
-                      { "pg_version": 16, "region_id": "azure-eastus2",          "platform": "neonvm-azure-captest-new",      "db_size": "50gb","runner": '"$runner_azure"',   "image": "neondatabase/build-tools:pinned" },
-                      { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-sharding-reuse", "db_size": "50gb","runner": '"$runner_default"', "image": "'"$image_default"'" }]
+          "include": [{ "platform": "neon-captest-freetier",         "db_size": "3gb"  },
+                      { "platform": "neon-captest-new",              "db_size": "50gb" },
+                      { "platform": "neonvm-captest-freetier",       "db_size": "3gb"  },
+                      { "platform": "neonvm-captest-new",            "db_size": "50gb" },
+                      { "platform": "neonvm-captest-sharding-reuse", "db_size": "50gb" }]
         }'
 
-        if [ "$(date +%A)" = "Saturday" ] || [ ${RUN_AWS_RDS_AND_AURORA} = "true" ]; then
-          matrix=$(echo "$matrix" | jq '.include += [{ "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "rds-postgres", "db_size": "10gb","runner": '"$runner_default"', "image": "'"$image_default"'" },
-                                                     { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "rds-aurora", "db_size": "10gb","runner": '"$runner_default"', "image": "'"$image_default"'" }]')
+        if [ "$(date +%A)" = "Saturday" ]; then
+          matrix=$(echo "$matrix" | jq '.include += [{ "platform": "rds-postgres", "db_size": "10gb"},
+                                                     { "platform": "rds-aurora",   "db_size": "50gb"}]')
         fi
 
         echo "matrix=$(echo "$matrix" | jq --compact-output '.')" >> $GITHUB_OUTPUT
@@ -304,7 +166,7 @@ jobs:
       run: |
         matrix='{
           "platform": [
-            "neonvm-captest-reuse"
+            "neon-captest-reuse"
           ]
         }'
 
@@ -320,7 +182,7 @@ jobs:
       run: |
         matrix='{
           "platform": [
-            "neonvm-captest-reuse"
+            "neon-captest-reuse"
           ],
           "scale": [
             "10"
@@ -334,17 +196,8 @@ jobs:
 
         echo "matrix=$(echo "$matrix" | jq --compact-output '.')" >> $GITHUB_OUTPUT
 
-  prepare_AWS_RDS_databases:
-    uses: ./.github/workflows/_benchmarking_preparation.yml
-    secrets: inherit
-
   pgbench-compare:
-    if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
-    needs: [ generate-matrices, prepare_AWS_RDS_databases ]
-    permissions:
-      contents: write
-      statuses: write
-      id-token: write # Required for OIDC authentication in azure runners
+    needs: [ generate-matrices ]
 
     strategy:
       fail-fast: false
@@ -354,15 +207,15 @@ jobs:
       TEST_PG_BENCH_DURATIONS_MATRIX: "60m"
       TEST_PG_BENCH_SCALES_MATRIX: ${{ matrix.db_size }}
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: ${{ matrix.pg_version }}
+      DEFAULT_PG_VERSION: 14
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
       PLATFORM: ${{ matrix.platform }}
 
-    runs-on: ${{ matrix.runner }}
+    runs-on: [ self-hosted, us-east-2, x64 ]
     container:
-      image: ${{ matrix.image }}
+      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:pinned
       options: --init
 
     # Increase timeout to 8h, default timeout is 6h
@@ -371,41 +224,40 @@ jobs:
     steps:
     - uses: actions/checkout@v4
 
-    - name: Configure AWS credentials # necessary on Azure runners
-      uses: aws-actions/configure-aws-credentials@v4
-      with:
-        aws-region: eu-central-1
-        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
-        role-duration-seconds: 18000 # 5 hours
-
     - name: Download Neon artifact
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
+        name: neon-${{ runner.os }}-release-artifact
         path: /tmp/neon/
         prefix: latest
 
+    - name: Add Postgres binaries to PATH
+      run: |
+        ${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/pgbench --version
+        echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
+
     - name: Create Neon Project
-      if: contains(fromJson('["neonvm-captest-new", "neonvm-captest-freetier", "neonvm-azure-captest-freetier", "neonvm-azure-captest-new"]'), matrix.platform)
+      if: contains(fromJson('["neon-captest-new", "neon-captest-freetier", "neonvm-captest-new", "neonvm-captest-freetier"]'), matrix.platform)
       id: create-neon-project
       uses: ./.github/actions/neon-project-create
       with:
-        region_id: ${{ matrix.region_id }}
+        region_id: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
         postgres_version: ${{ env.DEFAULT_PG_VERSION }}
         api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-        compute_units: ${{ (contains(matrix.platform, 'captest-freetier') && '[0.25, 0.25]') || '[1, 1]' }}
+        compute_units: ${{ (matrix.platform == 'neon-captest-freetier' && '[0.25, 0.25]') || '[1, 1]' }}
+        provisioner: ${{ (contains(matrix.platform, 'neonvm-') && 'k8s-neonvm') || 'k8s-pod' }}
 
     - name: Set up Connection String
       id: set-up-connstr
       run: |
         case "${PLATFORM}" in
-          neonvm-captest-reuse)
+          neon-captest-reuse)
             CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR }}
             ;;
           neonvm-captest-sharding-reuse)
             CONNSTR=${{ secrets.BENCHMARK_CAPTEST_SHARDING_CONNSTR }}
             ;;
-          neonvm-captest-new | neonvm-captest-freetier | neonvm-azure-captest-new | neonvm-azure-captest-freetier)
+          neon-captest-new | neon-captest-freetier | neonvm-captest-new | neonvm-captest-freetier)
             CONNSTR=${{ steps.create-neon-project.outputs.dsn }}
             ;;
           rds-aurora)
@@ -422,6 +274,16 @@ jobs:
 
         echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
 
+        QUERIES=("SELECT version()")
+        if [[ "${PLATFORM}" = "neon"* ]]; then
+          QUERIES+=("SHOW neon.tenant_id")
+          QUERIES+=("SHOW neon.timeline_id")
+        fi
+
+        for q in "${QUERIES[@]}"; do
+          psql ${CONNSTR} -c "${q}"
+        done
+
     - name: Benchmark init
       uses: ./.github/actions/run-python-test-set
       with:
@@ -430,7 +292,6 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_init
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
@@ -444,7 +305,6 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_simple_update
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
@@ -458,7 +318,6 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_select_only
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
@@ -472,7 +331,6 @@ jobs:
         api_key: ${{ secrets.NEON_STAGING_API_KEY }}
 
     - name: Create Allure report
-      id: create-allure-report
       if: ${{ !cancelled() }}
       uses: ./.github/actions/allure-report-generate
 
@@ -481,133 +339,7 @@ jobs:
       uses: slackapi/slack-github-action@v1
       with:
         channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: |
-          Periodic perf testing on ${{ matrix.platform }}: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
-      env:
-        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
-
-  pgbench-pgvector:
-    permissions:
-      contents: write
-      statuses: write
-      id-token: write # Required for OIDC authentication in azure runners
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - PLATFORM: "neonvm-captest-pgvector"
-            RUNNER: [ self-hosted, us-east-2, x64 ]
-            IMAGE: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:pinned
-          - PLATFORM: "azure-captest-pgvector"
-            RUNNER: [ self-hosted, eastus2, x64 ]
-            IMAGE: neondatabase/build-tools:pinned
-
-    env:
-      TEST_PG_BENCH_DURATIONS_MATRIX: "15m"
-      TEST_PG_BENCH_SCALES_MATRIX: "1"
-      POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
-      TEST_OUTPUT: /tmp/test_output
-      BUILD_TYPE: remote
-      LD_LIBRARY_PATH: /home/nonroot/pg/usr/lib/x86_64-linux-gnu
-      SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
-      PLATFORM: ${{ matrix.PLATFORM }}
-
-    runs-on: ${{ matrix.RUNNER }}
-    container:
-      image: ${{ matrix.IMAGE }}
-      options: --init
-
-    steps:
-    - uses: actions/checkout@v4
-
-    # until https://github.com/neondatabase/neon/issues/8275 is fixed we temporarily install postgresql-16
-    # instead of using Neon artifacts containing pgbench
-    - name: Install postgresql-16 where pytest expects it
-      run: |
-        cd /home/nonroot
-        wget -q https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-16/libpq5_16.4-1.pgdg110%2B1_amd64.deb
-        wget -q https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-16/postgresql-client-16_16.4-1.pgdg110%2B1_amd64.deb
-        wget -q https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-16/postgresql-16_16.4-1.pgdg110%2B1_amd64.deb 
-        dpkg -x libpq5_16.4-1.pgdg110+1_amd64.deb pg
-        dpkg -x postgresql-client-16_16.4-1.pgdg110+1_amd64.deb pg
-        dpkg -x postgresql-16_16.4-1.pgdg110+1_amd64.deb pg
-        mkdir -p /tmp/neon/pg_install/v16/bin
-        ln -s /home/nonroot/pg/usr/lib/postgresql/16/bin/pgbench /tmp/neon/pg_install/v16/bin/pgbench  
-        ln -s /home/nonroot/pg/usr/lib/postgresql/16/bin/psql /tmp/neon/pg_install/v16/bin/psql  
-        ln -s /home/nonroot/pg/usr/lib/x86_64-linux-gnu /tmp/neon/pg_install/v16/lib 
-        /tmp/neon/pg_install/v16/bin/pgbench --version
-        /tmp/neon/pg_install/v16/bin/psql --version
-
-    - name: Set up Connection String
-      id: set-up-connstr
-      run: |
-        case "${PLATFORM}" in
-          neonvm-captest-pgvector)
-            CONNSTR=${{ secrets.BENCHMARK_PGVECTOR_CONNSTR }}
-            ;;
-          azure-captest-pgvector)
-            CONNSTR=${{ secrets.BENCHMARK_PGVECTOR_CONNSTR_AZURE }}
-            ;;
-          *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}"
-            exit 1
-            ;;
-        esac
-
-        echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
-
-    - name: Configure AWS credentials # necessary on Azure runners to read/write from/to S3
-      uses: aws-actions/configure-aws-credentials@v4
-      with:
-        aws-region: eu-central-1
-        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
-        role-duration-seconds: 18000 # 5 hours
-
-    - name: Benchmark pgvector hnsw indexing
-      uses: ./.github/actions/run-python-test-set
-      with:
-        build_type: ${{ env.BUILD_TYPE }}
-        test_selection: performance/test_perf_olap.py
-        run_in_parallel: false
-        save_perf_report: ${{ env.SAVE_PERF_REPORT }}
-        extra_params: -m remote_cluster --timeout 21600 -k test_pgvector_indexing
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
-      env:
-        VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
-        PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
-        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
-
-    - name: Benchmark pgvector queries
-      uses: ./.github/actions/run-python-test-set
-      with:
-        build_type: ${{ env.BUILD_TYPE }}
-        test_selection: performance/test_perf_pgvector_queries.py
-        run_in_parallel: false
-        save_perf_report: ${{ env.SAVE_PERF_REPORT }}
-        extra_params: -m remote_cluster --timeout 21600
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
-        VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
-        PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
-
-    - name: Create Allure report
-      id: create-allure-report
-      if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
-
-    - name: Post to a Slack channel
-      if: ${{ github.event.schedule && failure() }}
-      uses: slackapi/slack-github-action@v1
-      with:
-        channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: |
-          Periodic perf testing on ${{ env.PLATFORM }}: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
+        slack-message: "Periodic perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
       env:
         SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
 
@@ -619,8 +351,8 @@ jobs:
     #
     # *_CLICKBENCH_CONNSTR: Genuine ClickBench DB with ~100M rows
     # *_CLICKBENCH_10M_CONNSTR: DB with the first 10M rows of ClickBench DB
-    if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
-    needs: [ generate-matrices, pgbench-compare, prepare_AWS_RDS_databases ]
+    if: ${{ !cancelled() }}
+    needs: [ generate-matrices, pgbench-compare ]
 
     strategy:
       fail-fast: false
@@ -628,7 +360,7 @@ jobs:
 
     env:
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      DEFAULT_PG_VERSION: 14
       TEST_OUTPUT: /tmp/test_output
       TEST_OLAP_COLLECT_EXPLAIN: ${{ github.event.inputs.collect_olap_explain }}
       TEST_OLAP_COLLECT_PG_STAT_STATEMENTS: ${{ github.event.inputs.collect_pg_stat_statements }}
@@ -647,15 +379,20 @@ jobs:
     - name: Download Neon artifact
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
+        name: neon-${{ runner.os }}-release-artifact
         path: /tmp/neon/
         prefix: latest
 
+    - name: Add Postgres binaries to PATH
+      run: |
+        ${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/pgbench --version
+        echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
+
     - name: Set up Connection String
       id: set-up-connstr
       run: |
         case "${PLATFORM}" in
-          neonvm-captest-reuse)
+          neon-captest-reuse)
             CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CLICKBENCH_10M_CONNSTR }}
             ;;
           rds-aurora)
@@ -665,13 +402,23 @@ jobs:
             CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_CLICKBENCH_10M_CONNSTR }}
             ;;
           *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neonvm-captest-reuse', 'rds-aurora', or 'rds-postgres'"
+            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'rds-aurora', or 'rds-postgres'"
             exit 1
             ;;
         esac
 
         echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
 
+        QUERIES=("SELECT version()")
+        if [[ "${PLATFORM}" = "neon"* ]]; then
+          QUERIES+=("SHOW neon.tenant_id")
+          QUERIES+=("SHOW neon.timeline_id")
+        fi
+
+        for q in "${QUERIES[@]}"; do
+          psql ${CONNSTR} -c "${q}"
+        done
+
     - name: ClickBench benchmark
       uses: ./.github/actions/run-python-test-set
       with:
@@ -680,7 +427,6 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_clickbench
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
         PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
@@ -690,7 +436,6 @@ jobs:
         TEST_OLAP_SCALE: 10
 
     - name: Create Allure report
-      id: create-allure-report
       if: ${{ !cancelled() }}
       uses: ./.github/actions/allure-report-generate
 
@@ -699,10 +444,7 @@ jobs:
       uses: slackapi/slack-github-action@v1
       with:
         channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: |
-          Periodic OLAP perf testing on ${{ matrix.platform }}: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
+        slack-message: "Periodic OLAP perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
       env:
         SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
 
@@ -713,8 +455,8 @@ jobs:
     # We might change it after https://github.com/neondatabase/neon/issues/2900.
     #
     # *_TPCH_S10_CONNSTR: DB generated with scale factor 10 (~10 GB)
-    if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
-    needs: [ generate-matrices, clickbench-compare, prepare_AWS_RDS_databases ]
+    if: ${{ !cancelled() }}
+    needs: [ generate-matrices, clickbench-compare ]
 
     strategy:
       fail-fast: false
@@ -722,7 +464,7 @@ jobs:
 
     env:
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      DEFAULT_PG_VERSION: 14
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -740,24 +482,29 @@ jobs:
     - name: Download Neon artifact
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
+        name: neon-${{ runner.os }}-release-artifact
         path: /tmp/neon/
         prefix: latest
 
+    - name: Add Postgres binaries to PATH
+      run: |
+        ${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/pgbench --version
+        echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
+
     - name: Get Connstring Secret Name
       run: |
         case "${PLATFORM}" in
-          neonvm-captest-reuse)
+          neon-captest-reuse)
             ENV_PLATFORM=CAPTEST_TPCH
             ;;
           rds-aurora)
             ENV_PLATFORM=RDS_AURORA_TPCH
             ;;
           rds-postgres)
-            ENV_PLATFORM=RDS_POSTGRES_TPCH
+            ENV_PLATFORM=RDS_AURORA_TPCH
             ;;
           *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neonvm-captest-reuse', 'rds-aurora', or 'rds-postgres'"
+            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'rds-aurora', or 'rds-postgres'"
             exit 1
             ;;
         esac
@@ -772,6 +519,16 @@ jobs:
 
         echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
 
+        QUERIES=("SELECT version()")
+        if [[ "${PLATFORM}" = "neon"* ]]; then
+          QUERIES+=("SHOW neon.tenant_id")
+          QUERIES+=("SHOW neon.timeline_id")
+        fi
+
+        for q in "${QUERIES[@]}"; do
+          psql ${CONNSTR} -c "${q}"
+        done
+
     - name: Run TPC-H benchmark
       uses: ./.github/actions/run-python-test-set
       with:
@@ -780,7 +537,6 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_tpch
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
         PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
@@ -788,7 +544,6 @@ jobs:
         TEST_OLAP_SCALE: ${{ matrix.scale }}
 
     - name: Create Allure report
-      id: create-allure-report
       if: ${{ !cancelled() }}
       uses: ./.github/actions/allure-report-generate
 
@@ -797,16 +552,13 @@ jobs:
       uses: slackapi/slack-github-action@v1
       with:
         channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: |
-          Periodic TPC-H perf testing on ${{ matrix.platform }}: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
+        slack-message: "Periodic TPC-H perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
       env:
         SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
 
   user-examples-compare:
-    if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
-    needs: [ generate-matrices, tpch-compare, prepare_AWS_RDS_databases ]
+    if: ${{ !cancelled() }}
+    needs: [ generate-matrices, tpch-compare ]
 
     strategy:
       fail-fast: false
@@ -814,7 +566,7 @@ jobs:
 
     env:
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      DEFAULT_PG_VERSION: 14
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -831,15 +583,20 @@ jobs:
     - name: Download Neon artifact
       uses: ./.github/actions/download
       with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
+        name: neon-${{ runner.os }}-release-artifact
         path: /tmp/neon/
         prefix: latest
 
+    - name: Add Postgres binaries to PATH
+      run: |
+        ${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/pgbench --version
+        echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
+
     - name: Set up Connection String
       id: set-up-connstr
       run: |
         case "${PLATFORM}" in
-          neonvm-captest-reuse)
+          neon-captest-reuse)
             CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_CAPTEST_CONNSTR }}
             ;;
           rds-aurora)
@@ -849,13 +606,23 @@ jobs:
             CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_RDS_POSTGRES_CONNSTR }}
             ;;
           *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neonvm-captest-reuse', 'rds-aurora', or 'rds-postgres'"
+            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'rds-aurora', or 'rds-postgres'"
             exit 1
             ;;
         esac
 
         echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
 
+        QUERIES=("SELECT version()")
+        if [[ "${PLATFORM}" = "neon"* ]]; then
+          QUERIES+=("SHOW neon.tenant_id")
+          QUERIES+=("SHOW neon.timeline_id")
+        fi
+
+        for q in "${QUERIES[@]}"; do
+          psql ${CONNSTR} -c "${q}"
+        done
+
     - name: Run user examples
       uses: ./.github/actions/run-python-test-set
       with:
@@ -864,14 +631,12 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_user_examples
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
         PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
 
     - name: Create Allure report
-      id: create-allure-report
       if: ${{ !cancelled() }}
       uses: ./.github/actions/allure-report-generate
 
@@ -880,10 +645,6 @@ jobs:
       uses: slackapi/slack-github-action@v1
       with:
         channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: |
-          Periodic TPC-H perf testing on ${{ matrix.platform }}: ${{ job.status }}
-          <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-          <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
-
+        slack-message: "Periodic User example perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
       env:
         SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

.github/workflows/build-build-tools-image.yml

@@ -21,7 +21,6 @@ defaults:
 
 concurrency:
   group: build-build-tools-image-${{ inputs.image-tag }}
-  cancel-in-progress: false
 
 # No permission for GITHUB_TOKEN by default; the **minimal required** set of permissions should be granted in each job.
 permissions: {}
@@ -30,6 +29,7 @@ jobs:
   check-image:
     uses: ./.github/workflows/check-build-tools-image.yml
 
+  # This job uses older version of GitHub Actions because it's run on gen2 runners, which don't support node 20 (for newer versions)
   build-image:
     needs: [ check-image ]
     if: needs.check-image.outputs.found == 'false'
@@ -38,7 +38,7 @@ jobs:
       matrix:
         arch: [ x64, arm64 ]
 
-    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
+    runs-on: ${{ fromJson(format('["self-hosted", "dev", "{0}"]', matrix.arch)) }}
 
     env:
       IMAGE_TAG: ${{ inputs.image-tag }}
@@ -54,38 +54,40 @@ jobs:
             exit 1
           fi
 
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v3
 
-      - uses: ./.github/actions/set-docker-config-dir
-      - uses: docker/setup-buildx-action@v3
-        with:
-          cache-binary: false
+      # Use custom DOCKER_CONFIG directory to avoid conflicts with default settings
+      # The default value is ~/.docker
+      - name: Set custom docker config directory
+        run: |
+          mkdir -p /tmp/.docker-custom
+          echo DOCKER_CONFIG=/tmp/.docker-custom >> $GITHUB_ENV
 
-      - uses: docker/login-action@v3
+      - uses: docker/setup-buildx-action@v2
+
+      - uses: docker/login-action@v2
         with:
           username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
           password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
 
-      - uses: docker/login-action@v3
-        with:
-          registry: cache.neon.build
-          username: ${{ secrets.NEON_CI_DOCKERCACHE_USERNAME }}
-          password: ${{ secrets.NEON_CI_DOCKERCACHE_PASSWORD }}
-
-      - uses: docker/build-push-action@v6
+      - uses: docker/build-push-action@v4
         with:
           context: .
           provenance: false
           push: true
           pull: true
           file: Dockerfile.build-tools
-          cache-from: type=registry,ref=cache.neon.build/build-tools:cache-${{ matrix.arch }}
-          cache-to: ${{ github.ref_name == 'main' && format('type=registry,ref=cache.neon.build/build-tools:cache-{0},mode=max', matrix.arch) || '' }}
+          cache-from: type=registry,ref=neondatabase/build-tools:cache-${{ matrix.arch }}
+          cache-to: type=registry,ref=neondatabase/build-tools:cache-${{ matrix.arch }},mode=max
           tags: neondatabase/build-tools:${{ inputs.image-tag }}-${{ matrix.arch }}
 
+      - name: Remove custom docker config directory
+        run: |
+          rm -rf /tmp/.docker-custom
+
   merge-images:
     needs: [ build-image ]
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     env:
       IMAGE_TAG: ${{ inputs.image-tag }}

.github/workflows/check-build-tools-image.yml

@@ -19,23 +19,30 @@ permissions: {}
 
 jobs:
   check-image:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     outputs:
       tag: ${{ steps.get-build-tools-tag.outputs.image-tag }}
       found: ${{ steps.check-image.outputs.found }}
 
     steps:
-      - uses: actions/checkout@v4
-
       - name: Get build-tools image tag for the current commit
         id: get-build-tools-tag
         env:
-          IMAGE_TAG: |
-            ${{ hashFiles('Dockerfile.build-tools',
-                          '.github/workflows/check-build-tools-image.yml',
-                          '.github/workflows/build-build-tools-image.yml') }}
+          COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
-          echo "image-tag=${IMAGE_TAG}" | tee -a $GITHUB_OUTPUT
+          LAST_BUILD_TOOLS_SHA=$(
+            gh api \
+              -H "Accept: application/vnd.github+json" \
+              -H "X-GitHub-Api-Version: 2022-11-28" \
+              --method GET \
+              --field path=Dockerfile.build-tools \
+              --field sha=${COMMIT_SHA} \
+              --field per_page=1 \
+              --jq ".[0].sha" \
+              "/repos/${GITHUB_REPOSITORY}/commits"
+          )
+          echo "image-tag=${LAST_BUILD_TOOLS_SHA}" | tee -a $GITHUB_OUTPUT
 
       - name: Check if such tag found in the registry
         id: check-image

.github/workflows/check-permissions.yml

@@ -16,7 +16,7 @@ permissions: {}
 
 jobs:
   check-permissions:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     steps:
     - name: Disallow CI runs on PRs from forks
       if: |

.github/workflows/cleanup-caches-by-a-branch.yml

@@ -9,7 +9,7 @@ on:
 
 jobs:
   cleanup:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     steps:
       - name: Cleanup
         run: |

.github/workflows/label-for-external-users.yml

@@ -1,54 +0,0 @@
-name: Add `external` label to issues and PRs created by external users
-
-on:
-  issues:
-    types:
-      - opened
-  pull_request_target:
-    types:
-      - opened
-
-# No permission for GITHUB_TOKEN by default; the **minimal required** set of permissions should be granted in each job.
-permissions: {}
-
-env:
-  LABEL: external
-
-jobs:
-  check-user:
-    runs-on: ubuntu-22.04
-
-    outputs:
-      is-member: ${{ steps.check-user.outputs.is-member }}
-
-    steps:
-    - name: Check whether `${{ github.actor }}` is a member of `${{ github.repository_owner }}`
-      id: check-user
-      env:
-        GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
-      run: |
-        if gh api -H "Accept: application/vnd.github+json" -H "X-GitHub-Api-Version: 2022-11-28" "/orgs/${GITHUB_REPOSITORY_OWNER}/members/${GITHUB_ACTOR}"; then
-          is_member=true
-        else
-          is_member=false
-        fi
-
-        echo "is-member=${is_member}" | tee -a ${GITHUB_OUTPUT}
-
-  add-label:
-    if: needs.check-user.outputs.is-member == 'false'
-    needs: [ check-user ]
-
-    runs-on: ubuntu-22.04
-    permissions:
-      pull-requests: write # for `gh pr edit`
-      issues: write        # for `gh issue edit`
-
-    steps:
-    - name: Add `${{ env.LABEL }}` label
-      env:
-        GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        ITEM_NUMBER: ${{ github.event[github.event_name == 'pull_request_target' && 'pull_request' || 'issue'].number }}
-        GH_CLI_COMMAND: ${{ github.event_name == 'pull_request_target' && 'pr' || 'issue' }}
-      run: |
-        gh ${GH_CLI_COMMAND} --repo ${GITHUB_REPOSITORY} edit --add-label=${LABEL} ${ITEM_NUMBER}

.github/workflows/neon_extra_builds.yml

@@ -56,6 +56,7 @@ jobs:
         uses: actions/checkout@v4
         with:
           submodules: true
+          fetch-depth: 1
 
       - name: Install macOS postgres dependencies
         run: brew install flex bison openssl protobuf icu4c pkg-config
@@ -132,13 +133,212 @@ jobs:
       - name: Check that no warnings are produced
         run: ./run_clippy.sh
 
+  check-linux-arm-build:
+    needs: [ check-permissions, build-build-tools-image ]
+    timeout-minutes: 90
+    runs-on: [ self-hosted, dev, arm64 ]
+
+    env:
+      # Use release build only, to have less debug info around
+      # Hence keeping target/ (and general cache size) smaller
+      BUILD_TYPE: release
+      CARGO_FEATURES: --features testing
+      CARGO_FLAGS: --release
+      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
+      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
+
+    container:
+      image: ${{ needs.build-build-tools-image.outputs.image }}
+      credentials:
+        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
+        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
+      options: --init
+
+    steps:
+      - name: Fix git ownership
+        run: |
+          # Workaround for `fatal: detected dubious ownership in repository at ...`
+          #
+          # Use both ${{ github.workspace }} and ${GITHUB_WORKSPACE} because they're different on host and in containers
+          #   Ref https://github.com/actions/checkout/issues/785
+          #
+          git config --global --add safe.directory ${{ github.workspace }}
+          git config --global --add safe.directory ${GITHUB_WORKSPACE}
+          for r in 14 15 16; do
+            git config --global --add safe.directory "${{ github.workspace }}/vendor/postgres-v$r"
+            git config --global --add safe.directory "${GITHUB_WORKSPACE}/vendor/postgres-v$r"
+          done
+
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          submodules: true
+          fetch-depth: 1
+
+      - name: Set pg 14 revision for caching
+        id: pg_v14_rev
+        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v14) >> $GITHUB_OUTPUT
+
+      - name: Set pg 15 revision for caching
+        id: pg_v15_rev
+        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v15) >> $GITHUB_OUTPUT
+
+      - name: Set pg 16 revision for caching
+        id: pg_v16_rev
+        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v16) >> $GITHUB_OUTPUT
+
+      - name: Set env variables
+        run: |
+          echo "CARGO_HOME=${GITHUB_WORKSPACE}/.cargo" >> $GITHUB_ENV
+
+      - name: Cache postgres v14 build
+        id: cache_pg_14
+        uses: actions/cache@v4
+        with:
+          path: pg_install/v14
+          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v14_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
+
+      - name: Cache postgres v15 build
+        id: cache_pg_15
+        uses: actions/cache@v4
+        with:
+          path: pg_install/v15
+          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v15_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
+
+      - name: Cache postgres v16 build
+        id: cache_pg_16
+        uses: actions/cache@v4
+        with:
+          path: pg_install/v16
+          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v16_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
+
+      - name: Build postgres v14
+        if: steps.cache_pg_14.outputs.cache-hit != 'true'
+        run: mold -run make postgres-v14 -j$(nproc)
+
+      - name: Build postgres v15
+        if: steps.cache_pg_15.outputs.cache-hit != 'true'
+        run: mold -run make postgres-v15 -j$(nproc)
+
+      - name: Build postgres v16
+        if: steps.cache_pg_16.outputs.cache-hit != 'true'
+        run: mold -run make postgres-v16 -j$(nproc)
+
+      - name: Build neon extensions
+        run: mold -run make neon-pg-ext -j$(nproc)
+
+      - name: Build walproposer-lib
+        run: mold -run make walproposer-lib -j$(nproc)
+
+      - name: Run cargo build
+        run: |
+          mold -run cargo build --locked $CARGO_FLAGS $CARGO_FEATURES --bins --tests
+
+      - name: Run cargo test
+        env:
+          NEXTEST_RETRIES: 3
+        run: |
+          cargo nextest run $CARGO_FEATURES
+
+          # Run separate tests for real S3
+          export ENABLE_REAL_S3_REMOTE_STORAGE=nonempty
+          export REMOTE_STORAGE_S3_BUCKET=neon-github-ci-tests
+          export REMOTE_STORAGE_S3_REGION=eu-central-1
+          # Avoid `$CARGO_FEATURES` since there's no `testing` feature in the e2e tests now
+          cargo nextest run --package remote_storage --test test_real_s3
+
+          # Run separate tests for real Azure Blob Storage
+          # XXX: replace region with `eu-central-1`-like region
+          export ENABLE_REAL_AZURE_REMOTE_STORAGE=y
+          export AZURE_STORAGE_ACCOUNT="${{ secrets.AZURE_STORAGE_ACCOUNT_DEV }}"
+          export AZURE_STORAGE_ACCESS_KEY="${{ secrets.AZURE_STORAGE_ACCESS_KEY_DEV }}"
+          export REMOTE_STORAGE_AZURE_CONTAINER="${{ vars.REMOTE_STORAGE_AZURE_CONTAINER }}"
+          export REMOTE_STORAGE_AZURE_REGION="${{ vars.REMOTE_STORAGE_AZURE_REGION }}"
+          # Avoid `$CARGO_FEATURES` since there's no `testing` feature in the e2e tests now
+          cargo nextest run --package remote_storage --test test_real_azure
+
+  check-codestyle-rust-arm:
+    needs: [ check-permissions, build-build-tools-image ]
+    timeout-minutes: 90
+    runs-on: [ self-hosted, dev, arm64 ]
+
+    container:
+      image: ${{ needs.build-build-tools-image.outputs.image }}
+      credentials:
+        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
+        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
+      options: --init
+
+    steps:
+      - name: Fix git ownership
+        run: |
+          # Workaround for `fatal: detected dubious ownership in repository at ...`
+          #
+          # Use both ${{ github.workspace }} and ${GITHUB_WORKSPACE} because they're different on host and in containers
+          #   Ref https://github.com/actions/checkout/issues/785
+          #
+          git config --global --add safe.directory ${{ github.workspace }}
+          git config --global --add safe.directory ${GITHUB_WORKSPACE}
+          for r in 14 15 16; do
+            git config --global --add safe.directory "${{ github.workspace }}/vendor/postgres-v$r"
+            git config --global --add safe.directory "${GITHUB_WORKSPACE}/vendor/postgres-v$r"
+          done
+
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          submodules: true
+          fetch-depth: 1
+
+      # Some of our rust modules use FFI and need those to be checked
+      - name: Get postgres headers
+        run: make postgres-headers -j$(nproc)
+
+      # cargo hack runs the given cargo subcommand (clippy in this case) for all feature combinations.
+      # This will catch compiler & clippy warnings in all feature combinations.
+      # TODO: use cargo hack for build and test as well, but, that's quite expensive.
+      # NB: keep clippy args in sync with ./run_clippy.sh
+      - run: |
+          CLIPPY_COMMON_ARGS="$( source .neon_clippy_args; echo "$CLIPPY_COMMON_ARGS")"
+          if [ "$CLIPPY_COMMON_ARGS" = "" ]; then
+            echo "No clippy args found in .neon_clippy_args"
+            exit 1
+          fi
+          echo "CLIPPY_COMMON_ARGS=${CLIPPY_COMMON_ARGS}" >> $GITHUB_ENV
+      - name: Run cargo clippy (debug)
+        run: cargo hack --feature-powerset clippy $CLIPPY_COMMON_ARGS
+      - name: Run cargo clippy (release)
+        run: cargo hack --feature-powerset clippy --release $CLIPPY_COMMON_ARGS
+
+      - name: Check documentation generation
+        run: cargo doc --workspace --no-deps --document-private-items
+        env:
+            RUSTDOCFLAGS: "-Dwarnings -Arustdoc::private_intra_doc_links"
+
+      # Use `${{ !cancelled() }}` to run quck tests after the longer clippy run
+      - name: Check formatting
+        if: ${{ !cancelled() }}
+        run: cargo fmt --all -- --check
+
+      # https://github.com/facebookincubator/cargo-guppy/tree/bec4e0eb29dcd1faac70b1b5360267fc02bf830e/tools/cargo-hakari#2-keep-the-workspace-hack-up-to-date-in-ci
+      - name: Check rust dependencies
+        if: ${{ !cancelled() }}
+        run: |
+          cargo hakari generate --diff  # workspace-hack Cargo.toml is up-to-date
+          cargo hakari manage-deps --dry-run  # all workspace crates depend on workspace-hack
+
+      # https://github.com/EmbarkStudios/cargo-deny
+      - name: Check rust licenses/bans/advisories/sources
+        if: ${{ !cancelled() }}
+        run: cargo deny check
+
   gather-rust-build-stats:
     needs: [ check-permissions, build-build-tools-image ]
     if: |
       contains(github.event.pull_request.labels.*.name, 'run-extra-build-stats') ||
       contains(github.event.pull_request.labels.*.name, 'run-extra-build-*') ||
       github.ref_name == 'main'
-    runs-on: [ self-hosted, large ]
+    runs-on: [ self-hosted, gen3, large ]
     container:
       image: ${{ needs.build-build-tools-image.outputs.image }}
       credentials:
@@ -148,6 +348,8 @@ jobs:
 
     env:
       BUILD_TYPE: release
+      # remove the cachepot wrapper and build without crate caches
+      RUSTC_WRAPPER: ""
       # build with incremental compilation produce partial results
       # so do not attempt to cache this build, also disable the incremental compilation
       CARGO_INCREMENTAL: 0
@@ -157,6 +359,7 @@ jobs:
         uses: actions/checkout@v4
         with:
           submodules: true
+          fetch-depth: 1
 
       # Some of our rust modules use FFI and need those to be checked
       - name: Get postgres headers
@@ -166,7 +369,7 @@ jobs:
         run: make walproposer-lib -j$(nproc)
 
       - name: Produce the build stats
-        run: PQ_LIB_DIR=$(pwd)/pg_install/v16/lib cargo build --all --release --timings -j$(nproc)
+        run: cargo build --all --release --timings
 
       - name: Upload the build stats
         id: upload-stats

.github/workflows/periodic_pagebench.yml

@@ -1,155 +0,0 @@
-name: Periodic pagebench performance test on dedicated EC2 machine in eu-central-1 region
-
-on:
-  schedule:
-    # * is a special character in YAML so you have to quote this string
-    #          ┌───────────── minute (0 - 59)
-    #          │ ┌───────────── hour (0 - 23)
-    #          │ │ ┌───────────── day of the month (1 - 31)
-    #          │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
-    #          │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
-    - cron:  '0 18 * * *' # Runs at 6 PM UTC every day
-  workflow_dispatch: # Allows manual triggering of the workflow
-    inputs:
-      commit_hash:
-        type: string
-        description: 'The long neon repo commit hash for the system under test (pageserver) to be tested.'
-        required: false
-        default: ''
-
-defaults:
-  run:
-    shell: bash -euo pipefail {0}
-
-concurrency:
-  group: ${{ github.workflow }}
-  cancel-in-progress: false
-
-jobs:
-  trigger_bench_on_ec2_machine_in_eu_central_1:
-    runs-on: [ self-hosted, small ]
-    container:
-      image: neondatabase/build-tools:pinned
-      credentials:
-        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
-        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
-      options: --init
-    timeout-minutes: 360  # Set the timeout to 6 hours
-    env:
-      API_KEY: ${{ secrets.PERIODIC_PAGEBENCH_EC2_RUNNER_API_KEY }}
-      RUN_ID: ${{ github.run_id }}
-      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_EC2_US_TEST_RUNNER_ACCESS_KEY_ID }}
-      AWS_SECRET_ACCESS_KEY : ${{ secrets.AWS_EC2_US_TEST_RUNNER_ACCESS_KEY_SECRET }}
-      AWS_DEFAULT_REGION : "eu-central-1"
-      AWS_INSTANCE_ID : "i-02a59a3bf86bc7e74"
-    steps:
-    # we don't need the neon source code because we run everything remotely
-    # however we still need the local github actions to run the allure step below
-    - uses: actions/checkout@v4
-
-    - name: Show my own (github runner) external IP address - usefull for IP allowlisting
-      run: curl https://ifconfig.me
-
-    - name: Start EC2 instance and wait for the instance to boot up
-      run: |
-        aws ec2 start-instances --instance-ids $AWS_INSTANCE_ID
-        aws ec2 wait instance-running --instance-ids $AWS_INSTANCE_ID
-        sleep 60 # sleep some time to allow cloudinit and our API server to start up
-
-    - name: Determine public IP of the EC2 instance and set env variable EC2_MACHINE_URL_US
-      run: |
-        public_ip=$(aws ec2 describe-instances --instance-ids $AWS_INSTANCE_ID --query 'Reservations[*].Instances[*].PublicIpAddress' --output text)
-        echo "Public IP of the EC2 instance: $public_ip"
-        echo "EC2_MACHINE_URL_US=https://${public_ip}:8443" >> $GITHUB_ENV
-
-    - name: Determine commit hash
-      env:
-        INPUT_COMMIT_HASH: ${{ github.event.inputs.commit_hash }}
-      run: |
-        if [ -z "$INPUT_COMMIT_HASH" ]; then
-          echo "COMMIT_HASH=$(curl -s https://api.github.com/repos/neondatabase/neon/commits/main | jq -r '.sha')" >> $GITHUB_ENV
-        else
-          echo "COMMIT_HASH=$INPUT_COMMIT_HASH" >> $GITHUB_ENV
-        fi
-
-    - name: Start Bench with run_id   
-      run: |
-        curl -k -X 'POST' \
-        "${EC2_MACHINE_URL_US}/start_test/${GITHUB_RUN_ID}" \
-        -H 'accept: application/json' \
-        -H 'Content-Type: application/json' \
-        -H "Authorization: Bearer $API_KEY" \
-        -d "{\"neonRepoCommitHash\": \"${COMMIT_HASH}\"}"
-
-    - name: Poll Test Status
-      id: poll_step
-      run: |
-        status=""
-        while [[ "$status" != "failure" && "$status" != "success" ]]; do
-          response=$(curl -k -X 'GET' \
-          "${EC2_MACHINE_URL_US}/test_status/${GITHUB_RUN_ID}" \
-          -H 'accept: application/json' \
-          -H "Authorization: Bearer $API_KEY")
-          echo "Response: $response"
-          set +x
-          status=$(echo $response | jq -r '.status')
-          echo "Test status: $status"
-          if [[ "$status" == "failure" ]]; then
-            echo "Test failed"
-            exit 1 # Fail the job step if status is failure
-          elif [[ "$status" == "success" || "$status" == "null" ]]; then
-            break
-          elif [[ "$status" == "too_many_runs" ]]; then
-            echo "Too many runs already running"
-            echo "too_many_runs=true" >> "$GITHUB_OUTPUT"
-            exit 1
-          fi
-
-          sleep 60 # Poll every 60 seconds
-        done
-
-    - name: Retrieve Test Logs
-      if: always() && steps.poll_step.outputs.too_many_runs != 'true'
-      run: |
-        curl -k -X 'GET' \
-        "${EC2_MACHINE_URL_US}/test_log/${GITHUB_RUN_ID}" \
-        -H 'accept: application/gzip' \
-        -H "Authorization: Bearer $API_KEY" \
-        --output "test_log_${GITHUB_RUN_ID}.gz"
-    
-    - name: Unzip Test Log and Print it into this job's log
-      if: always() && steps.poll_step.outputs.too_many_runs != 'true'
-      run: |
-        gzip -d "test_log_${GITHUB_RUN_ID}.gz"
-        cat "test_log_${GITHUB_RUN_ID}"
-
-    - name: Create Allure report
-      env:
-        AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
-        AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
-      if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
-
-    - name: Post to a Slack channel
-      if: ${{ github.event.schedule && failure() }}
-      uses: slackapi/slack-github-action@v1
-      with:
-        channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: "Periodic pagebench testing on dedicated hardware: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
-      env:
-        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
-
-    - name: Cleanup Test Resources
-      if: always() 
-      run: |
-        curl -k -X 'POST' \
-        "${EC2_MACHINE_URL_US}/cleanup_test/${GITHUB_RUN_ID}" \
-        -H 'accept: application/json' \
-        -H "Authorization: Bearer $API_KEY" \
-        -d ''
-
-    - name: Stop EC2 instance and wait for the instance to be stopped
-      if: always() && steps.poll_step.outputs.too_many_runs != 'true'
-      run: |
-        aws ec2 stop-instances --instance-ids $AWS_INSTANCE_ID
-        aws ec2 wait instance-stopped --instance-ids $AWS_INSTANCE_ID

.github/workflows/pg-clients.yml

@@ -1,211 +0,0 @@
-name: Test Postgres client libraries
-
-on:
-  schedule:
-    # * is a special character in YAML so you have to quote this string
-    #          ┌───────────── minute (0 - 59)
-    #          │ ┌───────────── hour (0 - 23)
-    #          │ │ ┌───────────── day of the month (1 - 31)
-    #          │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
-    #          │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
-    - cron:  '23 02 * * *' # run once a day, timezone is utc
-  pull_request:
-    paths:
-      - '.github/workflows/pg-clients.yml'
-      - 'test_runner/pg_clients/**'
-      - 'test_runner/logical_repl/**'
-      - 'poetry.lock'
-  workflow_dispatch:
-
-concurrency:
-  group: ${{ github.workflow }}-${{ github.ref_name }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
-
-defaults:
-  run:
-    shell: bash -euxo pipefail {0}
-
-env:
-  DEFAULT_PG_VERSION: 16
-  PLATFORM: neon-captest-new
-  AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
-  AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
-  AWS_DEFAULT_REGION: eu-central-1
-
-jobs:
-  check-permissions:
-    if: ${{ !contains(github.event.pull_request.labels.*.name, 'run-no-ci') }}
-    uses: ./.github/workflows/check-permissions.yml
-    with:
-      github-event-name: ${{ github.event_name }}
-
-  check-build-tools-image:
-    needs: [ check-permissions ]
-    uses: ./.github/workflows/check-build-tools-image.yml
-
-  build-build-tools-image:
-    needs: [ check-build-tools-image ]
-    uses: ./.github/workflows/build-build-tools-image.yml
-    with:
-      image-tag: ${{ needs.check-build-tools-image.outputs.image-tag }}
-    secrets: inherit
-
-  test-logical-replication:
-    needs: [ build-build-tools-image ]
-    runs-on: ubuntu-22.04
-
-    container:
-      image: ${{ needs.build-build-tools-image.outputs.image }}
-      credentials:
-        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
-        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
-      options: --init --user root
-    services:
-      clickhouse:
-        image: clickhouse/clickhouse-server:24.6.3.64
-        ports:
-          - 9000:9000
-          - 8123:8123
-      zookeeper:
-        image: quay.io/debezium/zookeeper:2.7
-        ports:
-          - 2181:2181
-      kafka:
-        image: quay.io/debezium/kafka:2.7
-        env:
-          ZOOKEEPER_CONNECT: "zookeeper:2181"
-          KAFKA_ADVERTISED_LISTENERS: PLAINTEXT://kafka:9092
-          KAFKA_BROKER_ID: 1
-          KAFKA_OFFSETS_TOPIC_REPLICATION_FACTOR: 1
-          KAFKA_JMX_PORT: 9991
-        ports:
-          - 9092:9092
-      debezium:
-        image: quay.io/debezium/connect:2.7
-        env:
-          BOOTSTRAP_SERVERS: kafka:9092
-          GROUP_ID: 1
-          CONFIG_STORAGE_TOPIC: debezium-config
-          OFFSET_STORAGE_TOPIC: debezium-offset
-          STATUS_STORAGE_TOPIC: debezium-status
-          DEBEZIUM_CONFIG_CONNECTOR_CLASS: io.debezium.connector.postgresql.PostgresConnector
-        ports:
-          - 8083:8083
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: Download Neon artifact
-        uses: ./.github/actions/download
-        with:
-          name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
-          path: /tmp/neon/
-          prefix: latest
-
-      - name: Create Neon Project
-        id: create-neon-project
-        uses: ./.github/actions/neon-project-create
-        with:
-          api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-          postgres_version: ${{ env.DEFAULT_PG_VERSION }}
-
-      - name: Run tests
-        uses: ./.github/actions/run-python-test-set
-        with:
-          build_type: remote
-          test_selection: logical_repl
-          run_in_parallel: false
-          extra_params: -m remote_cluster
-          pg_version: ${{ env.DEFAULT_PG_VERSION }}
-        env:
-          BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
-
-      - name: Delete Neon Project
-        if: always()
-        uses: ./.github/actions/neon-project-delete
-        with:
-          project_id: ${{ steps.create-neon-project.outputs.project_id }}
-          api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-
-      - name: Create Allure report
-        if: ${{ !cancelled() }}
-        id: create-allure-report
-        uses: ./.github/actions/allure-report-generate
-        with:
-          store-test-results-into-db: true
-        env:
-          REGRESS_TEST_RESULT_CONNSTR_NEW: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}
-
-      - name: Post to a Slack channel
-        if: github.event.schedule && failure()
-        uses: slackapi/slack-github-action@v1
-        with:
-          channel-id: "C06KHQVQ7U3" # on-call-qa-staging-stream
-          slack-message: |
-            Testing the logical replication: <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|${{ job.status }}> (<${{ steps.create-allure-report.outputs.report-url }}|test report>)
-        env:
-          SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
-
-  test-postgres-client-libs:
-    needs: [ build-build-tools-image ]
-    runs-on: ubuntu-22.04
-
-    container:
-      image: ${{ needs.build-build-tools-image.outputs.image }}
-      credentials:
-        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
-        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
-      options: --init --user root
-
-    steps:
-    - uses: actions/checkout@v4
-
-    - name: Download Neon artifact
-      uses: ./.github/actions/download
-      with:
-        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
-        path: /tmp/neon/
-        prefix: latest
-
-    - name: Create Neon Project
-      id: create-neon-project
-      uses: ./.github/actions/neon-project-create
-      with:
-        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
-
-    - name: Run tests
-      uses: ./.github/actions/run-python-test-set
-      with:
-        build_type: remote
-        test_selection: pg_clients
-        run_in_parallel: false
-        extra_params: -m remote_cluster
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
-      env:
-        BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
-
-    - name: Delete Neon Project
-      if: always()
-      uses: ./.github/actions/neon-project-delete
-      with:
-        project_id: ${{ steps.create-neon-project.outputs.project_id }}
-        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
-
-    - name: Create Allure report
-      if: ${{ !cancelled() }}
-      id: create-allure-report
-      uses: ./.github/actions/allure-report-generate
-      with:
-        store-test-results-into-db: true
-      env:
-        REGRESS_TEST_RESULT_CONNSTR_NEW: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}
-
-    - name: Post to a Slack channel
-      if: github.event.schedule && failure()
-      uses: slackapi/slack-github-action@v1
-      with:
-        channel-id: "C06KHQVQ7U3" # on-call-qa-staging-stream
-        slack-message: |
-          Testing Postgres clients: <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|${{ job.status }}> (<${{ steps.create-allure-report.outputs.report-url }}|test report>)
-      env:
-        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

.github/workflows/pg_clients.yml

@@ -0,0 +1,98 @@
+name: Test Postgres client libraries
+
+on:
+  schedule:
+    # * is a special character in YAML so you have to quote this string
+    #          ┌───────────── minute (0 - 59)
+    #          │ ┌───────────── hour (0 - 23)
+    #          │ │ ┌───────────── day of the month (1 - 31)
+    #          │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
+    #          │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
+    - cron:  '23 02 * * *' # run once a day, timezone is utc
+
+  workflow_dispatch:
+
+concurrency:
+  # Allow only one workflow per any non-`main` branch.
+  group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
+  cancel-in-progress: true
+
+jobs:
+  test-postgres-client-libs:
+    # TODO: switch to gen2 runner, requires docker
+    runs-on: [ ubuntu-latest ]
+
+    env:
+      DEFAULT_PG_VERSION: 14
+      TEST_OUTPUT: /tmp/test_output
+
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v4
+
+    - uses: actions/setup-python@v4
+      with:
+        python-version: 3.9
+
+    - name: Install Poetry
+      uses: snok/install-poetry@v1
+
+    - name: Cache poetry deps
+      uses: actions/cache@v4
+      with:
+        path: ~/.cache/pypoetry/virtualenvs
+        key: v2-${{ runner.os }}-python-deps-ubunutu-latest-${{ hashFiles('poetry.lock') }}
+
+    - name: Install Python deps
+      shell: bash -euxo pipefail {0}
+      run: ./scripts/pysync
+
+    - name: Create Neon Project
+      id: create-neon-project
+      uses: ./.github/actions/neon-project-create
+      with:
+        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
+        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
+
+    - name: Run pytest
+      env:
+        REMOTE_ENV: 1
+        BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
+        POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
+      shell: bash -euxo pipefail {0}
+      run: |
+        # Test framework expects we have psql binary;
+        # but since we don't really need it in this test, let's mock it
+        mkdir -p "$POSTGRES_DISTRIB_DIR/v${DEFAULT_PG_VERSION}/bin" && touch "$POSTGRES_DISTRIB_DIR/v${DEFAULT_PG_VERSION}/bin/psql";
+        ./scripts/pytest \
+          --junitxml=$TEST_OUTPUT/junit.xml \
+          --tb=short \
+          --verbose \
+          -m "remote_cluster" \
+          -rA "test_runner/pg_clients"
+
+    - name: Delete Neon Project
+      if: ${{ always() }}
+      uses: ./.github/actions/neon-project-delete
+      with:
+        project_id: ${{ steps.create-neon-project.outputs.project_id }}
+        api_key: ${{ secrets.NEON_STAGING_API_KEY }}
+
+    # We use GitHub's action upload-artifact because `ubuntu-latest` doesn't have configured AWS CLI.
+    # It will be fixed after switching to gen2 runner
+    - name: Upload python test logs
+      if: always()
+      uses: actions/upload-artifact@v4
+      with:
+        retention-days: 7
+        name: python-test-pg_clients-${{ runner.os }}-stage-logs
+        path: ${{ env.TEST_OUTPUT }}
+
+    - name: Post to a Slack channel
+      if: ${{ github.event.schedule && failure() }}
+      uses: slackapi/slack-github-action@v1
+      with:
+        channel-id: "C033QLM5P7D" # dev-staging-stream
+        slack-message: "Testing Postgres clients: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
+      env:
+        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

.github/workflows/pin-build-tools-image.yml

@@ -7,20 +7,12 @@ on:
         description: 'Source tag'
         required: true
         type: string
-      force:
-        description: 'Force the image to be pinned'
-        default: false
-        type: boolean
   workflow_call:
     inputs:
       from-tag:
         description: 'Source tag'
         required: true
         type: string
-      force:
-        description: 'Force the image to be pinned'
-        default: false
-        type: boolean
 
 defaults:
   run:
@@ -28,20 +20,16 @@ defaults:
 
 concurrency:
   group: pin-build-tools-image-${{ inputs.from-tag }}
-  cancel-in-progress: false
 
-# No permission for GITHUB_TOKEN by default; the **minimal required** set of permissions should be granted in each job.
 permissions: {}
 
-env:
-  FROM_TAG: ${{ inputs.from-tag }}
-  TO_TAG: pinned
-
 jobs:
-  check-manifests:
-    runs-on: ubuntu-22.04
-    outputs:
-      skip: ${{ steps.check-manifests.outputs.skip }}
+  tag-image:
+    runs-on: ubuntu-latest
+
+    env:
+      FROM_TAG: ${{ inputs.from-tag }}
+      TO_TAG: pinned
 
     steps:
       - name: Check if we really need to pin the image
@@ -58,44 +46,27 @@ jobs:
 
           echo "skip=${skip}" | tee -a $GITHUB_OUTPUT
 
-  tag-image:
-    needs: check-manifests
-
-    # use format(..) to catch both inputs.force = true AND inputs.force = 'true'
-    if: needs.check-manifests.outputs.skip == 'false' || format('{0}', inputs.force) == 'true'
-
-    runs-on: ubuntu-22.04
-
-    permissions:
-      id-token: write # for `azure/login`
-
-    steps:
       - uses: docker/login-action@v3
-
+        if: steps.check-manifests.outputs.skip == 'false'
         with:
           username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
           password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
 
+      - name: Tag build-tools with `${{ env.TO_TAG }}` in Docker Hub
+        if: steps.check-manifests.outputs.skip == 'false'
+        run: |
+          docker buildx imagetools create -t neondatabase/build-tools:${TO_TAG} \
+                                             neondatabase/build-tools:${FROM_TAG}
+
       - uses: docker/login-action@v3
+        if: steps.check-manifests.outputs.skip == 'false'
         with:
           registry: 369495373322.dkr.ecr.eu-central-1.amazonaws.com
           username: ${{ secrets.AWS_ACCESS_KEY_DEV }}
           password: ${{ secrets.AWS_SECRET_KEY_DEV }}
 
-      - name: Azure login
-        uses: azure/login@6c251865b4e6290e7b78be643ea2d005bc51f69a  # @v2.1.1
-        with:
-          client-id: ${{ secrets.AZURE_DEV_CLIENT_ID }}
-          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
-          subscription-id: ${{ secrets.AZURE_DEV_SUBSCRIPTION_ID }}
-
-      - name: Login to ACR
-        run: |
-          az acr login --name=neoneastus2
-
-      - name: Tag build-tools with `${{ env.TO_TAG }}` in Docker Hub, ECR, and ACR
+      - name: Tag build-tools with `${{ env.TO_TAG }}` in ECR
+        if: steps.check-manifests.outputs.skip == 'false'
         run: |
           docker buildx imagetools create -t 369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:${TO_TAG} \
-                                          -t neoneastus2.azurecr.io/neondatabase/build-tools:${TO_TAG} \
-                                          -t neondatabase/build-tools:${TO_TAG} \
                                              neondatabase/build-tools:${FROM_TAG}

.github/workflows/release-notify.yml

@@ -19,7 +19,7 @@ on:
 
 jobs:
   notify:
-    runs-on: ubuntu-22.04
+    runs-on: [ ubuntu-latest ]
 
     steps:
       - uses: neondatabase/dev-actions/release-pr-notify@main

.github/workflows/release.yml

@@ -26,7 +26,7 @@ defaults:
 jobs:
   create-storage-release-branch:
     if: ${{ github.event.schedule == '0 6 * * MON' || format('{0}', inputs.create-storage-release-branch) == 'true' }}
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     permissions:
       contents: write # for `git push`
@@ -52,22 +52,20 @@ jobs:
       env:
         GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
       run: |
-        TITLE="Storage & Compute release ${RELEASE_DATE}"
-
         cat << EOF > body.md
-          ## ${TITLE}
+          ## Release ${RELEASE_DATE}
 
           **Please merge this Pull Request using 'Create a merge commit' button**
         EOF
 
-        gh pr create --title "${TITLE}" \
+        gh pr create --title "Release ${RELEASE_DATE}" \
                      --body-file "body.md" \
                      --head "${RELEASE_BRANCH}" \
                      --base "release"
 
   create-proxy-release-branch:
     if: ${{ github.event.schedule == '0 6 * * THU' || format('{0}', inputs.create-proxy-release-branch) == 'true' }}
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     permissions:
       contents: write # for `git push`
@@ -93,15 +91,13 @@ jobs:
       env:
         GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
       run: |
-        TITLE="Proxy release ${RELEASE_DATE}"
-
         cat << EOF > body.md
-          ## ${TITLE}
+          ## Proxy release ${RELEASE_DATE}
 
           **Please merge this Pull Request using 'Create a merge commit' button**
         EOF
 
-        gh pr create --title "${TITLE}" \
+        gh pr create --title "Proxy release ${RELEASE_DATE}" \
                      --body-file "body.md" \
                      --head "${RELEASE_BRANCH}" \
                      --base "release-proxy"

.github/workflows/trigger-e2e-tests.yml

@@ -13,11 +13,13 @@ defaults:
 env:
   # A concurrency group that we use for e2e-tests runs, matches `concurrency.group` above with `github.repository` as a prefix
   E2E_CONCURRENCY_GROUP: ${{ github.repository }}-e2e-tests-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
+  AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
+  AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
 
 jobs:
   cancel-previous-e2e-tests:
     if: github.event_name == 'pull_request'
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
 
     steps:
       - name: Cancel previous e2e-tests runs for this PR
@@ -29,7 +31,7 @@ jobs:
               --field concurrency_group="${{ env.E2E_CONCURRENCY_GROUP }}"
 
   tag:
-    runs-on: ubuntu-22.04
+    runs-on: [ ubuntu-latest ]
     outputs:
       build-tag: ${{ steps.build-tag.outputs.tag }}
 
@@ -60,37 +62,21 @@ jobs:
 
   trigger-e2e-tests:
     needs: [ tag ]
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     env:
-      EVENT_ACTION: ${{ github.event.action }}
-      GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
       TAG: ${{ needs.tag.outputs.build-tag }}
     steps:
-      - name: Wait for `promote-images` job to finish
-        # It's important to have a timeout here, the script in the step can run infinitely
-        timeout-minutes: 60
+      - name: check if ecr image are present
+        env:
+          AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
+          AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
         run: |
-          if [ "${GITHUB_EVENT_NAME}" != "pull_request" ] || [ "${EVENT_ACTION}" != "ready_for_review" ]; then
-            exit 0
-          fi
-
-          # For PRs we use the run id as the tag
-          BUILD_AND_TEST_RUN_ID=${TAG}
-          while true; do
-            conclusion=$(gh run --repo ${GITHUB_REPOSITORY} view ${BUILD_AND_TEST_RUN_ID} --json jobs --jq '.jobs[] | select(.name == "promote-images") | .conclusion')
-            case "$conclusion" in
-              success)
-                break
-                ;;
-              failure | cancelled | skipped)
-                echo "The 'promote-images' job didn't succeed: '${conclusion}'. Exiting..."
-                exit 1
-                ;;
-              *)
-                echo "The 'promote-images' hasn't succeed yet. Waiting..."
-                sleep 60
-                ;;
-            esac
+          for REPO in neon compute-tools compute-node-v14 vm-compute-node-v14 compute-node-v15 vm-compute-node-v15 compute-node-v16 vm-compute-node-v16; do
+            OUTPUT=$(aws ecr describe-images --repository-name ${REPO} --region eu-central-1 --query "imageDetails[?imageTags[?contains(@, '${TAG}')]]" --output text)
+            if [ "$OUTPUT" == "" ]; then
+              echo "$REPO with image tag $TAG not found" >> $GITHUB_OUTPUT
+              exit 1
+            fi
           done
 
       - name: Set e2e-platforms

.neon_clippy_args

@@ -1,5 +1,4 @@
 # * `-A unknown_lints` – do not warn about unknown lint suppressions
 #                        that people with newer toolchains might use
 # * `-D warnings`      - fail on any warnings (`cargo` returns non-zero exit status)
-# * `-D clippy::todo`  - don't let `todo!()` slip into `main`
-export CLIPPY_COMMON_ARGS="--locked --workspace --all-targets -- -A unknown_lints -D warnings -D clippy::todo"
+export CLIPPY_COMMON_ARGS="--locked --workspace --all-targets -- -A unknown_lints -D warnings"

CODEOWNERS

@@ -1,13 +1,13 @@
 /compute_tools/ @neondatabase/control-plane @neondatabase/compute
 /storage_controller @neondatabase/storage
 /libs/pageserver_api/ @neondatabase/storage
-/libs/postgres_ffi/ @neondatabase/compute @neondatabase/storage
+/libs/postgres_ffi/ @neondatabase/compute @neondatabase/safekeepers
 /libs/remote_storage/ @neondatabase/storage
-/libs/safekeeper_api/ @neondatabase/storage
+/libs/safekeeper_api/ @neondatabase/safekeepers
 /libs/vm_monitor/ @neondatabase/autoscaling
 /pageserver/ @neondatabase/storage
 /pgxn/ @neondatabase/compute
-/pgxn/neon/ @neondatabase/compute @neondatabase/storage
+/pgxn/neon/ @neondatabase/compute @neondatabase/safekeepers
 /proxy/ @neondatabase/proxy
-/safekeeper/ @neondatabase/storage
+/safekeeper/ @neondatabase/safekeepers
 /vendor/ @neondatabase/compute

Cargo.toml

@@ -13,9 +13,9 @@ members = [
     "safekeeper",
     "storage_broker",
     "storage_controller",
-    "storage_controller/client",
-    "storage_scrubber",
+    "s3_scrubber",
     "workspace_hack",
+    "trace",
     "libs/compute_api",
     "libs/pageserver_api",
     "libs/postgres_ffi",
@@ -41,31 +41,28 @@ license = "Apache-2.0"
 
 ## All dependency versions, used in the project
 [workspace.dependencies]
-ahash = "0.8"
 anyhow = { version = "1.0", features = ["backtrace"] }
 arc-swap = "1.6"
 async-compression = { version = "0.4.0", features = ["tokio", "gzip", "zstd"] }
-atomic-take = "1.1.0"
-azure_core = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls", "hmac_rust"] }
-azure_identity = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls"] }
-azure_storage = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls"] }
-azure_storage_blobs = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls"] }
+azure_core = "0.18"
+azure_identity = "0.18"
+azure_storage = "0.18"
+azure_storage_blobs = "0.18"
 flate2 = "1.0.26"
 async-stream = "0.3"
 async-trait = "0.1"
-aws-config = { version = "1.3", default-features = false, features=["rustls"] }
-aws-sdk-s3 = "1.26"
+aws-config = { version = "1.1.4", default-features = false, features=["rustls"] }
+aws-sdk-s3 = "1.14"
 aws-sdk-iam = "1.15.0"
-aws-smithy-async = { version = "1.2.1", default-features = false, features=["rt-tokio"] }
-aws-smithy-types = "1.1.9"
-aws-credential-types = "1.2.0"
-aws-sigv4 = { version = "1.2.1", features = ["sign-http"] }
-aws-types = "1.2.0"
+aws-smithy-async = { version = "1.1.4", default-features = false, features=["rt-tokio"] }
+aws-smithy-types = "1.1.4"
+aws-credential-types = "1.1.4"
+aws-sigv4 = { version = "1.2.0", features = ["sign-http"] }
+aws-types = "1.1.7"
 axum = { version = "0.6.20", features = ["ws"] }
 base64 = "0.13.0"
 bincode = "1.3"
 bindgen = "0.65"
-bit_field = "0.10.2"
 bstr = "1.0"
 byteorder = "1.4"
 bytes = "1.0"
@@ -76,7 +73,6 @@ clap = { version = "4.0", features = ["derive"] }
 comfy-table = "6.1"
 const_format = "0.2"
 crc32c = "0.6"
-crossbeam-deque = "0.8.5"
 crossbeam-utils = "0.8.5"
 dashmap = { version = "5.5.0", features = ["raw-api"] }
 either = "1.8"
@@ -84,13 +80,13 @@ enum-map = "2.4.2"
 enumset = "1.0.12"
 fail = "0.5.0"
 fallible-iterator = "0.2"
-framed-websockets = { version = "0.1.0", git = "https://github.com/neondatabase/framed-websockets" }
+fs2 = "0.4.3"
 futures = "0.3"
 futures-core = "0.3"
 futures-util = "0.3"
 git-version = "0.3"
-hashbrown = "0.14"
-hashlink = "0.9.1"
+hashbrown = "0.13"
+hashlink = "0.8.4"
 hdrhistogram = "7.5.2"
 hex = "0.4"
 hex-literal = "0.4"
@@ -101,41 +97,41 @@ http-types = { version = "2", default-features = false }
 humantime = "2.1"
 humantime-serde = "1.1.1"
 hyper = "0.14"
-tokio-tungstenite = "0.20.0"
-indexmap = "2"
+hyper-tungstenite = "0.11"
 inotify = "0.10.2"
 ipnet = "2.9.0"
 itertools = "0.10"
 jsonwebtoken = "9"
 lasso = "0.7"
+leaky-bucket = "1.0.1"
 libc = "0.2"
 md5 = "0.7.0"
-measured = { version = "0.0.22", features=["lasso"] }
-measured-process = { version = "0.0.22" }
+measured = { version = "0.0.13", features=["default", "lasso"] }
 memoffset = "0.8"
-nix = { version = "0.27", features = ["dir", "fs", "process", "socket", "signal", "poll"] }
+native-tls = "0.2"
+nix = { version = "0.27", features = ["fs", "process", "socket", "signal", "poll"] }
 notify = "6.0.0"
 num_cpus = "1.15"
 num-traits = "0.2.15"
 once_cell = "1.13"
 opentelemetry = "0.20.0"
-opentelemetry-otlp = { version = "0.13.0", default-features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
+opentelemetry-otlp = { version = "0.13.0", default_features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
 opentelemetry-semantic-conventions = "0.12.0"
 parking_lot = "0.12"
-parquet = { version = "51.0.0", default-features = false, features = ["zstd"] }
-parquet_derive = "51.0.0"
+parquet = { version = "49.0.0", default-features = false, features = ["zstd"] }
+parquet_derive = "49.0.0"
 pbkdf2 = { version = "0.12.1", features = ["simple", "std"] }
 pin-project-lite = "0.2"
-procfs = "0.16"
-prometheus = {version = "0.13", default-features=false, features = ["process"]} # removes protobuf dependency
+procfs = "0.14"
+prometheus = {version = "0.13", default_features=false, features = ["process"]} # removes protobuf dependency
 prost = "0.11"
 rand = "0.8"
 redis = { version = "0.25.2", features = ["tokio-rustls-comp", "keep-alive"] }
 regex = "1.10.2"
-reqwest = { version = "0.12", default-features = false, features = ["rustls-tls"] }
-reqwest-tracing = { version = "0.5", features = ["opentelemetry_0_20"] }
-reqwest-middleware = "0.3.0"
-reqwest-retry = "0.5"
+reqwest = { version = "0.11", default-features = false, features = ["rustls-tls"] }
+reqwest-tracing = { version = "0.4.7", features = ["opentelemetry_0_20"] }
+reqwest-middleware = "0.2.0"
+reqwest-retry = "0.2.2"
 routerify = "3"
 rpds = "0.13"
 rustc-hash = "1.1.0"
@@ -145,8 +141,7 @@ rustls-split = "0.3"
 scopeguard = "1.1"
 sysinfo = "0.29.2"
 sd-notify = "0.4.1"
-send-future = "0.1.0"
-sentry = { version = "0.32", default-features = false, features = ["backtrace", "contexts", "panic", "rustls", "reqwest" ] }
+sentry = { version = "0.31", default-features = false, features = ["backtrace", "contexts", "panic", "rustls", "reqwest" ] }
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1"
 serde_path_to_error = "0.1"
@@ -160,8 +155,7 @@ socket2 = "0.5"
 strum = "0.24"
 strum_macros = "0.24"
 "subtle"  = "2.5.0"
-# Our PR https://github.com/nical/rust_debug/pull/4 has been merged but no new version released yet
-svg_fmt = { git = "https://github.com/nical/rust_debug", rev = "28a7d96eecff2f28e75b1ea09f2d499a60d0e3b4" }
+svg_fmt = "0.4.1"
 sync_wrapper = "0.1.2"
 tar = "0.4"
 task-local-extensions = "0.1.4"
@@ -180,21 +174,17 @@ tokio-util = { version = "0.7.10", features = ["io", "rt"] }
 toml = "0.7"
 toml_edit = "0.19"
 tonic = {version = "0.9", features = ["tls", "tls-roots"]}
-tower-service = "0.3.2"
 tracing = "0.1"
 tracing-error = "0.2.0"
-tracing-opentelemetry = "0.21.0"
-tracing-subscriber = { version = "0.3", default-features = false, features = ["smallvec", "fmt", "tracing-log", "std", "env-filter", "json"] }
-try-lock = "0.2.5"
+tracing-opentelemetry = "0.20.0"
+tracing-subscriber = { version = "0.3", default_features = false, features = ["smallvec", "fmt", "tracing-log", "std", "env-filter", "json"] }
 twox-hash = { version = "1.6.3", default-features = false }
-typed-json = "0.1"
 url = "2.2"
 urlencoding = "2.1"
 uuid = { version = "1.6.1", features = ["v4", "v7", "serde"] }
 walkdir = "2.3.2"
-rustls-native-certs = "0.7"
+webpki-roots = "0.25"
 x509-parser = "0.15"
-whoami = "1.5.1"
 
 ## TODO replace this with tracing
 env_logger = "0.10"
@@ -202,10 +192,14 @@ log = "0.4"
 
 ## Libraries from neondatabase/ git forks, ideally with changes to be upstreamed
 postgres = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
+postgres-native-tls = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
 postgres-protocol = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
 postgres-types = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
 tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
 
+## Other git libraries
+heapless = { default-features=false, features=[], git = "https://github.com/japaric/heapless.git", rev = "644653bf3b831c6bb4963be2de24804acf5e5001" } # upstream release pending
+
 ## Local libraries
 compute_api = { version = "0.1", path = "./libs/compute_api/" }
 consumption_metrics = { version = "0.1", path = "./libs/consumption_metrics/" }
@@ -221,7 +215,6 @@ remote_storage = { version = "0.1", path = "./libs/remote_storage/" }
 safekeeper_api = { version = "0.1", path = "./libs/safekeeper_api" }
 desim = { version = "0.1", path = "./libs/desim" }
 storage_broker = { version = "0.1", path = "./storage_broker/" } # Note: main broker code is inside the binary crate, so linking with the library shouldn't be heavy.
-storage_controller_client = { path = "./storage_controller/client" }
 tenant_size_model = { version = "0.1", path = "./libs/tenant_size_model/" }
 tracing-utils = { version = "0.1", path = "./libs/tracing-utils/" }
 utils = { version = "0.1", path = "./libs/utils/" }
@@ -240,12 +233,13 @@ tonic-build = "0.9"
 
 [patch.crates-io]
 
-# Needed to get `tokio-postgres-rustls` to depend on our fork.
+# This is only needed for proxy's tests.
+# TODO: we should probably fork `tokio-postgres-rustls` instead.
 tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
 
 # bug fixes for UUID
-parquet = { git = "https://github.com/apache/arrow-rs", branch = "master" }
-parquet_derive = { git = "https://github.com/apache/arrow-rs", branch = "master" }
+parquet = { git = "https://github.com/neondatabase/arrow-rs", branch = "neon-fix-bugs" }
+parquet_derive = { git = "https://github.com/neondatabase/arrow-rs", branch = "neon-fix-bugs" }
 
 ################# Binary contents sections
 

Dockerfile

@@ -17,7 +17,7 @@ COPY --chown=nonroot pgxn pgxn
 COPY --chown=nonroot Makefile Makefile
 COPY --chown=nonroot scripts/ninstall.sh scripts/ninstall.sh
 
-ENV BUILD_TYPE=release
+ENV BUILD_TYPE release
 RUN set -e \
     && mold -run make -j $(nproc) -s neon-pg-ext \
     && rm -rf pg_install/build \
@@ -29,15 +29,25 @@ WORKDIR /home/nonroot
 ARG GIT_VERSION=local
 ARG BUILD_TAG
 
+# Enable https://github.com/paritytech/cachepot to cache Rust crates' compilation results in Docker builds.
+# Set up cachepot to use an AWS S3 bucket for cache results, to reuse it between `docker build` invocations.
+# cachepot falls back to local filesystem if S3 is misconfigured, not failing the build
+ARG RUSTC_WRAPPER=cachepot
+ENV AWS_REGION=eu-central-1
+ENV CACHEPOT_S3_KEY_PREFIX=cachepot
+ARG CACHEPOT_BUCKET=neon-github-dev
+#ARG AWS_ACCESS_KEY_ID
+#ARG AWS_SECRET_ACCESS_KEY
+
 COPY --from=pg-build /home/nonroot/pg_install/v14/include/postgresql/server pg_install/v14/include/postgresql/server
 COPY --from=pg-build /home/nonroot/pg_install/v15/include/postgresql/server pg_install/v15/include/postgresql/server
 COPY --from=pg-build /home/nonroot/pg_install/v16/include/postgresql/server pg_install/v16/include/postgresql/server
-COPY --from=pg-build /home/nonroot/pg_install/v16/lib                       pg_install/v16/lib
 COPY --chown=nonroot . .
 
-ARG ADDITIONAL_RUSTFLAGS
+# Show build caching stats to check if it was used in the end.
+# Has to be the part of the same RUN since cachepot daemon is killed in the end of this RUN, losing the compilation stats.
 RUN set -e \
-    && PQ_LIB_DIR=$(pwd)/pg_install/v16/lib RUSTFLAGS="-Clinker=clang -Clink-arg=-fuse-ld=mold -Clink-arg=-Wl,--no-rosegment ${ADDITIONAL_RUSTFLAGS}" cargo build \
+    && RUSTFLAGS="-Clinker=clang -Clink-arg=-fuse-ld=mold -Clink-arg=-Wl,--no-rosegment" cargo build  \
       --bin pg_sni_router  \
       --bin pageserver  \
       --bin pagectl  \
@@ -46,8 +56,8 @@ RUN set -e \
       --bin storage_controller  \
       --bin proxy  \
       --bin neon_local \
-      --bin storage_scrubber \
-      --locked --release
+      --locked --release \
+    && cachepot -s
 
 # Build final image
 #
@@ -59,6 +69,8 @@ RUN set -e \
     && apt install -y \
         libreadline-dev \
         libseccomp-dev \
+        libicu67 \
+        openssl \
         ca-certificates \
     && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* \
     && useradd -d /data neon \
@@ -72,7 +84,6 @@ COPY --from=build --chown=neon:neon /home/nonroot/target/release/storage_broker
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/storage_controller  /usr/local/bin
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/proxy               /usr/local/bin
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/neon_local          /usr/local/bin
-COPY --from=build --chown=neon:neon /home/nonroot/target/release/storage_scrubber    /usr/local/bin
 
 COPY --from=pg-build /home/nonroot/pg_install/v14 /usr/local/v14/
 COPY --from=pg-build /home/nonroot/pg_install/v15 /usr/local/v15/
@@ -81,24 +92,20 @@ COPY --from=pg-build /home/nonroot/postgres_install.tar.gz /data/
 
 # By default, pageserver uses `.neon/` working directory in WORKDIR, so create one and fill it with the dummy config.
 # Now, when `docker run ... pageserver` is run, it can start without errors, yet will have some default dummy values.
-RUN mkdir -p /data/.neon/ && \
-  echo "id=1234" > "/data/.neon/identity.toml" && \
-  echo "broker_endpoint='http://storage_broker:50051'\n" \
-       "pg_distrib_dir='/usr/local/'\n" \
-       "listen_pg_addr='0.0.0.0:6400'\n" \
-       "listen_http_addr='0.0.0.0:9898'\n" \
-  > /data/.neon/pageserver.toml && \
-  chown -R neon:neon /data/.neon
+RUN mkdir -p /data/.neon/ && chown -R neon:neon /data/.neon/ \
+    && /usr/local/bin/pageserver -D /data/.neon/ --init \
+       -c "id=1234" \
+       -c "broker_endpoint='http://storage_broker:50051'" \
+       -c "pg_distrib_dir='/usr/local/'" \
+       -c "listen_pg_addr='0.0.0.0:6400'" \
+       -c "listen_http_addr='0.0.0.0:9898'"
 
 # When running a binary that links with libpq, default to using our most recent postgres version.  Binaries
 # that want a particular postgres version will select it explicitly: this is just a default.
-ENV LD_LIBRARY_PATH=/usr/local/v16/lib
+ENV LD_LIBRARY_PATH /usr/local/v16/lib
 
 
 VOLUME ["/data"]
 USER neon
 EXPOSE 6400
 EXPOSE 9898
-
-CMD ["/usr/local/bin/pageserver", "-D", "/data/.neon"]
-

Dockerfile.build-tools

@@ -1,13 +1,5 @@
 FROM debian:bullseye-slim
 
-# Use ARG as a build-time environment variable here to allow.
-# It's not supposed to be set outside.
-# Alternatively it can be obtained using the following command
-# ```
-# . /etc/os-release && echo "${VERSION_CODENAME}"
-# ```
-ARG DEBIAN_VERSION_CODENAME=bullseye
-
 # Add nonroot user
 RUN useradd -ms /bin/bash nonroot -b /home
 SHELL ["/bin/bash", "-c"]
@@ -34,6 +26,7 @@ RUN set -e \
         liblzma-dev \
         libncurses5-dev \
         libncursesw5-dev \
+        libpq-dev \
         libreadline-dev \
         libseccomp-dev \
         libsqlite3-dev \
@@ -58,40 +51,29 @@ RUN set -e \
     && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
 
 # protobuf-compiler (protoc)
-ENV PROTOC_VERSION=25.1
+ENV PROTOC_VERSION 25.1
 RUN curl -fsSL "https://github.com/protocolbuffers/protobuf/releases/download/v${PROTOC_VERSION}/protoc-${PROTOC_VERSION}-linux-$(uname -m | sed 's/aarch64/aarch_64/g').zip" -o "protoc.zip" \
     && unzip -q protoc.zip -d protoc \
     && mv protoc/bin/protoc /usr/local/bin/protoc \
     && mv protoc/include/google /usr/local/include/google \
     && rm -rf protoc.zip protoc
 
-# s5cmd
-ENV S5CMD_VERSION=2.2.2
-RUN curl -sL "https://github.com/peak/s5cmd/releases/download/v${S5CMD_VERSION}/s5cmd_${S5CMD_VERSION}_Linux-$(uname -m | sed 's/x86_64/64bit/g' | sed 's/aarch64/arm64/g').tar.gz" | tar zxvf - s5cmd \
-    && chmod +x s5cmd \
-    && mv s5cmd /usr/local/bin/s5cmd
-
 # LLVM
-ENV LLVM_VERSION=18
+ENV LLVM_VERSION=17
 RUN curl -fsSL 'https://apt.llvm.org/llvm-snapshot.gpg.key' | apt-key add - \
-    && echo "deb http://apt.llvm.org/${DEBIAN_VERSION_CODENAME}/ llvm-toolchain-${DEBIAN_VERSION_CODENAME}-${LLVM_VERSION} main" > /etc/apt/sources.list.d/llvm.stable.list \
+    && echo "deb http://apt.llvm.org/bullseye/ llvm-toolchain-bullseye-${LLVM_VERSION} main" > /etc/apt/sources.list.d/llvm.stable.list \
     && apt update \
     && apt install -y clang-${LLVM_VERSION} llvm-${LLVM_VERSION} \
     && bash -c 'for f in /usr/bin/clang*-${LLVM_VERSION} /usr/bin/llvm*-${LLVM_VERSION}; do ln -s "${f}" "${f%-${LLVM_VERSION}}"; done' \
     && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
 
-# Install docker
-RUN curl -fsSL https://download.docker.com/linux/ubuntu/gpg | gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg \
-    && echo "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/debian ${DEBIAN_VERSION_CODENAME} stable" > /etc/apt/sources.list.d/docker.list \
+# PostgreSQL 14
+RUN curl -fsSL 'https://www.postgresql.org/media/keys/ACCC4CF8.asc' | apt-key add - \
+    && echo 'deb http://apt.postgresql.org/pub/repos/apt bullseye-pgdg main' > /etc/apt/sources.list.d/pgdg.list \
     && apt update \
-    && apt install -y docker-ce docker-ce-cli \
+    && apt install -y postgresql-client-14 \
     && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
 
-# Configure sudo & docker
-RUN usermod -aG sudo nonroot && \
-    echo '%sudo ALL=(ALL) NOPASSWD:ALL' >> /etc/sudoers && \
-    usermod -aG docker nonroot
-
 # AWS CLI
 RUN curl "https://awscli.amazonaws.com/awscli-exe-linux-$(uname -m).zip" -o "awscliv2.zip" \
     && unzip -q awscliv2.zip \
@@ -99,7 +81,7 @@ RUN curl "https://awscli.amazonaws.com/awscli-exe-linux-$(uname -m).zip" -o "aws
     && rm awscliv2.zip
 
 # Mold: A Modern Linker
-ENV MOLD_VERSION=v2.33.0
+ENV MOLD_VERSION v2.4.0
 RUN set -e \
     && git clone https://github.com/rui314/mold.git \
     && mkdir mold/build \
@@ -124,51 +106,12 @@ RUN for package in Capture::Tiny DateTime Devel::Cover Digest::MD5 File::Spec JS
     && make install \
     && rm -rf ../lcov.tar.gz
 
-# Compile and install the static OpenSSL library
-ENV OPENSSL_VERSION=1.1.1w
-ENV OPENSSL_PREFIX=/usr/local/openssl
-RUN wget -O /tmp/openssl-${OPENSSL_VERSION}.tar.gz https://www.openssl.org/source/openssl-${OPENSSL_VERSION}.tar.gz && \
-    echo "cf3098950cb4d853ad95c0841f1f9c6d3dc102dccfcacd521d93925208b76ac8 /tmp/openssl-${OPENSSL_VERSION}.tar.gz" | sha256sum --check && \
-    cd /tmp && \
-    tar xzvf /tmp/openssl-${OPENSSL_VERSION}.tar.gz && \
-    rm /tmp/openssl-${OPENSSL_VERSION}.tar.gz && \
-    cd /tmp/openssl-${OPENSSL_VERSION} && \
-    ./config --prefix=${OPENSSL_PREFIX}  -static --static no-shared -fPIC && \
-    make -j "$(nproc)" && \
-    make install && \
-    cd /tmp && \
-    rm -rf /tmp/openssl-${OPENSSL_VERSION}
-
-# Use the same version of libicu as the compute nodes so that
-# clusters created using inidb on pageserver can be used by computes.
-#
-# TODO: at this time, Dockerfile.compute-node uses the debian bullseye libicu
-# package, which is 67.1. We're duplicating that knowledge here, and also, technically,
-# Debian has a few patches on top of 67.1 that we're not adding here.
-ENV ICU_VERSION=67.1
-ENV ICU_PREFIX=/usr/local/icu
-
-# Download and build static ICU
-RUN wget -O /tmp/libicu-${ICU_VERSION}.tgz https://github.com/unicode-org/icu/releases/download/release-${ICU_VERSION//./-}/icu4c-${ICU_VERSION//./_}-src.tgz && \
-    echo "94a80cd6f251a53bd2a997f6f1b5ac6653fe791dfab66e1eb0227740fb86d5dc /tmp/libicu-${ICU_VERSION}.tgz" | sha256sum --check && \
-    mkdir /tmp/icu && \
-    pushd /tmp/icu && \
-    tar -xzf /tmp/libicu-${ICU_VERSION}.tgz && \
-    pushd icu/source && \
-    ./configure --prefix=${ICU_PREFIX}  --enable-static --enable-shared=no CXXFLAGS="-fPIC" CFLAGS="-fPIC" && \
-    make -j "$(nproc)" && \
-    make install && \
-    popd && \
-    rm -rf icu && \
-    rm -f /tmp/libicu-${ICU_VERSION}.tgz && \
-    popd
-
 # Switch to nonroot user
 USER nonroot:nonroot
 WORKDIR /home/nonroot
 
 # Python
-ENV PYTHON_VERSION=3.9.19 \
+ENV PYTHON_VERSION=3.9.18 \
     PYENV_ROOT=/home/nonroot/.pyenv \
     PATH=/home/nonroot/.pyenv/shims:/home/nonroot/.pyenv/bin:/home/nonroot/.poetry/bin:$PATH
 RUN set -e \
@@ -192,14 +135,9 @@ WORKDIR /home/nonroot
 
 # Rust
 # Please keep the version of llvm (installed above) in sync with rust llvm (`rustc --version --verbose | grep LLVM`)
-ENV RUSTC_VERSION=1.80.1
+ENV RUSTC_VERSION=1.77.0
 ENV RUSTUP_HOME="/home/nonroot/.rustup"
 ENV PATH="/home/nonroot/.cargo/bin:${PATH}"
-ARG RUSTFILT_VERSION=0.2.1
-ARG CARGO_HAKARI_VERSION=0.9.30
-ARG CARGO_DENY_VERSION=0.16.1
-ARG CARGO_HACK_VERSION=0.6.31
-ARG CARGO_NEXTEST_VERSION=0.9.72
 RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux-gnu/rustup-init && whoami && \
 	chmod +x rustup-init && \
 	./rustup-init -y --default-toolchain ${RUSTC_VERSION} && \
@@ -208,13 +146,15 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
     . "$HOME/.cargo/env" && \
     cargo --version && rustup --version && \
     rustup component add llvm-tools-preview rustfmt clippy && \
-    cargo install rustfilt            --version ${RUSTFILT_VERSION} && \
-    cargo install cargo-hakari        --version ${CARGO_HAKARI_VERSION} && \
-    cargo install cargo-deny --locked --version ${CARGO_DENY_VERSION} && \
-    cargo install cargo-hack          --version ${CARGO_HACK_VERSION} && \
-    cargo install cargo-nextest       --version ${CARGO_NEXTEST_VERSION} && \
+    cargo install --git https://github.com/paritytech/cachepot && \
+    cargo install rustfilt && \
+    cargo install cargo-hakari && \
+    cargo install cargo-deny --locked && \
+    cargo install cargo-hack && \
+    cargo install cargo-nextest && \
     rm -rf /home/nonroot/.cargo/registry && \
     rm -rf /home/nonroot/.cargo/git
+ENV RUSTC_WRAPPER=cachepot
 
 # Show versions
 RUN whoami \
@@ -224,6 +164,3 @@ RUN whoami \
     && rustup --version --verbose \
     && rustc --version --verbose \
     && clang --version
-
-# Set following flag to check in Makefile if its running in Docker
-RUN touch /home/nonroot/.docker_build

Dockerfile.compute-node

@@ -89,16 +89,16 @@ RUN apt update && \
 # SFCGAL > 1.3 requires CGAL > 5.2, Bullseye's libcgal-dev is 5.2
 RUN wget https://gitlab.com/Oslandia/SFCGAL/-/archive/v1.3.10/SFCGAL-v1.3.10.tar.gz -O SFCGAL.tar.gz && \
     echo "4e39b3b2adada6254a7bdba6d297bb28e1a9835a9f879b74f37e2dab70203232 SFCGAL.tar.gz" | sha256sum --check && \
-    mkdir sfcgal-src && cd sfcgal-src && tar xzf ../SFCGAL.tar.gz --strip-components=1 -C . && \
+    mkdir sfcgal-src && cd sfcgal-src && tar xvzf ../SFCGAL.tar.gz --strip-components=1 -C . && \
     cmake -DCMAKE_BUILD_TYPE=Release . && make -j $(getconf _NPROCESSORS_ONLN) && \
     DESTDIR=/sfcgal make install -j $(getconf _NPROCESSORS_ONLN) && \
     make clean && cp -R /sfcgal/* /
 
-ENV PATH="/usr/local/pgsql/bin:$PATH"
+ENV PATH "/usr/local/pgsql/bin:$PATH"
 
 RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.3.tar.gz -O postgis.tar.gz && \
     echo "74eb356e3f85f14233791013360881b6748f78081cc688ff9d6f0f673a762d13 postgis.tar.gz" | sha256sum --check && \
-    mkdir postgis-src && cd postgis-src && tar xzf ../postgis.tar.gz --strip-components=1 -C . && \
+    mkdir postgis-src && cd postgis-src && tar xvzf ../postgis.tar.gz --strip-components=1 -C . && \
     find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
     ./autogen.sh && \
     ./configure --with-sfcgal=/usr/local/bin/sfcgal-config && \
@@ -124,7 +124,7 @@ RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.3.tar.gz -O postg
 
 RUN wget https://github.com/pgRouting/pgrouting/archive/v3.4.2.tar.gz -O pgrouting.tar.gz && \
     echo "cac297c07d34460887c4f3b522b35c470138760fe358e351ad1db4edb6ee306e pgrouting.tar.gz" | sha256sum --check && \
-    mkdir pgrouting-src && cd pgrouting-src && tar xzf ../pgrouting.tar.gz --strip-components=1 -C . && \
+    mkdir pgrouting-src && cd pgrouting-src && tar xvzf ../pgrouting.tar.gz --strip-components=1 -C . && \
     mkdir build && cd build && \
     cmake -DCMAKE_BUILD_TYPE=Release .. && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -149,7 +149,7 @@ RUN apt update && \
 
 RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.10.tar.gz -O plv8.tar.gz && \
     echo "7096c3290928561f0d4901b7a52794295dc47f6303102fae3f8e42dd575ad97d plv8.tar.gz" | sha256sum --check && \
-    mkdir plv8-src && cd plv8-src && tar xzf ../plv8.tar.gz --strip-components=1 -C . && \
+    mkdir plv8-src && cd plv8-src && tar xvzf ../plv8.tar.gz --strip-components=1 -C . && \
     # generate and copy upgrade scripts
     mkdir -p upgrade && ./generate_upgrade.sh 3.1.10 && \
     cp upgrade/* /usr/local/pgsql/share/extension/ && \
@@ -194,7 +194,7 @@ RUN case "$(uname -m)" in \
 
 RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz && \
     echo "ec99f1f5974846bde64f4513cf8d2ea1b8d172d2218ab41803bf6a63532272bc h3.tar.gz" | sha256sum --check && \
-    mkdir h3-src && cd h3-src && tar xzf ../h3.tar.gz --strip-components=1 -C . && \
+    mkdir h3-src && cd h3-src && tar xvzf ../h3.tar.gz --strip-components=1 -C . && \
     mkdir build && cd build && \
     cmake .. -DCMAKE_BUILD_TYPE=Release && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -204,7 +204,7 @@ RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz
 
 RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.1.3.tar.gz -O h3-pg.tar.gz && \
     echo "5c17f09a820859ffe949f847bebf1be98511fb8f1bd86f94932512c00479e324 h3-pg.tar.gz" | sha256sum --check && \
-    mkdir h3-pg-src && cd h3-pg-src && tar xzf ../h3-pg.tar.gz --strip-components=1 -C . && \
+    mkdir h3-pg-src && cd h3-pg-src && tar xvzf ../h3-pg.tar.gz --strip-components=1 -C . && \
     export PATH="/usr/local/pgsql/bin:$PATH" && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
@@ -222,7 +222,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -O postgresql-unit.tar.gz && \
     echo "411d05beeb97e5a4abf17572bfcfbb5a68d98d1018918feff995f6ee3bb03e79 postgresql-unit.tar.gz" | sha256sum --check && \
-    mkdir postgresql-unit-src && cd postgresql-unit-src && tar xzf ../postgresql-unit.tar.gz --strip-components=1 -C . && \
+    mkdir postgresql-unit-src && cd postgresql-unit-src && tar xvzf ../postgresql-unit.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     # unit extension's "create extension" script relies on absolute install path to fill some reference tables.
@@ -241,17 +241,11 @@ RUN wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -
 FROM build-deps AS vector-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-COPY patches/pgvector.patch /pgvector.patch
-
-# By default, pgvector Makefile uses `-march=native`. We don't want that,
-# because we build the images on different machines than where we run them.
-# Pass OPTFLAGS="" to remove it.
-RUN wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.7.2.tar.gz -O pgvector.tar.gz && \
-    echo "617fba855c9bcb41a2a9bc78a78567fd2e147c72afd5bf9d37b31b9591632b30 pgvector.tar.gz" | sha256sum --check && \
-    mkdir pgvector-src && cd pgvector-src && tar xzf ../pgvector.tar.gz --strip-components=1 -C . && \
-    patch -p1 < /pgvector.patch && \
-    make -j $(getconf _NPROCESSORS_ONLN) OPTFLAGS="" PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
-    make -j $(getconf _NPROCESSORS_ONLN) OPTFLAGS="" install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
+RUN wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.5.1.tar.gz -O pgvector.tar.gz && \
+    echo "cc7a8e034a96e30a819911ac79d32f6bc47bdd1aa2de4d7d4904e26b83209dc8 pgvector.tar.gz" | sha256sum --check && \
+    mkdir pgvector-src && cd pgvector-src && tar xvzf ../pgvector.tar.gz --strip-components=1 -C . && \
+    make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
+    make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/vector.control
 
 #########################################################################################
@@ -266,7 +260,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 # 9742dab1b2f297ad3811120db7b21451bca2d3c9 made on 13/11/2021
 RUN wget https://github.com/michelp/pgjwt/archive/9742dab1b2f297ad3811120db7b21451bca2d3c9.tar.gz -O pgjwt.tar.gz && \
     echo "cfdefb15007286f67d3d45510f04a6a7a495004be5b3aecb12cda667e774203f pgjwt.tar.gz" | sha256sum --check && \
-    mkdir pgjwt-src && cd pgjwt-src && tar xzf ../pgjwt.tar.gz --strip-components=1 -C . && \
+    mkdir pgjwt-src && cd pgjwt-src && tar xvzf ../pgjwt.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgjwt.control
 
@@ -281,7 +275,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/HypoPG/hypopg/archive/refs/tags/1.4.0.tar.gz -O hypopg.tar.gz && \
     echo "0821011743083226fc9b813c1f2ef5897a91901b57b6bea85a78e466187c6819 hypopg.tar.gz" | sha256sum --check && \
-    mkdir hypopg-src && cd hypopg-src && tar xzf ../hypopg.tar.gz --strip-components=1 -C . && \
+    mkdir hypopg-src && cd hypopg-src && tar xvzf ../hypopg.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/hypopg.control
@@ -297,7 +291,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/iCyberon/pg_hashids/archive/refs/tags/v1.2.1.tar.gz -O pg_hashids.tar.gz && \
     echo "74576b992d9277c92196dd8d816baa2cc2d8046fe102f3dcd7f3c3febed6822a pg_hashids.tar.gz" | sha256sum --check && \
-    mkdir pg_hashids-src && cd pg_hashids-src && tar xzf ../pg_hashids.tar.gz --strip-components=1 -C . && \
+    mkdir pg_hashids-src && cd pg_hashids-src && tar xvzf ../pg_hashids.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_hashids.control
@@ -311,12 +305,9 @@ RUN wget https://github.com/iCyberon/pg_hashids/archive/refs/tags/v1.2.1.tar.gz
 FROM build-deps AS rum-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-COPY patches/rum.patch /rum.patch
-
 RUN wget https://github.com/postgrespro/rum/archive/refs/tags/1.3.13.tar.gz -O rum.tar.gz && \
     echo "6ab370532c965568df6210bd844ac6ba649f53055e48243525b0b7e5c4d69a7d rum.tar.gz" | sha256sum --check && \
-    mkdir rum-src && cd rum-src && tar xzf ../rum.tar.gz --strip-components=1 -C . && \
-    patch -p1 < /rum.patch && \
+    mkdir rum-src && cd rum-src && tar xvzf ../rum.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/rum.control
@@ -332,7 +323,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/theory/pgtap/archive/refs/tags/v1.2.0.tar.gz -O pgtap.tar.gz && \
     echo "9c7c3de67ea41638e14f06da5da57bac6f5bd03fea05c165a0ec862205a5c052 pgtap.tar.gz" | sha256sum --check && \
-    mkdir pgtap-src && cd pgtap-src && tar xzf ../pgtap.tar.gz --strip-components=1 -C . && \
+    mkdir pgtap-src && cd pgtap-src && tar xvzf ../pgtap.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgtap.control
@@ -348,7 +339,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/RhodiumToad/ip4r/archive/refs/tags/2.4.2.tar.gz -O ip4r.tar.gz && \
     echo "0f7b1f159974f49a47842a8ab6751aecca1ed1142b6d5e38d81b064b2ead1b4b ip4r.tar.gz" | sha256sum --check && \
-    mkdir ip4r-src && cd ip4r-src && tar xzf ../ip4r.tar.gz --strip-components=1 -C . && \
+    mkdir ip4r-src && cd ip4r-src && tar xvzf ../ip4r.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/ip4r.control
@@ -364,7 +355,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/dimitri/prefix/archive/refs/tags/v1.2.10.tar.gz -O prefix.tar.gz && \
     echo "4342f251432a5f6fb05b8597139d3ccde8dcf87e8ca1498e7ee931ca057a8575 prefix.tar.gz" | sha256sum --check && \
-    mkdir prefix-src && cd prefix-src && tar xzf ../prefix.tar.gz --strip-components=1 -C . && \
+    mkdir prefix-src && cd prefix-src && tar xvzf ../prefix.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/prefix.control
@@ -380,7 +371,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/citusdata/postgresql-hll/archive/refs/tags/v2.18.tar.gz -O hll.tar.gz && \
     echo "e2f55a6f4c4ab95ee4f1b4a2b73280258c5136b161fe9d059559556079694f0e hll.tar.gz" | sha256sum --check && \
-    mkdir hll-src && cd hll-src && tar xzf ../hll.tar.gz --strip-components=1 -C . && \
+    mkdir hll-src && cd hll-src && tar xvzf ../hll.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/hll.control
@@ -396,7 +387,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN wget https://github.com/okbob/plpgsql_check/archive/refs/tags/v2.5.3.tar.gz -O plpgsql_check.tar.gz && \
     echo "6631ec3e7fb3769eaaf56e3dfedb829aa761abf163d13dba354b4c218508e1c0 plpgsql_check.tar.gz" | sha256sum --check && \
-    mkdir plpgsql_check-src && cd plpgsql_check-src && tar xzf ../plpgsql_check.tar.gz --strip-components=1 -C . && \
+    mkdir plpgsql_check-src && cd plpgsql_check-src && tar xvzf ../plpgsql_check.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/plpgsql_check.control
@@ -411,7 +402,7 @@ FROM build-deps AS timescaledb-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ARG PG_VERSION
-ENV PATH="/usr/local/pgsql/bin:$PATH"
+ENV PATH "/usr/local/pgsql/bin:$PATH"
 
 RUN case "${PG_VERSION}" in \
       "v14" | "v15") \
@@ -427,7 +418,7 @@ RUN case "${PG_VERSION}" in \
     apt-get install -y cmake && \
     wget https://github.com/timescale/timescaledb/archive/refs/tags/${TIMESCALEDB_VERSION}.tar.gz -O timescaledb.tar.gz && \
     echo "${TIMESCALEDB_CHECKSUM} timescaledb.tar.gz" | sha256sum --check && \
-    mkdir timescaledb-src && cd timescaledb-src && tar xzf ../timescaledb.tar.gz --strip-components=1 -C . && \
+    mkdir timescaledb-src && cd timescaledb-src && tar xvzf ../timescaledb.tar.gz --strip-components=1 -C . && \
     ./bootstrap -DSEND_TELEMETRY_DEFAULT:BOOL=OFF -DUSE_TELEMETRY:BOOL=OFF -DAPACHE_ONLY:BOOL=ON -DCMAKE_BUILD_TYPE=Release && \
     cd build && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -444,7 +435,7 @@ FROM build-deps AS pg-hint-plan-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ARG PG_VERSION
-ENV PATH="/usr/local/pgsql/bin:$PATH"
+ENV PATH "/usr/local/pgsql/bin:$PATH"
 
 RUN case "${PG_VERSION}" in \
       "v14") \
@@ -465,11 +456,36 @@ RUN case "${PG_VERSION}" in \
     esac && \
     wget https://github.com/ossc-db/pg_hint_plan/archive/refs/tags/REL${PG_HINT_PLAN_VERSION}.tar.gz -O pg_hint_plan.tar.gz && \
     echo "${PG_HINT_PLAN_CHECKSUM} pg_hint_plan.tar.gz" | sha256sum --check && \
-    mkdir pg_hint_plan-src && cd pg_hint_plan-src && tar xzf ../pg_hint_plan.tar.gz --strip-components=1 -C . && \
+    mkdir pg_hint_plan-src && cd pg_hint_plan-src && tar xvzf ../pg_hint_plan.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make install -j $(getconf _NPROCESSORS_ONLN) && \
     echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_hint_plan.control
 
+#########################################################################################
+#
+# Layer "kq-imcx-pg-build"
+# compile kq_imcx extension
+#
+#########################################################################################
+FROM build-deps AS kq-imcx-pg-build
+COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
+
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
+RUN apt-get update && \
+    apt-get install -y git libgtk2.0-dev libpq-dev libpam-dev libxslt-dev libkrb5-dev cmake && \
+    wget https://github.com/ketteq-neon/postgres-exts/archive/e0bd1a9d9313d7120c1b9c7bb15c48c0dede4c4e.tar.gz -O kq_imcx.tar.gz && \
+    echo "dc93a97ff32d152d32737ba7e196d9687041cda15e58ab31344c2f2de8855336 kq_imcx.tar.gz" | sha256sum --check && \
+    mkdir kq_imcx-src && cd kq_imcx-src && tar xvzf ../kq_imcx.tar.gz --strip-components=1 -C . && \
+    find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
+    mkdir build && cd build && \
+    cmake -DCMAKE_BUILD_TYPE=Release .. && \
+    make -j $(getconf _NPROCESSORS_ONLN) && \
+    make -j $(getconf _NPROCESSORS_ONLN) install && \
+    echo 'trusted = true' >> /usr/local/pgsql/share/extension/kq_imcx.control && \
+    find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /after.txt &&\
+    mkdir -p /extensions/kq_imcx && cp /usr/local/pgsql/share/extension/kq_imcx.control /extensions/kq_imcx && \
+    sort -o /before.txt /before.txt && sort -o /after.txt /after.txt && \
+    comm -13 /before.txt /after.txt | tar --directory=/usr/local/pgsql --zstd -cf /extensions/kq_imcx.tar.zst -T -
 
 #########################################################################################
 #
@@ -480,10 +496,10 @@ RUN case "${PG_VERSION}" in \
 FROM build-deps AS pg-cron-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/citusdata/pg_cron/archive/refs/tags/v1.6.0.tar.gz -O pg_cron.tar.gz && \
     echo "383a627867d730222c272bfd25cd5e151c578d73f696d32910c7db8c665cc7db pg_cron.tar.gz" | sha256sum --check && \
-    mkdir pg_cron-src && cd pg_cron-src && tar xzf ../pg_cron.tar.gz --strip-components=1 -C . && \
+    mkdir pg_cron-src && cd pg_cron-src && tar xvzf ../pg_cron.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_cron.control
@@ -506,10 +522,10 @@ RUN apt-get update && \
         libboost-system1.74-dev \
         libeigen3-dev
 
-ENV PATH="/usr/local/pgsql/bin/:/usr/local/pgsql/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:/usr/local/pgsql/:$PATH"
 RUN wget https://github.com/rdkit/rdkit/archive/refs/tags/Release_2023_03_3.tar.gz -O rdkit.tar.gz && \
     echo "bdbf9a2e6988526bfeb8c56ce3cdfe2998d60ac289078e2215374288185e8c8d rdkit.tar.gz" | sha256sum --check && \
-    mkdir rdkit-src && cd rdkit-src && tar xzf ../rdkit.tar.gz --strip-components=1 -C . && \
+    mkdir rdkit-src && cd rdkit-src && tar xvzf ../rdkit.tar.gz --strip-components=1 -C . && \
     cmake \
         -D RDK_BUILD_CAIRO_SUPPORT=OFF \
         -D RDK_BUILD_INCHI_SUPPORT=ON \
@@ -546,10 +562,10 @@ RUN wget https://github.com/rdkit/rdkit/archive/refs/tags/Release_2023_03_3.tar.
 FROM build-deps AS pg-uuidv7-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/fboulnois/pg_uuidv7/archive/refs/tags/v1.0.1.tar.gz -O pg_uuidv7.tar.gz && \
     echo "0d0759ab01b7fb23851ecffb0bce27822e1868a4a5819bfd276101c716637a7a pg_uuidv7.tar.gz" | sha256sum --check && \
-    mkdir pg_uuidv7-src && cd pg_uuidv7-src && tar xzf ../pg_uuidv7.tar.gz --strip-components=1 -C . && \
+    mkdir pg_uuidv7-src && cd pg_uuidv7-src && tar xvzf ../pg_uuidv7.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_uuidv7.control
@@ -563,10 +579,10 @@ RUN wget https://github.com/fboulnois/pg_uuidv7/archive/refs/tags/v1.0.1.tar.gz
 FROM build-deps AS pg-roaringbitmap-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/ChenHuajun/pg_roaringbitmap/archive/refs/tags/v0.5.4.tar.gz -O pg_roaringbitmap.tar.gz && \
     echo "b75201efcb1c2d1b014ec4ae6a22769cc7a224e6e406a587f5784a37b6b5a2aa pg_roaringbitmap.tar.gz" | sha256sum --check && \
-    mkdir pg_roaringbitmap-src && cd pg_roaringbitmap-src && tar xzf ../pg_roaringbitmap.tar.gz --strip-components=1 -C . && \
+    mkdir pg_roaringbitmap-src && cd pg_roaringbitmap-src && tar xvzf ../pg_roaringbitmap.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/roaringbitmap.control
@@ -580,10 +596,10 @@ RUN wget https://github.com/ChenHuajun/pg_roaringbitmap/archive/refs/tags/v0.5.4
 FROM build-deps AS pg-semver-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/theory/pg-semver/archive/refs/tags/v0.32.1.tar.gz -O pg_semver.tar.gz && \
     echo "fbdaf7512026d62eec03fad8687c15ed509b6ba395bff140acd63d2e4fbe25d7 pg_semver.tar.gz" | sha256sum --check && \
-    mkdir pg_semver-src && cd pg_semver-src && tar xzf ../pg_semver.tar.gz --strip-components=1 -C . && \
+    mkdir pg_semver-src && cd pg_semver-src && tar xvzf ../pg_semver.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/semver.control
@@ -598,7 +614,7 @@ FROM build-deps AS pg-embedding-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ARG PG_VERSION
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN case "${PG_VERSION}" in \
       "v14" | "v15") \
         export PG_EMBEDDING_VERSION=0.3.5 \
@@ -609,7 +625,7 @@ RUN case "${PG_VERSION}" in \
     esac && \
     wget https://github.com/neondatabase/pg_embedding/archive/refs/tags/${PG_EMBEDDING_VERSION}.tar.gz -O pg_embedding.tar.gz && \
     echo "${PG_EMBEDDING_CHECKSUM} pg_embedding.tar.gz" | sha256sum --check && \
-    mkdir pg_embedding-src && cd pg_embedding-src && tar xzf ../pg_embedding.tar.gz --strip-components=1 -C . && \
+    mkdir pg_embedding-src && cd pg_embedding-src && tar xvzf ../pg_embedding.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install
 
@@ -622,10 +638,10 @@ RUN case "${PG_VERSION}" in \
 FROM build-deps AS pg-anon-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget  https://github.com/neondatabase/postgresql_anonymizer/archive/refs/tags/neon_1.1.1.tar.gz -O pg_anon.tar.gz && \
     echo "321ea8d5c1648880aafde850a2c576e4a9e7b9933a34ce272efc839328999fa9  pg_anon.tar.gz" | sha256sum --check && \
-    mkdir pg_anon-src && cd pg_anon-src && tar xzf ../pg_anon.tar.gz --strip-components=1 -C . && \
+    mkdir pg_anon-src && cd pg_anon-src && tar xvzf ../pg_anon.tar.gz --strip-components=1 -C . && \
     find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/anon.control && \
@@ -657,7 +673,7 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
     chmod +x rustup-init && \
     ./rustup-init -y --no-modify-path --profile minimal --default-toolchain stable && \
     rm rustup-init && \
-    cargo install --locked --version 0.11.3 cargo-pgrx && \
+    cargo install --locked --version 0.10.2 cargo-pgrx && \
     /bin/bash -c 'cargo pgrx init --pg${PG_VERSION:1}=/usr/local/pgsql/bin/pg_config'
 
 USER root
@@ -672,15 +688,10 @@ USER root
 FROM rust-extensions-build AS pg-jsonschema-pg-build
 ARG PG_VERSION
 
-RUN wget https://github.com/supabase/pg_jsonschema/archive/refs/tags/v0.3.1.tar.gz -O pg_jsonschema.tar.gz && \
-    echo "61df3db1ed83cf24f6aa39c826f8818bfa4f0bd33b587fd6b2b1747985642297 pg_jsonschema.tar.gz" | sha256sum --check && \
-    mkdir pg_jsonschema-src && cd pg_jsonschema-src && tar xzf ../pg_jsonschema.tar.gz --strip-components=1 -C . && \
-    # see commit 252b3685a27a0f4c31a0f91e983c6314838e89e8
-    # `unsafe-postgres` feature allows to build pgx extensions
-    # against postgres forks that decided to change their ABI name (like us).
-    # With that we can build extensions without forking them and using stock
-    # pgx. As this feature is new few manual version bumps were required.
-    sed -i 's/pgrx = "0.11.3"/pgrx = { version = "0.11.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
+RUN wget https://github.com/supabase/pg_jsonschema/archive/refs/tags/v0.2.0.tar.gz -O pg_jsonschema.tar.gz && \
+    echo "9118fc508a6e231e7a39acaa6f066fcd79af17a5db757b47d2eefbe14f7794f0 pg_jsonschema.tar.gz" | sha256sum --check && \
+    mkdir pg_jsonschema-src && cd pg_jsonschema-src && tar xvzf ../pg_jsonschema.tar.gz --strip-components=1 -C . && \
+    sed -i 's/pgrx = "0.10.2"/pgrx = { version = "0.10.2", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
     cargo pgrx install --release && \
     echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_jsonschema.control
 
@@ -694,10 +705,10 @@ RUN wget https://github.com/supabase/pg_jsonschema/archive/refs/tags/v0.3.1.tar.
 FROM rust-extensions-build AS pg-graphql-pg-build
 ARG PG_VERSION
 
-RUN wget https://github.com/supabase/pg_graphql/archive/refs/tags/v1.5.7.tar.gz -O pg_graphql.tar.gz && \
-    echo "2b3e567a5b31019cb97ae0e33263c1bcc28580be5a444ac4c8ece5c4be2aea41 pg_graphql.tar.gz" | sha256sum --check && \
-    mkdir pg_graphql-src && cd pg_graphql-src && tar xzf ../pg_graphql.tar.gz --strip-components=1 -C . && \
-    sed -i 's/pgrx = "=0.11.3"/pgrx = { version = "0.11.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
+RUN wget https://github.com/supabase/pg_graphql/archive/refs/tags/v1.4.0.tar.gz -O pg_graphql.tar.gz && \
+    echo "bd8dc7230282b3efa9ae5baf053a54151ed0e66881c7c53750e2d0c765776edc pg_graphql.tar.gz" | sha256sum --check && \
+    mkdir pg_graphql-src && cd pg_graphql-src && tar xvzf ../pg_graphql.tar.gz --strip-components=1 -C . && \
+    sed -i 's/pgrx = "=0.10.2"/pgrx = { version = "0.10.2", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
     cargo pgrx install --release && \
     # it's needed to enable extension because it uses untrusted C language
     sed -i 's/superuser = false/superuser = true/g' /usr/local/pgsql/share/extension/pg_graphql.control && \
@@ -716,10 +727,7 @@ ARG PG_VERSION
 # 26806147b17b60763039c6a6878884c41a262318 made on 26/09/2023
 RUN wget https://github.com/kelvich/pg_tiktoken/archive/26806147b17b60763039c6a6878884c41a262318.tar.gz -O pg_tiktoken.tar.gz && \
     echo "e64e55aaa38c259512d3e27c572da22c4637418cf124caba904cd50944e5004e pg_tiktoken.tar.gz" | sha256sum --check && \
-    mkdir pg_tiktoken-src && cd pg_tiktoken-src && tar xzf ../pg_tiktoken.tar.gz --strip-components=1 -C . && \
-    # TODO update pgrx version in the pg_tiktoken repo and remove this line
-    sed -i 's/pgrx = { version = "=0.10.2",/pgrx = { version = "0.11.3",/g' Cargo.toml && \
-    sed -i 's/pgrx-tests = "=0.10.2"/pgrx-tests = "0.11.3"/g' Cargo.toml && \
+    mkdir pg_tiktoken-src && cd pg_tiktoken-src && tar xvzf ../pg_tiktoken.tar.gz --strip-components=1 -C . && \
     cargo pgrx install --release && \
     echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_tiktoken.control
 
@@ -733,10 +741,14 @@ RUN wget https://github.com/kelvich/pg_tiktoken/archive/26806147b17b60763039c6a6
 FROM rust-extensions-build AS pg-pgx-ulid-build
 ARG PG_VERSION
 
-RUN wget https://github.com/pksunkara/pgx_ulid/archive/refs/tags/v0.1.5.tar.gz -O pgx_ulid.tar.gz && \
-    echo "9d1659a2da65af0133d5451c454de31b37364e3502087dadf579f790bc8bef17 pgx_ulid.tar.gz" | sha256sum --check && \
-    mkdir pgx_ulid-src && cd pgx_ulid-src && tar xzf ../pgx_ulid.tar.gz --strip-components=1 -C . && \
-    sed -i 's/pgrx       = "^0.11.2"/pgrx = { version = "=0.11.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
+RUN wget https://github.com/pksunkara/pgx_ulid/archive/refs/tags/v0.1.3.tar.gz -O pgx_ulid.tar.gz && \
+    echo "ee5db82945d2d9f2d15597a80cf32de9dca67b897f605beb830561705f12683c pgx_ulid.tar.gz" | sha256sum --check && \
+    mkdir pgx_ulid-src && cd pgx_ulid-src && tar xvzf ../pgx_ulid.tar.gz --strip-components=1 -C . && \
+    echo "******************* Apply a patch for Postgres 16 support; delete in the next release ******************" && \
+    wget https://github.com/pksunkara/pgx_ulid/commit/f84954cf63fc8c80d964ac970d9eceed3c791196.patch && \
+    patch -p1 < f84954cf63fc8c80d964ac970d9eceed3c791196.patch && \
+    echo "********************************************************************************************************" && \
+    sed -i 's/pgrx       = "=0.10.2"/pgrx = { version = "=0.10.2", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
     cargo pgrx install --release && \
     echo "trusted = true" >> /usr/local/pgsql/share/extension/ulid.control
 
@@ -750,10 +762,10 @@ RUN wget https://github.com/pksunkara/pgx_ulid/archive/refs/tags/v0.1.5.tar.gz -
 FROM build-deps AS wal2json-pg-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/eulerto/wal2json/archive/refs/tags/wal2json_2_5.tar.gz && \
     echo "b516653575541cf221b99cf3f8be9b6821f6dbcfc125675c85f35090f824f00e wal2json_2_5.tar.gz" | sha256sum --check && \
-    mkdir wal2json-src && cd wal2json-src && tar xzf ../wal2json_2_5.tar.gz --strip-components=1 -C . && \
+    mkdir wal2json-src && cd wal2json-src && tar xvzf ../wal2json_2_5.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install
 
@@ -766,10 +778,10 @@ RUN wget https://github.com/eulerto/wal2json/archive/refs/tags/wal2json_2_5.tar.
 FROM build-deps AS pg-ivm-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/sraoss/pg_ivm/archive/refs/tags/v1.7.tar.gz -O pg_ivm.tar.gz && \
     echo "ebfde04f99203c7be4b0e873f91104090e2e83e5429c32ac242d00f334224d5e pg_ivm.tar.gz" | sha256sum --check && \
-    mkdir pg_ivm-src && cd pg_ivm-src && tar xzf ../pg_ivm.tar.gz --strip-components=1 -C . && \
+    mkdir pg_ivm-src && cd pg_ivm-src && tar xvzf ../pg_ivm.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_ivm.control
@@ -783,10 +795,10 @@ RUN wget https://github.com/sraoss/pg_ivm/archive/refs/tags/v1.7.tar.gz -O pg_iv
 FROM build-deps AS pg-partman-build
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-ENV PATH="/usr/local/pgsql/bin/:$PATH"
+ENV PATH "/usr/local/pgsql/bin/:$PATH"
 RUN wget https://github.com/pgpartman/pg_partman/archive/refs/tags/v5.0.1.tar.gz -O pg_partman.tar.gz && \
     echo "75b541733a9659a6c90dbd40fccb904a630a32880a6e3044d0c4c5f4c8a65525 pg_partman.tar.gz" | sha256sum --check && \
-    mkdir pg_partman-src && cd pg_partman-src && tar xzf ../pg_partman.tar.gz --strip-components=1 -C . && \
+    mkdir pg_partman-src && cd pg_partman-src && tar xvzf ../pg_partman.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_partman.control
@@ -822,6 +834,7 @@ COPY --from=hll-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=plpgsql-check-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=timescaledb-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg-hint-plan-pg-build /usr/local/pgsql/ /usr/local/pgsql/
+COPY --from=kq-imcx-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg-cron-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg-pgx-ulid-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=rdkit-pg-build /usr/local/pgsql/ /usr/local/pgsql/
@@ -909,70 +922,6 @@ RUN rm -r /usr/local/pgsql/include
 # if they were to be used by other libraries.
 RUN rm /usr/local/pgsql/lib/lib*.a
 
-
-#########################################################################################
-#
-# Layer neon-pg-ext-test
-#
-#########################################################################################
-
-FROM neon-pg-ext-build AS neon-pg-ext-test
-ARG PG_VERSION
-RUN mkdir /ext-src
-
-#COPY --from=postgis-build /postgis.tar.gz /ext-src/
-#COPY --from=postgis-build /sfcgal/* /usr
-COPY --from=plv8-build /plv8.tar.gz /ext-src/
-COPY --from=h3-pg-build /h3-pg.tar.gz /ext-src/
-COPY --from=unit-pg-build /postgresql-unit.tar.gz /ext-src/
-COPY --from=vector-pg-build /pgvector.tar.gz /ext-src/
-COPY --from=vector-pg-build /pgvector.patch /ext-src/
-COPY --from=pgjwt-pg-build /pgjwt.tar.gz /ext-src
-#COPY --from=pg-jsonschema-pg-build /home/nonroot/pg_jsonschema.tar.gz /ext-src
-#COPY --from=pg-graphql-pg-build /home/nonroot/pg_graphql.tar.gz /ext-src
-#COPY --from=pg-tiktoken-pg-build /home/nonroot/pg_tiktoken.tar.gz /ext-src
-COPY --from=hypopg-pg-build /hypopg.tar.gz /ext-src
-COPY --from=pg-hashids-pg-build /pg_hashids.tar.gz /ext-src
-COPY --from=rum-pg-build /rum.tar.gz /ext-src
-COPY patches/rum.patch /ext-src
-#COPY --from=pgtap-pg-build /pgtap.tar.gz /ext-src
-COPY --from=ip4r-pg-build /ip4r.tar.gz /ext-src
-COPY --from=prefix-pg-build /prefix.tar.gz /ext-src
-COPY --from=hll-pg-build /hll.tar.gz /ext-src
-COPY --from=plpgsql-check-pg-build /plpgsql_check.tar.gz /ext-src
-#COPY --from=timescaledb-pg-build /timescaledb.tar.gz /ext-src
-COPY --from=pg-hint-plan-pg-build /pg_hint_plan.tar.gz /ext-src
-COPY patches/pg_hint_plan.patch /ext-src
-COPY --from=pg-cron-pg-build /pg_cron.tar.gz /ext-src
-COPY patches/pg_cron.patch /ext-src
-#COPY --from=pg-pgx-ulid-build /home/nonroot/pgx_ulid.tar.gz /ext-src
-#COPY --from=rdkit-pg-build /rdkit.tar.gz /ext-src
-COPY --from=pg-uuidv7-pg-build /pg_uuidv7.tar.gz /ext-src
-COPY --from=pg-roaringbitmap-pg-build /pg_roaringbitmap.tar.gz /ext-src
-COPY --from=pg-semver-pg-build /pg_semver.tar.gz /ext-src
-#COPY --from=pg-embedding-pg-build /home/nonroot/pg_embedding-src/ /ext-src
-#COPY --from=wal2json-pg-build /wal2json_2_5.tar.gz /ext-src
-COPY --from=pg-anon-pg-build /pg_anon.tar.gz /ext-src
-COPY patches/pg_anon.patch /ext-src
-COPY --from=pg-ivm-build /pg_ivm.tar.gz /ext-src
-COPY --from=pg-partman-build /pg_partman.tar.gz /ext-src
-RUN cd /ext-src/ && for f in *.tar.gz; \
-    do echo $f; dname=$(echo $f | sed 's/\.tar.*//')-src; \
-    rm -rf $dname; mkdir $dname; tar xzf $f --strip-components=1 -C $dname \
-    || exit 1; rm -f $f; done
-RUN cd /ext-src/pgvector-src && patch -p1 <../pgvector.patch
-RUN cd /ext-src/rum-src && patch -p1 <../rum.patch
-# cmake is required for the h3 test
-RUN apt-get update && apt-get install -y cmake
-RUN cd /ext-src/pg_hint_plan-src && patch -p1 < /ext-src/pg_hint_plan.patch
-COPY --chmod=755 docker-compose/run-tests.sh /run-tests.sh
-RUN patch -p1 </ext-src/pg_anon.patch
-RUN patch -p1 </ext-src/pg_cron.patch
-ENV PATH=/usr/local/pgsql/bin:$PATH
-ENV PGHOST=compute
-ENV PGPORT=55433
-ENV PGUSER=cloud_admin
-ENV PGDATABASE=postgres
 #########################################################################################
 #
 # Final layer
@@ -1034,6 +983,6 @@ RUN apt update &&  \
     rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
     localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8
 
-ENV LANG=en_US.utf8
+ENV LANG en_US.utf8
 USER postgres
 ENTRYPOINT ["/usr/local/bin/compute_ctl"]

Makefile

@@ -3,9 +3,6 @@ ROOT_PROJECT_DIR := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
 # Where to install Postgres, default is ./pg_install, maybe useful for package managers
 POSTGRES_INSTALL_DIR ?= $(ROOT_PROJECT_DIR)/pg_install/
 
-OPENSSL_PREFIX_DIR := /usr/local/openssl
-ICU_PREFIX_DIR := /usr/local/icu
-
 #
 # We differentiate between release / debug build types using the BUILD_TYPE
 # environment variable.
@@ -23,31 +20,19 @@ else
 	$(error Bad build type '$(BUILD_TYPE)', see Makefile for options)
 endif
 
-ifeq ($(shell test -e /home/nonroot/.docker_build && echo -n yes),yes)
-	# Exclude static build openssl, icu for local build (MacOS, Linux)
-	# Only keep for build type release and debug
-	PG_CFLAGS += -I$(OPENSSL_PREFIX_DIR)/include
-	PG_CONFIGURE_OPTS += --with-icu
-	PG_CONFIGURE_OPTS += ICU_CFLAGS='-I/$(ICU_PREFIX_DIR)/include -DU_STATIC_IMPLEMENTATION'
-	PG_CONFIGURE_OPTS += ICU_LIBS='-L$(ICU_PREFIX_DIR)/lib -L$(ICU_PREFIX_DIR)/lib64 -licui18n -licuuc -licudata -lstdc++ -Wl,-Bdynamic -lm'
-	PG_CONFIGURE_OPTS += LDFLAGS='-L$(OPENSSL_PREFIX_DIR)/lib -L$(OPENSSL_PREFIX_DIR)/lib64 -L$(ICU_PREFIX_DIR)/lib -L$(ICU_PREFIX_DIR)/lib64 -Wl,-Bstatic -lssl -lcrypto -Wl,-Bdynamic -lrt -lm -ldl -lpthread'
-endif
-
 UNAME_S := $(shell uname -s)
 ifeq ($(UNAME_S),Linux)
 	# Seccomp BPF is only available for Linux
 	PG_CONFIGURE_OPTS += --with-libseccomp
 else ifeq ($(UNAME_S),Darwin)
-	ifndef DISABLE_HOMEBREW
-		# macOS with brew-installed openssl requires explicit paths
-		# It can be configured with OPENSSL_PREFIX variable
-		OPENSSL_PREFIX := $(shell brew --prefix openssl@3)
-		PG_CONFIGURE_OPTS += --with-includes=$(OPENSSL_PREFIX)/include --with-libraries=$(OPENSSL_PREFIX)/lib
-		PG_CONFIGURE_OPTS += PKG_CONFIG_PATH=$(shell brew --prefix icu4c)/lib/pkgconfig
-		# macOS already has bison and flex in the system, but they are old and result in postgres-v14 target failure
-		# brew formulae are keg-only and not symlinked into HOMEBREW_PREFIX, force their usage
-		EXTRA_PATH_OVERRIDES += $(shell brew --prefix bison)/bin/:$(shell brew --prefix flex)/bin/:
-	endif
+	# macOS with brew-installed openssl requires explicit paths
+	# It can be configured with OPENSSL_PREFIX variable
+	OPENSSL_PREFIX ?= $(shell brew --prefix openssl@3)
+	PG_CONFIGURE_OPTS += --with-includes=$(OPENSSL_PREFIX)/include --with-libraries=$(OPENSSL_PREFIX)/lib
+	PG_CONFIGURE_OPTS += PKG_CONFIG_PATH=$(shell brew --prefix icu4c)/lib/pkgconfig
+	# macOS already has bison and flex in the system, but they are old and result in postgres-v14 target failure
+	# brew formulae are keg-only and not symlinked into HOMEBREW_PREFIX, force their usage
+	EXTRA_PATH_OVERRIDES += $(shell brew --prefix bison)/bin/:$(shell brew --prefix flex)/bin/:
 endif
 
 # Use -C option so that when PostgreSQL "make install" installs the
@@ -69,8 +54,6 @@ CARGO_CMD_PREFIX += CARGO_TERM_PROGRESS_WHEN=never CI=1
 # Set PQ_LIB_DIR to make sure `storage_controller` get linked with bundled libpq (through diesel)
 CARGO_CMD_PREFIX += PQ_LIB_DIR=$(POSTGRES_INSTALL_DIR)/v16/lib
 
-CACHEDIR_TAG_CONTENTS := "Signature: 8a477f597d28d172789f06886806bc55"
-
 #
 # Top level Makefile to build Neon and PostgreSQL
 #
@@ -81,38 +64,26 @@ all: neon postgres neon-pg-ext
 #
 # The 'postgres_ffi' depends on the Postgres headers.
 .PHONY: neon
-neon: postgres-headers walproposer-lib cargo-target-dir
+neon: postgres-headers walproposer-lib
 	+@echo "Compiling Neon"
 	$(CARGO_CMD_PREFIX) cargo build $(CARGO_BUILD_FLAGS)
-.PHONY: cargo-target-dir
-cargo-target-dir:
-	# https://github.com/rust-lang/cargo/issues/14281
-	mkdir -p target
-	test -e target/CACHEDIR.TAG || echo "$(CACHEDIR_TAG_CONTENTS)" > target/CACHEDIR.TAG
 
 ### PostgreSQL parts
 # Some rules are duplicated for Postgres v14 and 15. We may want to refactor
 # to avoid the duplication in the future, but it's tolerable for now.
 #
 $(POSTGRES_INSTALL_DIR)/build/%/config.status:
-
-	mkdir -p $(POSTGRES_INSTALL_DIR)
-	test -e $(POSTGRES_INSTALL_DIR)/CACHEDIR.TAG || echo "$(CACHEDIR_TAG_CONTENTS)" > $(POSTGRES_INSTALL_DIR)/CACHEDIR.TAG
-
 	+@echo "Configuring Postgres $* build"
 	@test -s $(ROOT_PROJECT_DIR)/vendor/postgres-$*/configure || { \
 		echo "\nPostgres submodule not found in $(ROOT_PROJECT_DIR)/vendor/postgres-$*/, execute "; \
 		echo "'git submodule update --init --recursive --depth 2 --progress .' in project root.\n"; \
 		exit 1; }
 	mkdir -p $(POSTGRES_INSTALL_DIR)/build/$*
-
-	VERSION=$*; \
-	EXTRA_VERSION=$$(cd $(ROOT_PROJECT_DIR)/vendor/postgres-$$VERSION && git rev-parse HEAD); \
-	(cd $(POSTGRES_INSTALL_DIR)/build/$$VERSION && \
-	env PATH="$(EXTRA_PATH_OVERRIDES):$$PATH" $(ROOT_PROJECT_DIR)/vendor/postgres-$$VERSION/configure \
+	(cd $(POSTGRES_INSTALL_DIR)/build/$* && \
+	env PATH="$(EXTRA_PATH_OVERRIDES):$$PATH" $(ROOT_PROJECT_DIR)/vendor/postgres-$*/configure \
 		CFLAGS='$(PG_CFLAGS)' \
-		$(PG_CONFIGURE_OPTS) --with-extra-version=" ($$EXTRA_VERSION)" \
-		--prefix=$(abspath $(POSTGRES_INSTALL_DIR))/$$VERSION > configure.log)
+		$(PG_CONFIGURE_OPTS) \
+		--prefix=$(abspath $(POSTGRES_INSTALL_DIR))/$* > configure.log)
 
 # nicer alias to run 'configure'
 # Note: I've been unable to use templates for this part of our configuration.
@@ -148,8 +119,6 @@ postgres-%: postgres-configure-% \
 	$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pageinspect install
 	+@echo "Compiling amcheck $*"
 	$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/amcheck install
-	+@echo "Compiling test_decoding $*"
-	$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/test_decoding install
 
 .PHONY: postgres-clean-%
 postgres-clean-%:

README.md

@@ -1,6 +1,4 @@
-[![Neon](https://github.com/neondatabase/neon/assets/11527560/f15a17f0-836e-40c5-b35d-030606a6b660)](https://neon.tech)
-
-
+[![Neon](https://user-images.githubusercontent.com/13738772/236813940-dcfdcb5b-69d3-449b-a686-013febe834d4.png)](https://neon.tech)
 
 # Neon
 
@@ -126,7 +124,7 @@ make -j`sysctl -n hw.logicalcpu` -s
 To run the `psql` client, install the `postgresql-client` package or modify `PATH` and `LD_LIBRARY_PATH` to include `pg_install/bin` and `pg_install/lib`, respectively.
 
 To run the integration tests or Python scripts (not required to use the code), install
-Python (3.9 or higher), and install the python3 packages using `./scripts/pysync` (requires [poetry>=1.8](https://python-poetry.org/)) in the project directory.
+Python (3.9 or higher), and install the python3 packages using `./scripts/pysync` (requires [poetry>=1.3](https://python-poetry.org/)) in the project directory.
 
 
 #### Running neon database
@@ -262,7 +260,7 @@ By default, this runs both debug and release modes, and all supported postgres v
 testing locally, it is convenient to run just one set of permutations, like this:
 
 ```sh
-DEFAULT_PG_VERSION=16 BUILD_TYPE=release ./scripts/pytest
+DEFAULT_PG_VERSION=15 BUILD_TYPE=release ./scripts/pytest
 ```
 
 ## Flamegraphs

compute_tools/Cargo.toml

@@ -4,11 +4,6 @@ version = "0.1.0"
 edition.workspace = true
 license.workspace = true
 
-[features]
-default = []
-# Enables test specific features.
-testing = []
-
 [dependencies]
 anyhow.workspace = true
 async-compression.workspace = true
@@ -32,12 +27,10 @@ reqwest = { workspace = true, features = ["json"] }
 tokio = { workspace = true, features = ["rt", "rt-multi-thread"] }
 tokio-postgres.workspace = true
 tokio-util.workspace = true
-tokio-stream.workspace = true
 tracing.workspace = true
 tracing-opentelemetry.workspace = true
 tracing-subscriber.workspace = true
 tracing-utils.workspace = true
-thiserror.workspace = true
 url.workspace = true
 
 compute_api.workspace = true
@@ -49,4 +42,3 @@ vm_monitor = { version = "0.1", path = "../libs/vm_monitor/" }
 zstd = "0.13"
 bytes = "1.0"
 rust-ini = "0.20.0"
-rlimit = "0.10.1"

compute_tools/src/bin/compute_ctl.rs

@@ -6,7 +6,7 @@
 //! - Every start is a fresh start, so the data directory is removed and
 //!   initialized again on each run.
 //! - If remote_extension_config is provided, it will be used to fetch extensions list
-//!   and download `shared_preload_libraries` from the remote storage.
+//!  and download `shared_preload_libraries` from the remote storage.
 //! - Next it will put configuration files into the `PGDATA` directory.
 //! - Sync safekeepers and get commit LSN.
 //! - Get `basebackup` from pageserver using the returned on the previous step LSN.
@@ -33,6 +33,7 @@
 //!             -b /usr/local/bin/postgres \
 //!             -r http://pg-ext-s3-gateway \
 //! ```
+//!
 use std::collections::HashMap;
 use std::fs::File;
 use std::path::Path;
@@ -44,14 +45,12 @@ use std::{thread, time::Duration};
 use anyhow::{Context, Result};
 use chrono::Utc;
 use clap::Arg;
-use compute_tools::lsn_lease::launch_lsn_lease_bg_task_for_static;
 use signal_hook::consts::{SIGQUIT, SIGTERM};
 use signal_hook::{consts::SIGINT, iterator::Signals};
-use tracing::{error, info, warn};
+use tracing::{error, info};
 use url::Url;
 
 use compute_api::responses::ComputeStatus;
-use compute_api::spec::ComputeSpec;
 
 use compute_tools::compute::{
     forward_termination_signal, ComputeNode, ComputeState, ParsedSpec, PG_PID,
@@ -63,45 +62,12 @@ use compute_tools::logger::*;
 use compute_tools::monitor::launch_monitor;
 use compute_tools::params::*;
 use compute_tools::spec::*;
-use compute_tools::swap::resize_swap;
-use rlimit::{setrlimit, Resource};
 
 // this is an arbitrary build tag. Fine as a default / for testing purposes
 // in-case of not-set environment var
 const BUILD_TAG_DEFAULT: &str = "latest";
 
 fn main() -> Result<()> {
-    let (build_tag, clap_args) = init()?;
-
-    // enable core dumping for all child processes
-    setrlimit(Resource::CORE, rlimit::INFINITY, rlimit::INFINITY)?;
-
-    let (pg_handle, start_pg_result) = {
-        // Enter startup tracing context
-        let _startup_context_guard = startup_context_from_env();
-
-        let cli_args = process_cli(&clap_args)?;
-
-        let cli_spec = try_spec_from_cli(&clap_args, &cli_args)?;
-
-        let wait_spec_result = wait_spec(build_tag, cli_args, cli_spec)?;
-
-        start_postgres(&clap_args, wait_spec_result)?
-
-        // Startup is finished, exit the startup tracing span
-    };
-
-    // PostgreSQL is now running, if startup was successful. Wait until it exits.
-    let wait_pg_result = wait_postgres(pg_handle)?;
-
-    let delay_exit = cleanup_after_postgres_exit(start_pg_result)?;
-
-    maybe_delay_exit(delay_exit);
-
-    deinit_and_exit(wait_pg_result);
-}
-
-fn init() -> Result<(String, clap::ArgMatches)> {
     init_tracing_and_logging(DEFAULT_LOG_LEVEL)?;
 
     let mut signals = Signals::new([SIGINT, SIGTERM, SIGQUIT])?;
@@ -116,15 +82,9 @@ fn init() -> Result<(String, clap::ArgMatches)> {
         .to_string();
     info!("build_tag: {build_tag}");
 
-    Ok((build_tag, cli().get_matches()))
-}
-
-fn process_cli(matches: &clap::ArgMatches) -> Result<ProcessCliResult> {
-    let pgbin_default = "postgres";
-    let pgbin = matches
-        .get_one::<String>("pgbin")
-        .map(|s| s.as_str())
-        .unwrap_or(pgbin_default);
+    let matches = cli().get_matches();
+    let pgbin_default = String::from("postgres");
+    let pgbin = matches.get_one::<String>("pgbin").unwrap_or(&pgbin_default);
 
     let ext_remote_storage = matches
         .get_one::<String>("remote-ext-config")
@@ -150,32 +110,7 @@ fn process_cli(matches: &clap::ArgMatches) -> Result<ProcessCliResult> {
         .expect("Postgres connection string is required");
     let spec_json = matches.get_one::<String>("spec");
     let spec_path = matches.get_one::<String>("spec-path");
-    let resize_swap_on_bind = matches.get_flag("resize-swap-on-bind");
 
-    Ok(ProcessCliResult {
-        connstr,
-        pgdata,
-        pgbin,
-        ext_remote_storage,
-        http_port,
-        spec_json,
-        spec_path,
-        resize_swap_on_bind,
-    })
-}
-
-struct ProcessCliResult<'clap> {
-    connstr: &'clap str,
-    pgdata: &'clap str,
-    pgbin: &'clap str,
-    ext_remote_storage: Option<&'clap str>,
-    http_port: u16,
-    spec_json: Option<&'clap String>,
-    spec_path: Option<&'clap String>,
-    resize_swap_on_bind: bool,
-}
-
-fn startup_context_from_env() -> Option<opentelemetry::ContextGuard> {
     // Extract OpenTelemetry context for the startup actions from the
     // TRACEPARENT and TRACESTATE env variables, and attach it to the current
     // tracing context.
@@ -212,7 +147,7 @@ fn startup_context_from_env() -> Option<opentelemetry::ContextGuard> {
     if let Ok(val) = std::env::var("TRACESTATE") {
         startup_tracing_carrier.insert("tracestate".to_string(), val);
     }
-    if !startup_tracing_carrier.is_empty() {
+    let startup_context_guard = if !startup_tracing_carrier.is_empty() {
         use opentelemetry::propagation::TextMapPropagator;
         use opentelemetry::sdk::propagation::TraceContextPropagator;
         let guard = TraceContextPropagator::new()
@@ -222,17 +157,8 @@ fn startup_context_from_env() -> Option<opentelemetry::ContextGuard> {
         Some(guard)
     } else {
         None
-    }
-}
+    };
 
-fn try_spec_from_cli(
-    matches: &clap::ArgMatches,
-    ProcessCliResult {
-        spec_json,
-        spec_path,
-        ..
-    }: &ProcessCliResult,
-) -> Result<CliSpecParams> {
     let compute_id = matches.get_one::<String>("compute-id");
     let control_plane_uri = matches.get_one::<String>("control-plane-uri");
 
@@ -273,34 +199,6 @@ fn try_spec_from_cli(
         }
     };
 
-    Ok(CliSpecParams {
-        spec,
-        live_config_allowed,
-    })
-}
-
-struct CliSpecParams {
-    /// If a spec was provided via CLI or file, the [`ComputeSpec`]
-    spec: Option<ComputeSpec>,
-    live_config_allowed: bool,
-}
-
-fn wait_spec(
-    build_tag: String,
-    ProcessCliResult {
-        connstr,
-        pgdata,
-        pgbin,
-        ext_remote_storage,
-        resize_swap_on_bind,
-        http_port,
-        ..
-    }: ProcessCliResult,
-    CliSpecParams {
-        spec,
-        live_config_allowed,
-    }: CliSpecParams,
-) -> Result<WaitSpecResult> {
     let mut new_state = ComputeState::new();
     let spec_set;
 
@@ -328,17 +226,19 @@ fn wait_spec(
 
     // If this is a pooled VM, prewarm before starting HTTP server and becoming
     // available for binding. Prewarming helps Postgres start quicker later,
-    // because QEMU will already have its memory allocated from the host, and
+    // because QEMU will already have it's memory allocated from the host, and
     // the necessary binaries will already be cached.
     if !spec_set {
         compute.prewarm_postgres()?;
     }
 
-    // Launch http service first, so that we can serve control-plane requests
-    // while configuration is still in progress.
+    // Launch http service first, so we were able to serve control-plane
+    // requests, while configuration is still in progress.
     let _http_handle =
         launch_http_server(http_port, &compute).expect("cannot launch http endpoint thread");
 
+    let extension_server_port: u16 = http_port;
+
     if !spec_set {
         // No spec provided, hang waiting for it.
         info!("no compute spec provided, waiting");
@@ -353,47 +253,21 @@ fn wait_spec(
                 break;
             }
         }
-
-        // Record for how long we slept waiting for the spec.
-        let now = Utc::now();
-        state.metrics.wait_for_spec_ms = now
-            .signed_duration_since(state.start_time)
-            .to_std()
-            .unwrap()
-            .as_millis() as u64;
-
-        // Reset start time, so that the total startup time that is calculated later will
-        // not include the time that we waited for the spec.
-        state.start_time = now;
     }
 
-    launch_lsn_lease_bg_task_for_static(&compute);
-
-    Ok(WaitSpecResult {
-        compute,
-        http_port,
-        resize_swap_on_bind,
-    })
-}
-
-struct WaitSpecResult {
-    compute: Arc<ComputeNode>,
-    // passed through from ProcessCliResult
-    http_port: u16,
-    resize_swap_on_bind: bool,
-}
-
-fn start_postgres(
-    // need to allow unused because `matches` is only used if target_os = "linux"
-    #[allow(unused_variables)] matches: &clap::ArgMatches,
-    WaitSpecResult {
-        compute,
-        http_port,
-        resize_swap_on_bind,
-    }: WaitSpecResult,
-) -> Result<(Option<PostgresHandle>, StartPostgresResult)> {
     // We got all we need, update the state.
     let mut state = compute.state.lock().unwrap();
+
+    // Record for how long we slept waiting for the spec.
+    state.metrics.wait_for_spec_ms = Utc::now()
+        .signed_duration_since(state.start_time)
+        .to_std()
+        .unwrap()
+        .as_millis() as u64;
+    // Reset start time to the actual start of the configuration, so that
+    // total startup time was properly measured at the end.
+    state.start_time = Utc::now();
+
     state.status = ComputeStatus::Init;
     compute.state_changed.notify_all();
 
@@ -401,72 +275,33 @@ fn start_postgres(
         "running compute with features: {:?}",
         state.pspec.as_ref().unwrap().spec.features
     );
-    // before we release the mutex, fetch the swap size (if any) for later.
-    let swap_size_bytes = state.pspec.as_ref().unwrap().spec.swap_size_bytes;
     drop(state);
 
     // Launch remaining service threads
     let _monitor_handle = launch_monitor(&compute);
     let _configurator_handle = launch_configurator(&compute);
 
-    let mut prestartup_failed = false;
-    let mut delay_exit = false;
-
-    // Resize swap to the desired size if the compute spec says so
-    if let (Some(size_bytes), true) = (swap_size_bytes, resize_swap_on_bind) {
-        // To avoid 'swapoff' hitting postgres startup, we need to run resize-swap to completion
-        // *before* starting postgres.
-        //
-        // In theory, we could do this asynchronously if SkipSwapon was enabled for VMs, but this
-        // carries a risk of introducing hard-to-debug issues - e.g. if postgres sometimes gets
-        // OOM-killed during startup because swap wasn't available yet.
-        match resize_swap(size_bytes) {
-            Ok(()) => {
-                let size_gib = size_bytes as f32 / (1 << 20) as f32; // just for more coherent display.
-                info!(%size_bytes, %size_gib, "resized swap");
-            }
-            Err(err) => {
-                let err = err.context("failed to resize swap");
-                error!("{err:#}");
-
-                // Mark compute startup as failed; don't try to start postgres, and report this
-                // error to the control plane when it next asks.
-                prestartup_failed = true;
-                let mut state = compute.state.lock().unwrap();
-                state.error = Some(format!("{err:?}"));
-                state.status = ComputeStatus::Failed;
-                compute.state_changed.notify_all();
-                delay_exit = true;
-            }
-        }
-    }
-
-    let extension_server_port: u16 = http_port;
-
     // Start Postgres
-    let mut pg = None;
-    if !prestartup_failed {
-        pg = match compute.start_compute(extension_server_port) {
-            Ok(pg) => Some(pg),
-            Err(err) => {
-                error!("could not start the compute node: {:#}", err);
-                let mut state = compute.state.lock().unwrap();
-                state.error = Some(format!("{:?}", err));
-                state.status = ComputeStatus::Failed;
-                // Notify others that Postgres failed to start. In case of configuring the
-                // empty compute, it's likely that API handler is still waiting for compute
-                // state change. With this we will notify it that compute is in Failed state,
-                // so control plane will know about it earlier and record proper error instead
-                // of timeout.
-                compute.state_changed.notify_all();
-                drop(state); // unlock
-                delay_exit = true;
-                None
-            }
-        };
-    } else {
-        warn!("skipping postgres startup because pre-startup step failed");
-    }
+    let mut delay_exit = false;
+    let mut exit_code = None;
+    let pg = match compute.start_compute(extension_server_port) {
+        Ok(pg) => Some(pg),
+        Err(err) => {
+            error!("could not start the compute node: {:#}", err);
+            let mut state = compute.state.lock().unwrap();
+            state.error = Some(format!("{:?}", err));
+            state.status = ComputeStatus::Failed;
+            // Notify others that Postgres failed to start. In case of configuring the
+            // empty compute, it's likely that API handler is still waiting for compute
+            // state change. With this we will notify it that compute is in Failed state,
+            // so control plane will know about it earlier and record proper error instead
+            // of timeout.
+            compute.state_changed.notify_all();
+            drop(state); // unlock
+            delay_exit = true;
+            None
+        }
+    };
 
     // Start the vm-monitor if directed to. The vm-monitor only runs on linux
     // because it requires cgroups.
@@ -499,7 +334,7 @@ fn start_postgres(
             // This token is used internally by the monitor to clean up all threads
             let token = CancellationToken::new();
 
-            let vm_monitor = rt.as_ref().map(|rt| {
+            let vm_monitor = &rt.as_ref().map(|rt| {
                 rt.spawn(vm_monitor::start(
                     Box::leak(Box::new(vm_monitor::Args {
                         cgroup: cgroup.cloned(),
@@ -512,41 +347,12 @@ fn start_postgres(
         }
     }
 
-    Ok((
-        pg,
-        StartPostgresResult {
-            delay_exit,
-            compute,
-            #[cfg(target_os = "linux")]
-            rt,
-            #[cfg(target_os = "linux")]
-            token,
-            #[cfg(target_os = "linux")]
-            vm_monitor,
-        },
-    ))
-}
-
-type PostgresHandle = (std::process::Child, std::thread::JoinHandle<()>);
-
-struct StartPostgresResult {
-    delay_exit: bool,
-    // passed through from WaitSpecResult
-    compute: Arc<ComputeNode>,
-
-    #[cfg(target_os = "linux")]
-    rt: Option<tokio::runtime::Runtime>,
-    #[cfg(target_os = "linux")]
-    token: tokio_util::sync::CancellationToken,
-    #[cfg(target_os = "linux")]
-    vm_monitor: Option<tokio::task::JoinHandle<Result<()>>>,
-}
-
-fn wait_postgres(pg: Option<PostgresHandle>) -> Result<WaitPostgresResult> {
     // Wait for the child Postgres process forever. In this state Ctrl+C will
     // propagate to Postgres and it will be shut down as well.
-    let mut exit_code = None;
     if let Some((mut pg, logs_handle)) = pg {
+        // Startup is finished, exit the startup tracing span
+        drop(startup_context_guard);
+
         let ecode = pg
             .wait()
             .expect("failed to start waiting on Postgres process");
@@ -561,25 +367,6 @@ fn wait_postgres(pg: Option<PostgresHandle>) -> Result<WaitPostgresResult> {
         exit_code = ecode.code()
     }
 
-    Ok(WaitPostgresResult { exit_code })
-}
-
-struct WaitPostgresResult {
-    exit_code: Option<i32>,
-}
-
-fn cleanup_after_postgres_exit(
-    StartPostgresResult {
-        mut delay_exit,
-        compute,
-        #[cfg(target_os = "linux")]
-        vm_monitor,
-        #[cfg(target_os = "linux")]
-        token,
-        #[cfg(target_os = "linux")]
-        rt,
-    }: StartPostgresResult,
-) -> Result<bool> {
     // Terminate the vm_monitor so it releases the file watcher on
     // /sys/fs/cgroup/neon-postgres.
     // Note: the vm-monitor only runs on linux because it requires cgroups.
@@ -621,19 +408,13 @@ fn cleanup_after_postgres_exit(
         error!("error while checking for core dumps: {err:?}");
     }
 
-    Ok(delay_exit)
-}
-
-fn maybe_delay_exit(delay_exit: bool) {
     // If launch failed, keep serving HTTP requests for a while, so the cloud
     // control plane can get the actual error.
     if delay_exit {
         info!("giving control plane 30s to collect the error before shutdown");
         thread::sleep(Duration::from_secs(30));
     }
-}
 
-fn deinit_and_exit(WaitPostgresResult { exit_code }: WaitPostgresResult) -> ! {
     // Shutdown trace pipeline gracefully, so that it has a chance to send any
     // pending traces before we exit. Shutting down OTEL tracing provider may
     // hang for quite some time, see, for example:
@@ -741,15 +522,10 @@ fn cli() -> clap::Command {
             Arg::new("filecache-connstr")
                 .long("filecache-connstr")
                 .default_value(
-                    "host=localhost port=5432 dbname=postgres user=cloud_admin sslmode=disable application_name=vm-monitor",
+                    "host=localhost port=5432 dbname=postgres user=cloud_admin sslmode=disable",
                 )
                 .value_name("FILECACHE_CONNSTR"),
         )
-        .arg(
-            Arg::new("resize-swap-on-bind")
-                .long("resize-swap-on-bind")
-                .action(clap::ArgAction::SetTrue),
-        )
 }
 
 /// When compute_ctl is killed, send also termination signal to sync-safekeepers

compute_tools/src/catalog.rs

@@ -1,116 +0,0 @@
-use compute_api::{
-    responses::CatalogObjects,
-    spec::{Database, Role},
-};
-use futures::Stream;
-use postgres::{Client, NoTls};
-use std::{path::Path, process::Stdio, result::Result, sync::Arc};
-use tokio::{
-    io::{AsyncBufReadExt, BufReader},
-    process::Command,
-    task,
-};
-use tokio_stream::{self as stream, StreamExt};
-use tokio_util::codec::{BytesCodec, FramedRead};
-use tracing::warn;
-
-use crate::{
-    compute::ComputeNode,
-    pg_helpers::{get_existing_dbs, get_existing_roles},
-};
-
-pub async fn get_dbs_and_roles(compute: &Arc<ComputeNode>) -> anyhow::Result<CatalogObjects> {
-    let connstr = compute.connstr.clone();
-    task::spawn_blocking(move || {
-        let mut client = Client::connect(connstr.as_str(), NoTls)?;
-        let roles: Vec<Role>;
-        {
-            let mut xact = client.transaction()?;
-            roles = get_existing_roles(&mut xact)?;
-        }
-        let databases: Vec<Database> = get_existing_dbs(&mut client)?.values().cloned().collect();
-
-        Ok(CatalogObjects { roles, databases })
-    })
-    .await?
-}
-
-#[derive(Debug, thiserror::Error)]
-pub enum SchemaDumpError {
-    #[error("Database does not exist.")]
-    DatabaseDoesNotExist,
-    #[error("Failed to execute pg_dump.")]
-    IO(#[from] std::io::Error),
-}
-
-// It uses the pg_dump utility to dump the schema of the specified database.
-// The output is streamed back to the caller and supposed to be streamed via HTTP.
-//
-// Before return the result with the output, it checks that pg_dump produced any output.
-// If not, it tries to parse the stderr output to determine if the database does not exist
-// and special error is returned.
-//
-// To make sure that the process is killed when the caller drops the stream, we use tokio kill_on_drop feature.
-pub async fn get_database_schema(
-    compute: &Arc<ComputeNode>,
-    dbname: &str,
-) -> Result<impl Stream<Item = Result<bytes::Bytes, std::io::Error>>, SchemaDumpError> {
-    let pgbin = &compute.pgbin;
-    let basepath = Path::new(pgbin).parent().unwrap();
-    let pgdump = basepath.join("pg_dump");
-    let mut connstr = compute.connstr.clone();
-    connstr.set_path(dbname);
-    let mut cmd = Command::new(pgdump)
-        .arg("--schema-only")
-        .arg(connstr.as_str())
-        .stdout(Stdio::piped())
-        .stderr(Stdio::piped())
-        .kill_on_drop(true)
-        .spawn()?;
-
-    let stdout = cmd.stdout.take().ok_or_else(|| {
-        std::io::Error::new(std::io::ErrorKind::Other, "Failed to capture stdout.")
-    })?;
-
-    let stderr = cmd.stderr.take().ok_or_else(|| {
-        std::io::Error::new(std::io::ErrorKind::Other, "Failed to capture stderr.")
-    })?;
-
-    let mut stdout_reader = FramedRead::new(stdout, BytesCodec::new());
-    let stderr_reader = BufReader::new(stderr);
-
-    let first_chunk = match stdout_reader.next().await {
-        Some(Ok(bytes)) if !bytes.is_empty() => bytes,
-        Some(Err(e)) => {
-            return Err(SchemaDumpError::IO(e));
-        }
-        _ => {
-            let mut lines = stderr_reader.lines();
-            if let Some(line) = lines.next_line().await? {
-                if line.contains(&format!("FATAL:  database \"{}\" does not exist", dbname)) {
-                    return Err(SchemaDumpError::DatabaseDoesNotExist);
-                }
-                warn!("pg_dump stderr: {}", line)
-            }
-            tokio::spawn(async move {
-                while let Ok(Some(line)) = lines.next_line().await {
-                    warn!("pg_dump stderr: {}", line)
-                }
-            });
-
-            return Err(SchemaDumpError::IO(std::io::Error::new(
-                std::io::ErrorKind::Other,
-                "failed to start pg_dump",
-            )));
-        }
-    };
-    let initial_stream = stream::once(Ok(first_chunk.freeze()));
-    // Consume stderr and log warnings
-    tokio::spawn(async move {
-        let mut lines = stderr_reader.lines();
-        while let Ok(Some(line)) = lines.next_line().await {
-            warn!("pg_dump stderr: {}", line)
-        }
-    });
-    Ok(initial_stream.chain(stdout_reader.map(|res| res.map(|b| b.freeze()))))
-}

compute_tools/src/compute.rs

@@ -56,7 +56,6 @@ pub struct ComputeNode {
     /// - we push new spec and it does reconfiguration
     /// - but then something happens and compute pod / VM is destroyed,
     ///   so k8s controller starts it again with the **old** spec
-    ///
     /// and the same for empty computes:
     /// - we started compute without any spec
     /// - we push spec and it does configuration
@@ -400,15 +399,7 @@ impl ComputeNode {
     pub fn get_basebackup(&self, compute_state: &ComputeState, lsn: Lsn) -> Result<()> {
         let mut retry_period_ms = 500.0;
         let mut attempts = 0;
-        const DEFAULT_ATTEMPTS: u16 = 10;
-        #[cfg(feature = "testing")]
-        let max_attempts = if let Ok(v) = env::var("NEON_COMPUTE_TESTING_BASEBACKUP_RETRIES") {
-            u16::from_str(&v).unwrap()
-        } else {
-            DEFAULT_ATTEMPTS
-        };
-        #[cfg(not(feature = "testing"))]
-        let max_attempts = DEFAULT_ATTEMPTS;
+        let max_attempts = 10;
         loop {
             let result = self.try_get_basebackup(compute_state, lsn);
             match result {
@@ -807,11 +798,7 @@ impl ComputeNode {
         // In this case we need to connect with old `zenith_admin` name
         // and create new user. We cannot simply rename connected user,
         // but we can create a new one and grant it all privileges.
-        let mut connstr = self.connstr.clone();
-        connstr
-            .query_pairs_mut()
-            .append_pair("application_name", "apply_config");
-
+        let connstr = self.connstr.clone();
         let mut client = match Client::connect(connstr.as_str(), NoTls) {
             Err(e) => match e.code() {
                 Some(&SqlState::INVALID_PASSWORD)
@@ -831,15 +818,9 @@ impl ComputeNode {
                         Client::connect(zenith_admin_connstr.as_str(), NoTls)
                             .context("broken cloud_admin credential: tried connecting with cloud_admin but could not authenticate, and zenith_admin does not work either")?;
                     // Disable forwarding so that users don't get a cloud_admin role
-
-                    let mut func = || {
-                        client.simple_query("SET neon.forward_ddl = false")?;
-                        client.simple_query("CREATE USER cloud_admin WITH SUPERUSER")?;
-                        client.simple_query("GRANT zenith_admin TO cloud_admin")?;
-                        Ok::<_, anyhow::Error>(())
-                    };
-                    func().context("apply_config setup cloud_admin")?;
-
+                    client.simple_query("SET neon.forward_ddl = false")?;
+                    client.simple_query("CREATE USER cloud_admin WITH SUPERUSER")?;
+                    client.simple_query("GRANT zenith_admin TO cloud_admin")?;
                     drop(client);
 
                     // reconnect with connstring with expected name
@@ -851,48 +832,39 @@ impl ComputeNode {
         };
 
         // Disable DDL forwarding because control plane already knows about these roles/databases.
-        client
-            .simple_query("SET neon.forward_ddl = false")
-            .context("apply_config SET neon.forward_ddl = false")?;
+        client.simple_query("SET neon.forward_ddl = false")?;
 
         // Proceed with post-startup configuration. Note, that order of operations is important.
         let spec = &compute_state.pspec.as_ref().expect("spec must be set").spec;
-        create_neon_superuser(spec, &mut client).context("apply_config create_neon_superuser")?;
-        cleanup_instance(&mut client).context("apply_config cleanup_instance")?;
-        handle_roles(spec, &mut client).context("apply_config handle_roles")?;
-        handle_databases(spec, &mut client).context("apply_config handle_databases")?;
-        handle_role_deletions(spec, connstr.as_str(), &mut client)
-            .context("apply_config handle_role_deletions")?;
+        create_neon_superuser(spec, &mut client)?;
+        cleanup_instance(&mut client)?;
+        handle_roles(spec, &mut client)?;
+        handle_databases(spec, &mut client)?;
+        handle_role_deletions(spec, connstr.as_str(), &mut client)?;
         handle_grants(
             spec,
             &mut client,
             connstr.as_str(),
             self.has_feature(ComputeFeature::AnonExtension),
-        )
-        .context("apply_config handle_grants")?;
-        handle_extensions(spec, &mut client).context("apply_config handle_extensions")?;
-        handle_extension_neon(&mut client).context("apply_config handle_extension_neon")?;
-        create_availability_check_data(&mut client)
-            .context("apply_config create_availability_check_data")?;
+        )?;
+        handle_extensions(spec, &mut client)?;
+        handle_extension_neon(&mut client)?;
+        create_availability_check_data(&mut client)?;
 
         // 'Close' connection
         drop(client);
 
         // Run migrations separately to not hold up cold starts
         thread::spawn(move || {
-            let mut connstr = connstr.clone();
-            connstr
-                .query_pairs_mut()
-                .append_pair("application_name", "migrations");
-
             let mut client = Client::connect(connstr.as_str(), NoTls)?;
-            handle_migrations(&mut client).context("apply_config handle_migrations")
+            handle_migrations(&mut client)
         });
         Ok(())
     }
 
-    // Wrapped this around `pg_ctl reload`, but right now we don't use
-    // `pg_ctl` for start / stop.
+    // We could've wrapped this around `pg_ctl reload`, but right now we don't use
+    // `pg_ctl` for start / stop, so this just seems much easier to do as we already
+    // have opened connection to Postgres and superuser access.
     #[instrument(skip_all)]
     fn pg_reload_conf(&self) -> Result<()> {
         let pgctl_bin = Path::new(&self.pgbin).parent().unwrap().join("pg_ctl");
@@ -935,39 +907,38 @@ impl ComputeNode {
         // temporarily reset max_cluster_size in config
         // to avoid the possibility of hitting the limit, while we are reconfiguring:
         // creating new extensions, roles, etc...
-        config::with_compute_ctl_tmp_override(pgdata_path, "neon.max_cluster_size=-1", || {
-            self.pg_reload_conf()?;
+        config::compute_ctl_temp_override_create(pgdata_path, "neon.max_cluster_size=-1")?;
+        self.pg_reload_conf()?;
 
-            let mut client = Client::connect(self.connstr.as_str(), NoTls)?;
+        let mut client = Client::connect(self.connstr.as_str(), NoTls)?;
 
-            // Proceed with post-startup configuration. Note, that order of operations is important.
-            // Disable DDL forwarding because control plane already knows about these roles/databases.
-            if spec.mode == ComputeMode::Primary {
-                client.simple_query("SET neon.forward_ddl = false")?;
-                cleanup_instance(&mut client)?;
-                handle_roles(&spec, &mut client)?;
-                handle_databases(&spec, &mut client)?;
-                handle_role_deletions(&spec, self.connstr.as_str(), &mut client)?;
-                handle_grants(
-                    &spec,
-                    &mut client,
-                    self.connstr.as_str(),
-                    self.has_feature(ComputeFeature::AnonExtension),
-                )?;
-                handle_extensions(&spec, &mut client)?;
-                handle_extension_neon(&mut client)?;
-                // We can skip handle_migrations here because a new migration can only appear
-                // if we have a new version of the compute_ctl binary, which can only happen
-                // if compute got restarted, in which case we'll end up inside of apply_config
-                // instead of reconfigure.
-            }
+        // Proceed with post-startup configuration. Note, that order of operations is important.
+        // Disable DDL forwarding because control plane already knows about these roles/databases.
+        if spec.mode == ComputeMode::Primary {
+            client.simple_query("SET neon.forward_ddl = false")?;
+            cleanup_instance(&mut client)?;
+            handle_roles(&spec, &mut client)?;
+            handle_databases(&spec, &mut client)?;
+            handle_role_deletions(&spec, self.connstr.as_str(), &mut client)?;
+            handle_grants(
+                &spec,
+                &mut client,
+                self.connstr.as_str(),
+                self.has_feature(ComputeFeature::AnonExtension),
+            )?;
+            handle_extensions(&spec, &mut client)?;
+            handle_extension_neon(&mut client)?;
+            // We can skip handle_migrations here because a new migration can only appear
+            // if we have a new version of the compute_ctl binary, which can only happen
+            // if compute got restarted, in which case we'll end up inside of apply_config
+            // instead of reconfigure.
+        }
 
-            // 'Close' connection
-            drop(client);
-
-            Ok(())
-        })?;
+        // 'Close' connection
+        drop(client);
 
+        // reset max_cluster_size in config back to original value and reload config
+        config::compute_ctl_temp_override_remove(pgdata_path)?;
         self.pg_reload_conf()?;
 
         let unknown_op = "unknown".to_string();
@@ -1058,17 +1029,12 @@ impl ComputeNode {
                 // temporarily reset max_cluster_size in config
                 // to avoid the possibility of hitting the limit, while we are applying config:
                 // creating new extensions, roles, etc...
-                config::with_compute_ctl_tmp_override(
-                    pgdata_path,
-                    "neon.max_cluster_size=-1",
-                    || {
-                        self.pg_reload_conf()?;
+                config::compute_ctl_temp_override_create(pgdata_path, "neon.max_cluster_size=-1")?;
+                self.pg_reload_conf()?;
 
-                        self.apply_config(&compute_state)?;
+                self.apply_config(&compute_state)?;
 
-                        Ok(())
-                    },
-                )?;
+                config::compute_ctl_temp_override_remove(pgdata_path)?;
                 self.pg_reload_conf()?;
             }
             self.post_apply_config()?;
@@ -1125,7 +1091,7 @@ impl ComputeNode {
     // EKS worker nodes have following core dump settings:
     //   /proc/sys/kernel/core_pattern -> core
     //   /proc/sys/kernel/core_uses_pid -> 1
-    //   ulimit -c -> unlimited
+    //   ulimint -c -> unlimited
     // which results in core dumps being written to postgres data directory as core.<pid>.
     //
     // Use that as a default location and pattern, except macos where core dumps are written
@@ -1404,9 +1370,7 @@ pub fn forward_termination_signal() {
     let pg_pid = PG_PID.load(Ordering::SeqCst);
     if pg_pid != 0 {
         let pg_pid = nix::unistd::Pid::from_raw(pg_pid as i32);
-        // Use 'fast' shutdown (SIGINT) because it also creates a shutdown checkpoint, which is important for
-        // ROs to get a list of running xacts faster instead of going through the CLOG.
-        // See https://www.postgresql.org/docs/current/server-shutdown.html for the list of modes and signals.
-        kill(pg_pid, Signal::SIGINT).ok();
+        // use 'immediate' shutdown (SIGQUIT): https://www.postgresql.org/docs/current/server-shutdown.html
+        kill(pg_pid, Signal::SIGQUIT).ok();
     }
 }

compute_tools/src/config.rs

@@ -6,8 +6,8 @@ use std::path::Path;
 use anyhow::Result;
 
 use crate::pg_helpers::escape_conf_value;
-use crate::pg_helpers::{GenericOptionExt, PgOptionsSerialize};
-use compute_api::spec::{ComputeMode, ComputeSpec, GenericOption};
+use crate::pg_helpers::PgOptionsSerialize;
+use compute_api::spec::{ComputeMode, ComputeSpec};
 
 /// Check that `line` is inside a text file and put it there if it is not.
 /// Create file if it doesn't exist.
@@ -83,27 +83,12 @@ pub fn write_postgres_conf(
         ComputeMode::Replica => {
             // hot_standby is 'on' by default, but let's be explicit
             writeln!(file, "hot_standby=on")?;
-        }
-    }
 
-    if cfg!(target_os = "linux") {
-        // Check /proc/sys/vm/overcommit_memory -- if it equals 2 (i.e. linux memory overcommit is
-        // disabled), then the control plane has enabled swap and we should set
-        // dynamic_shared_memory_type = 'mmap'.
-        //
-        // This is (maybe?) temporary - for more, see https://github.com/neondatabase/cloud/issues/12047.
-        let overcommit_memory_contents = std::fs::read_to_string("/proc/sys/vm/overcommit_memory")
-            // ignore any errors - they may be expected to occur under certain situations (e.g. when
-            // not running in Linux).
-            .unwrap_or_else(|_| String::new());
-        if overcommit_memory_contents.trim() == "2" {
-            let opt = GenericOption {
-                name: "dynamic_shared_memory_type".to_owned(),
-                value: Some("mmap".to_owned()),
-                vartype: "enum".to_owned(),
-            };
-
-            write!(file, "{}", opt.to_pg_setting())?;
+            // Inform the replica about the primary state
+            // Default is 'false'
+            if let Some(primary_is_running) = spec.primary_is_running {
+                writeln!(file, "neon.primary_is_running={}", primary_is_running)?;
+            }
         }
     }
 
@@ -125,17 +110,18 @@ pub fn write_postgres_conf(
     Ok(())
 }
 
-pub fn with_compute_ctl_tmp_override<F>(pgdata_path: &Path, options: &str, exec: F) -> Result<()>
-where
-    F: FnOnce() -> Result<()>,
-{
+/// create file compute_ctl_temp_override.conf in pgdata_dir
+/// add provided options to this file
+pub fn compute_ctl_temp_override_create(pgdata_path: &Path, options: &str) -> Result<()> {
     let path = pgdata_path.join("compute_ctl_temp_override.conf");
     let mut file = File::create(path)?;
     write!(file, "{}", options)?;
-
-    let res = exec();
-
-    file.set_len(0)?;
-
-    res
+    Ok(())
+}
+
+/// remove file compute_ctl_temp_override.conf in pgdata_dir
+pub fn compute_ctl_temp_override_remove(pgdata_path: &Path) -> Result<()> {
+    let path = pgdata_path.join("compute_ctl_temp_override.conf");
+    std::fs::remove_file(path)?;
+    Ok(())
 }

compute_tools/src/http/api.rs

@@ -5,21 +5,17 @@ use std::net::SocketAddr;
 use std::sync::Arc;
 use std::thread;
 
-use crate::catalog::SchemaDumpError;
-use crate::catalog::{get_database_schema, get_dbs_and_roles};
 use crate::compute::forward_termination_signal;
 use crate::compute::{ComputeNode, ComputeState, ParsedSpec};
 use compute_api::requests::ConfigurationRequest;
 use compute_api::responses::{ComputeStatus, ComputeStatusResponse, GenericAPIError};
 
 use anyhow::Result;
-use hyper::header::CONTENT_TYPE;
 use hyper::service::{make_service_fn, service_fn};
 use hyper::{Body, Method, Request, Response, Server, StatusCode};
 use tokio::task;
-use tracing::{debug, error, info, warn};
+use tracing::{error, info, warn};
 use tracing_utils::http::OtelName;
-use utils::http::request::must_get_query_param;
 
 fn status_response_from_state(state: &ComputeState) -> ComputeStatusResponse {
     ComputeStatusResponse {
@@ -48,7 +44,7 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
     match (req.method(), req.uri().path()) {
         // Serialized compute state.
         (&Method::GET, "/status") => {
-            debug!("serving /status GET request");
+            info!("serving /status GET request");
             let state = compute.state.lock().unwrap();
             let status_response = status_response_from_state(&state);
             Response::new(Body::from(serde_json::to_string(&status_response).unwrap()))
@@ -137,34 +133,6 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
             }
         }
 
-        (&Method::GET, "/dbs_and_roles") => {
-            info!("serving /dbs_and_roles GET request",);
-            match get_dbs_and_roles(compute).await {
-                Ok(res) => render_json(Body::from(serde_json::to_string(&res).unwrap())),
-                Err(_) => {
-                    render_json_error("can't get dbs and roles", StatusCode::INTERNAL_SERVER_ERROR)
-                }
-            }
-        }
-
-        (&Method::GET, "/database_schema") => {
-            let database = match must_get_query_param(&req, "database") {
-                Err(e) => return e.into_response(),
-                Ok(database) => database,
-            };
-            info!("serving /database_schema GET request with database: {database}",);
-            match get_database_schema(compute, &database).await {
-                Ok(res) => render_plain(Body::wrap_stream(res)),
-                Err(SchemaDumpError::DatabaseDoesNotExist) => {
-                    render_json_error("database does not exist", StatusCode::NOT_FOUND)
-                }
-                Err(e) => {
-                    error!("can't get schema dump: {}", e);
-                    render_json_error("can't get schema dump", StatusCode::INTERNAL_SERVER_ERROR)
-                }
-            }
-        }
-
         // download extension files from remote extension storage on demand
         (&Method::POST, route) if route.starts_with("/extension_server/") => {
             info!("serving {:?} POST request", route);
@@ -335,25 +303,10 @@ fn render_json_error(e: &str, status: StatusCode) -> Response<Body> {
     };
     Response::builder()
         .status(status)
-        .header(CONTENT_TYPE, "application/json")
         .body(Body::from(serde_json::to_string(&error).unwrap()))
         .unwrap()
 }
 
-fn render_json(body: Body) -> Response<Body> {
-    Response::builder()
-        .header(CONTENT_TYPE, "application/json")
-        .body(body)
-        .unwrap()
-}
-
-fn render_plain(body: Body) -> Response<Body> {
-    Response::builder()
-        .header(CONTENT_TYPE, "text/plain")
-        .body(body)
-        .unwrap()
-}
-
 async fn handle_terminate_request(compute: &Arc<ComputeNode>) -> Result<(), (String, StatusCode)> {
     {
         let mut state = compute.state.lock().unwrap();

compute_tools/src/http/openapi_spec.yaml

@@ -68,51 +68,6 @@ paths:
               schema:
                 $ref: "#/components/schemas/Info"
 
-  /dbs_and_roles:
-    get:
-      tags:
-        - Info
-      summary: Get databases and roles in the catalog.
-      description: ""
-      operationId: getDbsAndRoles
-      responses:
-        200:
-          description: Compute schema objects
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/DbsAndRoles"
-
-  /database_schema:
-    get:
-      tags:
-        - Info
-      summary: Get schema dump
-      parameters:
-        - name: database
-          in: query
-          description: Database name to dump.
-          required: true
-          schema:
-            type: string
-          example: "postgres"
-      description: Get schema dump in SQL format.
-      operationId: getDatabaseSchema
-      responses:
-        200:
-          description: Schema dump
-          content:
-            text/plain:
-              schema:
-                type: string
-                description: Schema dump in SQL format.
-        404:
-          description: Non existing database.
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/GenericError"
-
   /check_writability:
     post:
       tags:
@@ -274,73 +229,6 @@ components:
         num_cpus:
           type: integer
 
-    DbsAndRoles:
-      type: object
-      description: Databases and Roles
-      required:
-        - roles
-        - databases
-      properties:
-        roles:
-          type: array
-          items:
-            $ref: "#/components/schemas/Role"
-        databases:
-          type: array
-          items:
-            $ref: "#/components/schemas/Database"
-
-    Database:
-      type: object
-      description: Database
-      required:
-        - name
-        - owner
-        - restrict_conn
-        - invalid
-      properties:
-        name:
-          type: string
-        owner:
-          type: string
-        options:
-          type: array
-          items:
-            $ref: "#/components/schemas/GenericOption"
-        restrict_conn:
-          type: boolean
-        invalid:
-          type: boolean
-
-    Role:
-      type: object
-      description: Role
-      required:
-        - name
-      properties:
-        name:
-          type: string
-        encrypted_password:
-          type: string
-        options:
-          type: array
-          items:
-            $ref: "#/components/schemas/GenericOption"
-
-    GenericOption:
-      type: object
-      description: Schema Generic option
-      required:
-        - name
-        - vartype
-      properties:
-        name:
-          type: string
-        value:
-          type: string
-        vartype:
-          type: string
-
     ComputeState:
       type: object
       required:

compute_tools/src/lib.rs

@@ -8,14 +8,10 @@ pub mod configurator;
 pub mod http;
 #[macro_use]
 pub mod logger;
-pub mod catalog;
 pub mod compute;
 pub mod extension_server;
-pub mod lsn_lease;
-mod migration;
 pub mod monitor;
 pub mod params;
 pub mod pg_helpers;
 pub mod spec;
-pub mod swap;
 pub mod sync_sk;

compute_tools/src/lsn_lease.rs

@@ -1,186 +0,0 @@
-use anyhow::bail;
-use anyhow::Result;
-use postgres::{NoTls, SimpleQueryMessage};
-use std::time::SystemTime;
-use std::{str::FromStr, sync::Arc, thread, time::Duration};
-use utils::id::TenantId;
-use utils::id::TimelineId;
-
-use compute_api::spec::ComputeMode;
-use tracing::{info, warn};
-use utils::{
-    lsn::Lsn,
-    shard::{ShardCount, ShardNumber, TenantShardId},
-};
-
-use crate::compute::ComputeNode;
-
-/// Spawns a background thread to periodically renew LSN leases for static compute.
-/// Do nothing if the compute is not in static mode.
-pub fn launch_lsn_lease_bg_task_for_static(compute: &Arc<ComputeNode>) {
-    let (tenant_id, timeline_id, lsn) = {
-        let state = compute.state.lock().unwrap();
-        let spec = state.pspec.as_ref().expect("Spec must be set");
-        match spec.spec.mode {
-            ComputeMode::Static(lsn) => (spec.tenant_id, spec.timeline_id, lsn),
-            _ => return,
-        }
-    };
-    let compute = compute.clone();
-
-    let span = tracing::info_span!("lsn_lease_bg_task", %tenant_id, %timeline_id, %lsn);
-    thread::spawn(move || {
-        let _entered = span.entered();
-        if let Err(e) = lsn_lease_bg_task(compute, tenant_id, timeline_id, lsn) {
-            // TODO: might need stronger error feedback than logging an warning.
-            warn!("Exited with error: {e}");
-        }
-    });
-}
-
-/// Renews lsn lease periodically so static compute are not affected by GC.
-fn lsn_lease_bg_task(
-    compute: Arc<ComputeNode>,
-    tenant_id: TenantId,
-    timeline_id: TimelineId,
-    lsn: Lsn,
-) -> Result<()> {
-    loop {
-        let valid_until = acquire_lsn_lease_with_retry(&compute, tenant_id, timeline_id, lsn)?;
-        let valid_duration = valid_until
-            .duration_since(SystemTime::now())
-            .unwrap_or(Duration::ZERO);
-
-        // Sleep for 60 seconds less than the valid duration but no more than half of the valid duration.
-        let sleep_duration = valid_duration
-            .saturating_sub(Duration::from_secs(60))
-            .max(valid_duration / 2);
-
-        info!(
-            "Succeeded, sleeping for {} seconds",
-            sleep_duration.as_secs()
-        );
-        thread::sleep(sleep_duration);
-    }
-}
-
-/// Acquires lsn lease in a retry loop. Returns the expiration time if a lease is granted.
-/// Returns an error if a lease is explicitly not granted. Otherwise, we keep sending requests.
-fn acquire_lsn_lease_with_retry(
-    compute: &Arc<ComputeNode>,
-    tenant_id: TenantId,
-    timeline_id: TimelineId,
-    lsn: Lsn,
-) -> Result<SystemTime> {
-    let mut attempts = 0usize;
-    let mut retry_period_ms: f64 = 500.0;
-    const MAX_RETRY_PERIOD_MS: f64 = 60.0 * 1000.0;
-
-    loop {
-        // Note: List of pageservers is dynamic, need to re-read configs before each attempt.
-        let configs = {
-            let state = compute.state.lock().unwrap();
-
-            let spec = state.pspec.as_ref().expect("spec must be set");
-
-            let conn_strings = spec.pageserver_connstr.split(',');
-
-            conn_strings
-                .map(|connstr| {
-                    let mut config = postgres::Config::from_str(connstr).expect("Invalid connstr");
-                    if let Some(storage_auth_token) = &spec.storage_auth_token {
-                        info!("Got storage auth token from spec file");
-                        config.password(storage_auth_token.clone());
-                    } else {
-                        info!("Storage auth token not set");
-                    }
-                    config
-                })
-                .collect::<Vec<_>>()
-        };
-
-        let result = try_acquire_lsn_lease(tenant_id, timeline_id, lsn, &configs);
-        match result {
-            Ok(Some(res)) => {
-                return Ok(res);
-            }
-            Ok(None) => {
-                bail!("Permanent error: lease could not be obtained, LSN is behind the GC cutoff");
-            }
-            Err(e) => {
-                warn!("Failed to acquire lsn lease: {e} (attempt {attempts}");
-
-                thread::sleep(Duration::from_millis(retry_period_ms as u64));
-                retry_period_ms *= 1.5;
-                retry_period_ms = retry_period_ms.min(MAX_RETRY_PERIOD_MS);
-            }
-        }
-        attempts += 1;
-    }
-}
-
-/// Tries to acquire an LSN lease through PS page_service API.
-fn try_acquire_lsn_lease(
-    tenant_id: TenantId,
-    timeline_id: TimelineId,
-    lsn: Lsn,
-    configs: &[postgres::Config],
-) -> Result<Option<SystemTime>> {
-    fn get_valid_until(
-        config: &postgres::Config,
-        tenant_shard_id: TenantShardId,
-        timeline_id: TimelineId,
-        lsn: Lsn,
-    ) -> Result<Option<SystemTime>> {
-        let mut client = config.connect(NoTls)?;
-        let cmd = format!("lease lsn {} {} {} ", tenant_shard_id, timeline_id, lsn);
-        let res = client.simple_query(&cmd)?;
-        let msg = match res.first() {
-            Some(msg) => msg,
-            None => bail!("empty response"),
-        };
-        let row = match msg {
-            SimpleQueryMessage::Row(row) => row,
-            _ => bail!("error parsing lsn lease response"),
-        };
-
-        // Note: this will be None if a lease is explicitly not granted.
-        let valid_until_str = row.get("valid_until");
-
-        let valid_until = valid_until_str.map(|s| {
-            SystemTime::UNIX_EPOCH
-                .checked_add(Duration::from_millis(u128::from_str(s).unwrap() as u64))
-                .expect("Time larger than max SystemTime could handle")
-        });
-        Ok(valid_until)
-    }
-
-    let shard_count = configs.len();
-
-    let valid_until = if shard_count > 1 {
-        configs
-            .iter()
-            .enumerate()
-            .map(|(shard_number, config)| {
-                let tenant_shard_id = TenantShardId {
-                    tenant_id,
-                    shard_count: ShardCount::new(shard_count as u8),
-                    shard_number: ShardNumber(shard_number as u8),
-                };
-                get_valid_until(config, tenant_shard_id, timeline_id, lsn)
-            })
-            .collect::<Result<Vec<Option<SystemTime>>>>()?
-            .into_iter()
-            .min()
-            .unwrap()
-    } else {
-        get_valid_until(
-            &configs[0],
-            TenantShardId::unsharded(tenant_id),
-            timeline_id,
-            lsn,
-        )?
-    };
-
-    Ok(valid_until)
-}

compute_tools/src/migration.rs

@@ -1,105 +0,0 @@
-use anyhow::{Context, Result};
-use postgres::Client;
-use tracing::info;
-
-pub(crate) struct MigrationRunner<'m> {
-    client: &'m mut Client,
-    migrations: &'m [&'m str],
-}
-
-impl<'m> MigrationRunner<'m> {
-    pub fn new(client: &'m mut Client, migrations: &'m [&'m str]) -> Self {
-        // The neon_migration.migration_id::id column is a bigint, which is equivalent to an i64
-        assert!(migrations.len() + 1 < i64::MAX as usize);
-
-        Self { client, migrations }
-    }
-
-    fn get_migration_id(&mut self) -> Result<i64> {
-        let query = "SELECT id FROM neon_migration.migration_id";
-        let row = self
-            .client
-            .query_one(query, &[])
-            .context("run_migrations get migration_id")?;
-
-        Ok(row.get::<&str, i64>("id"))
-    }
-
-    fn update_migration_id(&mut self, migration_id: i64) -> Result<()> {
-        let setval = format!("UPDATE neon_migration.migration_id SET id={}", migration_id);
-
-        self.client
-            .simple_query(&setval)
-            .context("run_migrations update id")?;
-
-        Ok(())
-    }
-
-    fn prepare_migrations(&mut self) -> Result<()> {
-        let query = "CREATE SCHEMA IF NOT EXISTS neon_migration";
-        self.client.simple_query(query)?;
-
-        let query = "CREATE TABLE IF NOT EXISTS neon_migration.migration_id (key INT NOT NULL PRIMARY KEY, id bigint NOT NULL DEFAULT 0)";
-        self.client.simple_query(query)?;
-
-        let query = "INSERT INTO neon_migration.migration_id VALUES (0, 0) ON CONFLICT DO NOTHING";
-        self.client.simple_query(query)?;
-
-        let query = "ALTER SCHEMA neon_migration OWNER TO cloud_admin";
-        self.client.simple_query(query)?;
-
-        let query = "REVOKE ALL ON SCHEMA neon_migration FROM PUBLIC";
-        self.client.simple_query(query)?;
-
-        Ok(())
-    }
-
-    pub fn run_migrations(mut self) -> Result<()> {
-        self.prepare_migrations()?;
-
-        let mut current_migration = self.get_migration_id()? as usize;
-        while current_migration < self.migrations.len() {
-            macro_rules! migration_id {
-                ($cm:expr) => {
-                    ($cm + 1) as i64
-                };
-            }
-
-            let migration = self.migrations[current_migration];
-
-            if migration.starts_with("-- SKIP") {
-                info!("Skipping migration id={}", migration_id!(current_migration));
-            } else {
-                info!(
-                    "Running migration id={}:\n{}\n",
-                    migration_id!(current_migration),
-                    migration
-                );
-
-                self.client
-                    .simple_query("BEGIN")
-                    .context("begin migration")?;
-
-                self.client.simple_query(migration).with_context(|| {
-                    format!(
-                        "run_migrations migration id={}",
-                        migration_id!(current_migration)
-                    )
-                })?;
-
-                // Migration IDs start at 1
-                self.update_migration_id(migration_id!(current_migration))?;
-
-                self.client
-                    .simple_query("COMMIT")
-                    .context("commit migration")?;
-
-                info!("Finished migration id={}", migration_id!(current_migration));
-            }
-
-            current_migration += 1;
-        }
-
-        Ok(())
-    }
-}

compute_tools/src/migrations/0001-neon_superuser_bypass_rls.sql

@@ -1 +0,0 @@
-ALTER ROLE neon_superuser BYPASSRLS;

compute_tools/src/migrations/0002-alter_roles.sql

@@ -1,18 +0,0 @@
-DO $$
-DECLARE
-    role_name text;
-BEGIN
-    FOR role_name IN SELECT rolname FROM pg_roles WHERE pg_has_role(rolname, 'neon_superuser', 'member')
-    LOOP
-        RAISE NOTICE 'EXECUTING ALTER ROLE % INHERIT', quote_ident(role_name);
-        EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' INHERIT';
-    END LOOP;
-
-    FOR role_name IN SELECT rolname FROM pg_roles
-        WHERE
-            NOT pg_has_role(rolname, 'neon_superuser', 'member') AND NOT starts_with(rolname, 'pg_')
-    LOOP
-        RAISE NOTICE 'EXECUTING ALTER ROLE % NOBYPASSRLS', quote_ident(role_name);
-        EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' NOBYPASSRLS';
-    END LOOP;
-END $$;

compute_tools/src/migrations/0003-grant_pg_create_subscription_to_neon_superuser.sql

@@ -1,6 +0,0 @@
-DO $$
-BEGIN
-    IF (SELECT setting::numeric >= 160000 FROM pg_settings WHERE name = 'server_version_num') THEN
-        EXECUTE 'GRANT pg_create_subscription TO neon_superuser';
-    END IF;
-END $$;

compute_tools/src/migrations/0004-grant_pg_monitor_to_neon_superuser.sql

@@ -1 +0,0 @@
-GRANT pg_monitor TO neon_superuser WITH ADMIN OPTION;

compute_tools/src/migrations/0005-grant_all_on_tables_to_neon_superuser.sql

@@ -1,4 +0,0 @@
--- SKIP: Deemed insufficient for allowing relations created by extensions to be
---       interacted with by neon_superuser without permission issues.
-
-ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON TABLES TO neon_superuser;

compute_tools/src/migrations/0006-grant_all_on_sequences_to_neon_superuser.sql

@@ -1,4 +0,0 @@
--- SKIP: Deemed insufficient for allowing relations created by extensions to be
---       interacted with by neon_superuser without permission issues.
-
-ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON SEQUENCES TO neon_superuser;

compute_tools/src/migrations/0007-grant_all_on_tables_to_neon_superuser_with_grant_option.sql

@@ -1,3 +0,0 @@
--- SKIP: Moved inline to the handle_grants() functions.
-
-ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON TABLES TO neon_superuser WITH GRANT OPTION;

compute_tools/src/migrations/0008-grant_all_on_sequences_to_neon_superuser_with_grant_option.sql

@@ -1,3 +0,0 @@
--- SKIP: Moved inline to the handle_grants() functions.
-
-ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON SEQUENCES TO neon_superuser WITH GRANT OPTION;

compute_tools/src/migrations/0009-revoke_replication_for_previously_allowed_roles.sql

@@ -1,13 +0,0 @@
--- SKIP: The original goal of this migration was to prevent creating
---       subscriptions, but this migration was insufficient.
-
-DO $$
-DECLARE
-    role_name TEXT;
-BEGIN
-    FOR role_name IN SELECT rolname FROM pg_roles WHERE rolreplication IS TRUE
-    LOOP
-        RAISE NOTICE 'EXECUTING ALTER ROLE % NOREPLICATION', quote_ident(role_name);
-        EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' NOREPLICATION';
-    END LOOP;
-END $$;

compute_tools/src/migrations/0010-grant_snapshot_synchronization_funcs_to_neon_superuser.sql

@@ -1,7 +0,0 @@
-DO $$
-BEGIN
-    IF (SELECT setting::numeric >= 160000 FROM pg_settings WHERE name = 'server_version_num') THEN
-       EXECUTE 'GRANT EXECUTE ON FUNCTION pg_export_snapshot TO neon_superuser';
-       EXECUTE 'GRANT EXECUTE ON FUNCTION pg_log_standby_snapshot TO neon_superuser';
-    END IF;
-END $$;

compute_tools/src/monitor.rs

@@ -17,11 +17,7 @@ const MONITOR_CHECK_INTERVAL: Duration = Duration::from_millis(500);
 // should be handled gracefully.
 fn watch_compute_activity(compute: &ComputeNode) {
     // Suppose that `connstr` doesn't change
-    let mut connstr = compute.connstr.clone();
-    connstr
-        .query_pairs_mut()
-        .append_pair("application_name", "compute_activity_monitor");
-    let connstr = connstr.as_str();
+    let connstr = compute.connstr.as_str();
 
     // During startup and configuration we connect to every Postgres database,
     // but we don't want to count this as some user activity. So wait until

compute_tools/src/pg_helpers.rs

@@ -44,7 +44,7 @@ pub fn escape_conf_value(s: &str) -> String {
     format!("'{}'", res)
 }
 
-pub trait GenericOptionExt {
+trait GenericOptionExt {
     fn to_pg_option(&self) -> String;
     fn to_pg_setting(&self) -> String;
 }
@@ -489,7 +489,7 @@ pub fn handle_postgres_logs(stderr: std::process::ChildStderr) -> JoinHandle<()>
 /// Read Postgres logs from `stderr` until EOF. Buffer is flushed on one of the following conditions:
 /// - next line starts with timestamp
 /// - EOF
-/// - no new lines were written for the last 100 milliseconds
+/// - no new lines were written for the last second
 async fn handle_postgres_logs_async(stderr: tokio::process::ChildStderr) -> Result<()> {
     let mut lines = tokio::io::BufReader::new(stderr).lines();
     let timeout_duration = Duration::from_millis(100);

compute_tools/src/spec.rs

@@ -2,7 +2,7 @@ use std::fs::File;
 use std::path::Path;
 use std::str::FromStr;
 
-use anyhow::{anyhow, bail, Context, Result};
+use anyhow::{anyhow, bail, Result};
 use postgres::config::Config;
 use postgres::{Client, NoTls};
 use reqwest::StatusCode;
@@ -10,7 +10,6 @@ use tracing::{error, info, info_span, instrument, span_enabled, warn, Level};
 
 use crate::config;
 use crate::logger::inlinify;
-use crate::migration::MigrationRunner;
 use crate::params::PG_HBA_ALL_MD5;
 use crate::pg_helpers::*;
 
@@ -491,7 +490,7 @@ pub fn handle_databases(spec: &ComputeSpec, client: &mut Client) -> Result<()> {
                 "rename_db" => {
                     let new_name = op.new_name.as_ref().unwrap();
 
-                    if existing_dbs.contains_key(&op.name) {
+                    if existing_dbs.get(&op.name).is_some() {
                         let query: String = format!(
                             "ALTER DATABASE {} RENAME TO {}",
                             op.name.pg_quote(),
@@ -699,8 +698,7 @@ pub fn handle_grants(
 
         // it is important to run this after all grants
         if enable_anon_extension {
-            handle_extension_anon(spec, &db.owner, &mut db_client, false)
-                .context("handle_grants handle_extension_anon")?;
+            handle_extension_anon(spec, &db.owner, &mut db_client, false)?;
         }
     }
 
@@ -775,27 +773,92 @@ pub fn handle_migrations(client: &mut Client) -> Result<()> {
     // !BE SURE TO ONLY ADD MIGRATIONS TO THE END OF THIS ARRAY. IF YOU DO NOT, VERY VERY BAD THINGS MAY HAPPEN!
     // !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 
-    // Add new migrations in numerical order.
     let migrations = [
-        include_str!("./migrations/0001-neon_superuser_bypass_rls.sql"),
-        include_str!("./migrations/0002-alter_roles.sql"),
-        include_str!("./migrations/0003-grant_pg_create_subscription_to_neon_superuser.sql"),
-        include_str!("./migrations/0004-grant_pg_monitor_to_neon_superuser.sql"),
-        include_str!("./migrations/0005-grant_all_on_tables_to_neon_superuser.sql"),
-        include_str!("./migrations/0006-grant_all_on_sequences_to_neon_superuser.sql"),
-        include_str!(
-            "./migrations/0007-grant_all_on_tables_to_neon_superuser_with_grant_option.sql"
-        ),
-        include_str!(
-            "./migrations/0008-grant_all_on_sequences_to_neon_superuser_with_grant_option.sql"
-        ),
-        include_str!("./migrations/0009-revoke_replication_for_previously_allowed_roles.sql"),
-        include_str!(
-            "./migrations/0010-grant_snapshot_synchronization_funcs_to_neon_superuser.sql"
-        ),
+        "ALTER ROLE neon_superuser BYPASSRLS",
+        r#"
+DO $$
+DECLARE
+    role_name text;
+BEGIN
+    FOR role_name IN SELECT rolname FROM pg_roles WHERE pg_has_role(rolname, 'neon_superuser', 'member')
+    LOOP
+        RAISE NOTICE 'EXECUTING ALTER ROLE % INHERIT', quote_ident(role_name);
+        EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' INHERIT';
+    END LOOP;
+
+    FOR role_name IN SELECT rolname FROM pg_roles
+        WHERE
+            NOT pg_has_role(rolname, 'neon_superuser', 'member') AND NOT starts_with(rolname, 'pg_')
+    LOOP
+        RAISE NOTICE 'EXECUTING ALTER ROLE % NOBYPASSRLS', quote_ident(role_name);
+        EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' NOBYPASSRLS';
+    END LOOP;
+END $$;
+"#,
+        r#"
+DO $$
+BEGIN
+    IF (SELECT setting::numeric >= 160000 FROM pg_settings WHERE name = 'server_version_num') THEN
+        EXECUTE 'GRANT pg_create_subscription TO neon_superuser';
+    END IF;
+END
+$$;"#,
+        "GRANT pg_monitor TO neon_superuser WITH ADMIN OPTION",
+        // Don't remove: these are some SQLs that we originally applied in migrations but turned out to execute somewhere else.
+        "",
+        "",
+        "",
+        "",
+        "",
+        // Add new migrations below.
     ];
 
-    MigrationRunner::new(client, &migrations).run_migrations()?;
+    let mut query = "CREATE SCHEMA IF NOT EXISTS neon_migration";
+    client.simple_query(query)?;
+
+    query = "CREATE TABLE IF NOT EXISTS neon_migration.migration_id (key INT NOT NULL PRIMARY KEY, id bigint NOT NULL DEFAULT 0)";
+    client.simple_query(query)?;
+
+    query = "INSERT INTO neon_migration.migration_id VALUES (0, 0) ON CONFLICT DO NOTHING";
+    client.simple_query(query)?;
+
+    query = "ALTER SCHEMA neon_migration OWNER TO cloud_admin";
+    client.simple_query(query)?;
+
+    query = "REVOKE ALL ON SCHEMA neon_migration FROM PUBLIC";
+    client.simple_query(query)?;
+
+    query = "SELECT id FROM neon_migration.migration_id";
+    let row = client.query_one(query, &[])?;
+    let mut current_migration: usize = row.get::<&str, i64>("id") as usize;
+    let starting_migration_id = current_migration;
+
+    query = "BEGIN";
+    client.simple_query(query)?;
+
+    while current_migration < migrations.len() {
+        let migration = &migrations[current_migration];
+        if migration.is_empty() {
+            info!("Skip migration id={}", current_migration);
+        } else {
+            info!("Running migration:\n{}\n", migration);
+            client.simple_query(migration)?;
+        }
+        current_migration += 1;
+    }
+    let setval = format!(
+        "UPDATE neon_migration.migration_id SET id={}",
+        migrations.len()
+    );
+    client.simple_query(&setval)?;
+
+    query = "COMMIT";
+    client.simple_query(query)?;
+
+    info!(
+        "Ran {} migrations",
+        (migrations.len() - starting_migration_id)
+    );
 
     Ok(())
 }

compute_tools/src/swap.rs

@@ -1,45 +0,0 @@
-use std::path::Path;
-
-use anyhow::{anyhow, Context};
-use tracing::warn;
-
-pub const RESIZE_SWAP_BIN: &str = "/neonvm/bin/resize-swap";
-
-pub fn resize_swap(size_bytes: u64) -> anyhow::Result<()> {
-    // run `/neonvm/bin/resize-swap --once {size_bytes}`
-    //
-    // Passing '--once' causes resize-swap to delete itself after successful completion, which
-    // means that if compute_ctl restarts later, we won't end up calling 'swapoff' while
-    // postgres is running.
-    //
-    // NOTE: resize-swap is not very clever. If present, --once MUST be the first arg.
-    let child_result = std::process::Command::new("/usr/bin/sudo")
-        .arg(RESIZE_SWAP_BIN)
-        .arg("--once")
-        .arg(size_bytes.to_string())
-        .spawn();
-
-    child_result
-        .context("spawn() failed")
-        .and_then(|mut child| child.wait().context("wait() failed"))
-        .and_then(|status| match status.success() {
-            true => Ok(()),
-            false => {
-                // The command failed. Maybe it was because the resize-swap file doesn't exist?
-                // The --once flag causes it to delete itself on success so we don't disable swap
-                // while postgres is running; maybe this is fine.
-                match Path::new(RESIZE_SWAP_BIN).try_exists() {
-                    Err(_) | Ok(true) => Err(anyhow!("process exited with {status}")),
-                    // The path doesn't exist; we're actually ok 
-                    Ok(false) => {
-                        warn!("ignoring \"not found\" error from resize-swap to avoid swapoff while compute is running");
-                        Ok(())
-                    },
-                }
-            }
-        })
-        // wrap any prior error with the overall context that we couldn't run the command
-        .with_context(|| {
-            format!("could not run `/usr/bin/sudo {RESIZE_SWAP_BIN} --once {size_bytes}`")
-        })
-}

control_plane/Cargo.toml

@@ -6,6 +6,7 @@ license.workspace = true
 
 [dependencies]
 anyhow.workspace = true
+async-trait.workspace = true
 camino.workspace = true
 clap.workspace = true
 comfy-table.workspace = true
@@ -16,7 +17,6 @@ nix.workspace = true
 once_cell.workspace = true
 postgres.workspace = true
 hex.workspace = true
-humantime-serde.workspace = true
 hyper.workspace = true
 regex.workspace = true
 reqwest = { workspace = true, features = ["blocking", "json"] }
@@ -27,7 +27,6 @@ serde_with.workspace = true
 tar.workspace = true
 thiserror.workspace = true
 toml.workspace = true
-toml_edit.workspace = true
 tokio.workspace = true
 tokio-postgres.workspace = true
 tokio-util.workspace = true
@@ -39,7 +38,6 @@ safekeeper_api.workspace = true
 postgres_connection.workspace = true
 storage_broker.workspace = true
 utils.workspace = true
-whoami.workspace = true
 
 compute_api.workspace = true
 workspace_hack.workspace = true

control_plane/src/background_process.rs

@@ -36,11 +36,11 @@ use utils::pid_file::{self, PidFileRead};
 // it's waiting. If the process hasn't started/stopped after 5 seconds,
 // it prints a notice that it's taking long, but keeps waiting.
 //
-const STOP_RETRY_TIMEOUT: Duration = Duration::from_secs(10);
-const STOP_RETRIES: u128 = STOP_RETRY_TIMEOUT.as_millis() / RETRY_INTERVAL.as_millis();
-const RETRY_INTERVAL: Duration = Duration::from_millis(100);
-const DOT_EVERY_RETRIES: u128 = 10;
-const NOTICE_AFTER_RETRIES: u128 = 50;
+const RETRY_UNTIL_SECS: u64 = 10;
+const RETRIES: u64 = (RETRY_UNTIL_SECS * 1000) / RETRY_INTERVAL_MILLIS;
+const RETRY_INTERVAL_MILLIS: u64 = 100;
+const DOT_EVERY_RETRIES: u64 = 10;
+const NOTICE_AFTER_RETRIES: u64 = 50;
 
 /// Argument to `start_process`, to indicate whether it should create pidfile or if the process creates
 /// it itself.
@@ -52,7 +52,6 @@ pub enum InitialPidFile {
 }
 
 /// Start a background child process using the parameters given.
-#[allow(clippy::too_many_arguments)]
 pub async fn start_process<F, Fut, AI, A, EI>(
     process_name: &str,
     datadir: &Path,
@@ -60,7 +59,6 @@ pub async fn start_process<F, Fut, AI, A, EI>(
     args: AI,
     envs: EI,
     initial_pid_file: InitialPidFile,
-    retry_timeout: &Duration,
     process_status_check: F,
 ) -> anyhow::Result<()>
 where
@@ -71,10 +69,6 @@ where
     // Not generic AsRef<OsStr>, otherwise empty `envs` prevents type inference
     EI: IntoIterator<Item = (String, String)>,
 {
-    let retries: u128 = retry_timeout.as_millis() / RETRY_INTERVAL.as_millis();
-    if !datadir.metadata().context("stat datadir")?.is_dir() {
-        anyhow::bail!("`datadir` must be a directory when calling this function: {datadir:?}");
-    }
     let log_path = datadir.join(format!("{process_name}.log"));
     let process_log_file = fs::OpenOptions::new()
         .create(true)
@@ -91,17 +85,8 @@ where
     let background_command = command
         .stdout(process_log_file)
         .stderr(same_file_for_stderr)
-        .args(args)
-        // spawn all child processes in their datadir, useful for all kinds of things,
-        // not least cleaning up child processes e.g. after an unclean exit from the test suite:
-        // ```
-        // lsof  -d cwd -a +D  Users/cs/src/neon/test_output
-        // ```
-        .current_dir(datadir);
-
-    let filled_cmd = fill_env_vars_prefixed_neon(fill_remote_storage_secrets_vars(
-        fill_rust_env_vars(background_command),
-    ));
+        .args(args);
+    let filled_cmd = fill_remote_storage_secrets_vars(fill_rust_env_vars(background_command));
     filled_cmd.envs(envs);
 
     let pid_file_to_check = match &initial_pid_file {
@@ -133,7 +118,7 @@ where
         .unwrap();
     });
 
-    for retries in 0..retries {
+    for retries in 0..RETRIES {
         match process_started(pid, pid_file_to_check, &process_status_check).await {
             Ok(true) => {
                 println!("\n{process_name} started and passed status check, pid: {pid}");
@@ -151,7 +136,7 @@ where
                     print!(".");
                     io::stdout().flush().unwrap();
                 }
-                thread::sleep(RETRY_INTERVAL);
+                thread::sleep(Duration::from_millis(RETRY_INTERVAL_MILLIS));
             }
             Err(e) => {
                 println!("error starting process {process_name:?}: {e:#}");
@@ -160,10 +145,9 @@ where
         }
     }
     println!();
-    anyhow::bail!(format!(
-        "{} did not start+pass status checks within {:?} seconds",
-        process_name, retry_timeout
-    ));
+    anyhow::bail!(
+        "{process_name} did not start+pass status checks within {RETRY_UNTIL_SECS} seconds"
+    );
 }
 
 /// Stops the process, using the pid file given. Returns Ok also if the process is already not running.
@@ -219,7 +203,7 @@ pub fn stop_process(
 }
 
 pub fn wait_until_stopped(process_name: &str, pid: Pid) -> anyhow::Result<()> {
-    for retries in 0..STOP_RETRIES {
+    for retries in 0..RETRIES {
         match process_has_stopped(pid) {
             Ok(true) => {
                 println!("\n{process_name} stopped");
@@ -235,7 +219,7 @@ pub fn wait_until_stopped(process_name: &str, pid: Pid) -> anyhow::Result<()> {
                     print!(".");
                     io::stdout().flush().unwrap();
                 }
-                thread::sleep(RETRY_INTERVAL);
+                thread::sleep(Duration::from_millis(RETRY_INTERVAL_MILLIS));
             }
             Err(e) => {
                 println!("{process_name} with pid {pid} failed to stop: {e:#}");
@@ -244,10 +228,7 @@ pub fn wait_until_stopped(process_name: &str, pid: Pid) -> anyhow::Result<()> {
         }
     }
     println!();
-    anyhow::bail!(format!(
-        "{} with pid {} did not stop in {:?} seconds",
-        process_name, pid, STOP_RETRY_TIMEOUT
-    ));
+    anyhow::bail!("{process_name} with pid {pid} did not stop in {RETRY_UNTIL_SECS} seconds");
 }
 
 fn fill_rust_env_vars(cmd: &mut Command) -> &mut Command {
@@ -287,15 +268,6 @@ fn fill_remote_storage_secrets_vars(mut cmd: &mut Command) -> &mut Command {
     cmd
 }
 
-fn fill_env_vars_prefixed_neon(mut cmd: &mut Command) -> &mut Command {
-    for (var, val) in std::env::vars() {
-        if var.starts_with("NEON_") {
-            cmd = cmd.env(var, val);
-        }
-    }
-    cmd
-}
-
 /// Add a `pre_exec` to the cmd that, inbetween fork() and exec(),
 /// 1. Claims a pidfile with a fcntl lock on it and
 /// 2. Sets up the pidfile's file descriptor so that it (and the lock)
@@ -379,7 +351,7 @@ where
     }
 }
 
-pub(crate) fn process_has_stopped(pid: Pid) -> anyhow::Result<bool> {
+fn process_has_stopped(pid: Pid) -> anyhow::Result<bool> {
     match kill(pid, None) {
         // Process exists, keep waiting
         Ok(_) => Ok(false),

control_plane/src/bin/neon_local.rs

@@ -9,25 +9,20 @@ use anyhow::{anyhow, bail, Context, Result};
 use clap::{value_parser, Arg, ArgAction, ArgMatches, Command, ValueEnum};
 use compute_api::spec::ComputeMode;
 use control_plane::endpoint::ComputeControlPlane;
-use control_plane::local_env::{
-    InitForceMode, LocalEnv, NeonBroker, NeonLocalInitConf, NeonLocalInitPageserverConf,
-    SafekeeperConf,
-};
-use control_plane::pageserver::PageServerNode;
+use control_plane::local_env::{InitForceMode, LocalEnv};
+use control_plane::pageserver::{PageServerNode, PAGESERVER_REMOTE_STORAGE_DIR};
 use control_plane::safekeeper::SafekeeperNode;
-use control_plane::storage_controller::{
-    NeonStorageControllerStartArgs, NeonStorageControllerStopArgs, StorageController,
-};
+use control_plane::storage_controller::StorageController;
 use control_plane::{broker, local_env};
-use pageserver_api::config::{
+use pageserver_api::controller_api::PlacementPolicy;
+use pageserver_api::models::{
+    ShardParameters, TenantCreateRequest, TimelineCreateRequest, TimelineInfo,
+};
+use pageserver_api::shard::{ShardCount, ShardStripeSize, TenantShardId};
+use pageserver_api::{
     DEFAULT_HTTP_LISTEN_PORT as DEFAULT_PAGESERVER_HTTP_PORT,
     DEFAULT_PG_LISTEN_PORT as DEFAULT_PAGESERVER_PG_PORT,
 };
-use pageserver_api::controller_api::{
-    NodeAvailabilityWrapper, PlacementPolicy, TenantCreateRequest,
-};
-use pageserver_api::models::{ShardParameters, TimelineCreateRequest, TimelineInfo};
-use pageserver_api::shard::{ShardCount, ShardStripeSize, TenantShardId};
 use postgres_backend::AuthType;
 use postgres_connection::parse_host_port;
 use safekeeper_api::{
@@ -38,7 +33,6 @@ use std::collections::{BTreeSet, HashMap};
 use std::path::PathBuf;
 use std::process::exit;
 use std::str::FromStr;
-use std::time::Duration;
 use storage_broker::DEFAULT_LISTEN_ADDR as DEFAULT_BROKER_ADDR;
 use url::Host;
 use utils::{
@@ -54,10 +48,48 @@ const DEFAULT_PAGESERVER_ID: NodeId = NodeId(1);
 const DEFAULT_BRANCH_NAME: &str = "main";
 project_git_version!(GIT_VERSION);
 
-const DEFAULT_PG_VERSION: &str = "16";
+const DEFAULT_PG_VERSION: &str = "15";
 
 const DEFAULT_PAGESERVER_CONTROL_PLANE_API: &str = "http://127.0.0.1:1234/upcall/v1/";
 
+fn default_conf(num_pageservers: u16) -> String {
+    let mut template = format!(
+        r#"
+# Default built-in configuration, defined in main.rs
+control_plane_api = '{DEFAULT_PAGESERVER_CONTROL_PLANE_API}'
+
+[broker]
+listen_addr = '{DEFAULT_BROKER_ADDR}'
+
+[[safekeepers]]
+id = {DEFAULT_SAFEKEEPER_ID}
+pg_port = {DEFAULT_SAFEKEEPER_PG_PORT}
+http_port = {DEFAULT_SAFEKEEPER_HTTP_PORT}
+
+"#,
+    );
+
+    for i in 0..num_pageservers {
+        let pageserver_id = NodeId(DEFAULT_PAGESERVER_ID.0 + i as u64);
+        let pg_port = DEFAULT_PAGESERVER_PG_PORT + i;
+        let http_port = DEFAULT_PAGESERVER_HTTP_PORT + i;
+
+        template += &format!(
+            r#"
+[[pageservers]]
+id = {pageserver_id}
+listen_pg_addr = '127.0.0.1:{pg_port}'
+listen_http_addr = '127.0.0.1:{http_port}'
+pg_auth_type = '{trust_auth}'
+http_auth_type = '{trust_auth}'
+"#,
+            trust_auth = AuthType::Trust,
+        )
+    }
+
+    template
+}
+
 ///
 /// Timelines tree element used as a value in the HashMap.
 ///
@@ -90,8 +122,7 @@ fn main() -> Result<()> {
         handle_init(sub_args).map(Some)
     } else {
         // all other commands need an existing config
-        let mut env =
-            LocalEnv::load_config(&local_env::base_path()).context("Error loading config")?;
+        let mut env = LocalEnv::load_config().context("Error loading config")?;
         let original_env = env.clone();
 
         let rt = tokio::runtime::Builder::new_current_thread()
@@ -102,7 +133,7 @@ fn main() -> Result<()> {
         let subcommand_result = match sub_name {
             "tenant" => rt.block_on(handle_tenant(sub_args, &mut env)),
             "timeline" => rt.block_on(handle_timeline(sub_args, &mut env)),
-            "start" => rt.block_on(handle_start_all(&env, get_start_timeout(sub_args))),
+            "start" => rt.block_on(handle_start_all(sub_args, &env)),
             "stop" => rt.block_on(handle_stop_all(sub_args, &env)),
             "pageserver" => rt.block_on(handle_pageserver(sub_args, &env)),
             "storage_controller" => rt.block_on(handle_storage_controller(sub_args, &env)),
@@ -121,7 +152,7 @@ fn main() -> Result<()> {
     };
 
     match subcommand_result {
-        Ok(Some(updated_env)) => updated_env.persist_config()?,
+        Ok(Some(updated_env)) => updated_env.persist_config(&updated_env.base_data_dir)?,
         Ok(None) => (),
         Err(e) => {
             eprintln!("command failed: {e:?}");
@@ -310,66 +341,48 @@ fn parse_timeline_id(sub_match: &ArgMatches) -> anyhow::Result<Option<TimelineId
 }
 
 fn handle_init(init_match: &ArgMatches) -> anyhow::Result<LocalEnv> {
-    let num_pageservers = init_match.get_one::<u16>("num-pageservers");
-
-    let force = init_match.get_one("force").expect("we set a default value");
-
-    // Create the in-memory `LocalEnv` that we'd normally load from disk in `load_config`.
-    let init_conf: NeonLocalInitConf = if let Some(config_path) =
-        init_match.get_one::<PathBuf>("config")
-    {
-        // User (likely the Python test suite) provided a description of the environment.
-        if num_pageservers.is_some() {
-            bail!("Cannot specify both --num-pageservers and --config, use key `pageservers` in the --config file instead");
-        }
+    let num_pageservers = init_match
+        .get_one::<u16>("num-pageservers")
+        .expect("num-pageservers arg has a default");
+    // Create config file
+    let toml_file: String = if let Some(config_path) = init_match.get_one::<PathBuf>("config") {
         // load and parse the file
-        let contents = std::fs::read_to_string(config_path).with_context(|| {
+        std::fs::read_to_string(config_path).with_context(|| {
             format!(
                 "Could not read configuration file '{}'",
                 config_path.display()
             )
-        })?;
-        toml_edit::de::from_str(&contents)?
+        })?
     } else {
-        // User (likely interactive) did not provide a description of the environment, give them the default
-        NeonLocalInitConf {
-            control_plane_api: Some(Some(DEFAULT_PAGESERVER_CONTROL_PLANE_API.parse().unwrap())),
-            broker: NeonBroker {
-                listen_addr: DEFAULT_BROKER_ADDR.parse().unwrap(),
-            },
-            safekeepers: vec![SafekeeperConf {
-                id: DEFAULT_SAFEKEEPER_ID,
-                pg_port: DEFAULT_SAFEKEEPER_PG_PORT,
-                http_port: DEFAULT_SAFEKEEPER_HTTP_PORT,
-                ..Default::default()
-            }],
-            pageservers: (0..num_pageservers.copied().unwrap_or(1))
-                .map(|i| {
-                    let pageserver_id = NodeId(DEFAULT_PAGESERVER_ID.0 + i as u64);
-                    let pg_port = DEFAULT_PAGESERVER_PG_PORT + i;
-                    let http_port = DEFAULT_PAGESERVER_HTTP_PORT + i;
-                    NeonLocalInitPageserverConf {
-                        id: pageserver_id,
-                        listen_pg_addr: format!("127.0.0.1:{pg_port}"),
-                        listen_http_addr: format!("127.0.0.1:{http_port}"),
-                        pg_auth_type: AuthType::Trust,
-                        http_auth_type: AuthType::Trust,
-                        other: Default::default(),
-                    }
-                })
-                .collect(),
-            pg_distrib_dir: None,
-            neon_distrib_dir: None,
-            default_tenant_id: TenantId::from_array(std::array::from_fn(|_| 0)),
-            storage_controller: None,
-            control_plane_compute_hook_api: None,
-        }
+        // Built-in default config
+        default_conf(*num_pageservers)
     };
 
-    LocalEnv::init(init_conf, force)
-        .context("materialize initial neon_local environment on disk")?;
-    Ok(LocalEnv::load_config(&local_env::base_path())
-        .expect("freshly written config should be loadable"))
+    let pg_version = init_match
+        .get_one::<u32>("pg-version")
+        .copied()
+        .context("Failed to parse postgres version from the argument string")?;
+
+    let mut env =
+        LocalEnv::parse_config(&toml_file).context("Failed to create neon configuration")?;
+    let force = init_match.get_one("force").expect("we set a default value");
+    env.init(pg_version, force)
+        .context("Failed to initialize neon repository")?;
+
+    // Create remote storage location for default LocalFs remote storage
+    std::fs::create_dir_all(env.base_data_dir.join(PAGESERVER_REMOTE_STORAGE_DIR))?;
+
+    // Initialize pageserver, create initial tenant and timeline.
+    for ps_conf in &env.pageservers {
+        PageServerNode::from_env(&env, ps_conf)
+            .initialize(&pageserver_config_overrides(init_match))
+            .unwrap_or_else(|e| {
+                eprintln!("pageserver init failed: {e:?}");
+                exit(1);
+            });
+    }
+
+    Ok(env)
 }
 
 /// The default pageserver is the one where CLI tenant/timeline operations are sent by default.
@@ -384,6 +397,15 @@ fn get_default_pageserver(env: &local_env::LocalEnv) -> PageServerNode {
     PageServerNode::from_env(env, ps_conf)
 }
 
+fn pageserver_config_overrides(init_match: &ArgMatches) -> Vec<&str> {
+    init_match
+        .get_many::<String>("pageserver-config-override")
+        .into_iter()
+        .flatten()
+        .map(String::as_str)
+        .collect()
+}
+
 async fn handle_tenant(
     tenant_match: &ArgMatches,
     env: &mut local_env::LocalEnv,
@@ -395,54 +417,6 @@ async fn handle_tenant(
                 println!("{} {:?}", t.id, t.state);
             }
         }
-        Some(("import", import_match)) => {
-            let tenant_id = parse_tenant_id(import_match)?.unwrap_or_else(TenantId::generate);
-
-            let storage_controller = StorageController::from_env(env);
-            let create_response = storage_controller.tenant_import(tenant_id).await?;
-
-            let shard_zero = create_response
-                .shards
-                .first()
-                .expect("Import response omitted shards");
-
-            let attached_pageserver_id = shard_zero.node_id;
-            let pageserver =
-                PageServerNode::from_env(env, env.get_pageserver_conf(attached_pageserver_id)?);
-
-            println!(
-                "Imported tenant {tenant_id}, attached to pageserver {attached_pageserver_id}"
-            );
-
-            let timelines = pageserver
-                .http_client
-                .list_timelines(shard_zero.shard_id)
-                .await?;
-
-            // Pick a 'main' timeline that has no ancestors, the rest will get arbitrary names
-            let main_timeline = timelines
-                .iter()
-                .find(|t| t.ancestor_timeline_id.is_none())
-                .expect("No timelines found")
-                .timeline_id;
-
-            let mut branch_i = 0;
-            for timeline in timelines.iter() {
-                let branch_name = if timeline.timeline_id == main_timeline {
-                    "main".to_string()
-                } else {
-                    branch_i += 1;
-                    format!("branch_{branch_i}")
-                };
-
-                println!(
-                    "Importing timeline {tenant_id}/{} as branch {branch_name}",
-                    timeline.timeline_id
-                );
-
-                env.register_branch_mapping(branch_name, tenant_id, timeline.timeline_id)?;
-            }
-        }
         Some(("create", create_match)) => {
             let tenant_conf: HashMap<_, _> = create_match
                 .get_many::<String>("config")
@@ -602,9 +576,13 @@ async fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::Local
         Some(("import", import_match)) => {
             let tenant_id = get_tenant_id(import_match, env)?;
             let timeline_id = parse_timeline_id(import_match)?.expect("No timeline id provided");
-            let branch_name = import_match
-                .get_one::<String>("branch-name")
-                .ok_or_else(|| anyhow!("No branch name provided"))?;
+            let name = import_match
+                .get_one::<String>("node-name")
+                .ok_or_else(|| anyhow!("No node name provided"))?;
+            let update_catalog = import_match
+                .get_one::<bool>("update-catalog")
+                .cloned()
+                .unwrap_or_default();
 
             // Parse base inputs
             let base_tarfile = import_match
@@ -631,11 +609,24 @@ async fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::Local
                 .copied()
                 .context("Failed to parse postgres version from the argument string")?;
 
+            let mut cplane = ComputeControlPlane::load(env.clone())?;
             println!("Importing timeline into pageserver ...");
             pageserver
                 .timeline_import(tenant_id, timeline_id, base, pg_wal, pg_version)
                 .await?;
-            env.register_branch_mapping(branch_name.to_string(), tenant_id, timeline_id)?;
+            env.register_branch_mapping(name.to_string(), tenant_id, timeline_id)?;
+
+            println!("Creating endpoint for imported timeline ...");
+            cplane.new_endpoint(
+                name,
+                tenant_id,
+                timeline_id,
+                None,
+                None,
+                pg_version,
+                ComputeMode::Primary,
+                !update_catalog,
+            )?;
             println!("Done");
         }
         Some(("branch", branch_match)) => {
@@ -798,8 +789,6 @@ async fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Re
                 .copied()
                 .unwrap_or(false);
 
-            let allow_multiple = sub_args.get_flag("allow-multiple");
-
             let mode = match (lsn, hot_standby) {
                 (Some(lsn), false) => ComputeMode::Static(lsn),
                 (None, true) => ComputeMode::Replica,
@@ -817,9 +806,7 @@ async fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Re
                 _ => {}
             }
 
-            if !allow_multiple {
-                cplane.check_conflicting_endpoints(mode, tenant_id, timeline_id)?;
-            }
+            cplane.check_conflicting_endpoints(mode, tenant_id, timeline_id)?;
 
             cplane.new_endpoint(
                 &endpoint_id,
@@ -848,15 +835,20 @@ async fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Re
 
             let remote_ext_config = sub_args.get_one::<String>("remote-ext-config");
 
-            let allow_multiple = sub_args.get_flag("allow-multiple");
-
-            // If --safekeepers argument is given, use only the listed
-            // safekeeper nodes; otherwise all from the env.
-            let safekeepers = if let Some(safekeepers) = parse_safekeepers(sub_args)? {
-                safekeepers
-            } else {
-                env.safekeepers.iter().map(|sk| sk.id).collect()
-            };
+            // If --safekeepers argument is given, use only the listed safekeeper nodes.
+            let safekeepers =
+                if let Some(safekeepers_str) = sub_args.get_one::<String>("safekeepers") {
+                    let mut safekeepers: Vec<NodeId> = Vec::new();
+                    for sk_id in safekeepers_str.split(',').map(str::trim) {
+                        let sk_id = NodeId(u64::from_str(sk_id).map_err(|_| {
+                            anyhow!("invalid node ID \"{sk_id}\" in --safekeepers list")
+                        })?);
+                        safekeepers.push(sk_id);
+                    }
+                    safekeepers
+                } else {
+                    env.safekeepers.iter().map(|sk| sk.id).collect()
+                };
 
             let endpoint = cplane
                 .endpoints
@@ -868,13 +860,11 @@ async fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Re
                 .cloned()
                 .unwrap_or_default();
 
-            if !allow_multiple {
-                cplane.check_conflicting_endpoints(
-                    endpoint.mode,
-                    endpoint.tenant_id,
-                    endpoint.timeline_id,
-                )?;
-            }
+            cplane.check_conflicting_endpoints(
+                endpoint.mode,
+                endpoint.tenant_id,
+                endpoint.timeline_id,
+            )?;
 
             let (pageservers, stripe_size) = if let Some(pageserver_id) = pageserver_id {
                 let conf = env.get_pageserver_conf(pageserver_id).unwrap();
@@ -960,10 +950,7 @@ async fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Re
                         })
                         .collect::<Vec<_>>()
                 };
-            // If --safekeepers argument is given, use only the listed
-            // safekeeper nodes; otherwise all from the env.
-            let safekeepers = parse_safekeepers(sub_args)?;
-            endpoint.reconfigure(pageservers, None, safekeepers).await?;
+            endpoint.reconfigure(pageservers, None).await?;
         }
         "stop" => {
             let endpoint_id = sub_args
@@ -985,23 +972,6 @@ async fn handle_endpoint(ep_match: &ArgMatches, env: &local_env::LocalEnv) -> Re
     Ok(())
 }
 
-/// Parse --safekeepers as list of safekeeper ids.
-fn parse_safekeepers(sub_args: &ArgMatches) -> Result<Option<Vec<NodeId>>> {
-    if let Some(safekeepers_str) = sub_args.get_one::<String>("safekeepers") {
-        let mut safekeepers: Vec<NodeId> = Vec::new();
-        for sk_id in safekeepers_str.split(',').map(str::trim) {
-            let sk_id = NodeId(
-                u64::from_str(sk_id)
-                    .map_err(|_| anyhow!("invalid node ID \"{sk_id}\" in --safekeepers list"))?,
-            );
-            safekeepers.push(sk_id);
-        }
-        Ok(Some(safekeepers))
-    } else {
-        Ok(None)
-    }
-}
-
 fn handle_mappings(sub_match: &ArgMatches, env: &mut local_env::LocalEnv) -> Result<()> {
     let (sub_name, sub_args) = match sub_match.subcommand() {
         Some(ep_subcommand_data) => ep_subcommand_data,
@@ -1047,48 +1017,11 @@ fn get_pageserver(env: &local_env::LocalEnv, args: &ArgMatches) -> Result<PageSe
     ))
 }
 
-fn get_start_timeout(args: &ArgMatches) -> &Duration {
-    let humantime_duration = args
-        .get_one::<humantime::Duration>("start-timeout")
-        .expect("invalid value for start-timeout");
-    humantime_duration.as_ref()
-}
-
-fn storage_controller_start_args(args: &ArgMatches) -> NeonStorageControllerStartArgs {
-    let maybe_instance_id = args.get_one::<u8>("instance-id");
-
-    let base_port = args.get_one::<u16>("base-port");
-
-    if maybe_instance_id.is_some() && base_port.is_none() {
-        panic!("storage-controller start specificied instance-id but did not provide base-port");
-    }
-
-    let start_timeout = args
-        .get_one::<humantime::Duration>("start-timeout")
-        .expect("invalid value for start-timeout");
-
-    NeonStorageControllerStartArgs {
-        instance_id: maybe_instance_id.copied().unwrap_or(1),
-        base_port: base_port.copied(),
-        start_timeout: *start_timeout,
-    }
-}
-
-fn storage_controller_stop_args(args: &ArgMatches) -> NeonStorageControllerStopArgs {
-    let maybe_instance_id = args.get_one::<u8>("instance-id");
-    let immediate = args.get_one::<String>("stop-mode").map(|s| s.as_str()) == Some("immediate");
-
-    NeonStorageControllerStopArgs {
-        instance_id: maybe_instance_id.copied().unwrap_or(1),
-        immediate,
-    }
-}
-
 async fn handle_pageserver(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
     match sub_match.subcommand() {
         Some(("start", subcommand_args)) => {
             if let Err(e) = get_pageserver(env, subcommand_args)?
-                .start(get_start_timeout(subcommand_args))
+                .start(&pageserver_config_overrides(subcommand_args))
                 .await
             {
                 eprintln!("pageserver start failed: {e}");
@@ -1116,7 +1049,10 @@ async fn handle_pageserver(sub_match: &ArgMatches, env: &local_env::LocalEnv) ->
                 exit(1);
             }
 
-            if let Err(e) = pageserver.start(get_start_timeout(sub_match)).await {
+            if let Err(e) = pageserver
+                .start(&pageserver_config_overrides(subcommand_args))
+                .await
+            {
                 eprintln!("pageserver start failed: {e}");
                 exit(1);
             }
@@ -1144,15 +1080,20 @@ async fn handle_storage_controller(
 ) -> Result<()> {
     let svc = StorageController::from_env(env);
     match sub_match.subcommand() {
-        Some(("start", start_match)) => {
-            if let Err(e) = svc.start(storage_controller_start_args(start_match)).await {
+        Some(("start", _start_match)) => {
+            if let Err(e) = svc.start().await {
                 eprintln!("start failed: {e}");
                 exit(1);
             }
         }
 
         Some(("stop", stop_match)) => {
-            if let Err(e) = svc.stop(storage_controller_stop_args(stop_match)).await {
+            let immediate = stop_match
+                .get_one::<String>("stop-mode")
+                .map(|s| s.as_str())
+                == Some("immediate");
+
+            if let Err(e) = svc.stop(immediate).await {
                 eprintln!("stop failed: {}", e);
                 exit(1);
             }
@@ -1199,10 +1140,7 @@ async fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) ->
         "start" => {
             let extra_opts = safekeeper_extra_opts(sub_args);
 
-            if let Err(e) = safekeeper
-                .start(extra_opts, get_start_timeout(sub_args))
-                .await
-            {
+            if let Err(e) = safekeeper.start(extra_opts).await {
                 eprintln!("safekeeper start failed: {}", e);
                 exit(1);
             }
@@ -1228,10 +1166,7 @@ async fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) ->
             }
 
             let extra_opts = safekeeper_extra_opts(sub_args);
-            if let Err(e) = safekeeper
-                .start(extra_opts, get_start_timeout(sub_args))
-                .await
-            {
+            if let Err(e) = safekeeper.start(extra_opts).await {
                 eprintln!("safekeeper start failed: {}", e);
                 exit(1);
             }
@@ -1244,23 +1179,15 @@ async fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) ->
     Ok(())
 }
 
-async fn handle_start_all(
-    env: &local_env::LocalEnv,
-    retry_timeout: &Duration,
-) -> anyhow::Result<()> {
+async fn handle_start_all(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow::Result<()> {
     // Endpoints are not started automatically
 
-    broker::start_broker_process(env, retry_timeout).await?;
+    broker::start_broker_process(env).await?;
 
     // Only start the storage controller if the pageserver is configured to need it
     if env.control_plane_api.is_some() {
         let storage_controller = StorageController::from_env(env);
-        if let Err(e) = storage_controller
-            .start(NeonStorageControllerStartArgs::with_default_instance_id(
-                (*retry_timeout).into(),
-            ))
-            .await
-        {
+        if let Err(e) = storage_controller.start().await {
             eprintln!("storage_controller start failed: {:#}", e);
             try_stop_all(env, true).await;
             exit(1);
@@ -1269,7 +1196,10 @@ async fn handle_start_all(
 
     for ps_conf in &env.pageservers {
         let pageserver = PageServerNode::from_env(env, ps_conf);
-        if let Err(e) = pageserver.start(retry_timeout).await {
+        if let Err(e) = pageserver
+            .start(&pageserver_config_overrides(sub_match))
+            .await
+        {
             eprintln!("pageserver {} start failed: {:#}", ps_conf.id, e);
             try_stop_all(env, true).await;
             exit(1);
@@ -1278,76 +1208,15 @@ async fn handle_start_all(
 
     for node in env.safekeepers.iter() {
         let safekeeper = SafekeeperNode::from_env(env, node);
-        if let Err(e) = safekeeper.start(vec![], retry_timeout).await {
+        if let Err(e) = safekeeper.start(vec![]).await {
             eprintln!("safekeeper {} start failed: {:#}", safekeeper.id, e);
             try_stop_all(env, false).await;
             exit(1);
         }
     }
-
-    neon_start_status_check(env, retry_timeout).await?;
-
     Ok(())
 }
 
-async fn neon_start_status_check(
-    env: &local_env::LocalEnv,
-    retry_timeout: &Duration,
-) -> anyhow::Result<()> {
-    const RETRY_INTERVAL: Duration = Duration::from_millis(100);
-    const NOTICE_AFTER_RETRIES: Duration = Duration::from_secs(5);
-
-    if env.control_plane_api.is_none() {
-        return Ok(());
-    }
-
-    let storcon = StorageController::from_env(env);
-
-    let retries = retry_timeout.as_millis() / RETRY_INTERVAL.as_millis();
-    let notice_after_retries = retry_timeout.as_millis() / NOTICE_AFTER_RETRIES.as_millis();
-
-    println!("\nRunning neon status check");
-
-    for retry in 0..retries {
-        if retry == notice_after_retries {
-            println!("\nNeon status check has not passed yet, continuing to wait")
-        }
-
-        let mut passed = true;
-        let mut nodes = storcon.node_list().await?;
-        let mut pageservers = env.pageservers.clone();
-
-        if nodes.len() != pageservers.len() {
-            continue;
-        }
-
-        nodes.sort_by_key(|ps| ps.id);
-        pageservers.sort_by_key(|ps| ps.id);
-
-        for (idx, pageserver) in pageservers.iter().enumerate() {
-            let node = &nodes[idx];
-            if node.id != pageserver.id {
-                passed = false;
-                break;
-            }
-
-            if !matches!(node.availability, NodeAvailabilityWrapper::Active) {
-                passed = false;
-                break;
-            }
-        }
-
-        if passed {
-            println!("\nNeon started and passed status check");
-            return Ok(());
-        }
-
-        tokio::time::sleep(RETRY_INTERVAL).await;
-    }
-
-    anyhow::bail!("\nNeon passed status check")
-}
-
 async fn handle_stop_all(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
     let immediate =
         sub_match.get_one::<String>("stop-mode").map(|s| s.as_str()) == Some("immediate");
@@ -1362,7 +1231,7 @@ async fn try_stop_all(env: &local_env::LocalEnv, immediate: bool) {
     match ComputeControlPlane::load(env.clone()) {
         Ok(cplane) => {
             for (_k, node) in cplane.endpoints {
-                if let Err(e) = node.stop(if immediate { "immediate" } else { "fast" }, false) {
+                if let Err(e) = node.stop(if immediate { "immediate" } else { "fast " }, false) {
                     eprintln!("postgres stop failed: {e:#}");
                 }
             }
@@ -1390,35 +1259,15 @@ async fn try_stop_all(env: &local_env::LocalEnv, immediate: bool) {
         eprintln!("neon broker stop failed: {e:#}");
     }
 
-    // Stop all storage controller instances. In the most common case there's only one,
-    // but iterate though the base data directory in order to discover the instances.
-    let storcon_instances = env
-        .storage_controller_instances()
-        .await
-        .expect("Must inspect data dir");
-    for (instance_id, _instance_dir_path) in storcon_instances {
+    if env.control_plane_api.is_some() {
         let storage_controller = StorageController::from_env(env);
-        let stop_args = NeonStorageControllerStopArgs {
-            instance_id,
-            immediate,
-        };
-
-        if let Err(e) = storage_controller.stop(stop_args).await {
-            eprintln!("Storage controller instance {instance_id} stop failed: {e:#}");
+        if let Err(e) = storage_controller.stop(immediate).await {
+            eprintln!("storage controller stop failed: {e:#}");
         }
     }
 }
 
 fn cli() -> Command {
-    let timeout_arg = Arg::new("start-timeout")
-        .long("start-timeout")
-        .short('t')
-        .global(true)
-        .help("timeout until we fail the command, e.g. 30s")
-        .value_parser(value_parser!(humantime::Duration))
-        .default_value("10s")
-        .required(false);
-
     let branch_name_arg = Arg::new("branch-name")
         .long("branch-name")
         .help("Name of the branch to be created or used as an alias for other services")
@@ -1491,6 +1340,13 @@ fn cli() -> Command {
         .required(false)
         .value_name("stop-mode");
 
+    let pageserver_config_args = Arg::new("pageserver-config-override")
+        .long("pageserver-config-override")
+        .num_args(1)
+        .action(ArgAction::Append)
+        .help("Additional pageserver's configuration options or overrides, refer to pageserver's 'config-override' CLI parameter docs for more")
+        .required(false);
+
     let remote_ext_config_args = Arg::new("remote-ext-config")
         .long("remote-ext-config")
         .num_args(1)
@@ -1524,7 +1380,9 @@ fn cli() -> Command {
     let num_pageservers_arg = Arg::new("num-pageservers")
         .value_parser(value_parser!(u16))
         .long("num-pageservers")
-        .help("How many pageservers to create (default 1)");
+        .help("How many pageservers to create (default 1)")
+        .required(false)
+        .default_value("1");
 
     let update_catalog = Arg::new("update-catalog")
         .value_parser(value_parser!(bool))
@@ -1538,37 +1396,20 @@ fn cli() -> Command {
         .help("If set, will create test user `user` and `neondb` database. Requires `update-catalog = true`")
         .required(false);
 
-    let allow_multiple = Arg::new("allow-multiple")
-        .help("Allow multiple primary endpoints running on the same branch. Shouldn't be used normally, but useful for tests.")
-        .long("allow-multiple")
-        .action(ArgAction::SetTrue)
-        .required(false);
-
-    let instance_id = Arg::new("instance-id")
-        .long("instance-id")
-        .help("Identifier used to distinguish storage controller instances (default 1)")
-        .value_parser(value_parser!(u8))
-        .required(false);
-
-    let base_port = Arg::new("base-port")
-        .long("base-port")
-        .help("Base port for the storage controller instance idenfified by instance-id (defaults to pagserver cplane api)")
-        .value_parser(value_parser!(u16))
-        .required(false);
-
     Command::new("Neon CLI")
         .arg_required_else_help(true)
         .version(GIT_VERSION)
         .subcommand(
             Command::new("init")
                 .about("Initialize a new Neon repository, preparing configs for services to start with")
+                .arg(pageserver_config_args.clone())
                 .arg(num_pageservers_arg.clone())
                 .arg(
                     Arg::new("config")
                         .long("config")
                         .required(false)
                         .value_parser(value_parser!(PathBuf))
-                        .value_name("config")
+                        .value_name("config"),
                 )
                 .arg(pg_version_arg.clone())
                 .arg(force_arg)
@@ -1576,7 +1417,6 @@ fn cli() -> Command {
         .subcommand(
             Command::new("timeline")
             .about("Manage timelines")
-            .arg_required_else_help(true)
             .subcommand(Command::new("list")
                 .about("List all timelines, available to this pageserver")
                 .arg(tenant_id_arg.clone()))
@@ -1599,7 +1439,8 @@ fn cli() -> Command {
                 .about("Import timeline from basebackup directory")
                 .arg(tenant_id_arg.clone())
                 .arg(timeline_id_arg.clone())
-                .arg(branch_name_arg.clone())
+                .arg(Arg::new("node-name").long("node-name")
+                    .help("Name to assign to the imported timeline"))
                 .arg(Arg::new("base-tarfile")
                     .long("base-tarfile")
                     .value_parser(value_parser!(PathBuf))
@@ -1615,6 +1456,7 @@ fn cli() -> Command {
                 .arg(Arg::new("end-lsn").long("end-lsn")
                     .help("Lsn the basebackup ends at"))
                 .arg(pg_version_arg.clone())
+                .arg(update_catalog.clone())
             )
         ).subcommand(
             Command::new("tenant")
@@ -1637,8 +1479,6 @@ fn cli() -> Command {
             .subcommand(Command::new("config")
                 .arg(tenant_id_arg.clone())
                 .arg(Arg::new("config").short('c').num_args(1).action(ArgAction::Append).required(false)))
-            .subcommand(Command::new("import").arg(tenant_id_arg.clone().required(true))
-                .about("Import a tenant that is present in remote storage, and create branches for its timelines"))
         )
         .subcommand(
             Command::new("pageserver")
@@ -1648,7 +1488,7 @@ fn cli() -> Command {
                 .subcommand(Command::new("status"))
                 .subcommand(Command::new("start")
                     .about("Start local pageserver")
-                    .arg(timeout_arg.clone())
+                    .arg(pageserver_config_args.clone())
                 )
                 .subcommand(Command::new("stop")
                     .about("Stop local pageserver")
@@ -1656,20 +1496,16 @@ fn cli() -> Command {
                 )
                 .subcommand(Command::new("restart")
                     .about("Restart local pageserver")
-                    .arg(timeout_arg.clone())
+                    .arg(pageserver_config_args.clone())
                 )
         )
         .subcommand(
             Command::new("storage_controller")
                 .arg_required_else_help(true)
                 .about("Manage storage_controller")
-                .subcommand(Command::new("start").about("Start storage controller")
-                            .arg(timeout_arg.clone())
-                            .arg(instance_id.clone())
-                            .arg(base_port))
-                .subcommand(Command::new("stop").about("Stop storage controller")
-                            .arg(stop_mode_arg.clone())
-                            .arg(instance_id))
+                .subcommand(Command::new("start").about("Start local pageserver").arg(pageserver_config_args.clone()))
+                .subcommand(Command::new("stop").about("Stop local pageserver")
+                            .arg(stop_mode_arg.clone()))
         )
         .subcommand(
             Command::new("safekeeper")
@@ -1679,7 +1515,6 @@ fn cli() -> Command {
                             .about("Start local safekeeper")
                             .arg(safekeeper_id_arg.clone())
                             .arg(safekeeper_extra_opt_arg.clone())
-                            .arg(timeout_arg.clone())
                 )
                 .subcommand(Command::new("stop")
                             .about("Stop local safekeeper")
@@ -1691,7 +1526,6 @@ fn cli() -> Command {
                             .arg(safekeeper_id_arg)
                             .arg(stop_mode_arg.clone())
                             .arg(safekeeper_extra_opt_arg)
-                            .arg(timeout_arg.clone())
                 )
         )
         .subcommand(
@@ -1716,22 +1550,18 @@ fn cli() -> Command {
                     .arg(pg_version_arg.clone())
                     .arg(hot_standby_arg.clone())
                     .arg(update_catalog)
-                    .arg(allow_multiple.clone())
                 )
                 .subcommand(Command::new("start")
                     .about("Start postgres.\n If the endpoint doesn't exist yet, it is created.")
                     .arg(endpoint_id_arg.clone())
                     .arg(endpoint_pageserver_id_arg.clone())
-                    .arg(safekeepers_arg.clone())
+                    .arg(safekeepers_arg)
                     .arg(remote_ext_config_args)
                     .arg(create_test_user)
-                    .arg(allow_multiple.clone())
-                    .arg(timeout_arg.clone())
                 )
                 .subcommand(Command::new("reconfigure")
                             .about("Reconfigure the endpoint")
                             .arg(endpoint_pageserver_id_arg)
-                            .arg(safekeepers_arg)
                             .arg(endpoint_id_arg.clone())
                             .arg(tenant_id_arg.clone())
                 )
@@ -1779,7 +1609,7 @@ fn cli() -> Command {
         .subcommand(
             Command::new("start")
                 .about("Start page server and safekeepers")
-                .arg(timeout_arg.clone())
+                .arg(pageserver_config_args)
         )
         .subcommand(
             Command::new("stop")

control_plane/src/broker.rs

@@ -1,22 +1,17 @@
 //! Code to manage the storage broker
 //!
-//! In the local test environment, the storage broker stores its data directly in
+//! In the local test environment, the data for each safekeeper is stored in
 //!
 //! ```text
-//!   .neon
+//!   .neon/safekeepers/<safekeeper id>
 //! ```
-use std::time::Duration;
-
 use anyhow::Context;
 
 use camino::Utf8PathBuf;
 
 use crate::{background_process, local_env};
 
-pub async fn start_broker_process(
-    env: &local_env::LocalEnv,
-    retry_timeout: &Duration,
-) -> anyhow::Result<()> {
+pub async fn start_broker_process(env: &local_env::LocalEnv) -> anyhow::Result<()> {
     let broker = &env.broker;
     let listen_addr = &broker.listen_addr;
 
@@ -32,7 +27,6 @@ pub async fn start_broker_process(
         args,
         [],
         background_process::InitialPidFile::Create(storage_broker_pid_file_path(env)),
-        retry_timeout,
         || async {
             let url = broker.client_url();
             let status_url = url.join("status").with_context(|| {

control_plane/src/endpoint.rs

@@ -499,23 +499,6 @@ impl Endpoint {
             .join(",")
     }
 
-    /// Map safekeepers ids to the actual connection strings.
-    fn build_safekeepers_connstrs(&self, sk_ids: Vec<NodeId>) -> Result<Vec<String>> {
-        let mut safekeeper_connstrings = Vec::new();
-        if self.mode == ComputeMode::Primary {
-            for sk_id in sk_ids {
-                let sk = self
-                    .env
-                    .safekeepers
-                    .iter()
-                    .find(|node| node.id == sk_id)
-                    .ok_or_else(|| anyhow!("safekeeper {sk_id} does not exist"))?;
-                safekeeper_connstrings.push(format!("127.0.0.1:{}", sk.get_compute_port()));
-            }
-        }
-        Ok(safekeeper_connstrings)
-    }
-
     pub async fn start(
         &self,
         auth_token: &Option<String>,
@@ -540,7 +523,18 @@ impl Endpoint {
         let pageserver_connstring = Self::build_pageserver_connstr(&pageservers);
         assert!(!pageserver_connstring.is_empty());
 
-        let safekeeper_connstrings = self.build_safekeepers_connstrs(safekeepers)?;
+        let mut safekeeper_connstrings = Vec::new();
+        if self.mode == ComputeMode::Primary {
+            for sk_id in safekeepers {
+                let sk = self
+                    .env
+                    .safekeepers
+                    .iter()
+                    .find(|node| node.id == sk_id)
+                    .ok_or_else(|| anyhow!("safekeeper {sk_id} does not exist"))?;
+                safekeeper_connstrings.push(format!("127.0.0.1:{}", sk.get_compute_port()));
+            }
+        }
 
         // check for file remote_extensions_spec.json
         // if it is present, read it and pass to compute_ctl
@@ -560,7 +554,6 @@ impl Endpoint {
             format_version: 1.0,
             operation_uuid: None,
             features: self.features.clone(),
-            swap_size_bytes: None,
             cluster: Cluster {
                 cluster_id: None, // project ID: not used
                 name: None,       // project name: not used
@@ -598,6 +591,7 @@ impl Endpoint {
             remote_extensions,
             pgbouncer_settings: None,
             shard_stripe_size: Some(shard_stripe_size),
+            primary_is_running: None,
         };
         let spec_path = self.endpoint_path().join("spec.json");
         std::fs::write(spec_path, serde_json::to_string_pretty(&spec)?)?;
@@ -746,7 +740,6 @@ impl Endpoint {
         &self,
         mut pageservers: Vec<(Host, u16)>,
         stripe_size: Option<ShardStripeSize>,
-        safekeepers: Option<Vec<NodeId>>,
     ) -> Result<()> {
         let mut spec: ComputeSpec = {
             let spec_path = self.endpoint_path().join("spec.json");
@@ -781,12 +774,6 @@ impl Endpoint {
             spec.shard_stripe_size = stripe_size.map(|s| s.0 as usize);
         }
 
-        // If safekeepers are not specified, don't change them.
-        if let Some(safekeepers) = safekeepers {
-            let safekeeper_connstrings = self.build_safekeepers_connstrs(safekeepers)?;
-            spec.safekeeper_connstrings = safekeeper_connstrings;
-        }
-
         let client = reqwest::Client::builder()
             .timeout(Duration::from_secs(30))
             .build()
@@ -824,12 +811,11 @@ impl Endpoint {
         // cleanup work to do after postgres stops, like syncing safekeepers,
         // etc.
         //
-        // If destroying or stop mode is immediate, send it SIGTERM before
-        // waiting. Sometimes we do *not* want this cleanup: tests intentionally
-        // do stop when majority of safekeepers is down, so sync-safekeepers
-        // would hang otherwise. This could be a separate flag though.
-        let send_sigterm = destroy || mode == "immediate";
-        self.wait_for_compute_ctl_to_exit(send_sigterm)?;
+        // If destroying, send it SIGTERM before waiting. Sometimes we do *not*
+        // want this cleanup: tests intentionally do stop when majority of
+        // safekeepers is down, so sync-safekeepers would hang otherwise. This
+        // could be a separate flag though.
+        self.wait_for_compute_ctl_to_exit(destroy)?;
         if destroy {
             println!(
                 "Destroying postgres data directory '{}'",

control_plane/src/local_env.rs

@@ -3,7 +3,7 @@
 //! Now it also provides init method which acts like a stub for proper installation
 //! script which will use local paths.
 
-use anyhow::{bail, Context};
+use anyhow::{bail, ensure, Context};
 
 use clap::ValueEnum;
 use postgres_backend::AuthType;
@@ -17,17 +17,14 @@ use std::net::Ipv4Addr;
 use std::net::SocketAddr;
 use std::path::{Path, PathBuf};
 use std::process::{Command, Stdio};
-use std::time::Duration;
 use utils::{
     auth::{encode_from_key_file, Claims},
     id::{NodeId, TenantId, TenantTimelineId, TimelineId},
 };
 
-use crate::pageserver::PageServerNode;
-use crate::pageserver::PAGESERVER_REMOTE_STORAGE_DIR;
 use crate::safekeeper::SafekeeperNode;
 
-pub const DEFAULT_PG_VERSION: u32 = 16;
+pub const DEFAULT_PG_VERSION: u32 = 15;
 
 //
 // This data structures represents neon_local CLI config
@@ -36,107 +33,63 @@ pub const DEFAULT_PG_VERSION: u32 = 16;
 // to 'neon_local init --config=<path>' option. See control_plane/simple.conf for
 // an example.
 //
-#[derive(PartialEq, Eq, Clone, Debug)]
+#[derive(Serialize, Deserialize, PartialEq, Eq, Clone, Debug)]
 pub struct LocalEnv {
     // Base directory for all the nodes (the pageserver, safekeepers and
     // compute endpoints).
     //
     // This is not stored in the config file. Rather, this is the path where the
-    // config file itself is. It is read from the NEON_REPO_DIR env variable which
-    // must be an absolute path. If the env var is not set, $PWD/.neon is used.
+    // config file itself is. It is read from the NEON_REPO_DIR env variable or
+    // '.neon' if not given.
+    #[serde(skip)]
     pub base_data_dir: PathBuf,
 
     // Path to postgres distribution. It's expected that "bin", "include",
     // "lib", "share" from postgres distribution are there. If at some point
     // in time we will be able to run against vanilla postgres we may split that
     // to four separate paths and match OS-specific installation layout.
+    #[serde(default)]
     pub pg_distrib_dir: PathBuf,
 
     // Path to pageserver binary.
+    #[serde(default)]
     pub neon_distrib_dir: PathBuf,
 
     // Default tenant ID to use with the 'neon_local' command line utility, when
     // --tenant_id is not explicitly specified.
+    #[serde(default)]
     pub default_tenant_id: Option<TenantId>,
 
     // used to issue tokens during e.g pg start
+    #[serde(default)]
     pub private_key_path: PathBuf,
 
     pub broker: NeonBroker,
 
-    // Configuration for the storage controller (1 per neon_local environment)
-    pub storage_controller: NeonStorageControllerConf,
-
     /// This Vec must always contain at least one pageserver
-    /// Populdated by [`Self::load_config`] from the individual `pageserver.toml`s.
-    /// NB: not used anymore except for informing users that they need to change their `.neon/config`.
     pub pageservers: Vec<PageServerConf>,
 
+    #[serde(default)]
     pub safekeepers: Vec<SafekeeperConf>,
 
     // Control plane upcall API for pageserver: if None, we will not run storage_controller  If set, this will
     // be propagated into each pageserver's configuration.
+    #[serde(default)]
     pub control_plane_api: Option<Url>,
 
     // Control plane upcall API for storage controller.  If set, this will be propagated into the
     // storage controller's configuration.
+    #[serde(default)]
     pub control_plane_compute_hook_api: Option<Url>,
 
     /// Keep human-readable aliases in memory (and persist them to config), to hide ZId hex strings from the user.
+    #[serde(default)]
     // A `HashMap<String, HashMap<TenantId, TimelineId>>` would be more appropriate here,
     // but deserialization into a generic toml object as `toml::Value::try_from` fails with an error.
     // https://toml.io/en/v1.0.0 does not contain a concept of "a table inside another table".
-    pub branch_name_mappings: HashMap<String, Vec<(TenantId, TimelineId)>>,
-}
-
-/// On-disk state stored in `.neon/config`.
-#[derive(PartialEq, Eq, Clone, Debug, Default, Serialize, Deserialize)]
-#[serde(default, deny_unknown_fields)]
-pub struct OnDiskConfig {
-    pub pg_distrib_dir: PathBuf,
-    pub neon_distrib_dir: PathBuf,
-    pub default_tenant_id: Option<TenantId>,
-    pub private_key_path: PathBuf,
-    pub broker: NeonBroker,
-    pub storage_controller: NeonStorageControllerConf,
-    #[serde(
-        skip_serializing,
-        deserialize_with = "fail_if_pageservers_field_specified"
-    )]
-    pub pageservers: Vec<PageServerConf>,
-    pub safekeepers: Vec<SafekeeperConf>,
-    pub control_plane_api: Option<Url>,
-    pub control_plane_compute_hook_api: Option<Url>,
     branch_name_mappings: HashMap<String, Vec<(TenantId, TimelineId)>>,
 }
 
-fn fail_if_pageservers_field_specified<'de, D>(_: D) -> Result<Vec<PageServerConf>, D::Error>
-where
-    D: serde::Deserializer<'de>,
-{
-    Err(serde::de::Error::custom(
-        "The 'pageservers' field is no longer used; pageserver.toml is now authoritative; \
-         Please remove the `pageservers` from your .neon/config.",
-    ))
-}
-
-/// The description of the neon_local env to be initialized by `neon_local init --config`.
-#[derive(Clone, Debug, Deserialize)]
-#[serde(deny_unknown_fields)]
-pub struct NeonLocalInitConf {
-    // TODO: do we need this? Seems unused
-    pub pg_distrib_dir: Option<PathBuf>,
-    // TODO: do we need this? Seems unused
-    pub neon_distrib_dir: Option<PathBuf>,
-    pub default_tenant_id: TenantId,
-    pub broker: NeonBroker,
-    pub storage_controller: Option<NeonStorageControllerConf>,
-    pub pageservers: Vec<NeonLocalInitPageserverConf>,
-    pub safekeepers: Vec<SafekeeperConf>,
-    pub control_plane_api: Option<Option<Url>>,
-    pub control_plane_compute_hook_api: Option<Option<Url>>,
-}
-
 /// Broker config for cluster internal communication.
 #[derive(Serialize, Deserialize, PartialEq, Eq, Clone, Debug)]
 #[serde(default)]
@@ -145,48 +98,6 @@ pub struct NeonBroker {
     pub listen_addr: SocketAddr,
 }
 
-/// Broker config for cluster internal communication.
-#[derive(Serialize, Deserialize, PartialEq, Eq, Clone, Debug)]
-#[serde(default)]
-pub struct NeonStorageControllerConf {
-    /// Heartbeat timeout before marking a node offline
-    #[serde(with = "humantime_serde")]
-    pub max_offline: Duration,
-
-    #[serde(with = "humantime_serde")]
-    pub max_warming_up: Duration,
-
-    pub start_as_candidate: bool,
-
-    /// Database url used when running multiple storage controller instances
-    pub database_url: Option<SocketAddr>,
-
-    /// Threshold for auto-splitting a tenant into shards
-    pub split_threshold: Option<u64>,
-
-    pub max_secondary_lag_bytes: Option<u64>,
-}
-
-impl NeonStorageControllerConf {
-    // Use a shorter pageserver unavailability interval than the default to speed up tests.
-    const DEFAULT_MAX_OFFLINE_INTERVAL: std::time::Duration = std::time::Duration::from_secs(10);
-
-    const DEFAULT_MAX_WARMING_UP_INTERVAL: std::time::Duration = std::time::Duration::from_secs(30);
-}
-
-impl Default for NeonStorageControllerConf {
-    fn default() -> Self {
-        Self {
-            max_offline: Self::DEFAULT_MAX_OFFLINE_INTERVAL,
-            max_warming_up: Self::DEFAULT_MAX_WARMING_UP_INTERVAL,
-            start_as_candidate: false,
-            database_url: None,
-            split_threshold: None,
-            max_secondary_lag_bytes: None,
-        }
-    }
-}
-
 // Dummy Default impl to satisfy Deserialize derive.
 impl Default for NeonBroker {
     fn default() -> Self {
@@ -202,18 +113,22 @@ impl NeonBroker {
     }
 }
 
-// neon_local needs to know this subset of pageserver configuration.
-// For legacy reasons, this information is duplicated from `pageserver.toml` into `.neon/config`.
-// It can get stale if `pageserver.toml` is changed.
-// TODO(christian): don't store this at all in `.neon/config`, always load it from `pageserver.toml`
 #[derive(Serialize, Deserialize, PartialEq, Eq, Clone, Debug)]
 #[serde(default, deny_unknown_fields)]
 pub struct PageServerConf {
+    // node id
     pub id: NodeId,
+
+    // Pageserver connection settings
     pub listen_pg_addr: String,
     pub listen_http_addr: String,
+
+    // auth type used for the PG and HTTP ports
     pub pg_auth_type: AuthType,
     pub http_auth_type: AuthType,
+
+    pub(crate) virtual_file_io_engine: Option<String>,
+    pub(crate) get_vectored_impl: Option<String>,
 }
 
 impl Default for PageServerConf {
@@ -224,40 +139,8 @@ impl Default for PageServerConf {
             listen_http_addr: String::new(),
             pg_auth_type: AuthType::Trust,
             http_auth_type: AuthType::Trust,
-        }
-    }
-}
-
-/// The toml that can be passed to `neon_local init --config`.
-/// This is a subset of the `pageserver.toml` configuration.
-// TODO(christian): use pageserver_api::config::ConfigToml (PR #7656)
-#[derive(Clone, Debug, serde::Deserialize, serde::Serialize)]
-pub struct NeonLocalInitPageserverConf {
-    pub id: NodeId,
-    pub listen_pg_addr: String,
-    pub listen_http_addr: String,
-    pub pg_auth_type: AuthType,
-    pub http_auth_type: AuthType,
-    #[serde(flatten)]
-    pub other: HashMap<String, toml::Value>,
-}
-
-impl From<&NeonLocalInitPageserverConf> for PageServerConf {
-    fn from(conf: &NeonLocalInitPageserverConf) -> Self {
-        let NeonLocalInitPageserverConf {
-            id,
-            listen_pg_addr,
-            listen_http_addr,
-            pg_auth_type,
-            http_auth_type,
-            other: _,
-        } = conf;
-        Self {
-            id: *id,
-            listen_pg_addr: listen_pg_addr.clone(),
-            listen_http_addr: listen_http_addr.clone(),
-            pg_auth_type: *pg_auth_type,
-            http_auth_type: *http_auth_type,
+            virtual_file_io_engine: None,
+            get_vectored_impl: None,
         }
     }
 }
@@ -273,7 +156,6 @@ pub struct SafekeeperConf {
     pub remote_storage: Option<String>,
     pub backup_threads: Option<u32>,
     pub auth_enabled: bool,
-    pub listen_addr: Option<String>,
 }
 
 impl Default for SafekeeperConf {
@@ -287,7 +169,6 @@ impl Default for SafekeeperConf {
             remote_storage: None,
             backup_threads: None,
             auth_enabled: false,
-            listen_addr: None,
         }
     }
 }
@@ -340,16 +221,11 @@ impl LocalEnv {
         }
     }
 
-    pub fn pg_dir(&self, pg_version: u32, dir_name: &str) -> anyhow::Result<PathBuf> {
-        Ok(self.pg_distrib_dir(pg_version)?.join(dir_name))
-    }
-
     pub fn pg_bin_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
-        self.pg_dir(pg_version, "bin")
+        Ok(self.pg_distrib_dir(pg_version)?.join("bin"))
     }
-
     pub fn pg_lib_dir(&self, pg_version: u32) -> anyhow::Result<PathBuf> {
-        self.pg_dir(pg_version, "lib")
+        Ok(self.pg_distrib_dir(pg_version)?.join("lib"))
     }
 
     pub fn pageserver_bin(&self) -> PathBuf {
@@ -399,36 +275,6 @@ impl LocalEnv {
         }
     }
 
-    /// Inspect the base data directory and extract the instance id and instance directory path
-    /// for all storage controller instances
-    pub async fn storage_controller_instances(&self) -> std::io::Result<Vec<(u8, PathBuf)>> {
-        let mut instances = Vec::default();
-
-        let dir = std::fs::read_dir(self.base_data_dir.clone())?;
-        for dentry in dir {
-            let dentry = dentry?;
-            let is_dir = dentry.metadata()?.is_dir();
-            let filename = dentry.file_name().into_string().unwrap();
-            let parsed_instance_id = match filename.strip_prefix("storage_controller_") {
-                Some(suffix) => suffix.parse::<u8>().ok(),
-                None => None,
-            };
-
-            let is_instance_dir = is_dir && parsed_instance_id.is_some();
-
-            if !is_instance_dir {
-                continue;
-            }
-
-            instances.push((
-                parsed_instance_id.expect("Checked previously"),
-                dentry.path(),
-            ));
-        }
-
-        Ok(instances)
-    }
-
     pub fn register_branch_mapping(
         &mut self,
         branch_name: String,
@@ -480,8 +326,44 @@ impl LocalEnv {
             .collect()
     }
 
-    ///  Construct `Self` from on-disk state.
-    pub fn load_config(repopath: &Path) -> anyhow::Result<Self> {
+    /// Create a LocalEnv from a config file.
+    ///
+    /// Unlike 'load_config', this function fills in any defaults that are missing
+    /// from the config file.
+    pub fn parse_config(toml: &str) -> anyhow::Result<Self> {
+        let mut env: LocalEnv = toml::from_str(toml)?;
+
+        // Find postgres binaries.
+        // Follow POSTGRES_DISTRIB_DIR if set, otherwise look in "pg_install".
+        // Note that later in the code we assume, that distrib dirs follow the same pattern
+        // for all postgres versions.
+        if env.pg_distrib_dir == Path::new("") {
+            if let Some(postgres_bin) = env::var_os("POSTGRES_DISTRIB_DIR") {
+                env.pg_distrib_dir = postgres_bin.into();
+            } else {
+                let cwd = env::current_dir()?;
+                env.pg_distrib_dir = cwd.join("pg_install")
+            }
+        }
+
+        // Find neon binaries.
+        if env.neon_distrib_dir == Path::new("") {
+            env.neon_distrib_dir = env::current_exe()?.parent().unwrap().to_owned();
+        }
+
+        if env.pageservers.is_empty() {
+            anyhow::bail!("Configuration must contain at least one pageserver");
+        }
+
+        env.base_data_dir = base_path();
+
+        Ok(env)
+    }
+
+    /// Locate and load config
+    pub fn load_config() -> anyhow::Result<Self> {
+        let repopath = base_path();
+
         if !repopath.exists() {
             bail!(
                 "Neon config is not found in {}. You need to run 'neon_local init' first",
@@ -492,140 +374,38 @@ impl LocalEnv {
         // TODO: check that it looks like a neon repository
 
         // load and parse file
-        let config_file_contents = fs::read_to_string(repopath.join("config"))?;
-        let on_disk_config: OnDiskConfig = toml::from_str(config_file_contents.as_str())?;
-        let mut env = {
-            let OnDiskConfig {
-                pg_distrib_dir,
-                neon_distrib_dir,
-                default_tenant_id,
-                private_key_path,
-                broker,
-                storage_controller,
-                pageservers,
-                safekeepers,
-                control_plane_api,
-                control_plane_compute_hook_api,
-                branch_name_mappings,
-            } = on_disk_config;
-            LocalEnv {
-                base_data_dir: repopath.to_owned(),
-                pg_distrib_dir,
-                neon_distrib_dir,
-                default_tenant_id,
-                private_key_path,
-                broker,
-                storage_controller,
-                pageservers,
-                safekeepers,
-                control_plane_api,
-                control_plane_compute_hook_api,
-                branch_name_mappings,
-            }
-        };
+        let config = fs::read_to_string(repopath.join("config"))?;
+        let mut env: LocalEnv = toml::from_str(config.as_str())?;
 
-        // The source of truth for pageserver configuration is the pageserver.toml.
-        assert!(
-            env.pageservers.is_empty(),
-            "we ensure this during deserialization"
-        );
-        env.pageservers = {
-            let iter = std::fs::read_dir(repopath).context("open dir")?;
-            let mut pageservers = Vec::new();
-            for res in iter {
-                let dentry = res?;
-                const PREFIX: &str = "pageserver_";
-                let dentry_name = dentry
-                    .file_name()
-                    .into_string()
-                    .ok()
-                    .with_context(|| format!("non-utf8 dentry: {:?}", dentry.path()))
-                    .unwrap();
-                if !dentry_name.starts_with(PREFIX) {
-                    continue;
-                }
-                if !dentry.file_type().context("determine file type")?.is_dir() {
-                    anyhow::bail!("expected a directory, got {:?}", dentry.path());
-                }
-                let id = dentry_name[PREFIX.len()..]
-                    .parse::<NodeId>()
-                    .with_context(|| format!("parse id from {:?}", dentry.path()))?;
-                // TODO(christian): use pageserver_api::config::ConfigToml (PR #7656)
-                #[derive(serde::Serialize, serde::Deserialize)]
-                // (allow unknown fields, unlike PageServerConf)
-                struct PageserverConfigTomlSubset {
-                    listen_pg_addr: String,
-                    listen_http_addr: String,
-                    pg_auth_type: AuthType,
-                    http_auth_type: AuthType,
-                }
-                let config_toml_path = dentry.path().join("pageserver.toml");
-                let config_toml: PageserverConfigTomlSubset = toml_edit::de::from_str(
-                    &std::fs::read_to_string(&config_toml_path)
-                        .with_context(|| format!("read {:?}", config_toml_path))?,
-                )
-                .context("parse pageserver.toml")?;
-                let identity_toml_path = dentry.path().join("identity.toml");
-                #[derive(serde::Serialize, serde::Deserialize)]
-                struct IdentityTomlSubset {
-                    id: NodeId,
-                }
-                let identity_toml: IdentityTomlSubset = toml_edit::de::from_str(
-                    &std::fs::read_to_string(&identity_toml_path)
-                        .with_context(|| format!("read {:?}", identity_toml_path))?,
-                )
-                .context("parse identity.toml")?;
-                let PageserverConfigTomlSubset {
-                    listen_pg_addr,
-                    listen_http_addr,
-                    pg_auth_type,
-                    http_auth_type,
-                } = config_toml;
-                let IdentityTomlSubset {
-                    id: identity_toml_id,
-                } = identity_toml;
-                let conf = PageServerConf {
-                    id: {
-                        anyhow::ensure!(
-                            identity_toml_id == id,
-                            "id mismatch: identity.toml:id={identity_toml_id} pageserver_(.*) id={id}",
-                        );
-                        id
-                    },
-                    listen_pg_addr,
-                    listen_http_addr,
-                    pg_auth_type,
-                    http_auth_type,
-                };
-                pageservers.push(conf);
-            }
-            pageservers
-        };
+        env.base_data_dir = repopath;
 
         Ok(env)
     }
 
-    pub fn persist_config(&self) -> anyhow::Result<()> {
-        Self::persist_config_impl(
-            &self.base_data_dir,
-            &OnDiskConfig {
-                pg_distrib_dir: self.pg_distrib_dir.clone(),
-                neon_distrib_dir: self.neon_distrib_dir.clone(),
-                default_tenant_id: self.default_tenant_id,
-                private_key_path: self.private_key_path.clone(),
-                broker: self.broker.clone(),
-                storage_controller: self.storage_controller.clone(),
-                pageservers: vec![], // it's skip_serializing anyway
-                safekeepers: self.safekeepers.clone(),
-                control_plane_api: self.control_plane_api.clone(),
-                control_plane_compute_hook_api: self.control_plane_compute_hook_api.clone(),
-                branch_name_mappings: self.branch_name_mappings.clone(),
-            },
-        )
-    }
+    pub fn persist_config(&self, base_path: &Path) -> anyhow::Result<()> {
+        // Currently, the user first passes a config file with 'neon_local init --config=<path>'
+        // We read that in, in `create_config`, and fill any missing defaults. Then it's saved
+        // to .neon/config. TODO: We lose any formatting and comments along the way, which is
+        // a bit sad.
+        let mut conf_content = r#"# This file describes a local deployment of the page server
+# and safekeeeper node. It is read by the 'neon_local' command-line
+# utility.
+"#
+        .to_string();
+
+        // Convert the LocalEnv to a toml file.
+        //
+        // This could be as simple as this:
+        //
+        // conf_content += &toml::to_string_pretty(env)?;
+        //
+        // But it results in a "values must be emitted before tables". I'm not sure
+        // why, AFAICS the table, i.e. 'safekeepers: Vec<SafekeeperConf>' is last.
+        // Maybe rust reorders the fields to squeeze avoid padding or something?
+        // In any case, converting to toml::Value first, and serializing that, works.
+        // See https://github.com/alexcrichton/toml-rs/issues/142
+        conf_content += &toml::to_string_pretty(&toml::Value::try_from(self)?)?;
 
-    pub fn persist_config_impl(base_path: &Path, config: &OnDiskConfig) -> anyhow::Result<()> {
-        let conf_content = &toml::to_string_pretty(config)?;
         let target_config_path = base_path.join("config");
         fs::write(&target_config_path, conf_content).with_context(|| {
             format!(
@@ -650,13 +430,17 @@ impl LocalEnv {
         }
     }
 
-    /// Materialize the [`NeonLocalInitConf`] to disk. Called during [`neon_local init`].
-    pub fn init(conf: NeonLocalInitConf, force: &InitForceMode) -> anyhow::Result<()> {
-        let base_path = base_path();
-        assert_ne!(base_path, Path::new(""));
-        let base_path = &base_path;
+    //
+    // Initialize a new Neon repository
+    //
+    pub fn init(&mut self, pg_version: u32, force: &InitForceMode) -> anyhow::Result<()> {
+        // check if config already exists
+        let base_path = &self.base_data_dir;
+        ensure!(
+            base_path != Path::new(""),
+            "repository base path is missing"
+        );
 
-        // create base_path dir
         if base_path.exists() {
             match force {
                 InitForceMode::MustNotExist => {
@@ -688,115 +472,74 @@ impl LocalEnv {
                 }
             }
         }
+
+        if !self.pg_bin_dir(pg_version)?.join("postgres").exists() {
+            bail!(
+                "Can't find postgres binary at {}",
+                self.pg_bin_dir(pg_version)?.display()
+            );
+        }
+        for binary in ["pageserver", "safekeeper"] {
+            if !self.neon_distrib_dir.join(binary).exists() {
+                bail!(
+                    "Can't find binary '{binary}' in neon distrib dir '{}'",
+                    self.neon_distrib_dir.display()
+                );
+            }
+        }
+
         if !base_path.exists() {
             fs::create_dir(base_path)?;
         }
 
-        let NeonLocalInitConf {
-            pg_distrib_dir,
-            neon_distrib_dir,
-            default_tenant_id,
-            broker,
-            storage_controller,
-            pageservers,
-            safekeepers,
-            control_plane_api,
-            control_plane_compute_hook_api,
-        } = conf;
-
-        // Find postgres binaries.
-        // Follow POSTGRES_DISTRIB_DIR if set, otherwise look in "pg_install".
-        // Note that later in the code we assume, that distrib dirs follow the same pattern
-        // for all postgres versions.
-        let pg_distrib_dir = pg_distrib_dir.unwrap_or_else(|| {
-            if let Some(postgres_bin) = env::var_os("POSTGRES_DISTRIB_DIR") {
-                postgres_bin.into()
-            } else {
-                let cwd = env::current_dir().unwrap();
-                cwd.join("pg_install")
-            }
-        });
-
-        // Find neon binaries.
-        let neon_distrib_dir = neon_distrib_dir
-            .unwrap_or_else(|| env::current_exe().unwrap().parent().unwrap().to_owned());
-
         // Generate keypair for JWT.
         //
         // The keypair is only needed if authentication is enabled in any of the
         // components. For convenience, we generate the keypair even if authentication
         // is not enabled, so that you can easily enable it after the initialization
-        // step.
-        generate_auth_keys(
-            base_path.join("auth_private_key.pem").as_path(),
-            base_path.join("auth_public_key.pem").as_path(),
-        )
-        .context("generate auth keys")?;
-        let private_key_path = PathBuf::from("auth_private_key.pem");
-
-        // create the runtime type because the remaining initialization code below needs
-        // a LocalEnv instance op operation
-        // TODO: refactor to avoid this, LocalEnv should only be constructed from on-disk state
-        let env = LocalEnv {
-            base_data_dir: base_path.clone(),
-            pg_distrib_dir,
-            neon_distrib_dir,
-            default_tenant_id: Some(default_tenant_id),
-            private_key_path,
-            broker,
-            storage_controller: storage_controller.unwrap_or_default(),
-            pageservers: pageservers.iter().map(Into::into).collect(),
-            safekeepers,
-            control_plane_api: control_plane_api.unwrap_or_default(),
-            control_plane_compute_hook_api: control_plane_compute_hook_api.unwrap_or_default(),
-            branch_name_mappings: Default::default(),
-        };
-
-        // create endpoints dir
-        fs::create_dir_all(env.endpoints_path())?;
-
-        // create safekeeper dirs
-        for safekeeper in &env.safekeepers {
-            fs::create_dir_all(SafekeeperNode::datadir_path_by_id(&env, safekeeper.id))?;
+        // step. However, if the key generation fails, we treat it as non-fatal if
+        // authentication was not enabled.
+        if self.private_key_path == PathBuf::new() {
+            match generate_auth_keys(
+                base_path.join("auth_private_key.pem").as_path(),
+                base_path.join("auth_public_key.pem").as_path(),
+            ) {
+                Ok(()) => {
+                    self.private_key_path = PathBuf::from("auth_private_key.pem");
+                }
+                Err(e) => {
+                    if !self.auth_keys_needed() {
+                        eprintln!("Could not generate keypair for JWT authentication: {e}");
+                        eprintln!("Continuing anyway because authentication was not enabled");
+                        self.private_key_path = PathBuf::from("auth_private_key.pem");
+                    } else {
+                        return Err(e);
+                    }
+                }
+            }
         }
 
-        // initialize pageserver state
-        for (i, ps) in pageservers.into_iter().enumerate() {
-            let runtime_ps = &env.pageservers[i];
-            assert_eq!(&PageServerConf::from(&ps), runtime_ps);
-            fs::create_dir(env.pageserver_data_dir(ps.id))?;
-            PageServerNode::from_env(&env, runtime_ps)
-                .initialize(ps)
-                .context("pageserver init failed")?;
+        fs::create_dir_all(self.endpoints_path())?;
+
+        for safekeeper in &self.safekeepers {
+            fs::create_dir_all(SafekeeperNode::datadir_path_by_id(self, safekeeper.id))?;
         }
 
-        // setup remote remote location for default LocalFs remote storage
-        std::fs::create_dir_all(env.base_data_dir.join(PAGESERVER_REMOTE_STORAGE_DIR))?;
+        self.persist_config(base_path)
+    }
 
-        env.persist_config()
+    fn auth_keys_needed(&self) -> bool {
+        self.pageservers.iter().any(|ps| {
+            ps.pg_auth_type == AuthType::NeonJWT || ps.http_auth_type == AuthType::NeonJWT
+        }) || self.safekeepers.iter().any(|sk| sk.auth_enabled)
     }
 }
 
-pub fn base_path() -> PathBuf {
-    let path = match std::env::var_os("NEON_REPO_DIR") {
-        Some(val) => {
-            let path = PathBuf::from(val);
-            if !path.is_absolute() {
-                // repeat the env var in the error because our default is always absolute
-                panic!("NEON_REPO_DIR must be an absolute path, got {path:?}");
-            }
-            path
-        }
-        None => {
-            let pwd = std::env::current_dir()
-                // technically this can fail but it's quite unlikeley
-                .expect("determine current directory");
-            let pwd_abs = pwd.canonicalize().expect("canonicalize current directory");
-            pwd_abs.join(".neon")
-        }
-    };
-    assert!(path.is_absolute());
-    path
+fn base_path() -> PathBuf {
+    match std::env::var_os("NEON_REPO_DIR") {
+        Some(val) => PathBuf::from(val),
+        None => PathBuf::from(".neon"),
+    }
 }
 
 /// Generate a public/private key pair for JWT authentication
@@ -835,3 +578,31 @@ fn generate_auth_keys(private_key_path: &Path, public_key_path: &Path) -> anyhow
     }
     Ok(())
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn simple_conf_parsing() {
+        let simple_conf_toml = include_str!("../simple.conf");
+        let simple_conf_parse_result = LocalEnv::parse_config(simple_conf_toml);
+        assert!(
+            simple_conf_parse_result.is_ok(),
+            "failed to parse simple config {simple_conf_toml}, reason: {simple_conf_parse_result:?}"
+        );
+
+        let string_to_replace = "listen_addr = '127.0.0.1:50051'";
+        let spoiled_url_str = "listen_addr = '!@$XOXO%^&'";
+        let spoiled_url_toml = simple_conf_toml.replace(string_to_replace, spoiled_url_str);
+        assert!(
+            spoiled_url_toml.contains(spoiled_url_str),
+            "Failed to replace string {string_to_replace} in the toml file {simple_conf_toml}"
+        );
+        let spoiled_url_parse_result = LocalEnv::parse_config(&spoiled_url_toml);
+        assert!(
+            spoiled_url_parse_result.is_err(),
+            "expected toml with invalid Url {spoiled_url_toml} to fail the parsing, but got {spoiled_url_parse_result:?}"
+        );
+    }
+}

control_plane/src/pageserver.rs

@@ -1,37 +1,36 @@
 //! Code to manage pageservers
 //!
-//! In the local test environment, the data for each pageserver is stored in
+//! In the local test environment, the pageserver stores its data directly in
 //!
-//! ```text
-//!   .neon/pageserver_<pageserver_id>
-//! ```
+//!   .neon/
 //!
+use std::borrow::Cow;
 use std::collections::HashMap;
 
 use std::io;
 use std::io::Write;
 use std::num::NonZeroU64;
 use std::path::PathBuf;
-use std::str::FromStr;
+use std::process::Command;
 use std::time::Duration;
 
 use anyhow::{bail, Context};
 use camino::Utf8PathBuf;
+use futures::SinkExt;
 use pageserver_api::models::{
-    self, AuxFilePolicy, LocationConfig, TenantHistorySize, TenantInfo, TimelineInfo,
+    self, LocationConfig, ShardParameters, TenantHistorySize, TenantInfo, TimelineInfo,
 };
 use pageserver_api::shard::TenantShardId;
 use pageserver_client::mgmt_api;
 use postgres_backend::AuthType;
 use postgres_connection::{parse_host_port, PgConnectionConfig};
 use utils::auth::{Claims, Scope};
-use utils::id::NodeId;
 use utils::{
     id::{TenantId, TimelineId},
     lsn::Lsn,
 };
 
-use crate::local_env::{NeonLocalInitPageserverConf, PageServerConf};
+use crate::local_env::PageServerConf;
 use crate::{background_process, local_env::LocalEnv};
 
 /// Directory within .neon which will be used by default for LocalFs remote storage.
@@ -75,27 +74,57 @@ impl PageServerNode {
         }
     }
 
-    fn pageserver_make_identity_toml(&self, node_id: NodeId) -> toml_edit::Document {
-        toml_edit::Document::from_str(&format!("id={node_id}")).unwrap()
-    }
-
-    fn pageserver_init_make_toml(
-        &self,
-        conf: NeonLocalInitPageserverConf,
-    ) -> anyhow::Result<toml_edit::Document> {
-        assert_eq!(&PageServerConf::from(&conf), &self.conf, "during neon_local init, we derive the runtime state of ps conf (self.conf) from the --config flag fully");
-
-        // TODO(christian): instead of what we do here, create a pageserver_api::config::ConfigToml (PR #7656)
-
+    /// Merge overrides provided by the user on the command line with our default overides derived from neon_local configuration.
+    ///
+    /// These all end up on the command line of the `pageserver` binary.
+    fn neon_local_overrides(&self, cli_overrides: &[&str]) -> Vec<String> {
         // FIXME: the paths should be shell-escaped to handle paths with spaces, quotas etc.
         let pg_distrib_dir_param = format!(
             "pg_distrib_dir='{}'",
             self.env.pg_distrib_dir_raw().display()
         );
 
+        let PageServerConf {
+            id,
+            listen_pg_addr,
+            listen_http_addr,
+            pg_auth_type,
+            http_auth_type,
+            virtual_file_io_engine,
+            get_vectored_impl,
+        } = &self.conf;
+
+        let id = format!("id={}", id);
+
+        let http_auth_type_param = format!("http_auth_type='{}'", http_auth_type);
+        let listen_http_addr_param = format!("listen_http_addr='{}'", listen_http_addr);
+
+        let pg_auth_type_param = format!("pg_auth_type='{}'", pg_auth_type);
+        let listen_pg_addr_param = format!("listen_pg_addr='{}'", listen_pg_addr);
+        let virtual_file_io_engine = if let Some(virtual_file_io_engine) = virtual_file_io_engine {
+            format!("virtual_file_io_engine='{virtual_file_io_engine}'")
+        } else {
+            String::new()
+        };
+        let get_vectored_impl = if let Some(get_vectored_impl) = get_vectored_impl {
+            format!("get_vectored_impl='{get_vectored_impl}'")
+        } else {
+            String::new()
+        };
+
         let broker_endpoint_param = format!("broker_endpoint='{}'", self.env.broker.client_url());
 
-        let mut overrides = vec![pg_distrib_dir_param, broker_endpoint_param];
+        let mut overrides = vec![
+            id,
+            pg_distrib_dir_param,
+            http_auth_type_param,
+            pg_auth_type_param,
+            listen_http_addr_param,
+            listen_pg_addr_param,
+            broker_endpoint_param,
+            virtual_file_io_engine,
+            get_vectored_impl,
+        ];
 
         if let Some(control_plane_api) = &self.env.control_plane_api {
             overrides.push(format!(
@@ -105,7 +134,7 @@ impl PageServerNode {
 
             // Storage controller uses the same auth as pageserver: if JWT is enabled
             // for us, we will also need it to talk to them.
-            if matches!(conf.http_auth_type, AuthType::NeonJWT) {
+            if matches!(http_auth_type, AuthType::NeonJWT) {
                 let jwt_token = self
                     .env
                     .generate_auth_token(&Claims::new(None, Scope::GenerationsApi))
@@ -114,43 +143,31 @@ impl PageServerNode {
             }
         }
 
-        if !conf.other.contains_key("remote_storage") {
+        if !cli_overrides
+            .iter()
+            .any(|c| c.starts_with("remote_storage"))
+        {
             overrides.push(format!(
                 "remote_storage={{local_path='../{PAGESERVER_REMOTE_STORAGE_DIR}'}}"
             ));
         }
 
-        if conf.http_auth_type != AuthType::Trust || conf.pg_auth_type != AuthType::Trust {
+        if *http_auth_type != AuthType::Trust || *pg_auth_type != AuthType::Trust {
             // Keys are generated in the toplevel repo dir, pageservers' workdirs
             // are one level below that, so refer to keys with ../
             overrides.push("auth_validation_public_key_path='../auth_public_key.pem'".to_owned());
         }
 
         // Apply the user-provided overrides
-        overrides.push({
-            let mut doc =
-                toml_edit::ser::to_document(&conf).expect("we deserialized this from toml earlier");
-            // `id` is written out to `identity.toml` instead of `pageserver.toml`
-            doc.remove("id").expect("it's part of the struct");
-            doc.to_string()
-        });
+        overrides.extend(cli_overrides.iter().map(|&c| c.to_owned()));
 
-        // Turn `overrides` into a toml document.
-        // TODO: above code is legacy code, it should be refactored to use toml_edit directly.
-        let mut config_toml = toml_edit::Document::new();
-        for fragment_str in overrides {
-            let fragment = toml_edit::Document::from_str(&fragment_str)
-                .expect("all fragments in `overrides` are valid toml documents, this function controls that");
-            for (key, item) in fragment.iter() {
-                config_toml.insert(key, item.clone());
-            }
-        }
-        Ok(config_toml)
+        overrides
     }
 
     /// Initializes a pageserver node by creating its config with the overrides provided.
-    pub fn initialize(&self, conf: NeonLocalInitPageserverConf) -> anyhow::Result<()> {
-        self.pageserver_init(conf)
+    pub fn initialize(&self, config_overrides: &[&str]) -> anyhow::Result<()> {
+        // First, run `pageserver --init` and wait for it to write a config into FS and exit.
+        self.pageserver_init(config_overrides)
             .with_context(|| format!("Failed to run init for pageserver node {}", self.conf.id))
     }
 
@@ -166,11 +183,11 @@ impl PageServerNode {
             .expect("non-Unicode path")
     }
 
-    pub async fn start(&self, retry_timeout: &Duration) -> anyhow::Result<()> {
-        self.start_node(retry_timeout).await
+    pub async fn start(&self, config_overrides: &[&str]) -> anyhow::Result<()> {
+        self.start_node(config_overrides, false).await
     }
 
-    fn pageserver_init(&self, conf: NeonLocalInitPageserverConf) -> anyhow::Result<()> {
+    fn pageserver_init(&self, config_overrides: &[&str]) -> anyhow::Result<()> {
         let datadir = self.repo_path();
         let node_id = self.conf.id;
         println!(
@@ -181,33 +198,29 @@ impl PageServerNode {
         );
         io::stdout().flush()?;
 
-        let config = self
-            .pageserver_init_make_toml(conf)
-            .context("make pageserver toml")?;
-        let config_file_path = datadir.join("pageserver.toml");
-        let mut config_file = std::fs::OpenOptions::new()
-            .create_new(true)
-            .write(true)
-            .open(&config_file_path)
-            .with_context(|| format!("open pageserver toml for write: {config_file_path:?}"))?;
-        config_file
-            .write_all(config.to_string().as_bytes())
-            .context("write pageserver toml")?;
-        drop(config_file);
+        if !datadir.exists() {
+            std::fs::create_dir(&datadir)?;
+        }
 
-        let identity_file_path = datadir.join("identity.toml");
-        let mut identity_file = std::fs::OpenOptions::new()
-            .create_new(true)
-            .write(true)
-            .open(identity_file_path)
-            .with_context(|| format!("open identity toml for write: {config_file_path:?}"))?;
-        let identity_toml = self.pageserver_make_identity_toml(node_id);
-        identity_file
-            .write_all(identity_toml.to_string().as_bytes())
-            .context("write identity toml")?;
-        drop(identity_toml);
+        let datadir_path_str = datadir.to_str().with_context(|| {
+            format!("Cannot start pageserver node {node_id} in path that has no string representation: {datadir:?}")
+        })?;
+        let mut args = self.pageserver_basic_args(config_overrides, datadir_path_str);
+        args.push(Cow::Borrowed("--init"));
 
-        // TODO: invoke a TBD config-check command to validate that pageserver will start with the written config
+        let init_output = Command::new(self.env.pageserver_bin())
+            .args(args.iter().map(Cow::as_ref))
+            .envs(self.pageserver_env_variables()?)
+            .output()
+            .with_context(|| format!("Failed to run pageserver init for node {node_id}"))?;
+
+        anyhow::ensure!(
+            init_output.status.success(),
+            "Pageserver init for node {} did not finish successfully, stdout: {}, stderr: {}",
+            node_id,
+            String::from_utf8_lossy(&init_output.stdout),
+            String::from_utf8_lossy(&init_output.stderr),
+        );
 
         // Write metadata file, used by pageserver on startup to register itself with
         // the storage controller
@@ -221,13 +234,12 @@ impl PageServerNode {
         // situation: the metadata is written by some other script.
         std::fs::write(
             metadata_path,
-            serde_json::to_vec(&pageserver_api::config::NodeMetadata {
-                postgres_host: "localhost".to_string(),
-                postgres_port: self.pg_connection_config.port(),
-                http_host: "localhost".to_string(),
-                http_port,
-                other: HashMap::new(),
-            })
+            serde_json::to_vec(&serde_json::json!({
+                "host": "localhost",
+                "port": self.pg_connection_config.port(),
+                "http_host": "localhost",
+                "http_port": http_port,
+            }))
             .unwrap(),
         )
         .expect("Failed to write metadata file");
@@ -235,15 +247,18 @@ impl PageServerNode {
         Ok(())
     }
 
-    async fn start_node(&self, retry_timeout: &Duration) -> anyhow::Result<()> {
+    async fn start_node(
+        &self,
+        config_overrides: &[&str],
+        update_config: bool,
+    ) -> anyhow::Result<()> {
         // TODO: using a thread here because start_process() is not async but we need to call check_status()
         let datadir = self.repo_path();
         print!(
-            "Starting pageserver node {} at '{}' in {:?}, retrying for {:?}",
+            "Starting pageserver node {} at '{}' in {:?}",
             self.conf.id,
             self.pg_connection_config.raw_address(),
-            datadir,
-            retry_timeout
+            datadir
         );
         io::stdout().flush().context("flush stdout")?;
 
@@ -253,15 +268,17 @@ impl PageServerNode {
                 self.conf.id, datadir,
             )
         })?;
-        let args = vec!["-D", datadir_path_str];
+        let mut args = self.pageserver_basic_args(config_overrides, datadir_path_str);
+        if update_config {
+            args.push(Cow::Borrowed("--update-config"));
+        }
         background_process::start_process(
             "pageserver",
             &datadir,
             &self.env.pageserver_bin(),
-            args,
+            args.iter().map(Cow::as_ref),
             self.pageserver_env_variables()?,
             background_process::InitialPidFile::Expect(self.pid_file()),
-            retry_timeout,
             || async {
                 let st = self.check_status().await;
                 match st {
@@ -276,6 +293,22 @@ impl PageServerNode {
         Ok(())
     }
 
+    fn pageserver_basic_args<'a>(
+        &self,
+        config_overrides: &'a [&'a str],
+        datadir_path_str: &'a str,
+    ) -> Vec<Cow<'a, str>> {
+        let mut args = vec![Cow::Borrowed("-D"), Cow::Borrowed(datadir_path_str)];
+
+        let overrides = self.neon_local_overrides(config_overrides);
+        for config_override in overrides {
+            args.push(Cow::Borrowed("-c"));
+            args.push(Cow::Owned(config_override));
+        }
+
+        args
+    }
+
     fn pageserver_env_variables(&self) -> anyhow::Result<Vec<(String, String)>> {
         // FIXME: why is this tied to pageserver's auth type? Whether or not the safekeeper
         // needs a token, and how to generate that token, seems independent to whether
@@ -372,6 +405,11 @@ impl PageServerNode {
                 .map(|x| x.parse::<NonZeroU64>())
                 .transpose()
                 .context("Failed to parse 'max_lsn_wal_lag' as non zero integer")?,
+            trace_read_requests: settings
+                .remove("trace_read_requests")
+                .map(|x| x.parse::<bool>())
+                .transpose()
+                .context("Failed to parse 'trace_read_requests' as bool")?,
             eviction_policy: settings
                 .remove("eviction_policy")
                 .map(serde_json::from_str)
@@ -396,15 +434,6 @@ impl PageServerNode {
                 .map(serde_json::from_str)
                 .transpose()
                 .context("parse `timeline_get_throttle` from json")?,
-            switch_aux_file_policy: settings
-                .remove("switch_aux_file_policy")
-                .map(|x| x.parse::<AuxFilePolicy>())
-                .transpose()
-                .context("Failed to parse 'switch_aux_file_policy'")?,
-            lsn_lease_length: settings.remove("lsn_lease_length").map(|x| x.to_string()),
-            lsn_lease_length_for_ts: settings
-                .remove("lsn_lease_length_for_ts")
-                .map(|x| x.to_string()),
         };
         if !settings.is_empty() {
             bail!("Unrecognized tenant settings: {settings:?}")
@@ -413,6 +442,28 @@ impl PageServerNode {
         }
     }
 
+    pub async fn tenant_create(
+        &self,
+        new_tenant_id: TenantId,
+        generation: Option<u32>,
+        settings: HashMap<&str, &str>,
+    ) -> anyhow::Result<TenantId> {
+        let config = Self::parse_config(settings.clone())?;
+
+        let request = models::TenantCreateRequest {
+            new_tenant_id: TenantShardId::unsharded(new_tenant_id),
+            generation,
+            config,
+            shard_parameters: ShardParameters::default(),
+            // Placement policy is not meaningful for creations not done via storage controller
+            placement_policy: None,
+        };
+        if !settings.is_empty() {
+            bail!("Unrecognized tenant settings: {settings:?}")
+        }
+        Ok(self.http_client.tenant_create(&request).await?)
+    }
+
     pub async fn tenant_config(
         &self,
         tenant_id: TenantId,
@@ -472,6 +523,11 @@ impl PageServerNode {
                     .map(|x| x.parse::<NonZeroU64>())
                     .transpose()
                     .context("Failed to parse 'max_lsn_wal_lag' as non zero integer")?,
+                trace_read_requests: settings
+                    .remove("trace_read_requests")
+                    .map(|x| x.parse::<bool>())
+                    .transpose()
+                    .context("Failed to parse 'trace_read_requests' as bool")?,
                 eviction_policy: settings
                     .remove("eviction_policy")
                     .map(serde_json::from_str)
@@ -496,15 +552,6 @@ impl PageServerNode {
                     .map(serde_json::from_str)
                     .transpose()
                     .context("parse `timeline_get_throttle` from json")?,
-                switch_aux_file_policy: settings
-                    .remove("switch_aux_file_policy")
-                    .map(|x| x.parse::<AuxFilePolicy>())
-                    .transpose()
-                    .context("Failed to parse 'switch_aux_file_policy'")?,
-                lsn_lease_length: settings.remove("lsn_lease_length").map(|x| x.to_string()),
-                lsn_lease_length_for_ts: settings
-                    .remove("lsn_lease_length_for_ts")
-                    .map(|x| x.to_string()),
             }
         };
 
@@ -578,39 +625,60 @@ impl PageServerNode {
         pg_wal: Option<(Lsn, PathBuf)>,
         pg_version: u32,
     ) -> anyhow::Result<()> {
+        let (client, conn) = self.page_server_psql_client().await?;
+        // The connection object performs the actual communication with the database,
+        // so spawn it off to run on its own.
+        tokio::spawn(async move {
+            if let Err(e) = conn.await {
+                eprintln!("connection error: {}", e);
+            }
+        });
+        let client = std::pin::pin!(client);
+
         // Init base reader
         let (start_lsn, base_tarfile_path) = base;
         let base_tarfile = tokio::fs::File::open(base_tarfile_path).await?;
-        let base_tarfile =
-            mgmt_api::ReqwestBody::wrap_stream(tokio_util::io::ReaderStream::new(base_tarfile));
+        let base_tarfile = tokio_util::io::ReaderStream::new(base_tarfile);
 
         // Init wal reader if necessary
         let (end_lsn, wal_reader) = if let Some((end_lsn, wal_tarfile_path)) = pg_wal {
             let wal_tarfile = tokio::fs::File::open(wal_tarfile_path).await?;
-            let wal_reader =
-                mgmt_api::ReqwestBody::wrap_stream(tokio_util::io::ReaderStream::new(wal_tarfile));
+            let wal_reader = tokio_util::io::ReaderStream::new(wal_tarfile);
             (end_lsn, Some(wal_reader))
         } else {
             (start_lsn, None)
         };
 
-        // Import base
-        self.http_client
-            .import_basebackup(
-                tenant_id,
-                timeline_id,
-                start_lsn,
-                end_lsn,
-                pg_version,
-                base_tarfile,
-            )
-            .await?;
+        let copy_in = |reader, cmd| {
+            let client = &client;
+            async move {
+                let writer = client.copy_in(&cmd).await?;
+                let writer = std::pin::pin!(writer);
+                let mut writer = writer.sink_map_err(|e| {
+                    std::io::Error::new(std::io::ErrorKind::Other, format!("{e}"))
+                });
+                let mut reader = std::pin::pin!(reader);
+                writer.send_all(&mut reader).await?;
+                writer.into_inner().finish().await?;
+                anyhow::Ok(())
+            }
+        };
 
+        // Import base
+        copy_in(
+            base_tarfile,
+            format!(
+                "import basebackup {tenant_id} {timeline_id} {start_lsn} {end_lsn} {pg_version}"
+            ),
+        )
+        .await?;
         // Import wal if necessary
         if let Some(wal_reader) = wal_reader {
-            self.http_client
-                .import_wal(tenant_id, timeline_id, start_lsn, end_lsn, wal_reader)
-                .await?;
+            copy_in(
+                wal_reader,
+                format!("import wal {tenant_id} {timeline_id} {start_lsn} {end_lsn}"),
+            )
+            .await?;
         }
 
         Ok(())

control_plane/src/safekeeper.rs

@@ -5,10 +5,8 @@
 //! ```text
 //!   .neon/safekeepers/<safekeeper id>
 //! ```
-use std::future::Future;
 use std::io::Write;
 use std::path::PathBuf;
-use std::time::Duration;
 use std::{io, result};
 
 use anyhow::Context;
@@ -16,7 +14,6 @@ use camino::Utf8PathBuf;
 use postgres_connection::PgConnectionConfig;
 use reqwest::{IntoUrl, Method};
 use thiserror::Error;
-use utils::auth::{Claims, Scope};
 use utils::{http::error::HttpErrorBody, id::NodeId};
 
 use crate::{
@@ -35,10 +32,12 @@ pub enum SafekeeperHttpError {
 
 type Result<T> = result::Result<T, SafekeeperHttpError>;
 
-pub(crate) trait ResponseErrorMessageExt: Sized {
-    fn error_from_body(self) -> impl Future<Output = Result<Self>> + Send;
+#[async_trait::async_trait]
+pub trait ResponseErrorMessageExt: Sized {
+    async fn error_from_body(self) -> Result<Self>;
 }
 
+#[async_trait::async_trait]
 impl ResponseErrorMessageExt for reqwest::Response {
     async fn error_from_body(self) -> Result<Self> {
         let status = self.status();
@@ -71,31 +70,24 @@ pub struct SafekeeperNode {
     pub pg_connection_config: PgConnectionConfig,
     pub env: LocalEnv,
     pub http_client: reqwest::Client,
-    pub listen_addr: String,
     pub http_base_url: String,
 }
 
 impl SafekeeperNode {
     pub fn from_env(env: &LocalEnv, conf: &SafekeeperConf) -> SafekeeperNode {
-        let listen_addr = if let Some(ref listen_addr) = conf.listen_addr {
-            listen_addr.clone()
-        } else {
-            "127.0.0.1".to_string()
-        };
         SafekeeperNode {
             id: conf.id,
             conf: conf.clone(),
-            pg_connection_config: Self::safekeeper_connection_config(&listen_addr, conf.pg_port),
+            pg_connection_config: Self::safekeeper_connection_config(conf.pg_port),
             env: env.clone(),
             http_client: reqwest::Client::new(),
-            http_base_url: format!("http://{}:{}/v1", listen_addr, conf.http_port),
-            listen_addr,
+            http_base_url: format!("http://127.0.0.1:{}/v1", conf.http_port),
         }
     }
 
     /// Construct libpq connection string for connecting to this safekeeper.
-    fn safekeeper_connection_config(addr: &str, port: u16) -> PgConnectionConfig {
-        PgConnectionConfig::new_host_port(url::Host::parse(addr).unwrap(), port)
+    fn safekeeper_connection_config(port: u16) -> PgConnectionConfig {
+        PgConnectionConfig::new_host_port(url::Host::parse("127.0.0.1").unwrap(), port)
     }
 
     pub fn datadir_path_by_id(env: &LocalEnv, sk_id: NodeId) -> PathBuf {
@@ -111,21 +103,16 @@ impl SafekeeperNode {
             .expect("non-Unicode path")
     }
 
-    pub async fn start(
-        &self,
-        extra_opts: Vec<String>,
-        retry_timeout: &Duration,
-    ) -> anyhow::Result<()> {
+    pub async fn start(&self, extra_opts: Vec<String>) -> anyhow::Result<()> {
         print!(
-            "Starting safekeeper at '{}' in '{}', retrying for {:?}",
+            "Starting safekeeper at '{}' in '{}'",
             self.pg_connection_config.raw_address(),
-            self.datadir_path().display(),
-            retry_timeout,
+            self.datadir_path().display()
         );
         io::stdout().flush().unwrap();
 
-        let listen_pg = format!("{}:{}", self.listen_addr, self.conf.pg_port);
-        let listen_http = format!("{}:{}", self.listen_addr, self.conf.http_port);
+        let listen_pg = format!("127.0.0.1:{}", self.conf.pg_port);
+        let listen_http = format!("127.0.0.1:{}", self.conf.http_port);
         let id = self.id;
         let datadir = self.datadir_path();
 
@@ -152,7 +139,7 @@ impl SafekeeperNode {
             availability_zone,
         ];
         if let Some(pg_tenant_only_port) = self.conf.pg_tenant_only_port {
-            let listen_pg_tenant_only = format!("{}:{}", self.listen_addr, pg_tenant_only_port);
+            let listen_pg_tenant_only = format!("127.0.0.1:{}", pg_tenant_only_port);
             args.extend(["--listen-pg-tenant-only".to_owned(), listen_pg_tenant_only]);
         }
         if !self.conf.sync {
@@ -203,9 +190,8 @@ impl SafekeeperNode {
             &datadir,
             &self.env.safekeeper_bin(),
             &args,
-            self.safekeeper_env_variables()?,
+            [],
             background_process::InitialPidFile::Expect(self.pid_file()),
-            retry_timeout,
             || async {
                 match self.check_status().await {
                     Ok(()) => Ok(true),
@@ -217,18 +203,6 @@ impl SafekeeperNode {
         .await
     }
 
-    fn safekeeper_env_variables(&self) -> anyhow::Result<Vec<(String, String)>> {
-        // Generate a token to connect from safekeeper to peers
-        if self.conf.auth_enabled {
-            let token = self
-                .env
-                .generate_auth_token(&Claims::new(None, Scope::SafekeeperData))?;
-            Ok(vec![("SAFEKEEPER_AUTH_TOKEN".to_owned(), token)])
-        } else {
-            Ok(Vec::new())
-        }
-    }
-
     ///
     /// Stop the server.
     ///

control_plane/src/storage_controller.rs

@@ -1,26 +1,21 @@
-use crate::{
-    background_process,
-    local_env::{LocalEnv, NeonStorageControllerConf},
-};
+use crate::{background_process, local_env::LocalEnv};
 use camino::{Utf8Path, Utf8PathBuf};
-use hyper::Uri;
-use nix::unistd::Pid;
+use hyper::Method;
 use pageserver_api::{
     controller_api::{
-        NodeConfigureRequest, NodeDescribeResponse, NodeRegisterRequest, TenantCreateRequest,
-        TenantCreateResponse, TenantLocateResponse, TenantShardMigrateRequest,
-        TenantShardMigrateResponse,
+        NodeConfigureRequest, NodeRegisterRequest, TenantCreateResponse, TenantLocateResponse,
+        TenantShardMigrateRequest, TenantShardMigrateResponse,
     },
     models::{
-        TenantShardSplitRequest, TenantShardSplitResponse, TimelineCreateRequest, TimelineInfo,
+        TenantCreateRequest, TenantShardSplitRequest, TenantShardSplitResponse,
+        TimelineCreateRequest, TimelineInfo,
     },
     shard::{ShardStripeSize, TenantShardId},
 };
 use pageserver_client::mgmt_api::ResponseErrorMessageExt;
 use postgres_backend::AuthType;
-use reqwest::Method;
 use serde::{de::DeserializeOwned, Deserialize, Serialize};
-use std::{fs, net::SocketAddr, path::PathBuf, str::FromStr, sync::OnceLock};
+use std::{fs, str::FromStr};
 use tokio::process::Command;
 use tracing::instrument;
 use url::Url;
@@ -31,57 +26,25 @@ use utils::{
 
 pub struct StorageController {
     env: LocalEnv,
+    listen: String,
+    path: Utf8PathBuf,
     private_key: Option<Vec<u8>>,
     public_key: Option<String>,
+    postgres_port: u16,
     client: reqwest::Client,
-    config: NeonStorageControllerConf,
-
-    // The listen addresses is learned when starting the storage controller,
-    // hence the use of OnceLock to init it at the right time.
-    listen: OnceLock<SocketAddr>,
 }
 
 const COMMAND: &str = "storage_controller";
 
 const STORAGE_CONTROLLER_POSTGRES_VERSION: u32 = 16;
 
-const DB_NAME: &str = "storage_controller";
-
-pub struct NeonStorageControllerStartArgs {
-    pub instance_id: u8,
-    pub base_port: Option<u16>,
-    pub start_timeout: humantime::Duration,
-}
-
-impl NeonStorageControllerStartArgs {
-    pub fn with_default_instance_id(start_timeout: humantime::Duration) -> Self {
-        Self {
-            instance_id: 1,
-            base_port: None,
-            start_timeout,
-        }
-    }
-}
-
-pub struct NeonStorageControllerStopArgs {
-    pub instance_id: u8,
-    pub immediate: bool,
-}
-
-impl NeonStorageControllerStopArgs {
-    pub fn with_default_instance_id(immediate: bool) -> Self {
-        Self {
-            instance_id: 1,
-            immediate,
-        }
-    }
-}
+// Use a shorter pageserver unavailability interval than the default to speed up tests.
+const NEON_LOCAL_MAX_UNAVAILABLE_INTERVAL: std::time::Duration = std::time::Duration::from_secs(10);
 
 #[derive(Serialize, Deserialize)]
 pub struct AttachHookRequest {
     pub tenant_shard_id: TenantShardId,
     pub node_id: Option<NodeId>,
-    pub generation_override: Option<i32>,
 }
 
 #[derive(Serialize, Deserialize)]
@@ -101,6 +64,27 @@ pub struct InspectResponse {
 
 impl StorageController {
     pub fn from_env(env: &LocalEnv) -> Self {
+        let path = Utf8PathBuf::from_path_buf(env.base_data_dir.clone())
+            .unwrap()
+            .join("attachments.json");
+
+        // Makes no sense to construct this if pageservers aren't going to use it: assume
+        // pageservers have control plane API set
+        let listen_url = env.control_plane_api.clone().unwrap();
+
+        let listen = format!(
+            "{}:{}",
+            listen_url.host_str().unwrap(),
+            listen_url.port().unwrap()
+        );
+
+        // Convention: NeonEnv in python tests reserves the next port after the control_plane_api
+        // port, for use by our captive postgres.
+        let postgres_port = listen_url
+            .port()
+            .expect("Control plane API setting should always have a port")
+            + 1;
+
         // Assume all pageservers have symmetric auth configuration: this service
         // expects to use one JWT token to talk to all of them.
         let ps_conf = env
@@ -143,28 +127,20 @@ impl StorageController {
 
         Self {
             env: env.clone(),
+            path,
+            listen,
             private_key,
             public_key,
+            postgres_port,
             client: reqwest::ClientBuilder::new()
                 .build()
                 .expect("Failed to construct http client"),
-            config: env.storage_controller.clone(),
-            listen: OnceLock::default(),
         }
     }
 
-    fn storage_controller_instance_dir(&self, instance_id: u8) -> PathBuf {
-        self.env
-            .base_data_dir
-            .join(format!("storage_controller_{}", instance_id))
-    }
-
-    fn pid_file(&self, instance_id: u8) -> Utf8PathBuf {
-        Utf8PathBuf::from_path_buf(
-            self.storage_controller_instance_dir(instance_id)
-                .join("storage_controller.pid"),
-        )
-        .expect("non-Unicode path")
+    fn pid_file(&self) -> Utf8PathBuf {
+        Utf8PathBuf::from_path_buf(self.env.base_data_dir.join("storage_controller.pid"))
+            .expect("non-Unicode path")
     }
 
     /// PIDFile for the postgres instance used to store storage controller state
@@ -177,16 +153,16 @@ impl StorageController {
         .expect("non-Unicode path")
     }
 
-    /// Find the directory containing postgres subdirectories, such `bin` and `lib`
+    /// Find the directory containing postgres binaries, such as `initdb` and `pg_ctl`
     ///
     /// This usually uses STORAGE_CONTROLLER_POSTGRES_VERSION of postgres, but will fall back
     /// to other versions if that one isn't found.  Some automated tests create circumstances
     /// where only one version is available in pg_distrib_dir, such as `test_remote_extensions`.
-    async fn get_pg_dir(&self, dir_name: &str) -> anyhow::Result<Utf8PathBuf> {
+    pub async fn get_pg_bin_dir(&self) -> anyhow::Result<Utf8PathBuf> {
         let prefer_versions = [STORAGE_CONTROLLER_POSTGRES_VERSION, 15, 14];
 
         for v in prefer_versions {
-            let path = Utf8PathBuf::from_path_buf(self.env.pg_dir(v, dir_name)?).unwrap();
+            let path = Utf8PathBuf::from_path_buf(self.env.pg_bin_dir(v)?).unwrap();
             if tokio::fs::try_exists(&path).await? {
                 return Ok(path);
             }
@@ -194,38 +170,30 @@ impl StorageController {
 
         // Fall through
         anyhow::bail!(
-            "Postgres directory '{}' not found in {}",
-            dir_name,
-            self.env.pg_distrib_dir.display(),
+            "Postgres binaries not found in {}",
+            self.env.pg_distrib_dir.display()
         );
     }
 
-    pub async fn get_pg_bin_dir(&self) -> anyhow::Result<Utf8PathBuf> {
-        self.get_pg_dir("bin").await
-    }
-
-    pub async fn get_pg_lib_dir(&self) -> anyhow::Result<Utf8PathBuf> {
-        self.get_pg_dir("lib").await
-    }
-
     /// Readiness check for our postgres process
-    async fn pg_isready(&self, pg_bin_dir: &Utf8Path, postgres_port: u16) -> anyhow::Result<bool> {
+    async fn pg_isready(&self, pg_bin_dir: &Utf8Path) -> anyhow::Result<bool> {
         let bin_path = pg_bin_dir.join("pg_isready");
-        let args = ["-h", "localhost", "-p", &format!("{}", postgres_port)];
+        let args = ["-h", "localhost", "-p", &format!("{}", self.postgres_port)];
         let exitcode = Command::new(bin_path).args(args).spawn()?.wait().await?;
 
         Ok(exitcode.success())
     }
 
-    /// Create our database if it doesn't exist
+    /// Create our database if it doesn't exist, and run migrations.
     ///
     /// This function is equivalent to the `diesel setup` command in the diesel CLI.  We implement
     /// the same steps by hand to avoid imposing a dependency on installing diesel-cli for developers
     /// who just want to run `cargo neon_local` without knowing about diesel.
     ///
     /// Returns the database url
-    pub async fn setup_database(&self, postgres_port: u16) -> anyhow::Result<String> {
-        let database_url = format!("postgresql://localhost:{}/{DB_NAME}", postgres_port);
+    pub async fn setup_database(&self) -> anyhow::Result<String> {
+        const DB_NAME: &str = "storage_controller";
+        let database_url = format!("postgresql://localhost:{}/{DB_NAME}", self.postgres_port);
 
         let pg_bin_dir = self.get_pg_bin_dir().await?;
         let createdb_path = pg_bin_dir.join("createdb");
@@ -234,7 +202,7 @@ impl StorageController {
                 "-h",
                 "localhost",
                 "-p",
-                &format!("{}", postgres_port),
+                &format!("{}", self.postgres_port),
                 DB_NAME,
             ])
             .output()
@@ -253,211 +221,78 @@ impl StorageController {
         Ok(database_url)
     }
 
-    pub async fn connect_to_database(
-        &self,
-        postgres_port: u16,
-    ) -> anyhow::Result<(
-        tokio_postgres::Client,
-        tokio_postgres::Connection<tokio_postgres::Socket, tokio_postgres::tls::NoTlsStream>,
-    )> {
-        tokio_postgres::Config::new()
-            .host("localhost")
-            .port(postgres_port)
-            // The user is the ambient operating system user name.
-            // That is an impurity which we want to fix in => TODO https://github.com/neondatabase/neon/issues/8400
-            //
-            // Until we get there, use the ambient operating system user name.
-            // Recent tokio-postgres versions default to this if the user isn't specified.
-            // But tokio-postgres fork doesn't have this upstream commit:
-            // https://github.com/sfackler/rust-postgres/commit/cb609be758f3fb5af537f04b584a2ee0cebd5e79
-            // => we should rebase our fork => TODO https://github.com/neondatabase/neon/issues/8399
-            .user(&whoami::username())
-            .dbname(DB_NAME)
-            .connect(tokio_postgres::NoTls)
-            .await
-            .map_err(anyhow::Error::new)
-    }
+    pub async fn start(&self) -> anyhow::Result<()> {
+        // Start a vanilla Postgres process used by the storage controller for persistence.
+        let pg_data_path = Utf8PathBuf::from_path_buf(self.env.base_data_dir.clone())
+            .unwrap()
+            .join("storage_controller_db");
+        let pg_bin_dir = self.get_pg_bin_dir().await?;
+        let pg_log_path = pg_data_path.join("postgres.log");
 
-    pub async fn start(&self, start_args: NeonStorageControllerStartArgs) -> anyhow::Result<()> {
-        let instance_dir = self.storage_controller_instance_dir(start_args.instance_id);
-        if let Err(err) = tokio::fs::create_dir(&instance_dir).await {
-            if err.kind() != std::io::ErrorKind::AlreadyExists {
-                panic!("Failed to create instance dir {instance_dir:?}");
+        if !tokio::fs::try_exists(&pg_data_path).await? {
+            // Initialize empty database
+            let initdb_path = pg_bin_dir.join("initdb");
+            let mut child = Command::new(&initdb_path)
+                .args(["-D", pg_data_path.as_ref()])
+                .spawn()
+                .expect("Failed to spawn initdb");
+            let status = child.wait().await?;
+            if !status.success() {
+                anyhow::bail!("initdb failed with status {status}");
             }
-        }
 
-        let (listen, postgres_port) = {
-            if let Some(base_port) = start_args.base_port {
-                (
-                    format!("127.0.0.1:{base_port}"),
-                    self.config
-                        .database_url
-                        .expect("--base-port requires NeonStorageControllerConf::database_url")
-                        .port(),
-                )
-            } else {
-                let listen_url = self.env.control_plane_api.clone().unwrap();
-
-                let listen = format!(
-                    "{}:{}",
-                    listen_url.host_str().unwrap(),
-                    listen_url.port().unwrap()
-                );
-
-                (listen, listen_url.port().unwrap() + 1)
-            }
-        };
-
-        let socket_addr = listen
-            .parse()
-            .expect("listen address is a valid socket address");
-        self.listen
-            .set(socket_addr)
-            .expect("StorageController::listen is only set here");
-
-        // Do we remove the pid file on stop?
-        let pg_started = self.is_postgres_running().await?;
-        let pg_lib_dir = self.get_pg_lib_dir().await?;
-
-        if !pg_started {
-            // Start a vanilla Postgres process used by the storage controller for persistence.
-            let pg_data_path = Utf8PathBuf::from_path_buf(self.env.base_data_dir.clone())
-                .unwrap()
-                .join("storage_controller_db");
-            let pg_bin_dir = self.get_pg_bin_dir().await?;
-            let pg_log_path = pg_data_path.join("postgres.log");
-
-            if !tokio::fs::try_exists(&pg_data_path).await? {
-                // Initialize empty database
-                let initdb_path = pg_bin_dir.join("initdb");
-                let mut child = Command::new(&initdb_path)
-                    .envs(vec![
-                        ("LD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
-                        ("DYLD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
-                    ])
-                    .args(["-D", pg_data_path.as_ref()])
-                    .spawn()
-                    .expect("Failed to spawn initdb");
-                let status = child.wait().await?;
-                if !status.success() {
-                    anyhow::bail!("initdb failed with status {status}");
-                }
-            };
-
-            // Write a minimal config file:
-            // - Specify the port, since this is chosen dynamically
-            // - Switch off fsync, since we're running on lightweight test environments and when e.g. scale testing
-            //   the storage controller we don't want a slow local disk to interfere with that.
-            //
-            // NB: it's important that we rewrite this file on each start command so we propagate changes
-            // from `LocalEnv`'s config file (`.neon/config`).
             tokio::fs::write(
                 &pg_data_path.join("postgresql.conf"),
-                format!("port = {}\nfsync=off\n", postgres_port),
+                format!("port = {}", self.postgres_port),
             )
             .await?;
-
-            println!("Starting storage controller database...");
-            let db_start_args = [
-                "-w",
-                "-D",
-                pg_data_path.as_ref(),
-                "-l",
-                pg_log_path.as_ref(),
-                "start",
-            ];
-
-            background_process::start_process(
-                "storage_controller_db",
-                &self.env.base_data_dir,
-                pg_bin_dir.join("pg_ctl").as_std_path(),
-                db_start_args,
-                vec![
-                    ("LD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
-                    ("DYLD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
-                ],
-                background_process::InitialPidFile::Create(self.postgres_pid_file()),
-                &start_args.start_timeout,
-                || self.pg_isready(&pg_bin_dir, postgres_port),
-            )
-            .await?;
-
-            self.setup_database(postgres_port).await?;
-        }
-
-        let database_url = format!("postgresql://localhost:{}/{DB_NAME}", postgres_port);
-
-        // We support running a startup SQL script to fiddle with the database before we launch storcon.
-        // This is used by the test suite.
-        let startup_script_path = self
-            .env
-            .base_data_dir
-            .join("storage_controller_db.startup.sql");
-        let startup_script = match tokio::fs::read_to_string(&startup_script_path).await {
-            Ok(script) => {
-                tokio::fs::remove_file(startup_script_path).await?;
-                script
-            }
-            Err(e) => {
-                if e.kind() == std::io::ErrorKind::NotFound {
-                    // always run some startup script so that this code path doesn't bit rot
-                    "BEGIN; COMMIT;".to_string()
-                } else {
-                    anyhow::bail!("Failed to read startup script: {e}")
-                }
-            }
         };
-        let (mut client, conn) = self.connect_to_database(postgres_port).await?;
-        let conn = tokio::spawn(conn);
-        let tx = client.build_transaction();
-        let tx = tx.start().await?;
-        tx.batch_execute(&startup_script).await?;
-        tx.commit().await?;
-        drop(client);
-        conn.await??;
 
-        let listen = self
-            .listen
-            .get()
-            .expect("cell is set earlier in this function");
-        let address_for_peers = Uri::builder()
-            .scheme("http")
-            .authority(format!("{}:{}", listen.ip(), listen.port()))
-            .path_and_query("")
-            .build()
-            .unwrap();
+        println!("Starting storage controller database...");
+        let db_start_args = [
+            "-w",
+            "-D",
+            pg_data_path.as_ref(),
+            "-l",
+            pg_log_path.as_ref(),
+            "start",
+        ];
+
+        background_process::start_process(
+            "storage_controller_db",
+            &self.env.base_data_dir,
+            pg_bin_dir.join("pg_ctl").as_std_path(),
+            db_start_args,
+            [],
+            background_process::InitialPidFile::Create(self.postgres_pid_file()),
+            || self.pg_isready(&pg_bin_dir),
+        )
+        .await?;
+
+        // Run migrations on every startup, in case something changed.
+        let database_url = self.setup_database().await?;
+
+        let max_unavailable: humantime::Duration = NEON_LOCAL_MAX_UNAVAILABLE_INTERVAL.into();
 
         let mut args = vec![
             "-l",
-            &listen.to_string(),
+            &self.listen,
+            "-p",
+            self.path.as_ref(),
             "--dev",
             "--database-url",
             &database_url,
-            "--max-offline-interval",
-            &humantime::Duration::from(self.config.max_offline).to_string(),
-            "--max-warming-up-interval",
-            &humantime::Duration::from(self.config.max_warming_up).to_string(),
-            "--address-for-peers",
-            &address_for_peers.to_string(),
+            "--max-unavailable-interval",
+            &max_unavailable.to_string(),
         ]
         .into_iter()
         .map(|s| s.to_string())
         .collect::<Vec<_>>();
-
-        if self.config.start_as_candidate {
-            args.push("--start-as-candidate".to_string());
-        }
-
         if let Some(private_key) = &self.private_key {
             let claims = Claims::new(None, Scope::PageServerApi);
             let jwt_token =
                 encode_from_key_file(&claims, private_key).expect("failed to generate jwt token");
             args.push(format!("--jwt-token={jwt_token}"));
-
-            let peer_claims = Claims::new(None, Scope::Admin);
-            let peer_jwt_token = encode_from_key_file(&peer_claims, private_key)
-                .expect("failed to generate jwt token");
-            args.push(format!("--peer-jwt-token={peer_jwt_token}"));
         }
 
         if let Some(public_key) = &self.public_key {
@@ -470,30 +305,16 @@ impl StorageController {
             ));
         }
 
-        if let Some(split_threshold) = self.config.split_threshold.as_ref() {
-            args.push(format!("--split-threshold={split_threshold}"))
-        }
-
-        if let Some(lag) = self.config.max_secondary_lag_bytes.as_ref() {
-            args.push(format!("--max-secondary-lag-bytes={lag}"))
-        }
-
-        args.push(format!(
-            "--neon-local-repo-dir={}",
-            self.env.base_data_dir.display()
-        ));
-
         background_process::start_process(
             COMMAND,
-            &instance_dir,
+            &self.env.base_data_dir,
             &self.env.storage_controller_bin(),
             args,
-            vec![
-                ("LD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
-                ("DYLD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
-            ],
-            background_process::InitialPidFile::Create(self.pid_file(start_args.instance_id)),
-            &start_args.start_timeout,
+            [(
+                "NEON_REPO_DIR".to_string(),
+                self.env.base_data_dir.to_string_lossy().to_string(),
+            )],
+            background_process::InitialPidFile::Create(self.pid_file()),
             || async {
                 match self.ready().await {
                     Ok(_) => Ok(true),
@@ -506,35 +327,8 @@ impl StorageController {
         Ok(())
     }
 
-    pub async fn stop(&self, stop_args: NeonStorageControllerStopArgs) -> anyhow::Result<()> {
-        background_process::stop_process(
-            stop_args.immediate,
-            COMMAND,
-            &self.pid_file(stop_args.instance_id),
-        )?;
-
-        let storcon_instances = self.env.storage_controller_instances().await?;
-        for (instance_id, instanced_dir_path) in storcon_instances {
-            if instance_id == stop_args.instance_id {
-                continue;
-            }
-
-            let pid_file = instanced_dir_path.join("storage_controller.pid");
-            let pid = tokio::fs::read_to_string(&pid_file)
-                .await
-                .map_err(|err| {
-                    anyhow::anyhow!("Failed to read storcon pid file at {pid_file:?}: {err}")
-                })?
-                .parse::<i32>()
-                .expect("pid is valid i32");
-
-            let other_proc_alive = !background_process::process_has_stopped(Pid::from_raw(pid))?;
-            if other_proc_alive {
-                // There is another storage controller instance running, so we return
-                // and leave the database running.
-                return Ok(());
-            }
-        }
+    pub async fn stop(&self, immediate: bool) -> anyhow::Result<()> {
+        background_process::stop_process(immediate, COMMAND, &self.pid_file())?;
 
         let pg_data_path = self.env.base_data_dir.join("storage_controller_db");
         let pg_bin_dir = self.get_pg_bin_dir().await?;
@@ -547,51 +341,27 @@ impl StorageController {
             .wait()
             .await?;
         if !stop_status.success() {
-            match self.is_postgres_running().await {
-                Ok(false) => {
-                    println!("Storage controller database is already stopped");
-                    return Ok(());
-                }
-                Ok(true) => {
-                    anyhow::bail!("Failed to stop storage controller database");
-                }
-                Err(err) => {
-                    anyhow::bail!("Failed to stop storage controller database: {err}");
-                }
+            let pg_status_args = ["-D", &pg_data_path.to_string_lossy(), "status"];
+            let status_exitcode = Command::new(pg_bin_dir.join("pg_ctl"))
+                .args(pg_status_args)
+                .spawn()?
+                .wait()
+                .await?;
+
+            // pg_ctl status returns this exit code if postgres is not running: in this case it is
+            // fine that stop failed.  Otherwise it is an error that stop failed.
+            const PG_STATUS_NOT_RUNNING: i32 = 3;
+            if Some(PG_STATUS_NOT_RUNNING) == status_exitcode.code() {
+                println!("Storage controller database is already stopped");
+                return Ok(());
+            } else {
+                anyhow::bail!("Failed to stop storage controller database: {stop_status}")
             }
         }
 
         Ok(())
     }
 
-    async fn is_postgres_running(&self) -> anyhow::Result<bool> {
-        let pg_data_path = self.env.base_data_dir.join("storage_controller_db");
-        let pg_bin_dir = self.get_pg_bin_dir().await?;
-
-        let pg_status_args = ["-D", &pg_data_path.to_string_lossy(), "status"];
-        let status_exitcode = Command::new(pg_bin_dir.join("pg_ctl"))
-            .args(pg_status_args)
-            .spawn()?
-            .wait()
-            .await?;
-
-        // pg_ctl status returns this exit code if postgres is not running: in this case it is
-        // fine that stop failed.  Otherwise it is an error that stop failed.
-        const PG_STATUS_NOT_RUNNING: i32 = 3;
-        const PG_NO_DATA_DIR: i32 = 4;
-        const PG_STATUS_RUNNING: i32 = 0;
-        match status_exitcode.code() {
-            Some(PG_STATUS_NOT_RUNNING) => Ok(false),
-            Some(PG_NO_DATA_DIR) => Ok(false),
-            Some(PG_STATUS_RUNNING) => Ok(true),
-            Some(code) => Err(anyhow::anyhow!(
-                "pg_ctl status returned unexpected status code: {:?}",
-                code
-            )),
-            None => Err(anyhow::anyhow!("pg_ctl status returned no status code")),
-        }
-    }
-
     fn get_claims_for_path(path: &str) -> anyhow::Result<Option<Claims>> {
         let category = match path.find('/') {
             Some(idx) => &path[..idx],
@@ -609,7 +379,7 @@ impl StorageController {
     /// Simple HTTP request wrapper for calling into storage controller
     async fn dispatch<RQ, RS>(
         &self,
-        method: reqwest::Method,
+        method: hyper::Method,
         path: String,
         body: Option<RQ>,
     ) -> anyhow::Result<RS>
@@ -617,31 +387,15 @@ impl StorageController {
         RQ: Serialize + Sized,
         RS: DeserializeOwned + Sized,
     {
-        // In the special case of the `storage_controller start` subcommand, we wish
-        // to use the API endpoint of the newly started storage controller in order
-        // to pass the readiness check. In this scenario [`Self::listen`] will be set
-        // (see [`Self::start`]).
-        //
-        // Otherwise, we infer the storage controller api endpoint from the configured
-        // control plane API.
-        let url = if let Some(socket_addr) = self.listen.get() {
-            Url::from_str(&format!(
-                "http://{}:{}/{path}",
-                socket_addr.ip().to_canonical(),
-                socket_addr.port()
-            ))
-            .unwrap()
-        } else {
-            // The configured URL has the /upcall path prefix for pageservers to use: we will strip that out
-            // for general purpose API access.
-            let listen_url = self.env.control_plane_api.clone().unwrap();
-            Url::from_str(&format!(
-                "http://{}:{}/{path}",
-                listen_url.host_str().unwrap(),
-                listen_url.port().unwrap()
-            ))
-            .unwrap()
-        };
+        // The configured URL has the /upcall path prefix for pageservers to use: we will strip that out
+        // for general purpose API access.
+        let listen_url = self.env.control_plane_api.clone().unwrap();
+        let url = Url::from_str(&format!(
+            "http://{}:{}/{path}",
+            listen_url.host_str().unwrap(),
+            listen_url.port().unwrap()
+        ))
+        .unwrap();
 
         let mut builder = self.client.request(method, url);
         if let Some(body) = body {
@@ -678,7 +432,6 @@ impl StorageController {
         let request = AttachHookRequest {
             tenant_shard_id,
             node_id: Some(pageserver_id),
-            generation_override: None,
         };
 
         let response = self
@@ -719,16 +472,6 @@ impl StorageController {
             .await
     }
 
-    #[instrument(skip(self))]
-    pub async fn tenant_import(&self, tenant_id: TenantId) -> anyhow::Result<TenantCreateResponse> {
-        self.dispatch::<(), TenantCreateResponse>(
-            Method::POST,
-            format!("debug/v1/tenant/{tenant_id}/import"),
-            None,
-        )
-        .await
-    }
-
     #[instrument(skip(self))]
     pub async fn tenant_locate(&self, tenant_id: TenantId) -> anyhow::Result<TenantLocateResponse> {
         self.dispatch::<(), _>(
@@ -790,15 +533,6 @@ impl StorageController {
         .await
     }
 
-    pub async fn node_list(&self) -> anyhow::Result<Vec<NodeDescribeResponse>> {
-        self.dispatch::<(), Vec<NodeDescribeResponse>>(
-            Method::GET,
-            "control/v1/node".to_string(),
-            None,
-        )
-        .await
-    }
-
     #[instrument(skip(self))]
     pub async fn ready(&self) -> anyhow::Result<()> {
         self.dispatch::<(), ()>(Method::GET, "ready".to_string(), None)

control_plane/storcon_cli/Cargo.toml

@@ -9,15 +9,12 @@ license.workspace = true
 anyhow.workspace = true
 clap.workspace = true
 comfy-table.workspace = true
-futures.workspace = true
-humantime.workspace = true
 hyper.workspace = true
 pageserver_api.workspace = true
 pageserver_client.workspace = true
 reqwest.workspace = true
 serde.workspace = true
 serde_json = { workspace = true, features = ["raw_value"] }
-storage_controller_client.workspace = true
 thiserror.workspace = true
 tokio.workspace = true
 tracing.workspace = true

control_plane/storcon_cli/src/main.rs

@@ -1,28 +1,27 @@
-use futures::StreamExt;
-use std::{str::FromStr, time::Duration};
+use std::{collections::HashMap, str::FromStr};
 
 use clap::{Parser, Subcommand};
+use hyper::Method;
 use pageserver_api::{
     controller_api::{
-        NodeAvailabilityWrapper, NodeDescribeResponse, ShardSchedulingPolicy, TenantCreateRequest,
+        NodeAvailabilityWrapper, NodeDescribeResponse, ShardSchedulingPolicy,
         TenantDescribeResponse, TenantPolicyRequest,
     },
     models::{
-        EvictionPolicy, EvictionPolicyLayerAccessThreshold, LocationConfigSecondary,
-        ShardParameters, TenantConfig, TenantConfigRequest, TenantShardSplitRequest,
-        TenantShardSplitResponse,
+        ShardParameters, TenantConfig, TenantConfigRequest, TenantCreateRequest,
+        TenantShardSplitRequest, TenantShardSplitResponse,
     },
     shard::{ShardStripeSize, TenantShardId},
 };
-use pageserver_client::mgmt_api::{self};
-use reqwest::{Method, StatusCode, Url};
+use pageserver_client::mgmt_api::{self, ResponseErrorMessageExt};
+use reqwest::Url;
+use serde::{de::DeserializeOwned, Serialize};
 use utils::id::{NodeId, TenantId};
 
 use pageserver_api::controller_api::{
     NodeConfigureRequest, NodeRegisterRequest, NodeSchedulingPolicy, PlacementPolicy,
-    TenantShardMigrateRequest, TenantShardMigrateResponse,
+    TenantLocateResponse, TenantShardMigrateRequest, TenantShardMigrateResponse,
 };
-use storage_controller_client::control_api::Client;
 
 #[derive(Subcommand, Debug)]
 enum Command {
@@ -41,8 +40,6 @@ enum Command {
         listen_http_addr: String,
         #[arg(long)]
         listen_http_port: u16,
-        #[arg(long)]
-        availability_zone_id: String,
     },
 
     /// Modify a node's configuration in the storage controller
@@ -58,10 +55,6 @@ enum Command {
         #[arg(long)]
         scheduling: Option<NodeSchedulingPolicy>,
     },
-    NodeDelete {
-        #[arg(long)]
-        node_id: NodeId,
-    },
     /// Modify a tenant's policies in the storage controller
     TenantPolicy {
         #[arg(long)]
@@ -116,83 +109,17 @@ enum Command {
         #[arg(long)]
         config: String,
     },
+    /// Attempt to balance the locations for a tenant across pageservers.  This is a client-side
+    /// alternative to the storage controller's scheduling optimization behavior.
+    TenantScatter {
+        #[arg(long)]
+        tenant_id: TenantId,
+    },
     /// Print details about a particular tenant, including all its shards' states.
     TenantDescribe {
         #[arg(long)]
         tenant_id: TenantId,
     },
-    /// For a tenant which hasn't been onboarded to the storage controller yet, add it in secondary
-    /// mode so that it can warm up content on a pageserver.
-    TenantWarmup {
-        #[arg(long)]
-        tenant_id: TenantId,
-    },
-    /// Uncleanly drop a tenant from the storage controller: this doesn't delete anything from pageservers. Appropriate
-    /// if you e.g. used `tenant-warmup` by mistake on a tenant ID that doesn't really exist, or is in some other region.
-    TenantDrop {
-        #[arg(long)]
-        tenant_id: TenantId,
-        #[arg(long)]
-        unclean: bool,
-    },
-    NodeDrop {
-        #[arg(long)]
-        node_id: NodeId,
-        #[arg(long)]
-        unclean: bool,
-    },
-    TenantSetTimeBasedEviction {
-        #[arg(long)]
-        tenant_id: TenantId,
-        #[arg(long)]
-        period: humantime::Duration,
-        #[arg(long)]
-        threshold: humantime::Duration,
-    },
-    // Migrate away from a set of specified pageservers by moving the primary attachments to pageservers
-    // outside of the specified set.
-    BulkMigrate {
-        // Set of pageserver node ids to drain.
-        #[arg(long)]
-        nodes: Vec<NodeId>,
-        // Optional: migration concurrency (default is 8)
-        #[arg(long)]
-        concurrency: Option<usize>,
-        // Optional: maximum number of shards to migrate
-        #[arg(long)]
-        max_shards: Option<usize>,
-        // Optional: when set to true, nothing is migrated, but the plan is printed to stdout
-        #[arg(long)]
-        dry_run: Option<bool>,
-    },
-    /// Start draining the specified pageserver.
-    /// The drain is complete when the schedulling policy returns to active.
-    StartDrain {
-        #[arg(long)]
-        node_id: NodeId,
-    },
-    /// Cancel draining the specified pageserver and wait for `timeout`
-    /// for the operation to be canceled. May be retried.
-    CancelDrain {
-        #[arg(long)]
-        node_id: NodeId,
-        #[arg(long)]
-        timeout: humantime::Duration,
-    },
-    /// Start filling the specified pageserver.
-    /// The drain is complete when the schedulling policy returns to active.
-    StartFill {
-        #[arg(long)]
-        node_id: NodeId,
-    },
-    /// Cancel filling the specified pageserver and wait for `timeout`
-    /// for the operation to be canceled. May be retried.
-    CancelFill {
-        #[arg(long)]
-        node_id: NodeId,
-        #[arg(long)]
-        timeout: humantime::Duration,
-    },
 }
 
 #[derive(Parser)]
@@ -279,32 +206,62 @@ impl FromStr for NodeAvailabilityArg {
     }
 }
 
-async fn wait_for_scheduling_policy<F>(
-    client: Client,
-    node_id: NodeId,
-    timeout: Duration,
-    f: F,
-) -> anyhow::Result<NodeSchedulingPolicy>
-where
-    F: Fn(NodeSchedulingPolicy) -> bool,
-{
-    let waiter = tokio::time::timeout(timeout, async move {
-        loop {
-            let node = client
-                .dispatch::<(), NodeDescribeResponse>(
-                    Method::GET,
-                    format!("control/v1/node/{node_id}"),
-                    None,
-                )
-                .await?;
+struct Client {
+    base_url: Url,
+    jwt_token: Option<String>,
+    client: reqwest::Client,
+}
 
-            if f(node.scheduling) {
-                return Ok::<NodeSchedulingPolicy, mgmt_api::Error>(node.scheduling);
-            }
+impl Client {
+    fn new(base_url: Url, jwt_token: Option<String>) -> Self {
+        Self {
+            base_url,
+            jwt_token,
+            client: reqwest::ClientBuilder::new()
+                .build()
+                .expect("Failed to construct http client"),
         }
-    });
+    }
 
-    Ok(waiter.await??)
+    /// Simple HTTP request wrapper for calling into storage controller
+    async fn dispatch<RQ, RS>(
+        &self,
+        method: hyper::Method,
+        path: String,
+        body: Option<RQ>,
+    ) -> mgmt_api::Result<RS>
+    where
+        RQ: Serialize + Sized,
+        RS: DeserializeOwned + Sized,
+    {
+        // The configured URL has the /upcall path prefix for pageservers to use: we will strip that out
+        // for general purpose API access.
+        let url = Url::from_str(&format!(
+            "http://{}:{}/{path}",
+            self.base_url.host_str().unwrap(),
+            self.base_url.port().unwrap()
+        ))
+        .unwrap();
+
+        let mut builder = self.client.request(method, url);
+        if let Some(body) = body {
+            builder = builder.json(&body)
+        }
+        if let Some(jwt_token) = &self.jwt_token {
+            builder = builder.header(
+                reqwest::header::AUTHORIZATION,
+                format!("Bearer {jwt_token}"),
+            );
+        }
+
+        let response = builder.send().await.map_err(mgmt_api::Error::ReceiveBody)?;
+        let response = response.error_from_body().await?;
+
+        response
+            .json()
+            .await
+            .map_err(pageserver_client::mgmt_api::Error::ReceiveBody)
+    }
 }
 
 #[tokio::main]
@@ -324,7 +281,6 @@ async fn main() -> anyhow::Result<()> {
             listen_pg_port,
             listen_http_addr,
             listen_http_port,
-            availability_zone_id,
         } => {
             storcon_client
                 .dispatch::<_, ()>(
@@ -336,24 +292,19 @@ async fn main() -> anyhow::Result<()> {
                         listen_pg_port,
                         listen_http_addr,
                         listen_http_port,
-                        availability_zone_id: Some(availability_zone_id),
                     }),
                 )
                 .await?;
         }
         Command::TenantCreate { tenant_id } => {
-            storcon_client
-                .dispatch::<_, ()>(
-                    Method::POST,
-                    "v1/tenant".to_string(),
-                    Some(TenantCreateRequest {
-                        new_tenant_id: TenantShardId::unsharded(tenant_id),
-                        generation: None,
-                        shard_parameters: ShardParameters::default(),
-                        placement_policy: Some(PlacementPolicy::Attached(1)),
-                        config: TenantConfig::default(),
-                    }),
-                )
+            vps_client
+                .tenant_create(&TenantCreateRequest {
+                    new_tenant_id: TenantShardId::unsharded(tenant_id),
+                    generation: None,
+                    shard_parameters: ShardParameters::default(),
+                    placement_policy: Some(PlacementPolicy::Attached(1)),
+                    config: TenantConfig::default(),
+                })
                 .await?;
         }
         Command::TenantDelete { tenant_id } => {
@@ -363,16 +314,13 @@ async fn main() -> anyhow::Result<()> {
             tracing::info!("Delete status: {}", status);
         }
         Command::Nodes {} => {
-            let mut resp = storcon_client
+            let resp = storcon_client
                 .dispatch::<(), Vec<NodeDescribeResponse>>(
                     Method::GET,
                     "control/v1/node".to_string(),
                     None,
                 )
                 .await?;
-
-            resp.sort_by(|a, b| a.listen_http_addr.cmp(&b.listen_http_addr));
-
             let mut table = comfy_table::Table::new();
             table.set_header(["Id", "Hostname", "Scheduling", "Availability"]);
             for node in resp {
@@ -404,16 +352,13 @@ async fn main() -> anyhow::Result<()> {
                 .await?;
         }
         Command::Tenants {} => {
-            let mut resp = storcon_client
+            let resp = storcon_client
                 .dispatch::<(), Vec<TenantDescribeResponse>>(
                     Method::GET,
                     "control/v1/tenant".to_string(),
                     None,
                 )
                 .await?;
-
-            resp.sort_by(|a, b| a.tenant_id.cmp(&b.tenant_id));
-
             let mut table = comfy_table::Table::new();
             table.set_header([
                 "TenantId",
@@ -508,6 +453,88 @@ async fn main() -> anyhow::Result<()> {
                 })
                 .await?;
         }
+        Command::TenantScatter { tenant_id } => {
+            // Find the shards
+            let locate_response = storcon_client
+                .dispatch::<(), TenantLocateResponse>(
+                    Method::GET,
+                    format!("control/v1/tenant/{tenant_id}/locate"),
+                    None,
+                )
+                .await?;
+            let shards = locate_response.shards;
+
+            let mut node_to_shards: HashMap<NodeId, Vec<TenantShardId>> = HashMap::new();
+            let shard_count = shards.len();
+            for s in shards {
+                let entry = node_to_shards.entry(s.node_id).or_default();
+                entry.push(s.shard_id);
+            }
+
+            // Load list of available nodes
+            let nodes_resp = storcon_client
+                .dispatch::<(), Vec<NodeDescribeResponse>>(
+                    Method::GET,
+                    "control/v1/node".to_string(),
+                    None,
+                )
+                .await?;
+
+            for node in nodes_resp {
+                if matches!(node.availability, NodeAvailabilityWrapper::Active) {
+                    node_to_shards.entry(node.id).or_default();
+                }
+            }
+
+            let max_shard_per_node = shard_count / node_to_shards.len();
+
+            loop {
+                let mut migrate_shard = None;
+                for shards in node_to_shards.values_mut() {
+                    if shards.len() > max_shard_per_node {
+                        // Pick the emptiest
+                        migrate_shard = Some(shards.pop().unwrap());
+                    }
+                }
+                let Some(migrate_shard) = migrate_shard else {
+                    break;
+                };
+
+                // Pick the emptiest node to migrate to
+                let mut destinations = node_to_shards
+                    .iter()
+                    .map(|(k, v)| (k, v.len()))
+                    .collect::<Vec<_>>();
+                destinations.sort_by_key(|i| i.1);
+                let (destination_node, destination_count) = *destinations.first().unwrap();
+                if destination_count + 1 > max_shard_per_node {
+                    // Even the emptiest destination doesn't have space: we're done
+                    break;
+                }
+                let destination_node = *destination_node;
+
+                node_to_shards
+                    .get_mut(&destination_node)
+                    .unwrap()
+                    .push(migrate_shard);
+
+                println!("Migrate {} -> {} ...", migrate_shard, destination_node);
+
+                storcon_client
+                    .dispatch::<TenantShardMigrateRequest, TenantShardMigrateResponse>(
+                        Method::PUT,
+                        format!("control/v1/tenant/{migrate_shard}/migrate"),
+                        Some(TenantShardMigrateRequest {
+                            tenant_shard_id: migrate_shard,
+                            node_id: destination_node,
+                        }),
+                    )
+                    .await?;
+                println!("Migrate {} -> {} OK", migrate_shard, destination_node);
+            }
+
+            // Spread the shards across the nodes
+        }
         Command::TenantDescribe { tenant_id } => {
             let describe_response = storcon_client
                 .dispatch::<(), TenantDescribeResponse>(
@@ -554,389 +581,6 @@ async fn main() -> anyhow::Result<()> {
             }
             println!("{table}");
         }
-        Command::TenantWarmup { tenant_id } => {
-            let describe_response = storcon_client
-                .dispatch::<(), TenantDescribeResponse>(
-                    Method::GET,
-                    format!("control/v1/tenant/{tenant_id}"),
-                    None,
-                )
-                .await;
-            match describe_response {
-                Ok(describe) => {
-                    if matches!(describe.policy, PlacementPolicy::Secondary) {
-                        // Fine: it's already known to controller in secondary mode: calling
-                        // again to put it into secondary mode won't cause problems.
-                    } else {
-                        anyhow::bail!("Tenant already present with policy {:?}", describe.policy);
-                    }
-                }
-                Err(mgmt_api::Error::ApiError(StatusCode::NOT_FOUND, _)) => {
-                    // Fine: this tenant isn't know to the storage controller yet.
-                }
-                Err(e) => {
-                    // Unexpected API error
-                    return Err(e.into());
-                }
-            }
-
-            vps_client
-                .location_config(
-                    TenantShardId::unsharded(tenant_id),
-                    pageserver_api::models::LocationConfig {
-                        mode: pageserver_api::models::LocationConfigMode::Secondary,
-                        generation: None,
-                        secondary_conf: Some(LocationConfigSecondary { warm: true }),
-                        shard_number: 0,
-                        shard_count: 0,
-                        shard_stripe_size: ShardParameters::DEFAULT_STRIPE_SIZE.0,
-                        tenant_conf: TenantConfig::default(),
-                    },
-                    None,
-                    true,
-                )
-                .await?;
-
-            let describe_response = storcon_client
-                .dispatch::<(), TenantDescribeResponse>(
-                    Method::GET,
-                    format!("control/v1/tenant/{tenant_id}"),
-                    None,
-                )
-                .await?;
-
-            let secondary_ps_id = describe_response
-                .shards
-                .first()
-                .unwrap()
-                .node_secondary
-                .first()
-                .unwrap();
-
-            println!("Tenant {tenant_id} warming up on pageserver {secondary_ps_id}");
-            loop {
-                let (status, progress) = vps_client
-                    .tenant_secondary_download(
-                        TenantShardId::unsharded(tenant_id),
-                        Some(Duration::from_secs(10)),
-                    )
-                    .await?;
-                println!(
-                    "Progress: {}/{} layers, {}/{} bytes",
-                    progress.layers_downloaded,
-                    progress.layers_total,
-                    progress.bytes_downloaded,
-                    progress.bytes_total
-                );
-                match status {
-                    StatusCode::OK => {
-                        println!("Download complete");
-                        break;
-                    }
-                    StatusCode::ACCEPTED => {
-                        // Loop
-                    }
-                    _ => {
-                        anyhow::bail!("Unexpected download status: {status}");
-                    }
-                }
-            }
-        }
-        Command::TenantDrop { tenant_id, unclean } => {
-            if !unclean {
-                anyhow::bail!("This command is not a tenant deletion, and uncleanly drops all controller state for the tenant.  If you know what you're doing, add `--unclean` to proceed.")
-            }
-            storcon_client
-                .dispatch::<(), ()>(
-                    Method::POST,
-                    format!("debug/v1/tenant/{tenant_id}/drop"),
-                    None,
-                )
-                .await?;
-        }
-        Command::NodeDrop { node_id, unclean } => {
-            if !unclean {
-                anyhow::bail!("This command is not a clean node decommission, and uncleanly drops all controller state for the node, without checking if any tenants still refer to it.  If you know what you're doing, add `--unclean` to proceed.")
-            }
-            storcon_client
-                .dispatch::<(), ()>(Method::POST, format!("debug/v1/node/{node_id}/drop"), None)
-                .await?;
-        }
-        Command::NodeDelete { node_id } => {
-            storcon_client
-                .dispatch::<(), ()>(Method::DELETE, format!("control/v1/node/{node_id}"), None)
-                .await?;
-        }
-        Command::TenantSetTimeBasedEviction {
-            tenant_id,
-            period,
-            threshold,
-        } => {
-            vps_client
-                .tenant_config(&TenantConfigRequest {
-                    tenant_id,
-                    config: TenantConfig {
-                        eviction_policy: Some(EvictionPolicy::LayerAccessThreshold(
-                            EvictionPolicyLayerAccessThreshold {
-                                period: period.into(),
-                                threshold: threshold.into(),
-                            },
-                        )),
-                        heatmap_period: Some("300s".to_string()),
-                        ..Default::default()
-                    },
-                })
-                .await?;
-        }
-        Command::BulkMigrate {
-            nodes,
-            concurrency,
-            max_shards,
-            dry_run,
-        } => {
-            // Load the list of nodes, split them up into the drained and filled sets,
-            // and validate that draining is possible.
-            let node_descs = storcon_client
-                .dispatch::<(), Vec<NodeDescribeResponse>>(
-                    Method::GET,
-                    "control/v1/node".to_string(),
-                    None,
-                )
-                .await?;
-
-            let mut node_to_drain_descs = Vec::new();
-            let mut node_to_fill_descs = Vec::new();
-
-            for desc in node_descs {
-                let to_drain = nodes.iter().any(|id| *id == desc.id);
-                if to_drain {
-                    node_to_drain_descs.push(desc);
-                } else {
-                    node_to_fill_descs.push(desc);
-                }
-            }
-
-            if nodes.len() != node_to_drain_descs.len() {
-                anyhow::bail!("Bulk migration requested away from node which doesn't exist.")
-            }
-
-            node_to_fill_descs.retain(|desc| {
-                matches!(desc.availability, NodeAvailabilityWrapper::Active)
-                    && matches!(
-                        desc.scheduling,
-                        NodeSchedulingPolicy::Active | NodeSchedulingPolicy::Filling
-                    )
-            });
-
-            if node_to_fill_descs.is_empty() {
-                anyhow::bail!("There are no nodes to migrate to")
-            }
-
-            // Set the node scheduling policy to draining for the nodes which
-            // we plan to drain.
-            for node_desc in node_to_drain_descs.iter() {
-                let req = NodeConfigureRequest {
-                    node_id: node_desc.id,
-                    availability: None,
-                    scheduling: Some(NodeSchedulingPolicy::Draining),
-                };
-
-                storcon_client
-                    .dispatch::<_, ()>(
-                        Method::PUT,
-                        format!("control/v1/node/{}/config", node_desc.id),
-                        Some(req),
-                    )
-                    .await?;
-            }
-
-            // Perform the migration: move each tenant shard scheduled on a node to
-            // be drained to a node which is being filled. A simple round robin
-            // strategy is used to pick the new node.
-            let tenants = storcon_client
-                .dispatch::<(), Vec<TenantDescribeResponse>>(
-                    Method::GET,
-                    "control/v1/tenant".to_string(),
-                    None,
-                )
-                .await?;
-
-            let mut selected_node_idx = 0;
-
-            struct MigrationMove {
-                tenant_shard_id: TenantShardId,
-                from: NodeId,
-                to: NodeId,
-            }
-
-            let mut moves: Vec<MigrationMove> = Vec::new();
-
-            let shards = tenants
-                .into_iter()
-                .flat_map(|tenant| tenant.shards.into_iter());
-            for shard in shards {
-                if let Some(max_shards) = max_shards {
-                    if moves.len() >= max_shards {
-                        println!(
-                            "Stop planning shard moves since the requested maximum was reached"
-                        );
-                        break;
-                    }
-                }
-
-                let should_migrate = {
-                    if let Some(attached_to) = shard.node_attached {
-                        node_to_drain_descs
-                            .iter()
-                            .map(|desc| desc.id)
-                            .any(|id| id == attached_to)
-                    } else {
-                        false
-                    }
-                };
-
-                if !should_migrate {
-                    continue;
-                }
-
-                moves.push(MigrationMove {
-                    tenant_shard_id: shard.tenant_shard_id,
-                    from: shard
-                        .node_attached
-                        .expect("We only migrate attached tenant shards"),
-                    to: node_to_fill_descs[selected_node_idx].id,
-                });
-                selected_node_idx = (selected_node_idx + 1) % node_to_fill_descs.len();
-            }
-
-            let total_moves = moves.len();
-
-            if dry_run == Some(true) {
-                println!("Dryrun requested. Planned {total_moves} moves:");
-                for mv in &moves {
-                    println!("{}: {} -> {}", mv.tenant_shard_id, mv.from, mv.to)
-                }
-
-                return Ok(());
-            }
-
-            const DEFAULT_MIGRATE_CONCURRENCY: usize = 8;
-            let mut stream = futures::stream::iter(moves)
-                .map(|mv| {
-                    let client = Client::new(cli.api.clone(), cli.jwt.clone());
-                    async move {
-                        client
-                            .dispatch::<TenantShardMigrateRequest, TenantShardMigrateResponse>(
-                                Method::PUT,
-                                format!("control/v1/tenant/{}/migrate", mv.tenant_shard_id),
-                                Some(TenantShardMigrateRequest {
-                                    tenant_shard_id: mv.tenant_shard_id,
-                                    node_id: mv.to,
-                                }),
-                            )
-                            .await
-                            .map_err(|e| (mv.tenant_shard_id, mv.from, mv.to, e))
-                    }
-                })
-                .buffered(concurrency.unwrap_or(DEFAULT_MIGRATE_CONCURRENCY));
-
-            let mut success = 0;
-            let mut failure = 0;
-
-            while let Some(res) = stream.next().await {
-                match res {
-                    Ok(_) => {
-                        success += 1;
-                    }
-                    Err((tenant_shard_id, from, to, error)) => {
-                        failure += 1;
-                        println!(
-                            "Failed to migrate {} from node {} to node {}: {}",
-                            tenant_shard_id, from, to, error
-                        );
-                    }
-                }
-
-                if (success + failure) % 20 == 0 {
-                    println!(
-                        "Processed {}/{} shards: {} succeeded, {} failed",
-                        success + failure,
-                        total_moves,
-                        success,
-                        failure
-                    );
-                }
-            }
-
-            println!(
-                "Processed {}/{} shards: {} succeeded, {} failed",
-                success + failure,
-                total_moves,
-                success,
-                failure
-            );
-        }
-        Command::StartDrain { node_id } => {
-            storcon_client
-                .dispatch::<(), ()>(
-                    Method::PUT,
-                    format!("control/v1/node/{node_id}/drain"),
-                    None,
-                )
-                .await?;
-            println!("Drain started for {node_id}");
-        }
-        Command::CancelDrain { node_id, timeout } => {
-            storcon_client
-                .dispatch::<(), ()>(
-                    Method::DELETE,
-                    format!("control/v1/node/{node_id}/drain"),
-                    None,
-                )
-                .await?;
-
-            println!("Waiting for node {node_id} to quiesce on scheduling policy ...");
-
-            let final_policy =
-                wait_for_scheduling_policy(storcon_client, node_id, *timeout, |sched| {
-                    use NodeSchedulingPolicy::*;
-                    matches!(sched, Active | PauseForRestart)
-                })
-                .await?;
-
-            println!(
-                "Drain was cancelled for node {node_id}. Schedulling policy is now {final_policy:?}"
-            );
-        }
-        Command::StartFill { node_id } => {
-            storcon_client
-                .dispatch::<(), ()>(Method::PUT, format!("control/v1/node/{node_id}/fill"), None)
-                .await?;
-
-            println!("Fill started for {node_id}");
-        }
-        Command::CancelFill { node_id, timeout } => {
-            storcon_client
-                .dispatch::<(), ()>(
-                    Method::DELETE,
-                    format!("control/v1/node/{node_id}/fill"),
-                    None,
-                )
-                .await?;
-
-            println!("Waiting for node {node_id} to quiesce on scheduling policy ...");
-
-            let final_policy =
-                wait_for_scheduling_policy(storcon_client, node_id, *timeout, |sched| {
-                    use NodeSchedulingPolicy::*;
-                    matches!(sched, Active)
-                })
-                .await?;
-
-            println!(
-                "Fill was cancelled for node {node_id}. Schedulling policy is now {final_policy:?}"
-            );
-        }
     }
 
     Ok(())

deny.toml

@@ -4,7 +4,6 @@
 # to your expectations and requirements.
 
 # Root options
-[graph]
 targets = [
     { triple = "x86_64-unknown-linux-gnu" },
     { triple = "aarch64-unknown-linux-gnu" },
@@ -13,7 +12,6 @@ targets = [
 ]
 all-features = false
 no-default-features = false
-[output]
 feature-depth = 1
 
 # This section is considered when running `cargo deny check advisories`
@@ -21,16 +19,17 @@ feature-depth = 1
 # https://embarkstudios.github.io/cargo-deny/checks/advisories/cfg.html
 [advisories]
 db-urls = ["https://github.com/rustsec/advisory-db"]
+vulnerability = "deny"
+unmaintained = "warn"
 yanked = "warn"
-
-[[advisories.ignore]]
-id = "RUSTSEC-2023-0071"
-reason = "the marvin attack only affects private key decryption, not public key signature verification"
+notice = "warn"
+ignore = []
 
 # This section is considered when running `cargo deny check licenses`
 # More documentation for the licenses section can be found here:
 # https://embarkstudios.github.io/cargo-deny/checks/licenses/cfg.html
 [licenses]
+unlicensed = "deny"
 allow = [
     "Apache-2.0",
     "Artistic-2.0",
@@ -43,6 +42,10 @@ allow = [
     "OpenSSL",
     "Unicode-DFS-2016",
 ]
+deny = []
+copyleft = "warn"
+allow-osi-fsf-free = "neither"
+default = "deny"
 confidence-threshold = 0.8
 exceptions = [
     # Zlib license has some restrictions if we decide to change sth
@@ -96,13 +99,6 @@ name = "async-executor"
 [[bans.deny]]
 name = "smol"
 
-[[bans.deny]]
-# We want to use rustls instead of the platform's native tls implementation.
-name = "native-tls"
-
-[[bans.deny]]
-name = "openssl"
-
 # This section is considered when running `cargo deny check sources`.
 # More documentation about the 'sources' section can be found here:
 # https://embarkstudios.github.io/cargo-deny/checks/sources/cfg.html

docker-compose/README.md

@@ -1,10 +0,0 @@
-
-# Example docker compose configuration
-
-The configuration in this directory is used for testing Neon docker images: it is
-not intended for deploying a usable system.  To run a development environment where
-you can experiment with a minature Neon system, use `cargo neon` rather than container images.
-
-This configuration does not start the storage controller, because the controller
-needs a way to reconfigure running computes, and no such thing exists in this setup.
-

docker-compose/compute_wrapper/Dockerfile

@@ -1,4 +1,4 @@
-ARG REPOSITORY=neondatabase
+ARG REPOSITORY=369495373322.dkr.ecr.eu-central-1.amazonaws.com
 ARG COMPUTE_IMAGE=compute-node-v14
 ARG TAG=latest
 
@@ -8,11 +8,6 @@ USER root
 RUN apt-get update &&       \
     apt-get install -y curl \
                        jq   \
-                       python3-pip \
                        netcat
-#Faker is required for the pg_anon test
-RUN pip3 install Faker
-#This is required for the pg_hintplan test
-RUN mkdir -p /ext-src/pg_hint_plan-src && chown postgres /ext-src/pg_hint_plan-src 
 
-USER postgres
+USER postgres

docker-compose/compute_wrapper/shell/compute.sh

@@ -23,17 +23,18 @@ echo "Page server is ready."
 echo "Create a tenant and timeline"
 generate_id tenant_id
 PARAMS=(
-     -X PUT
+     -sb 
+     -X POST
      -H "Content-Type: application/json"
-     -d "{\"mode\": \"AttachedSingle\", \"generation\": 1, \"tenant_conf\": {}}"
-     "http://pageserver:9898/v1/tenant/${tenant_id}/location_config"
+     -d "{\"new_tenant_id\": \"${tenant_id}\"}"
+     http://pageserver:9898/v1/tenant/
 )
 result=$(curl "${PARAMS[@]}")
 echo $result | jq .
 
 generate_id timeline_id
 PARAMS=(
-     -sbf
+     -sb 
      -X POST
      -H "Content-Type: application/json"
      -d "{\"new_timeline_id\": \"${timeline_id}\", \"pg_version\": ${PG_VERSION}}"

docker-compose/compute_wrapper/var/db/postgres/specs/spec.json

@@ -95,7 +95,7 @@
             },
             {
                 "name": "shared_preload_libraries",
-                "value": "neon,pg_cron,timescaledb,pg_stat_statements",
+                "value": "neon",
                 "vartype": "string"
             },
             {
@@ -127,16 +127,6 @@
                 "name": "max_replication_flush_lag",
                 "value": "10GB",
                 "vartype": "string"
-            },
-            {
-                "name": "cron.database",
-                "value": "postgres",
-                "vartype": "string"
-            },
-            {
-                "name": "session_preload_libraries",
-                "value": "anon",
-                "vartype": "string"
             }
         ]
     },

docker-compose/docker-compose.yml

@@ -1,3 +1,5 @@
+version: '3'
+
 services:
   minio:
     restart: always
@@ -31,14 +33,25 @@ services:
     restart: always
     image: ${REPOSITORY:-neondatabase}/neon:${TAG:-latest}
     environment:
+      - BROKER_ENDPOINT='http://storage_broker:50051'
       - AWS_ACCESS_KEY_ID=minio
       - AWS_SECRET_ACCESS_KEY=password
       #- RUST_BACKTRACE=1
     ports:
        #- 6400:6400  # pg protocol handler
        - 9898:9898 # http endpoints
-    volumes:
-      - ./pageserver_config:/data/.neon/
+    entrypoint:
+      - "/bin/sh"
+      - "-c"
+    command:
+      - "/usr/local/bin/pageserver -D /data/.neon/
+                                   -c \"broker_endpoint=$$BROKER_ENDPOINT\"
+                                   -c \"listen_pg_addr='0.0.0.0:6400'\"
+                                   -c \"listen_http_addr='0.0.0.0:9898'\"
+                                   -c \"remote_storage={endpoint='http://minio:9000',
+                                                        bucket_name='neon',
+                                                        bucket_region='eu-north-1',
+                                                        prefix_in_bucket='/pageserver/'}\""
     depends_on:
       - storage_broker
       - minio_create_buckets
@@ -148,12 +161,12 @@ services:
       context: ./compute_wrapper/
       args:
         - REPOSITORY=${REPOSITORY:-neondatabase}
-        - COMPUTE_IMAGE=compute-node-v${PG_VERSION:-16}
+        - COMPUTE_IMAGE=compute-node-v${PG_VERSION:-14}
         - TAG=${TAG:-latest}
         - http_proxy=$http_proxy
         - https_proxy=$https_proxy
     environment:
-      - PG_VERSION=${PG_VERSION:-16}
+      - PG_VERSION=${PG_VERSION:-14}
       #- RUST_BACKTRACE=1
     # Mount the test files directly, for faster editing cycle.
     volumes:
@@ -181,14 +194,3 @@ services:
          done"
     depends_on:
       - compute
-
-  neon-test-extensions:
-    profiles: ["test-extensions"]
-    image: ${REPOSITORY:-neondatabase}/neon-test-extensions-v${PG_TEST_VERSION:-16}:${TAG:-latest}
-    entrypoint:
-      - "/bin/bash"
-      - "-c"
-    command:
-      - sleep 1800
-    depends_on:
-      - compute

docker-compose/docker_compose_test.sh

@@ -7,94 +7,54 @@
 # Implicitly accepts `REPOSITORY` and `TAG` env vars that are passed into the compose file
 # Their defaults point at DockerHub `neondatabase/neon:latest` image.`,
 # to verify custom image builds (e.g pre-published ones).
-#
-# A test script for postgres extensions
-# Currently supports only v16
-#
+
+# XXX: Current does not work on M1 macs due to x86_64 Docker images compiled only, and no seccomp support in M1 Docker emulation layer.
+
 set -eux -o pipefail
 
-COMPOSE_FILE='docker-compose.yml'
-cd $(dirname $0)
+SCRIPT_DIR="$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
+COMPOSE_FILE=$SCRIPT_DIR/docker-compose.yml
+
 COMPUTE_CONTAINER_NAME=docker-compose-compute-1
-TEST_CONTAINER_NAME=docker-compose-neon-test-extensions-1
-PSQL_OPTION="-h localhost -U cloud_admin -p 55433 -d postgres"
-: ${http_proxy:=}
-: ${https_proxy:=}
-export http_proxy https_proxy
+SQL="CREATE TABLE t(key int primary key, value text); insert into t values(1,1); select * from t;"
+PSQL_OPTION="-h localhost -U cloud_admin -p 55433 -c '$SQL' postgres"
 
 cleanup() {
     echo "show container information"
     docker ps
-    docker compose --profile test-extensions -f $COMPOSE_FILE logs
+    docker compose -f $COMPOSE_FILE logs
     echo "stop containers..."
-    docker compose --profile test-extensions -f $COMPOSE_FILE down
+    docker compose -f $COMPOSE_FILE down
 }
 
+echo "clean up containers if exists"
+cleanup
+
 for pg_version in 14 15 16; do
-    echo "clean up containers if exists"
-    cleanup
-    PG_TEST_VERSION=$(($pg_version < 16 ? 16 : $pg_version))
-    PG_VERSION=$pg_version PG_TEST_VERSION=$PG_TEST_VERSION docker compose --profile test-extensions -f $COMPOSE_FILE up --build -d
+    echo "start containers (pg_version=$pg_version)."
+    PG_VERSION=$pg_version docker compose -f $COMPOSE_FILE up --build -d
 
     echo "wait until the compute is ready. timeout after 60s. "
     cnt=0
-    while sleep 3; do
+    while sleep 1; do
         # check timeout
-        cnt=`expr $cnt + 3`
+        cnt=`expr $cnt + 1`
         if [ $cnt -gt 60 ]; then
             echo "timeout before the compute is ready."
             cleanup
             exit 1
         fi
-        if docker compose --profile test-extensions -f $COMPOSE_FILE logs "compute_is_ready" | grep -q "accepting connections"; then
+
+        # check if the compute is ready
+        set +o pipefail
+        result=`docker compose -f $COMPOSE_FILE logs "compute_is_ready" | grep "accepting connections" | wc -l`
+        set -o pipefail
+        if [ $result -eq 1 ]; then
             echo "OK. The compute is ready to connect."
             echo "execute simple queries."
             docker exec $COMPUTE_CONTAINER_NAME /bin/bash -c "psql $PSQL_OPTION"
+            cleanup
             break
         fi
     done
-
-    if [ $pg_version -ge 16 ]
-    then
-        echo Enabling trust connection
-        docker exec $COMPUTE_CONTAINER_NAME bash -c "sed -i '\$d' /var/db/postgres/compute/pg_hba.conf && echo -e 'host\t all\t all\t all\t trust' >> /var/db/postgres/compute/pg_hba.conf && psql $PSQL_OPTION -c 'select pg_reload_conf()' "
-        echo Adding postgres role
-        docker exec $COMPUTE_CONTAINER_NAME psql $PSQL_OPTION -c "CREATE ROLE postgres SUPERUSER LOGIN"
-        # This is required for the pg_hint_plan test, to prevent flaky log message causing the test to fail
-        # It cannot be moved to Dockerfile now because the database directory is created after the start of the container
-        echo Adding dummy config
-        docker exec $COMPUTE_CONTAINER_NAME touch /var/db/postgres/compute/compute_ctl_temp_override.conf
-        # This block is required for the pg_anon extension test.
-        # The test assumes that it is running on the same host with the postgres engine.
-        # In our case it's not true, that's why we are copying files to the compute node
-        TMPDIR=$(mktemp -d)
-        docker cp $TEST_CONTAINER_NAME:/ext-src/pg_anon-src/data $TMPDIR/data
-        echo -e '1\t too \t many \t tabs' > $TMPDIR/data/bad.csv
-        docker cp $TMPDIR/data $COMPUTE_CONTAINER_NAME:/tmp/tmp_anon_alternate_data
-        rm -rf $TMPDIR
-        TMPDIR=$(mktemp -d)
-        # The following block does the same for the pg_hintplan test
-        docker cp $TEST_CONTAINER_NAME:/ext-src/pg_hint_plan-src/data $TMPDIR/data
-        docker cp $TMPDIR/data $COMPUTE_CONTAINER_NAME:/ext-src/pg_hint_plan-src/
-        rm -rf $TMPDIR
-        # We are running tests now
-        if docker exec -e SKIP=timescaledb-src,rdkit-src,postgis-src,pgx_ulid-src,pgtap-src,pg_tiktoken-src,pg_jsonschema-src,pg_graphql-src,kq_imcx-src,wal2json_2_5-src \
-            $TEST_CONTAINER_NAME /run-tests.sh | tee testout.txt
-        then
-            cleanup
-        else
-            FAILED=$(tail -1 testout.txt)
-            for d in $FAILED
-            do
-                mkdir $d
-                docker cp $TEST_CONTAINER_NAME:/ext-src/$d/regression.diffs $d || true
-                docker cp $TEST_CONTAINER_NAME:/ext-src/$d/regression.out $d || true
-                cat $d/regression.out $d/regression.diffs || true
-            done
-        rm -rf $FAILED
-        cleanup
-        exit 1
-        fi
-    fi
-    cleanup
 done

docker-compose/pageserver_config/identity.toml

@@ -1 +0,0 @@
-id=1234

docker-compose/pageserver_config/pageserver.toml

@@ -1,5 +0,0 @@
-broker_endpoint='http://storage_broker:50051'
-pg_distrib_dir='/usr/local/'
-listen_pg_addr='0.0.0.0:6400'
-listen_http_addr='0.0.0.0:9898'
-remote_storage={ endpoint='http://minio:9000', bucket_name='neon', bucket_region='eu-north-1', prefix_in_bucket='/pageserver' }

docker-compose/run-tests.sh

@@ -1,15 +0,0 @@
-#!/bin/bash
-set -x
-
-cd /ext-src || exit 2
-FAILED=
-LIST=$( (echo -e "${SKIP//","/"\n"}"; ls -d -- *-src) | sort | uniq -u)
-for d in ${LIST}
-do
-       [ -d "${d}" ] || continue
-    psql -c "select 1" >/dev/null || break
-       USE_PGXS=1 make -C "${d}" installcheck || FAILED="${d} ${FAILED}"
-done
-[ -z "${FAILED}" ] && exit 0
-echo "${FAILED}"
-exit 1

docs/SUMMARY.md

@@ -1,18 +1,13 @@
 # Summary
 
-# Looking for `neon.tech` docs?
-
-This page linkes to a selection of technical content about the open source code in this repository.
-
-Please visit https://neon.tech/docs for documentation about using the Neon service, which is based on the code
-in this repository.
-
-# Architecture
-
 [Introduction]()
 - [Separation of Compute and Storage](./separation-compute-storage.md)
 
+# Architecture
+
 - [Compute]()
+  - [WAL proposer]()
+  - [WAL Backpressure]()
   - [Postgres changes](./core_changes.md)
 
 - [Pageserver](./pageserver.md)
@@ -21,15 +16,33 @@ in this repository.
     - [WAL Redo](./pageserver-walredo.md)
     - [Page cache](./pageserver-pagecache.md)
     - [Storage](./pageserver-storage.md)
+        - [Datadir mapping]()
+        - [Layer files]()
+        - [Branching]()
+        - [Garbage collection]()
+    - [Cloud Storage]()
     - [Processing a GetPage request](./pageserver-processing-getpage.md)
     - [Processing WAL](./pageserver-processing-wal.md)
+	- [Management API]()
+	- [Tenant Rebalancing]()
 
 - [WAL Service](walservice.md)
   - [Consensus protocol](safekeeper-protocol.md)
+  - [Management API]()
+  - [Rebalancing]()
+
+- [Control Plane]()
+
+- [Proxy]()
 
 - [Source view](./sourcetree.md)
   - [docker.md](./docker.md) — Docker images and building pipeline.
   - [Error handling and logging](./error-handling.md)
+  - [Testing]()
+    - [Unit testing]()
+    - [Integration testing]()
+    - [Benchmarks]()
+
 
 - [Glossary](./glossary.md)
 
@@ -45,6 +58,28 @@ in this repository.
 
 # RFCs
 
-Major changes are documented in RFCS:
-- See [RFCs](./rfcs/README.md) for more information
-- view the RFCs at https://github.com/neondatabase/neon/tree/main/docs/rfcs
+- [RFCs](./rfcs/README.md)
+
+- [002-storage](rfcs/002-storage.md)
+- [003-laptop-cli](rfcs/003-laptop-cli.md)
+- [004-durability](rfcs/004-durability.md)
+- [005-zenith_local](rfcs/005-zenith_local.md)
+- [006-laptop-cli-v2-CLI](rfcs/006-laptop-cli-v2-CLI.md)
+- [006-laptop-cli-v2-repository-structure](rfcs/006-laptop-cli-v2-repository-structure.md)
+- [007-serverless-on-laptop](rfcs/007-serverless-on-laptop.md)
+- [008-push-pull](rfcs/008-push-pull.md)
+- [009-snapshot-first-storage-cli](rfcs/009-snapshot-first-storage-cli.md)
+- [009-snapshot-first-storage](rfcs/009-snapshot-first-storage.md)
+- [009-snapshot-first-storage-pitr](rfcs/009-snapshot-first-storage-pitr.md)
+- [010-storage_details](rfcs/010-storage_details.md)
+- [011-retention-policy](rfcs/011-retention-policy.md)
+- [012-background-tasks](rfcs/012-background-tasks.md)
+- [013-term-history](rfcs/013-term-history.md)
+- [014-safekeepers-gossip](rfcs/014-safekeepers-gossip.md)
+- [014-storage-lsm](rfcs/014-storage-lsm.md)
+- [015-storage-messaging](rfcs/015-storage-messaging.md)
+- [016-connection-routing](rfcs/016-connection-routing.md)
+- [017-timeline-data-management](rfcs/017-timeline-data-management.md)
+- [018-storage-messaging-2](rfcs/018-storage-messaging-2.md)
+- [019-tenant-timeline-lifecycles](rfcs/019-tenant-timeline-lifecycles.md)
+- [cluster-size-limits](rfcs/cluster-size-limits.md)

docs/core_changes.md

@@ -11,28 +11,15 @@ page server. We currently use the same binary for both, with --wal-redo runtime
 the WAL redo mode. Some PostgreSQL changes are needed in the compute node, while others are just for
 the WAL redo process.
 
-In addition to core PostgreSQL changes, there is a Neon extension in the pgxn/neon directory that
-hooks into the smgr interface, and rmgr extension in pgxn/neon_rmgr. The extensions are loaded into
-the Postgres processes with shared_preload_libraries. Most of the Neon-specific code is in the
-extensions, and for any new features, that is preferred over modifying core PostgreSQL code.
+In addition to core PostgreSQL changes, there is a Neon extension in contrib/neon, to hook into the
+smgr interface. Once all the core changes have been submitted to upstream or eliminated some other
+way, the extension could live outside the postgres repository and build against vanilla PostgreSQL.
 
 Below is a list of all the PostgreSQL source code changes, categorized into changes needed for
 compute, and changes needed for the WAL redo process:
 
 # Changes for Compute node
 
-## Prefetching
-
-There are changes in many places to perform prefetching, for example for sequential scans. Neon
-doesn't benefit from OS readahead, and the latency to pageservers is quite high compared to local
-disk, so prefetching is critical for performance, also for sequential scans.
-
-### How to get rid of the patch
-
-Upcoming "streaming read" work in v17 might simplify this. And async I/O work in v18 will hopefully
-do more.
-
-
 ## Add t_cid to heap WAL records
 
 ```
@@ -50,11 +37,54 @@ The problem is that the XLOG_HEAP_INSERT record does not include the command id
 
 Bite the bullet and submit the patch to PostgreSQL, to add the t_cid to the WAL records. It makes the WAL records larger, which could make this unpopular in the PostgreSQL community. However, it might simplify some logical decoding code; Andres Freund briefly mentioned in PGCon 2022 discussion on Heikki's Neon presentation that logical decoding currently needs to jump through some hoops to reconstruct the same information.
 
-Update from Heikki (2024-04-17): I tried to write an upstream patch for that, to use the t_cid field for logical decoding, but it was not as straightforward as it first sounded.
 
 ### Alternatives
 Perhaps we could write an extra WAL record with the t_cid information, when a page is evicted that contains rows that were touched a transaction that's still running. However, that seems very complicated.
 
+## ginfast.c
+
+```
+diff --git a/src/backend/access/gin/ginfast.c b/src/backend/access/gin/ginfast.c
+index e0d9940946..2d964c02e9 100644
+--- a/src/backend/access/gin/ginfast.c
++++ b/src/backend/access/gin/ginfast.c
+@@ -285,6 +285,17 @@ ginHeapTupleFastInsert(GinState *ginstate, GinTupleCollector *collector)
+                memset(&sublist, 0, sizeof(GinMetaPageData));
+                makeSublist(index, collector->tuples, collector->ntuples, &sublist);
+ 
++               if (metadata->head != InvalidBlockNumber)
++               {
++                       /*
++                        * ZENITH: Get buffer before XLogBeginInsert() to avoid recursive call
++                        * of XLogBeginInsert(). Reading a new buffer might evict a dirty page from
++                        * the buffer cache, and if that page happens to be an FSM or VM page, zenith_write()
++                        * will try to WAL-log an image of the page.
++                        */
++                       buffer = ReadBuffer(index, metadata->tail);
++               }
++
+                if (needWal)
+                        XLogBeginInsert();
+ 
+@@ -316,7 +327,6 @@ ginHeapTupleFastInsert(GinState *ginstate, GinTupleCollector *collector)
+                        data.prevTail = metadata->tail;
+                        data.newRightlink = sublist.head;
+ 
+-                       buffer = ReadBuffer(index, metadata->tail);
+                        LockBuffer(buffer, GIN_EXCLUSIVE);
+                        page = BufferGetPage(buffer);
+```
+
+The problem is explained in the comment above
+
+### How to get rid of the patch
+
+Can we stop WAL-logging FSM or VM pages? Or delay the WAL logging until we're out of the critical
+section or something.
+
+Maybe some bigger rewrite of FSM and VM would help to avoid WAL-logging FSM and VM page images?
+
+
 ## Mark index builds that use buffer manager without logging explicitly
 
 ```
@@ -65,8 +95,6 @@ Perhaps we could write an extra WAL record with the t_cid information, when a pa
 also some changes in src/backend/storage/smgr/smgr.c
 ```
 
-pgvector 0.6.0 also needs a similar change, which would be very nice to get rid of too.
-
 When a GIN index is built, for example, it is built by inserting the entries into the index more or
 less normally, but without WAL-logging anything. After the index has been built, we iterate through
 all pages and write them to the WAL. That doesn't work for Neon, because if a page is not WAL-logged
@@ -81,10 +109,6 @@ an operation: `smgr_start_unlogged_build`, `smgr_finish_unlogged_build_phase_1`
 I think it would make sense to be more explicit about that in PostgreSQL too. So extract these
 changes to a patch and post to pgsql-hackers.
 
-Perhaps we could deduce that an unlogged index build has started when we see a page being evicted
-with zero LSN. How to be sure it's an unlogged index build rather than a bug? Currently we have a
-check for that and PANIC if we see page with zero LSN being evicted. And how do we detect when the
-index build has finished? See https://github.com/neondatabase/neon/pull/7440 for an attempt at that.
 
 ## Track last-written page LSN
 
@@ -116,6 +140,57 @@ The old method is still available, though.
 Wait until v15?
 
 
+## Cache relation sizes
+
+The Neon extension contains a little cache for smgrnblocks() and smgrexists() calls, to avoid going
+to the page server every time. It might be useful to cache those in PostgreSQL, maybe in the
+relcache? (I think we do cache nblocks in relcache already, check why that's not good enough for
+Neon)
+
+
+## Use buffer manager when extending VM or FSM
+
+```
+ src/backend/storage/freespace/freespace.c                   |   14 +-
+ src/backend/access/heap/visibilitymap.c                     |   15 +-
+
+diff --git a/src/backend/access/heap/visibilitymap.c b/src/backend/access/heap/visibilitymap.c
+index e198df65d8..addfe93eac 100644
+--- a/src/backend/access/heap/visibilitymap.c
++++ b/src/backend/access/heap/visibilitymap.c
+@@ -652,10 +652,19 @@ vm_extend(Relation rel, BlockNumber vm_nblocks)
+        /* Now extend the file */
+        while (vm_nblocks_now < vm_nblocks)
+        {
+-               PageSetChecksumInplace((Page) pg.data, vm_nblocks_now);
++               /*
++                * ZENITH: Initialize VM pages through buffer cache to prevent loading
++                * them from pageserver.
++                */
++               Buffer  buffer = ReadBufferExtended(rel, VISIBILITYMAP_FORKNUM, P_NEW,
++                                                                                       RBM_ZERO_AND_LOCK, NULL);
++               Page    page = BufferGetPage(buffer);
++
++               PageInit((Page) page, BLCKSZ, 0);
++               PageSetChecksumInplace(page, vm_nblocks_now);
++               MarkBufferDirty(buffer);
++               UnlockReleaseBuffer(buffer);
+ 
+-               smgrextend(rel->rd_smgr, VISIBILITYMAP_FORKNUM, vm_nblocks_now,
+-                                  pg.data, false);
+                vm_nblocks_now++;
+        }
+```
+
+### Problem we're trying to solve
+
+???
+
+### How to get rid of the patch
+
+Maybe this would be a reasonable change in PostgreSQL too?
+
+
 ## Allow startup without reading checkpoint record
 
 In Neon, the compute node is stateless. So when we are launching compute node, we need to provide
@@ -156,7 +231,7 @@ index 0415df9ccb..9f9db3c8bc 100644
   * crash we can lose (skip over) as many values as we pre-logged.
   */
 -#define SEQ_LOG_VALS   32
-+/* Neon XXX: to ensure sequence order of sequence in Zenith we need to WAL log each sequence update. */
++/* Zenith XXX: to ensure sequence order of sequence in Zenith we need to WAL log each sequence update. */
 +/* #define SEQ_LOG_VALS        32 */
 +#define SEQ_LOG_VALS   0
 ```
@@ -175,6 +250,66 @@ would be weird if the sequence moved backwards though, think of PITR.
 Or add a GUC for the amount to prefix to PostgreSQL, and force it to 1 in Neon.
 
 
+## Walproposer
+
+```
+ src/Makefile                                                |    1 +
+ src/backend/replication/libpqwalproposer/Makefile           |   37 +
+ src/backend/replication/libpqwalproposer/libpqwalproposer.c |  416 ++++++++++++
+ src/backend/postmaster/bgworker.c                           |    4 +
+ src/backend/postmaster/postmaster.c                         |    6 +
+ src/backend/replication/Makefile                            |    4 +-
+ src/backend/replication/walproposer.c                       | 2350 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+ src/backend/replication/walproposer_utils.c                 |  402 +++++++++++
+ src/backend/replication/walreceiver.c                       |    7 +
+ src/backend/replication/walsender.c                         |  320 ++++++---
+ src/backend/storage/ipc/ipci.c                              |    6 +
+ src/include/replication/walproposer.h                       |  565 ++++++++++++++++
+```
+
+WAL proposer is communicating with safekeeper and ensures WAL durability by quorum writes.  It is
+currently implemented as patch to standard WAL sender.
+
+### How to get rid of the patch
+
+Refactor into an extension. Submit hooks or APIs into upstream if necessary.
+
+@MMeent did some work on this already: https://github.com/neondatabase/postgres/pull/96
+
+## Ignore unexpected data beyond EOF in bufmgr.c
+
+```
+@@ -922,11 +928,14 @@ ReadBuffer_common(SMgrRelation smgr, char relpersistence, ForkNumber forkNum,
+                 */
+                bufBlock = isLocalBuf ? LocalBufHdrGetBlock(bufHdr) : BufHdrGetBlock(bufHdr);
+                if (!PageIsNew((Page) bufBlock))
+-                       ereport(ERROR,
++               {
++                        // XXX-ZENITH
++                        MemSet((char *) bufBlock, 0, BLCKSZ);
++                        ereport(DEBUG1,
+                                        (errmsg("unexpected data beyond EOF in block %u of relation %s",
+                                                        blockNum, relpath(smgr->smgr_rnode, forkNum)),
+                                         errhint("This has been seen to occur with buggy kernels; consider updating your system.")));
+-
++               }
+                /*
+                 * We *must* do smgrextend before succeeding, else the page will not
+                 * be reserved by the kernel, and the next P_NEW call will decide to
+```
+
+PostgreSQL is a bit sloppy with extending relations. Usually, the relation is extended with zeros
+first, then the page is filled, and finally the new page WAL-logged. But if multiple backends extend
+a relation at the same time, the pages can be WAL-logged in different order.
+
+I'm not sure what scenario exactly required this change in Neon, though.
+
+### How to get rid of the patch
+
+Submit patches to pgsql-hackers, to tighten up the WAL-logging around relation extension. It's a bit
+confusing even in PostgreSQL. Maybe WAL log the intention to extend first, then extend the relation,
+and finally WAL-log that the extension succeeded.
+
 ## Make smgr interface available to extensions
 
 ```
@@ -186,8 +321,6 @@ Or add a GUC for the amount to prefix to PostgreSQL, and force it to 1 in Neon.
 
 Submit to upstream. This could be useful for the Disk Encryption patches too, or for compression.
 
-We have submitted this to upstream, but it's moving at glacial a speed.
-https://commitfest.postgresql.org/47/4428/
 
 ## Added relpersistence argument to smgropen()
 
@@ -311,148 +444,6 @@ Ignore it. This is only needed for disaster recovery, so once we've eliminated a
 patches, we can just keep it around as a patch or as separate branch in a repo.
 
 
-## pg_waldump flags to ignore errors
-
-After creating a new project or branch in Neon, the first timeline can begin in the middle of a WAL segment. pg_waldump chokes on that, so we added some flags to make it possible to ignore errors.
-
-### How to get rid of the patch
-
-Like previous one, ignore it.
-
-
-
-## Backpressure if pageserver doesn't ingest WAL fast enough
-
-```
-@@ -3200,6 +3202,7 @@ ProcessInterrupts(void)
-                return;
-        InterruptPending = false;
- 
-+retry:
-        if (ProcDiePending)
-        {
-                ProcDiePending = false;
-@@ -3447,6 +3450,13 @@ ProcessInterrupts(void)
- 
-        if (ParallelApplyMessagePending)
-                HandleParallelApplyMessages();
-+
-+       /* Call registered callback if any */
-+       if (ProcessInterruptsCallback)
-+       {
-+               if (ProcessInterruptsCallback())
-+                       goto retry;
-+       }
- }
-```
-
-
-### How to get rid of the patch
-
-Submit a patch to upstream, for a hook in ProcessInterrupts. Could be useful for other extensions
-too.
-
-
-## SLRU on-demand download
-
-```
- src/backend/access/transam/slru.c | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------------
- 1 file changed, 92 insertions(+), 13 deletions(-)
-```
-
-### Problem we're trying to solve
-
-Previously, SLRU files were included in the basebackup, but the total size of them can be large,
-several GB, and downloading them all made the startup time too long.
-
-### Alternatives
-
-FUSE hook or LD_PRELOAD trick to intercept the reads on SLRU files
-
-
-## WAL-log an all-zeros page as one large hole
-
-- In XLogRecordAssemble()
-
-### Problem we're trying to solve
-
-This change was made in v16. Starting with v16, when PostgreSQL extends a relation, it first extends
-it with zeros, and it can extend the relation more than one block at a time. The all-zeros page is WAL-ogged, but it's very wasteful to include 8 kB of zeros in the WAL for that. This hack was made so that we WAL logged a compact record with a whole-page "hole". However, PostgreSQL has assertions that prevent that such WAL records from being replayed, so this breaks compatibility such that unmodified PostreSQL cannot process Neon-generated WAL.
-
-### How to get rid of the patch
-
-Find another compact representation for a full-page image of an all-zeros page. A compressed image perhaps.
-
-
-## Shut down walproposer after checkpointer
-
-```
-+                       /* Neon: Also allow walproposer background worker to be treated like a WAL sender, so that it's shut down last */
-+                       if ((bp->bkend_type == BACKEND_TYPE_NORMAL || bp->bkend_type == BACKEND_TYPE_BGWORKER) &&
-```
-
-This changes was needed so that postmaster shuts down the walproposer process only after the shutdown checkpoint record is written. Otherwise, the shutdown record will never make it to the safekeepers.
-
-### How to get rid of the patch
-
-Do a bigger refactoring of the postmaster state machine, such that a background worker can specify
-the shutdown ordering by itself. The postmaster state machine has grown pretty complicated, and
-would benefit from a refactoring for the sake of readability anyway.
-
-
-## EXPLAIN changes for prefetch and LFC
-
-### How to get rid of the patch
-
-Konstantin submitted a patch to -hackers already: https://commitfest.postgresql.org/47/4643/. Get that into a committable state.
-
-
-## On-demand download of extensions
-
-### How to get rid of the patch
-
-FUSE or LD_PRELOAD trickery to intercept reads?
-
-
-## Publication superuser checks
-
-We have hacked CreatePublication so that also neon_superuser can create them.
-
-### How to get rid of the patch
-
-Create an upstream patch with more fine-grained privileges for publications CREATE/DROP that can be GRANTed to users.
-
-
-## WAL log replication slots
-
-### How to get rid of the patch
-
-Utilize the upcoming v17 "slot sync worker", or a similar neon-specific background worker process, to periodically WAL-log the slots, or to export them somewhere else.
-
-
-## WAL-log replication snapshots
-
-### How to get rid of the patch
-
-WAL-log them periodically, from a backgound worker.
-
-
-## WAL-log relmapper files
-
-Similarly to replications snapshot files, the CID mapping files generated during VACUUM FULL of a catalog table are WAL-logged
-
-### How to get rid of the patch
-
-WAL-log them periodically, from a backgound worker.
-
-
-## XLogWaitForReplayOf()
-
-??
-
-
-
-
 # Not currently committed but proposed
 
 ## Disable ring buffer buffer manager strategies
@@ -481,10 +472,23 @@ hint bits are set. Wal logging hint bits updates requires FPI which significantl
 
 Add special WAL record for setting page hints.
 
+## Prefetching
+
+### Why?
+
+As far as pages in Neon are loaded on demand, to reduce node startup time
+and also speedup some massive queries we need some mechanism for bulk loading to
+reduce page request round-trip overhead.
+
+Currently Postgres is supporting prefetching only for bitmap scan.
+In Neon we should also use prefetch for sequential and index scans, because the OS is not doing it for us.
+For sequential scan we could prefetch some number of following pages. For index scan we could prefetch pages
+of heap relation addressed by TIDs.
+
 ## Prewarming
 
 ### Why?
 
-Short downtime (or, in other words, fast compute node restart time) is one of the key feature of Neon.
+Short downtime (or, in other words, fast compute node restart time) is one of the key feature of Zenith.
 But overhead of request-response round-trip for loading pages on demand can make started node warm-up quite slow.
 We can capture state of compute node buffer cache and send bulk request for this pages at startup.

docs/docker.md

@@ -4,18 +4,18 @@
 
 Currently we build two main images:
 
-- [neondatabase/neon](https://hub.docker.com/repository/docker/neondatabase/neon) — image with pre-built `pageserver`, `safekeeper` and `proxy` binaries and all the required runtime dependencies. Built from [/Dockerfile](/Dockerfile).
-- [neondatabase/compute-node-v16](https://hub.docker.com/repository/docker/neondatabase/compute-node-v16) — compute node image with pre-built Postgres binaries from [neondatabase/postgres](https://github.com/neondatabase/postgres). Similar images exist for v15 and v14.
+- [neondatabase/neon](https://hub.docker.com/repository/docker/zenithdb/zenith) — image with pre-built `pageserver`, `safekeeper` and `proxy` binaries and all the required runtime dependencies. Built from [/Dockerfile](/Dockerfile).
+- [neondatabase/compute-node](https://hub.docker.com/repository/docker/zenithdb/compute-node) — compute node image with pre-built Postgres binaries from [neondatabase/postgres](https://github.com/neondatabase/postgres).
 
 And additional intermediate image:
 
 - [neondatabase/compute-tools](https://hub.docker.com/repository/docker/neondatabase/compute-tools) — compute node configuration management tools.
 
-## Build pipeline
+## Building pipeline
 
 We build all images after a successful `release` tests run and push automatically to Docker Hub with two parallel CI jobs
 
-1. `neondatabase/compute-tools` and `neondatabase/compute-node-v16` (and -v15 and -v14)
+1. `neondatabase/compute-tools` and `neondatabase/compute-node`
 
 2. `neondatabase/neon`
 
@@ -34,12 +34,12 @@ You can see a [docker compose](https://docs.docker.com/compose/) example to crea
 1. create containers
 
 You can specify version of neon cluster using following environment values.
-- PG_VERSION: postgres version for compute (default is 16 as of this writing)
-- TAG: the tag version of [docker image](https://registry.hub.docker.com/r/neondatabase/neon/tags), which is tagged in [CI test](/.github/workflows/build_and_test.yml). Default is 'latest'
+- PG_VERSION: postgres version for compute (default is 14)
+- TAG: the tag version of [docker image](https://registry.hub.docker.com/r/neondatabase/neon/tags) (default is latest), which is tagged in [CI test](/.github/workflows/build_and_test.yml)
 ```
 $ cd docker-compose/
 $ docker-compose down   # remove the containers if exists
-$ PG_VERSION=16 TAG=latest docker-compose up --build -d  # You can specify the postgres and image version
+$ PG_VERSION=15 TAG=2937 docker-compose up --build -d  # You can specify the postgres and image version
 Creating network "dockercompose_default" with the default driver
 Creating docker-compose_storage_broker_1       ... done
 (...omit...)
@@ -47,31 +47,29 @@ Creating docker-compose_storage_broker_1       ... done
 
 2. connect compute node
 ```
-$ psql postgresql://cloud_admin:cloud_admin@localhost:55433/postgres
-psql (16.3)
-Type "help" for help.
-
+$ echo "localhost:55433:postgres:cloud_admin:cloud_admin" >> ~/.pgpass
+$ chmod 600 ~/.pgpass
+$ psql -h localhost -p 55433 -U cloud_admin
 postgres=# CREATE TABLE t(key int primary key, value text);
 CREATE TABLE
-postgres=# insert into t values(1, 1);
+postgres=# insert into t values(1,1);
 INSERT 0 1
 postgres=# select * from t;
- key | value 
+ key | value
 -----+-------
    1 | 1
 (1 row)
-
 ```
 
 3. If you want to see the log, you can use `docker-compose logs` command.
 ```
 # check the container name you want to see
 $ docker ps
-CONTAINER ID   IMAGE                                              COMMAND                  CREATED         STATUS         PORTS                                                                                      NAMES
-3582f6d76227   docker-compose_compute                             "/shell/compute.sh"      2 minutes ago   Up 2 minutes   0.0.0.0:3080->3080/tcp, :::3080->3080/tcp, 0.0.0.0:55433->55433/tcp, :::55433->55433/tcp   docker-compose_compute_1
+CONTAINER ID   IMAGE                                              COMMAND                  CREATED         STATUS         PORTS                                                                                                                                  NAMES
+d6968a5ae912   dockercompose_compute                              "/shell/compute.sh"      5 minutes ago   Up 5 minutes   0.0.0.0:3080->3080/tcp, 0.0.0.0:55433->55433/tcp                                                                                       dockercompose_compute_1
 (...omit...)
 
-$ docker logs -f docker-compose_compute_1
+$ docker logs -f dockercompose_compute_1
 2022-10-21 06:15:48.757 GMT [56] LOG:  connection authorized: user=cloud_admin database=postgres application_name=psql
 2022-10-21 06:17:00.307 GMT [56] LOG:  [NEON_SMGR] libpagestore: connected to 'host=pageserver port=6400'
 (...omit...)

docs/pageserver-pagecache.md

@@ -5,3 +5,4 @@ TODO:
 - shared across tenants
 - store pages from layer files
 - store pages from "in-memory layer"
+- store materialized pages

docs/pageserver-services.md

@@ -101,12 +101,11 @@ or
 ```toml
 [remote_storage]
 container_name = 'some-container-name'
-storage_account = 'somestorageaccnt'
 container_region = 'us-east'
 prefix_in_container = '/test-prefix/'
 ```
 
-The `AZURE_STORAGE_ACCESS_KEY` env variable can be used to specify the azure credentials if needed.
+`AZURE_STORAGE_ACCOUNT` and `AZURE_STORAGE_ACCESS_KEY` env variables can be used to specify the azure credentials if needed.
 
 ## Repository background tasks
 

docs/rfcs/033-storage-controller-drain-and-fill.md

@@ -1,345 +0,0 @@
-# Graceful Restarts of Storage Controller Managed Clusters
-
-## Summary
-This RFC describes new storage controller APIs for draining and filling tenant shards from/on pageserver nodes.
-It also covers how these new APIs should be used by an orchestrator (e.g. Ansible) in order to implement
-graceful cluster restarts.
-
-## Motivation
-
-Pageserver restarts cause read availablity downtime for tenants.
-
-For example pageserver-3 @ us-east-1 was unavailable for a randomly
-picked tenant (which requested on-demand activation) for around 30 seconds
-during the restart at 2024-04-03 16:37 UTC.
-
-Note that lots of shutdowns on loaded pageservers do not finish within the
-[10 second systemd enforced timeout](https://github.com/neondatabase/infra/blob/0a5280b383e43c063d43cbf87fa026543f6d6ad4/.github/ansible/systemd/pageserver.service#L16). This means we are shutting down without flushing ephemeral layers
-and have to reingest data in order to serve requests after restarting, potentially making first request latencies worse.
-
-This problem is not yet very acutely felt in storage controller managed pageservers since
-tenant density is much lower there. However, we are planning on eventually migrating all
-pageservers to storage controller management, so it makes sense to solve the issue proactively.
-
-## Requirements
-
-- Pageserver re-deployments cause minimal downtime for tenants
-- The storage controller exposes HTTP API hooks for draining and filling tenant shards
-from a given pageserver. Said hooks can be used by an orchestrator proces or a human operator.
-- The storage controller exposes some HTTP API to cancel draining and filling background operations.
-- Failures to drain or fill the node should not be fatal. In such cases, cluster restarts should proceed
-as usual (with downtime).
-- Progress of draining/filling is visible through metrics
-
-## Non Goals
-
-- Integration with the control plane
-- Graceful restarts for large non-HA tenants.
-
-## Impacted Components
-
-- storage controller
-- deployment orchestrator (i.e. Ansible)
-- pageserver (indirectly)
-
-## Terminology
-
-** Draining ** is the process through which all tenant shards that can be migrated from a given pageserver
-are distributed across the rest of the cluster.
-
-** Filling ** is the symmetric opposite of draining. In this process tenant shards are migrated onto a given
-pageserver until the cluster reaches a resonable, quiescent distribution of tenant shards across pageservers.
-
-** Node scheduling policies ** act as constraints to the scheduler. For instance, when a
-node is set in the `Paused` policy, no further shards will be scheduled on it.
-
-** Node ** is a pageserver. Term is used interchangeably in this RFC.
-
-** Deployment orchestrator ** is a generic term for whatever drives our deployments.
-Currently, it's an Ansible playbook.
-
-## Background
-
-### Storage Controller Basics (skip if already familiar)
-
-Fundamentally, the storage controller is a reconciler which aims to move from the observed mapping between pageservers and tenant shards to an intended mapping. Pageserver nodes and tenant shards metadata is durably persisted in a database, but note that the mapping between the two entities is not durably persisted. Instead, this mapping (*observed state*) is constructed at startup by sending `GET location_config` requests to registered pageservers.
-
-An internal scheduler maps tenant shards to pageservers while respecting certain constraints. The result of scheduling is the *intent state*. When the intent state changes, a *reconciliation* will inform pageservers about the new assigment via `PUT location_config` requests and will notify the compute via the configured hook.
-
-### Background Optimizations
-
-The storage controller performs scheduling optimizations in the background. It will
-migrate attachments to warm secondaries and replace secondaries in order to balance
-the cluster out.
-
-### Reconciliations Concurrency Limiting
-
-There's a hard limit on the number of reconciles that the storage controller
-can have in flight at any given time. To get an idea of scales, the limit is
-128 at the time of writing.
-
-## Implementation
-
-Note: this section focuses on the core functionality of the graceful restart process.
-It doesn't neccesarily describe the most efficient approach. Optimizations are described
-separately in a later section.
-
-### Overall Flow
-
-This section describes how to implement graceful restarts from the perspective
-of Ansible, the deployment orchestrator. Pageservers are already restarted sequentially.
-The orchestrator shall implement the following epilogue and prologue steps for each
-pageserver restart:
-
-#### Prologue
-
-The orchestrator shall first fetch the pageserver node id from the control plane or
-the pageserver it aims to restart directly. Next, it issues an HTTP request
-to the storage controller in order to start the drain of said pageserver node.
-All error responses are retried with a short back-off. When a 202 (Accepted)
-HTTP code is returned, the drain has started. Now the orchestrator polls the
-node status endpoint exposed by the storage controller in order to await the
-end of the drain process. When the `policy` field of the node status response
-becomes `PauseForRestart`, the drain has completed and the orchestrator can
-proceed with restarting the pageserver.
-
-The prologue is subject to an overall timeout. It will have a value in the ballpark
-of minutes. As storage controller managed pageservers become more loaded this timeout
-will likely have to increase.
-
-#### Epilogue
-
-After restarting the pageserver, the orchestrator issues an HTTP request
-to the storage controller to kick off the filling process. This API call
-may be retried for all error codes with a short backoff. This also serves
-as a synchronization primitive as the fill will be refused if the pageserver
-has not yet re-attached to the storage controller. When a 202(Accepted) HTTP
-code is returned, the fill has started. Now the orchestrator polls the node
-status endpoint exposed by the storage controller in order to await the end of
-the filling process. When the `policy` field of the node status response becomes
-`Active`, the fill has completed and the orchestrator may proceed to the next pageserver.
-
-Again, the epilogue is subject to an overall timeout. We can start off with
-using the same timeout as for the prologue, but can also consider relying on
-the storage controller's background optimizations with a shorter timeout.
-
-In the case that the deployment orchestrator times out, it attempts to cancel
-the fill. This operation shall be retried with a short back-off. If it ultimately
-fails it will require manual intervention to set the nodes scheduling policy to
-`NodeSchedulingPolicy::Active`. Not doing that is not immediately problematic,
-but it constrains the scheduler as mentioned previously.
-
-### Node Scheduling Policy State Machine
-
-The state machine below encodes the behaviours discussed above and
-the various failover situations described in a later section.
-
-Assuming no failures and/or timeouts the flow should be:
-`Active -> Draining -> PauseForRestart -> Active -> Filling -> Active`
-
-```
-                          Operator requested drain
-               +-----------------------------------------+
-               |                                         |
-       +-------+-------+                         +-------v-------+
-       |               |                         |               |
-       |     Pause     |             +----------->    Draining   +----------+
-       |               |             |           |               |          |
-       +---------------+             |           +-------+-------+          |
-                                     |                   |                  |
-                                     |                   |                  |
-                      Drain requested|                   |                  |
-                                     |                   |Drain complete    | Drain failed
-                                     |                   |                  | Cancelled/PS reattach/Storcon restart
-                                     |                   |                  |
-                             +-------+-------+           |                  |
-                             |               |           |                  |
-               +-------------+    Active     <-----------+------------------+
-               |             |               |           |
-Fill requested |             +---^---^-------+           |
-               |                 |   |                   |
-               |                 |   |                   |
-               |                 |   |                   |
-               |   Fill completed|   |                   |
-               |                 |   |PS reattach        |
-               |                 |   |after restart      |
-       +-------v-------+         |   |           +-------v-------+
-       |               |         |   |           |               |
-       |    Filling    +---------+   +-----------+PauseForRestart|
-       |               |                         |               |
-       +---------------+                         +---------------+
-```
-
-### Draining/Filling APIs
-
-The storage controller API to trigger the draining of a given node is:
-`PUT /v1/control/node/:node_id/{drain,fill}`.
-
-The following HTTP non-success return codes are used.
-All of them are safely retriable from the perspective of the storage controller.
-- 404: Requested node was not found
-- 503: Requested node is known to the storage controller, but unavailable
-- 412: Drain precondition failed: there is no other node to drain to or the node's schedulling policy forbids draining
-- 409: A {drain, fill} is already in progress. Only one such background operation
-is allowed per node.
-
-When the drain is accepted and commenced a 202 HTTP code is returned.
-
-Drains and fills shall be cancellable by the deployment orchestrator or a
-human operator via: `DELETE /v1/control/node/:node_id/{drain,fill}`. A 200
-response is returned when the cancelation is successful. Errors are retriable.
-
-### Drain Process
-
-Before accpeting a drain request the following validations is applied:
-* Ensure that the node is known the storage controller
-* Ensure that the schedulling policy is `NodeSchedulingPolicy::Active` or `NodeSchedulingPolicy::Pause`
-* Ensure that another drain or fill is not already running on the node
-* Ensure that a drain is possible (i.e. check that there is at least one
-schedulable node to drain to)
-
-After accepting the drain, the scheduling policy of the node is set to
-`NodeSchedulingPolicy::Draining` and persisted in both memory and the database.
-This disallows the optimizer from adding or removing shards from the node which
-is desirable to avoid them racing.
-
-Next, a separate Tokio task is spawned to manage the draining. For each tenant
-shard attached to the node being drained, demote the node to a secondary and
-attempt to schedule the node away. Scheduling might fail due to unsatisfiable
-constraints, but that is fine. Draining is a best effort process since it might
-not always be possible to cut over all shards.
-
-Importantly, this task manages the concurrency of issued reconciles in order to
-avoid drowning out the target pageservers and to allow other important reconciles
-to proceed.
-
-Once the triggered reconciles have finished or timed out, set the node's scheduling
-policy to `NodeSchedulingPolicy::PauseForRestart` to signal the end of the drain.
-
-A note on non HA tenants: These tenants do not have secondaries, so by the description
-above, they would not be migrated. It makes sense to skip them (especially the large ones)
-since, depending on tenant size, this might be more disruptive than the restart since the
-pageserver we've moved to do will need to on-demand download the entire working set for the tenant.
-We can consider expanding to small non-HA tenants in the future.
-
-### Fill Process
-
-Before accpeting a fill request the following validations is applied:
-* Ensure that the node is known the storage controller
-* Ensure that the schedulling policy is `NodeSchedulingPolicy::Active`.
-This is the only acceptable policy for the fill starting state. When a node re-attaches,
-it set the scheduling policy to `NodeSchedulingPolicy::Active` if it was equal to
-`NodeSchedulingPolicy::PauseForRestart` or `NodeSchedulingPolicy::Draining` (possible end states for a node drain).
-* Ensure that another drain or fill is not already running on the node
-
-After accepting the drain, the scheduling policy of the node is set to
-`NodeSchedulingPolicy::Filling` and persisted in both memory and the database.
-This disallows the optimizer from adding or removing shards from the node which
-is desirable to avoid them racing.
-
-Next, a separate Tokio task is spawned to manage the draining. For each tenant
-shard where the filled node is a secondary, promote the secondary. This is done
-until we run out of shards or the counts of attached shards become balanced across
-the cluster.
-
-Like for draining, the concurrency of spawned reconciles is limited.
-
-### Failure Modes & Handling
-
-Failures are generally handled by transition back into the `Active`
-(neutral) state. This simplifies the implementation greatly at the
-cost of adding transitions to the state machine. For example, we
-could detect the `Draining` state upon restart and proceed with a drain,
-but how should the storage controller know that's what the orchestrator
-needs still?
-
-#### Storage Controller Crash
-
-When the storage controller starts up reset the node scheduling policy
-of all nodes in states `Draining`, `Filling` or `PauseForRestart` to
-`Active`. The rationale is that when the storage controller restarts,
-we have lost context of what the deployment orchestrator wants. It also
-has the benefit of making things easier to reason about.
-
-#### Pageserver Crash During Drain
-
-The pageserver will attempt to re-attach during restart at which
-point the node scheduling policy will be set back to `Active`, thus
-reenabling the scheduler to use the node.
-
-#### Non-drained Pageserver Crash During Drain
-
-What should happen when a pageserver we are draining to crashes during the
-process. Two reasonable options are: cancel the drain and focus on the failover
-*or* do both, but prioritise failover. Since the number of concurrent reconciles
-produced by drains/fills are limited, we get the later behaviour for free.
-My suggestion is we take this approach, but the cancellation option is trivial
-to implement as well.
-
-#### Pageserver Crash During Fill
-
-The pageserver will attempt to re-attach during restart at which
-point the node scheduling policy will be set back to `Active`, thus
-reenabling the scheduler to use the node.
-
-#### Pageserver Goes unavailable During Drain/Fill
-
-The drain and fill jobs handle this by stopping early. When the pageserver
-is detected as online by storage controller heartbeats, reset its scheduling
-policy to `Active`. If a restart happens instead, see the pageserver crash
-failure mode.
-
-#### Orchestrator Drain Times Out
-
-Orchestrator will still proceed with the restart.
-When the pageserver re-attaches, the scheduling policy is set back to
-`Active`.
-
-#### Orchestrator Fill Times Out
-
-Orchestrator will attempt to cancel the fill operation. If that fails,
-the fill will continue until it quiesces and the node will be left
-in the `Filling` scheduling policy. This hinders the scheduler, but is
-otherwise harmless. A human operator can handle this by setting the scheduling
-policy to `Active`, or we can bake in a fill timeout into the storage controller.
-
-## Optimizations
-
-### Location Warmth
-
-When cutting over to a secondary, the storage controller will wait for it to
-become "warm" (i.e. download enough of the tenants data). This means that some
-reconciliations can take significantly longer than others and hold up precious
-reconciliations units. As an optimization, the drain stage can only cut over
-tenants that are already "warm". Similarly, the fill stage can prioritise the
-"warmest" tenants in the fill.
-
-Given that the number of tenants by the storage controller will be fairly low
-for the foreseable future, the first implementation could simply query the tenants
-for secondary status. This doesn't scale well with increasing tenant counts, so
-eventually we will need new pageserver API endpoints to report the sets of
-"warm" and "cold" nodes.
-
-## Alternatives Considered
-
-### Draining and Filling Purely as Scheduling Constraints
-
-At its core, the storage controller is a big background loop that detects changes
-in the environment and reacts on them. One could express draining and filling
-of nodes purely in terms of constraining the scheduler (as opposed to having
-such background tasks).
-
-While theoretically nice, I think that's harder to implement and more importantly operate and reason about.
-Consider cancellation of a drain/fill operation. We would have to update the scheduler state, create
-an entirely new schedule (intent state) and start work on applying that. It gets trickier if we wish
-to cancel the reconciliation tasks spawned by drain/fill nodes. How would we know which ones belong
-to the conceptual drain/fill? One could add labels to reconciliations, but it gets messy in my opinion.
-
-It would also mean that reconciliations themselves have side effects that persist in the database
-(persist something to the databse when the drain is done), which I'm not conceptually fond of.
-
-## Proof of Concept
-
-This RFC is accompanied by a POC which implements nearly everything mentioned here
-apart from the optimizations and some of the failure handling:
-https://github.com/neondatabase/neon/pull/7682

docs/rfcs/034-ancestor-deletion.md

@@ -1,252 +0,0 @@
-# Ancestor Timeline Deletion
-
-Created on: 2024-02-23
-
-Author: John Spray
-
-# Summary
-
-When a tenant creates a new timeline that they will treat as their 'main' history,
-it is awkward to permanently retain an 'old main' timeline as its ancestor. Currently
-this is necessary because it is forbidden to delete a timeline which has descendents.
-
-A new pageserver API is proposed to 'adopt' data from a parent timeline into
-one of its children, such that the link between ancestor and child can be severed,
-leaving the parent in a state where it may then be deleted.
-
-# Motivation
-
-Retaining parent timelines currently has two costs:
-
-- Cognitive load on users, who have to remember which is the "real" main timeline.
-- Storage capacity cost, as the parent timeline will retain layers up to the
-  child's timeline point, even if the child fully covers its keyspace with image
-  layers and will never actually read from the parent.
-
-# Solution
-
-A new pageserver API `PUT /v1/tenant/:tenant_id/timeline/:timeline_id/detach_ancestor`
-will be added. The `timeline_id` in this URL is that of the _child_ timeline that we
-wish to detach from its parent.
-
-On success, this API will leave the following state:
-
-- The detached child timeline will no longer have an ancestor, and will contain all
-  the data needed to service reads without recursing into an ancestor.
-- Any other children of the parent whose timeline points were at a lower LSN than
-  the detached child timeline will be modified to have the child timeline as their
-  new parent.
-- The parent timeline will still exist, but the child will no longer have it as an
-  ancestor. If this was the last timeline that depended on the parent, then the
-  parent will become deletable.
-
-This API's implementation will consist of a series of retryable steps, such that
-on failures/timeout it can safely be called again to reach the target state.
-
-## Example
-
-### Before
-
-The user has "rolled back" their project to LSN X, resulting in a "new main"
-timeline. The parent "old main" timeline still exists, and they would like
-to clean it up.
-
-They have two other timelines A and B. A is from before the rollback point,
-and B is from after the rollback point.
-
-```
-----"old main" timeline-------X-------------------------------------------->
-                |             |                         |
-                |-> child A   |                         |
-                              |-> "new main" timeline   |
-                                                        -> child B
-
-```
-
-### After calling detach ancestor API
-
-The "new main" timeline is no longer dependent on old main, and neither
-is child A, because it had a branch point before X.
-
-The user may now choose to delete child B and "old main" to get to
-a pristine state. Child B is likely to be unwanted since the user
-chose to roll back to X, and it branches from after X. However, we
-don't assume this in the API; it is up to the user to delete it.
-
-```
-|----"old main" timeline---------------------------------------------------->
-                                                         |
-                                                         |
-                                                         |
-                                                         -> child B
-
-|----"new main" timeline--------->
-                 |
-                 |-> child A
-
-
-```
-
-### After removing timelines
-
-We end up with a totally clean state that leaves no trace that a rollback
-ever happened: there is only one root timeline.
-
-```
-| ----"new main" timeline----------->
-                |
-                |-> child A
-
-
-```
-
-## Caveats
-
-Important things for API users to bear in mind:
-
-- this API does not delete the parent timeline: you must still do that explicitly.
-- if there are other child timelines ahead of the branch point of the detached
-  child, the parent won't be deletable: you must either delete or detach those
-  children.
-- do _not_ simply loop over all children and detach them all: this can have an
-  extremely high storage cost. The detach ancestor API is intended for use on a single
-  timeline to make it the new "main".
-- The detach ancestor API should also not be
-  exposed directly to the user as button/API, because they might decide
-  to click it for all the children and thereby generate many copies of the
-  parent's data -- the detach ancestor API should be used as part
-  of a high level "clean up after rollback" feature.
-
-## `detach_ancestor` API implementation
-
-Terms used in the following sections:
-
-- "the child": the timeline whose ID is specified in the detach ancestor API URL, also
-  called "new main" in the example.
-- "the parent": the parent of "the child". Also called "old main" in the example.
-- "the branch point" the ancestor_lsn of "the child"
-
-### Phase 1: write out adopted layers to S3
-
-The child will "adopt" layers from the parent, such that its end state contains
-all the parent's history as well as its own.
-
-For all layers in the parent's layer map whose high LSN is below the branch
-point, issue S3 CopyObject requests to duplicate them into the child timeline's
-prefix. Do not add them to the child's layer map yet.
-
-For delta layers in the parent's layer map which straddle the branch point, read them
-and write out only content up to the branch point into new layer objects.
-
-This is a long running operation if the parent has many layers: it should be
-implemented in a way that resumes rather than restarting from scratch, if the API
-times out and is called again.
-
-As an optimization, if there are no other timelines that will be adopted into
-the child, _and_ the child's image layers already full cover the branch LSN,
-then we may skip adopting layers.
-
-### Phase 2: update the child's index
-
-Having written out all needed layers in phase 1, atomically link them all
-into the child's IndexPart and upload to S3. This may be done while the
-child Timeline is still running.
-
-### Phase 3: modify timelines ancestry
-
-Modify the child's ancestor to None, and upload its IndexPart to persist the change.
-
-For all timelines which have the same parent as the child, and have a branch
-point lower than our branch point, switch their ancestor_timeline to the child,
-and upload their IndexPart to persist the change.
-
-## Alternatives considered
-
-### Generate full image layer on child, rather than adopting parent deltas
-
-This would work for the case of a single child, but would prevent re-targeting
-other timelines that depended on the parent. If we detached many children this
-way, the storage cost would become prohibitive (consider a 1TB database with
-100 child timelines: it would cost 100TiB if they all generated their own image layers).
-
-### Don't rewrite anything: just fake it in the API
-
-We could add a layer of indirection that let a child "pretend" that it had no
-ancestor, when in reality it still had the parent. The pageserver API could
-accept deletion of ancestor timelines, and just update child metadata to make
-them look like they have no ancestor.
-
-This would not achieve the desired reduction in storage cost, and may well be more
-complex to maintain than simply implementing the API described in this RFC.
-
-### Avoid copying objects: enable child index to use parent layers directly
-
-We could teach IndexPart to store a TimelineId for each layer, such that a child
-timeline could reference a parent's layers directly, rather than copying them
-into the child's prefix.
-
-This would impose a cost for the normal case of indices that only target the
-timeline's own layers, add complexity, and break the useful simplifying
-invariant that timelines "own" their own path. If child timelines were
-referencing layers from the parent, we would have to ensure that the parent
-never runs GC/compaction again, which would make the API less flexible (the
-proposal in this RFC enables deletion of the parent but doesn't require it.)
-
-## Performance
-
-### Adopting layers
-
-- CopyObject is a relatively cheap operation, but we may need to issue tens of thousands
-  of such requests: this can take up to tens of seconds and will compete for RemoteStorage
-  semaphore units with other activity on the pageserver.
-- If we are running on storage backend that doesn't implement CopyObject, then
-  this part will be much more expensive as we would stream all layer content
-  through the pageserver. This is no different to issuing a lot
-  of reads to a timeline that does not have a warm local cache: it will move
-  a lot of gigabytes, but that shouldn't break anything.
-- Generating truncated layers for delta that straddle the branch point will
-  require streaming read/write of all the layers in question.
-
-### Updating timeline ancestry
-
-The simplest way to update timeline ancestry will probably be to stop and start
-all the Timeline objects: this is preferable to the complexity of making their
-ancestry mutable at runtime.
-
-There will be a corresponding "stutter" in the availability of the timelines,
-of the order 10-100ms, which is the time taken to upload their IndexPart, and
-restart the Timeline.
-
-# Interaction with other features
-
-## Concurrent timeline creation
-
-If new historic timelines are created using the parent as an ancestor while the
-detach ancestor API is running, they will not be re-parented to the child. This
-doesn't break anything, but it leaves the parent in a state where it might not
-be possible to delete it.
-
-Since timeline creations are an explicit user action, this is not something we need to
-worry about as the storage layer: a user who wants to delete their parent timeline will not create
-new children, and if they do, they can choose to delete those children to
-enable deleting the parent.
-
-For the least surprise to the user, before starting the detach ancestor branch
-operation, the control plane should wait until all branches are created and not
-allow any branches to be created before the branch point on the ancestor branch
-while the operation is ongoing.
-
-## WAL based disaster recovery
-
-WAL based disaster recovery currently supports only restoring of the main
-branch. Enabling WAL based disaster recovery in the future requires that we
-keep a record which timeline generated the WAL and at which LSN was a parent
-detached. Keep a list of timeline ids and the LSN in which they were detached in
-the `index_part.json`. Limit the size of the list to 100 first entries, after
-which the WAL disaster recovery will not be possible.
-
-## Sharded tenants
-
-For sharded tenants, calls to the detach ancestor API will pass through the storage
-controller, which will handle them the same as timeline creations: invoke first
-on shard zero, and then on all the other shards.

docs/rfcs/035-safekeeper-dynamic-membership-change.md

@@ -1,495 +0,0 @@
-# Safekeeper dynamic membership change
-
-To quickly recover from safekeeper node failures and do rebalancing we need to
-be able to change set of safekeepers the timeline resides on. The procedure must
-be safe (not lose committed log) regardless of safekeepers and compute state. It
-should be able to progress if any majority of old safekeeper set, any majority
-of new safekeeper set and compute are up and connected. This is known as a
-consensus membership change. It always involves two phases: 1) switch old
-majority to old + new configuration, preventing commits without acknowledge from
-the new set 2) bootstrap the new set by ensuring majority of the new set has all
-data which ever could have been committed before the first phase completed;
-after that switch is safe to finish. Without two phases switch to the new set
-which quorum might not intersect with quorum of the old set (and typical case of
-ABC -> ABD switch is an example of that, because quorums AC and BD don't
-intersect). Furthermore, procedure is typically carried out by the consensus
-leader, and so enumeration of configurations which establishes order between
-them is done through consensus log.
-
-In our case consensus leader is compute (walproposer), and we don't want to wake
-up all computes for the change. Neither we want to fully reimplement the leader
-logic second time outside compute. Because of that the proposed algorithm relies
-for issuing configurations on the external fault tolerant (distributed) strongly
-consisent storage with simple API: CAS (compare-and-swap) on the single key.
-Properly configured postgres suits this.
-
-In the system consensus is implemented at the timeline level, so algorithm below
-applies to the single timeline.
-
-## Algorithm
-
-### Definitions
-
-A configuration is
-
-```
-struct Configuration {
-    generation: Generation, // a number uniquely identifying configuration
-    sk_set: Vec<NodeId>, // current safekeeper set
-    new_sk_set: Optional<Vec<NodeId>>,
-}
-```
-
-Configuration with `new_set` present is used for the intermediate step during
-the change and called joint configuration. Generations establish order of
-generations: we say `c1` is higher than `c2` if `c1.generation` >
-`c2.generation`.
-
-### Persistently stored data changes
-
-Safekeeper starts storing its current configuration in the control file. Update
-of is atomic, so in-memory value always matches the persistent one.
-
-External CAS providing storage (let's call it configuration storage here) also
-stores configuration for each timeline. It is initialized with generation 1 and
-initial set of safekeepers during timeline creation. Executed CAS on it must
-never be lost.
-
-### Compute <-> safekeeper protocol changes
-
-`ProposerGreeting` message carries walproposer's configuration if it is already
-established (see below), else null.  `AcceptorGreeting` message carries
-safekeeper's current `Configuration`. All further messages (`VoteRequest`,
-`VoteResponse`, `ProposerElected`, `AppendRequest`, `AppendResponse`) carry
-generation number, of walproposer in case of wp->sk message or of safekeeper in
-case of sk->wp message.
-
-### Safekeeper changes
-
-Basic rule: once safekeeper observes configuration higher than his own it
-immediately switches to it. It must refuse all messages with lower generation
-that his. It also refuses messages if it is not member of the current generation
-(that is, of either `sk_set` of `sk_new_set`), though it is likely not unsafe to
-process them (walproposer should ignore result anyway).
-
-If there is non null configuration in `ProposerGreeting` and it is higher than
-current safekeeper one, safekeeper switches to it.
-
-Safekeeper sends its current configuration in its first message to walproposer
-`AcceptorGreeting`. It refuses all other walproposer messages if the
-configuration generation in them is less than its current one. Namely, it
-refuses to vote, to truncate WAL in `handle_elected` and to accept WAL. In
-response it sends its current configuration generation to let walproposer know.
-
-Safekeeper gets `PUT /v1/tenants/{tenant_id}/timelines/{timeline_id}/configuration` 
-accepting `Configuration`. Safekeeper switches to the given conf it is higher than its
-current one and ignores it otherwise. In any case it replies with
-```
-struct ConfigurationSwitchResponse {
-    conf: Configuration,
-    term: Term,
-    last_log_term: Term,
-    flush_lsn: Lsn,
-}
-```
-
-### Compute (walproposer) changes
-
-Basic rule is that joint configuration requires votes from majorities in the
-both `set` and `new_sk_set`.
-
-Compute receives list of safekeepers to connect to from the control plane as
-currently and tries to communicate with all of them. However, the list does not
-define consensus members. Instead, on start walproposer tracks highest
-configuration it receives from `AcceptorGreeting`s. Once it assembles greetings
-from majority of `sk_set` and majority of `new_sk_set` (if it is present), it
-establishes this configuration as its own and moves to voting. 
-
-It should stop talking to safekeepers not listed in the configuration at this
-point, though it is not unsafe to continue doing so.
-
-To be elected it must receive votes from both majorites if `new_sk_set` is present.
-Similarly, to commit WAL it must receive flush acknowledge from both majorities.
-
-If walproposer hears from safekeeper configuration higher than his own (i.e.
-refusal to accept due to configuration change) it simply restarts.
-
-### Change algorithm
-
-The following algorithm can be executed anywhere having access to configuration
-storage and safekeepers. It is safe to interrupt / restart it and run multiple
-instances of it concurrently, though likely one of them won't make
-progress then. It accepts `desired_set: Vec<NodeId>` as input. 
-
-Algorithm will refuse to make the change if it encounters previous interrupted
-change attempt, but in this case it will try to finish it.
-
-It will eventually converge if old majority, new majority and configuration
-storage are reachable.
-
-1) Fetch current timeline configuration from the configuration storage.
-2) If it is already joint one and `new_set` is different from `desired_set`
-   refuse to change. However, assign join conf to (in memory) var
-   `join_conf` and proceed to step 4 to finish the ongoing change.
-3) Else, create joint `joint_conf: Configuration`: increment current conf number
-   `n` and put `desired_set` to `new_sk_set`. Persist it in the configuration
-   storage by doing CAS on the current generation: change happens only if
-   current configuration number is still `n`. Apart from guaranteeing uniqueness
-   of configurations, CAS linearizes them, ensuring that new configuration is
-   created only following the previous one when we know that the transition is
-   safe. Failed CAS aborts the procedure.
-4) Call `PUT` `configuration` on safekeepers from the current set,
-   delivering them `joint_conf`. Collecting responses from majority is required
-   to proceed. If any response returned generation higher than 
-   `joint_conf.generation`, abort (another switch raced us). Otherwise, choose
-   max `<last_log_term, flush_lsn>` among responses and establish it as
-   (in memory) `sync_position`. Also choose max `term` and establish it as (in
-   memory) `sync_term`. We can't finish the switch until majority of the new set
-   catches up to this `sync_position` because data before it could be committed
-   without ack from the new set. Similarly, we'll bump term on new majority
-   to `sync_term` so that two computes with the same term are never elected.
-4) Initialize timeline on safekeeper(s) from `new_sk_set` where it
-   doesn't exist yet by doing `pull_timeline` from the majority of the 
-   current set. Doing that on majority of `new_sk_set` is enough to
-   proceed, but it is reasonable to ensure that all `new_sk_set` members
-   are initialized -- if some of them are down why are we migrating there?
-5) Call `POST` `bump_term(sync_term)` on safekeepers from the new set. 
-   Success on majority is enough.
-6) Repeatedly call `PUT` `configuration` on safekeepers from the new set,
-   delivering them `joint_conf` and collecting their positions. This will
-   switch them to the `joint_conf` which generally won't be needed 
-   because `pull_timeline` already includes it and plus additionally would be
-   broadcast by compute. More importantly, we may proceed to the next step
-   only when `<last_log_term, flush_lsn>` on the majority of the new set reached 
-   `sync_position`. Similarly, on the happy path no waiting is not needed because 
-   `pull_timeline` already includes it. However, we should double
-    check to be safe. For example, timeline could have been created earlier e.g.
-    manually or after try-to-migrate, abort, try-to-migrate-again sequence. 
-7) Create `new_conf: Configuration` incrementing `join_conf` generation and having new 
-   safekeeper set as `sk_set` and None `new_sk_set`. Write it to configuration 
-   storage under one more CAS.
-8) Call `PUT` `configuration` on safekeepers from the new set,
-   delivering them `new_conf`. It is enough to deliver it to the majority 
-   of the new set; the rest can be updated by compute.
-
-I haven't put huge effort to make the description above very precise, because it
-is natural language prone to interpretations anyway. Instead I'd like to make TLA+
-spec of it.
-
-Description above focuses on safety. To make the flow practical and live, here a few more 
-considerations.
-1) It makes sense to ping new set to ensure it we are migrating to live node(s) before 
-  step 3.
-2) If e.g. accidentally wrong new sk set has been specified, before CAS in step `6` is completed 
-   it is safe to rollback to the old conf with one more CAS.
-3) On step 4 timeline might be already created on members of the new set for various reasons; 
-   the simplest is the procedure restart. There are more complicated scenarious like mentioned
-   in step 5. Deleting and re-doing `pull_timeline` is generally unsafe without involving 
-   generations, so seems simpler to treat existing timeline as success. However, this also 
-   has a disadvantage: you might imagine an surpassingly unlikely schedule where condition in
-   the step 5 is never reached until compute is (re)awaken up to synchronize new member(s).
-   I don't think we'll observe this in practice, but can add waking up compute if needed.
-4) In the end timeline should be locally deleted on the safekeeper(s) which are
-   in the old set but not in the new one, unless they are unreachable. To be
-   safe this also should be done under generation number (deletion proceeds only if 
-   current configuration is <= than one in request and safekeeper is not memeber of it).
-5) If current conf fetched on step 1 is already not joint and members equal to `desired_set`,
-   jump to step 7, using it as `new_conf`.
-
-## Implementation
-
-The procedure ought to be driven from somewhere. Obvious candidates are control
-plane and storage_controller; and as each of them already has db we don't want
-yet another storage. I propose to manage safekeepers in storage_controller
-because 1) since it is in rust it simplifies simulation testing (more on this
-below) 2) it already manages pageservers. 
-
-This assumes that migration will be fully usable only after we migrate all
-tenants/timelines to storage_controller. It is discussible whether we want also
-to manage pageserver attachments for all of these, but likely we do.
-
-This requires us to define storcon <-> cplane interface.
-
-### storage_controller <-> control plane interface
-
-First of all, control plane should
-[change](https://neondb.slack.com/archives/C03438W3FLZ/p1719226543199829)
-storing safekeepers per timeline instead of per tenant because we can't migrate
-tenants atomically. 
-
-The important question is how updated configuration is delivered from
-storage_controller to control plane to provide it to computes. As always, there
-are two options, pull and push. Let's do it the same push as with pageserver
-`/notify-attach` because 1) it keeps storage_controller out of critical compute
-start path 2) provides easier upgrade: there won't be such a thing as 'timeline
-managed by control plane / storcon', cplane just takes the value out of its db
-when needed 3) uniformity. It makes storage_controller responsible for retrying notifying
-control plane until it succeeds.
-
-So, cplane `/notify-safekeepers` for the timeline accepts `Configuration` and
-updates it in the db if the provided conf generation is higher (the cplane db
-should also store generations for this). Similarly to [`/notify-attach`](https://www.notion.so/neondatabase/Storage-Controller-Control-Plane-interface-6de56dd310a043bfa5c2f5564fa98365), it
-should update db which makes the call successful, and then try to schedule
-`apply_config` if possible, it is ok if not. storage_controller 
-should rate limit calling the endpoint, but likely this won't be needed, as migration
-throughput is limited by `pull_timeline`.
-
-Timeline (branch) creation in cplane should call storage_controller POST
-`tenant/:tenant_id/timeline` like it currently does for sharded tenants.
-Response should be augmented with `safekeeper_conf: Configuration`. The call
-should be retried until succeeds.
-
-Timeline deletion and tenant deletion in cplane should call appropriate
-storage_controller endpoints like it currently does for sharded tenants. The
-calls should be retried until they succeed.
-
-### storage_controller implementation
-
-Current 'load everything on startup and keep in memory' easy design is fine.
-Single timeline shouldn't take more than 100 bytes (it's 16 byte tenant_id, 16
-byte timeline_id, int generation, vec of ~3 safekeeper ids plus some flags), so
-10^6 of timelines shouldn't take more than 100MB.
-
-Similar to pageserver attachment Intents storage_controller would have in-memory
-`MigrationRequest` (or its absense) for each timeline and pool of tasks trying
-to make these request reality; this ensures one instance of storage_controller
-won't do several migrations on the same timeline concurrently. In the first
-version it is simpler to have more manual control and no retries, i.e. migration
-failure removes the request. Later we can build retries and automatic
-scheduling/migration. `MigrationRequest` is
-```
-enum MigrationRequest {
-    To(Vec<NodeId>),
-    FinishPending,
-}
-```
-
-`FinishPending` requests to run the procedure to ensure state is clean: current
-configuration is not joint and majority of safekeepers are aware of it, but do
-not attempt to migrate anywhere. If current configuration fetched on step 1 is
-not joint it jumps to step 7. It should be run at startup for all timelines (but
-similarly, in the first version it is ok to trigger it manually).
-
-#### Schema
-
-`safekeepers` table mirroring current `nodes` should be added, except that for
-`scheduling_policy` field (seems like `status` is a better name for it): it is enough
-to have at least in the beginning only 3 fields: 1) `active` 2) `offline` 3)
-`decomissioned`.
-
-`timelines` table:
-```
-table! {
-    // timeline_id is primary key
-    timelines (tenant_id, timeline_id) {
-        timeline_id -> Varchar,
-        tenant_id -> Varchar,
-        generation -> Int4,
-        sk_set -> Array<Int4>, // list of safekeeper ids
-        new_sk_set -> Nullable<Array<Int4>>, // list of safekeeper ids, null if not joint conf
-        cplane_notified_generation -> Int4,
-    }
-}
-```
-
-#### API
-
-Node management is similar to pageserver:
-1) POST `/control/v1/safekeepers` upserts safekeeper.
-2) GET `/control/v1/safekeepers` lists safekeepers.
-3) GET `/control/v1/safekeepers/:node_id` gets safekeeper.
-4) PUT `/control/v1/safekepers/:node_id/status` changes status to e.g.
-   `offline` or `decomissioned`. Initially it is simpler not to schedule any
-    migrations here.
-
-Safekeeper deploy scripts should register safekeeper at storage_contorller as
-they currently do with cplane, under the same id.
-
-Timeline creation/deletion: already existing POST `tenant/:tenant_id/timeline`
-would 1) choose initial set of safekeepers; 2) write to the db initial
-`Configuration` with `INSERT ON CONFLICT DO NOTHING` returning existing row in
-case of conflict; 3) create timeline on the majority of safekeepers (already
-created is ok).
-
-We don't want to block timeline creation when one safekeeper is down. Currently
-this is solved by compute implicitly creating timeline on any safekeeper it is
-connected to. This creates ugly timeline state on safekeeper when timeline is
-created, but start LSN is not defined yet. It would be nice to remove this; to
-do that, controller can in the background retry to create timeline on
-safekeeper(s) which missed that during initial creation call. It can do that
-through `pull_timeline` from majority so it doesn't need to remember
-`parent_lsn` in its db.
-
-Timeline deletion removes the row from the db and forwards deletion to the
-current configuration members. Without additional actions deletions might leak,
-see below on this; initially let's ignore these, reporting to cplane success if
-at least one safekeeper deleted the timeline (this will remove s3 data).
-
-Tenant deletion repeats timeline deletion for all timelines.
-
-Migration API: the first version is the simplest and the most imperative:
-1) PUT `/control/v1/safekeepers/migrate` schedules `MigrationRequest`s to move
-all timelines from one safekeeper to another. It accepts json
-```
-{
-    "src_sk": u32,
-    "dst_sk": u32,
-    "limit": Optional<u32>,
-}
-```
-
-Returns list of scheduled requests.
-
-2) PUT `/control/v1/tenant/:tenant_id/timeline/:timeline_id/safekeeper_migrate` schedules `MigrationRequest`
-   to move single timeline to given set of safekeepers:
-```
-{
-    "desired_set": Vec<u32>,
-}
-```
-
-Returns scheduled request.
-
-Similar call should be added for the tenant.
-
-It would be great to have some way of subscribing to the results (apart from
-looking at logs/metrics).
-
-Migration is executed as described above. One subtlety is that (local) deletion on
-source safekeeper might fail, which is not a problem if we are going to
-decomission the node but leaves garbage otherwise. I'd propose in the first version
-1) Don't attempt deletion at all if node status is `offline`.
-2) If it failed, just issue warning.
-And add PUT `/control/v1/safekeepers/:node_id/scrub` endpoint which would find and 
-remove garbage timelines for manual use. It will 1) list all timelines on the 
-safekeeper 2) compare each one against configuration storage: if timeline 
-doesn't exist at all (had been deleted), it can be deleted. Otherwise, it can 
-be deleted under generation number if node is not member of current generation.
-
-Automating this is untrivial; we'd need to register all potential missing
-deletions <tenant_id, timeline_id, generation, node_id> in the same transaction
-which switches configurations. Similarly when timeline is fully deleted to
-prevent cplane operation from blocking when some safekeeper is not available
-deletion should be also registered.
-
-One more task pool should infinitely retry notifying control plane about changed
-safekeeper sets.
-
-3) GET `/control/v1/tenant/:tenant_id/timeline/:timeline_id/` should return
-   current in memory state of the timeline and pending `MigrationRequest`,
-   if any.
-
-4) PUT `/control/v1/tenant/:tenant_id/timeline/:timeline_id/safekeeper_migrate_abort` tries to abort the
-   migration by switching configuration from the joint to the one with (previous) `sk_set` under CAS
-   (incrementing generation as always).
-
-#### Dealing with multiple instances of storage_controller
-
-Operations described above executed concurrently might create some errors but do
-not prevent progress, so while we normally don't want to run multiple instances
-of storage_controller it is fine to have it temporarily, e.g. during redeploy.
-
-Any interactions with db update in-memory controller state, e.g. if migration
-request failed because different one is in progress, controller remembers that
-and tries to finish it.
-
-## Testing
-
-`neon_local` should be switched to use storage_controller, playing role of
-control plane.
-
-There should be following layers of tests:
-1) Model checked TLA+ spec specifies the algorithm and verifies its basic safety.
-
-2) To cover real code and at the same time test many schedules we should have
-   simulation tests. For that, configuration storage, storage_controller <->
-   safekeeper communication and pull_timeline need to be mocked and main switch
-   procedure wrapped to as a node (thread) in simulation tests, using these
-   mocks. Test would inject migrations like it currently injects
-   safekeeper/walproposer restars. Main assert is the same -- committed WAL must
-   not be lost.
-
-3) Since simulation testing injects at relatively high level points (not
-   syscalls), it omits some code, in particular `pull_timeline`. Thus it is
-   better to have basic tests covering whole system as well. Extended version of
-   `test_restarts_under_load` would do: start background load and do migration 
-   under it, then restart endpoint and check that no reported commits 
-   had been lost. I'd also add one more creating classic network split scenario, with
-   one compute talking to AC and another to BD while migration from nodes ABC to ABD
-   happens.
-
-4) Simple e2e test should ensure that full flow including cplane notification works.
-
-## Order of implementation and rollout
-
-Note that 
-- Control plane parts and integration with it is fully independent from everything else
-  (tests would use simulation and neon_local).
-- There is a lot of infra work making storage_controller aware of timelines and safekeepers
-  and its impl/rollout should be separate from migration itself.
-- Initially walproposer can just stop working while it observers joint configuration.
-  Such window would be typically very short anyway.
-
-To rollout smoothly, both walproposer and safekeeper should have flag
-`configurations_enabled`; when set to false, they would work as currently, i.e.
-walproposer is able to commit on whatever safekeeper set it is provided. Until
-all timelines are managed by storcon we'd need to use current script to migrate
-and update/drop entries in the storage_controller database if it has any.
-
-Safekeepers would need to be able to talk both current and new protocol version
-with compute to reduce number of computes restarted in prod once v2 protocol is
-deployed (though before completely switching we'd need to force this).
-
-Let's have the following rollout order:
-- storage_controller becomes aware of safekeepers;
-- storage_controller gets timeline creation for new timelines and deletion requests, but
-  doesn't manage all timelines yet. Migration can be tested on these new timelines.
-  To keep control plane and storage_controller databases in sync while control 
-  plane still chooses the safekeepers initially (until all timelines are imported
-  it can choose better), `TimelineCreateRequest` can get optional safekeepers
-  field with safekeepers chosen by cplane.
-- Then we can import all existing timelines from control plane to
-  storage_controller and gradually enable configurations region by region.
-
-
-Very rough implementation order:
-- Add concept of configurations to safekeepers (including control file),
-  implement v3 protocol.
-- Implement walproposer changes, including protocol.
-- Implement storconn part. Use it in neon_local (and pytest).
-- Make cplane store safekeepers per timeline instead of per tenant.
-- Implement cplane/storcon integration. Route branch creation/deletion 
-  through storcon. Then we can test migration of new branches.
-- Finally import existing branches. Then we can drop cplane 
-  safekeeper selection code. Gradually enable configurations at 
-  computes and safekeepers. Before that, all computes must talk only
-  v3 protocol version.
-
-## Integration with evicted timelines
-
-Currently, `pull_timeline` doesn't work correctly with evicted timelines because
-copy would point to original partial file. To fix let's just do s3 copy of the
-file. It is a bit stupid as generally unnecessary work, but it makes sense to
-implement proper migration before doing smarter timeline archival. [Issue](https://github.com/neondatabase/neon/issues/8542)
-
-## Possible optimizations
-
-Steps above suggest walproposer restart (with re-election) and thus reconnection
-to safekeepers. Since by bumping term on new majority we ensure that leader
-terms are unique even across generation switches it is possible to preserve
-connections. However, it is more complicated, reconnection is very fast and it
-is much more important to avoid compute restart than millisecond order of write
-stall.
-
-Multiple joint consensus: algorithm above rejects attempt to change membership
-while another attempt is in progress. It is possible to overlay them and AFAIK
-Aurora does this but similarly I don't think this is needed.
-
-## Misc
-
-We should use Compute <-> safekeeper protocol change to include other (long
-yearned) modifications:
-- send data in network order to make arm work.
-- remove term_start_lsn from AppendRequest
-- add horizon to TermHistory
-- add to ProposerGreeting number of connection from this wp to sk

docs/rfcs/035-timeline-archive.md

@@ -1,507 +0,0 @@
-# Timeline Archival
-
-## Summary
-
-This RFC describes a mechanism for pageservers to eliminate local storage + compute work
-for timelines which are not in use, in response to external API calls to "archive" a timeline.
-
-The archived state roughly corresponds to fully offloading a timeline to object storage, such
-that its cost is purely the cost of that object storage.
-
-## Motivation
-
-Archived timelines serve multiple purposes:
-- Act as a 'snapshot' for workloads that would like to retain restorable copies of their
-  database from longer ago than their PITR window.
-- Enable users to create huge numbers of branches (e.g. one per github PR) without having
-  to diligently clean them up later to avoid overloading the pageserver (currently we support
-  up to ~500 branches per tenant).
-
-### Prior art
-
-Most storage and database systems have some form of snapshot, which can be implemented several ways:
-1. full copies of data (e.g. an EBS snapshot to S3)
-2. shallow snapshots which are CoW relative to the original version of the data, e.g. on a typical NFS appliance, or a filesystem like CephFS.
-3. a series of snapshots which are CoW or de-duplicated relative to one another.
-
-Today's Neon branches are approximately like `2.`, although due to implementation details branches
-often end up storing much more data than they really need, as parent branches assume that all data
-at the branch point is needed.  The layers pinned in the parent branch may have a much larger size
-than the physical size of a compressed image layer representing the data at the branch point.
-
-## Requirements
-
-- Enter & exit the archived state in response to external admin API calls
-- API calls to modify the archived state are atomic and durable
-- An archived timeline should eventually (once out of PITR window) use an efficient compressed
-  representation, and avoid retaining arbitrarily large data in its parent branch.
-- Remote object GETs during tenant start may be O(N) with the number of _active_ branches,
-  but must not scale with the number of _archived_ branches.
-- Background I/O for archived branches should only be done a limited number of times to evolve them
-  to a long-term-efficient state (e.g. rewriting to image layers).  There should be no ongoing "housekeeping"
-  overhead for archived branches, including operations related to calculating sizes for billing.
-- The pageserver should put no load on the safekeeper for archived branches.
-- Performance of un-archiving a branch must make good use of S3/disk bandwidth to restore the branch
-  to a performant state in a short time (linear with the branch's logical size)
-
-## Non Goals
-
-- Archived branches are not a literal `fullbackup` postgres snapshot: they are still stored
-  in Neon's internal format.
-- Compute cold starts after activating an archived branch will not have comparable performance to
-  cold starts on an active branch.
-- Archived branches will not use any new/additional compression or de-duplication beyond what
-  is already implemented for image layers (zstd per page).
-- The pageserver will not "auto start" archived branches in response to page_service API requests: they
-  are only activated explicitly via the HTTP API.
-- We will not implement a total offload of archived timelines from safekeepers: their control file (small) will
-  remain on local disk, although existing eviction mechanisms will remove any segments from local disk.
-- We will not expose any prometheus metrics for archived timelines, or make them visible in any
-  detailed HTTP APIs other than the specific API for listing archived timelines.
-- A parent branch may not be archived unless all its children are.
-
-## Impacted Components
-
-pageserver, storage controller
-
-## Terminology
-
-**Archived**: a branch is _archived_ when an HTTP API request to archive it has succeeded: the caller
-may assume that this branch is now very cheap to store, although this may not be physically so until the
-branch proceeds to the offloaded state.
-
-**Active** branches are branches which are available for use by page_service clients, and have a relatively
-high cost due to consuming local storage.
-
-**Offloaded** branches are a subset of _archived_ branches, which have had their local state removed such
-that they now consume minimal runtime resources and have a cost similar to the cost of object storage.
-
-**Activate** (verb): transition from Archived to Active
-
-**Archive** (verb): transition from Active to Archived
-
-**Offload** (verb): transition from Archived to Offloaded
-
-**Offload manifest**: an object stored in S3 that describes timelines which pageservers do not load.
-
-**Warm up** (verb): operation done on an active branch, by downloading its active layers.  Once a branch is
-warmed up, good performance will be available to page_service clients.
-
-## Implementation
-
-### High level flow
-
-We may think of a timeline which is archived and then activated as proceeding through a series of states:
-
-```mermaid
-stateDiagram
-  [*] --> Active(warm)
-  Active(warm) --> Archived
-  Archived --> Offloaded
-  Archived --> Active(warm)
-  Offloaded --> Active(cold)
-  Active(cold) --> Active(warm)
-```
-
-Note that the transition from Archived to Active(warm) is expected to be fairly rare: the most common lifecycles
-of branches will be:
-- Very frequent: Short lived branches: Active -> Deleted
-- Frequent: Long-lived branches: Active -> Archived -> Offloaded -> Deleted
-- Rare: Branches used to restore old state: Active ->Archived -> Offloaded -> Active
-
-These states are _not_ all stored as a single physical state on the timeline, but rather represent the combination
-of:
-- the timeline's lifecycle state: active or archived, stored in the timeline's index
-- its offload state: whether pageserver has chosen to drop local storage of the timeline and write it into the
-  manifest of offloaded timelines.
-- cache state (whether it's warm or cold).
-
-### Storage format changes
-
-There are two storage format changes:
-1. `index_part.json` gets a new attribute `state` that describes whether the timeline is to
-   be considered active or archived.
-2. A new tenant-level _manifest_ object `tenant_manifest-v1.json` describes which timelines a tenant does not need to load
-   at startup (and is available for storing other small, rarely changing tenant-wide attributes in future)
-
-The manifest object will have a format like this:
-```
-{
-  "offload_timelines": [
-    {
-      "timeline_id": ...
-      "last_record_lsn": ...
-      "last_record_lsn_time": ...
-      "pitr_interval": ...
-      "last_gc_lsn": ...  # equal to last_record_lsn if this branch has no history (i.e. a snapshot)
-      "logical_size": ...  # The size at last_record_lsn
-      "physical_size" ...
-      "parent": Option<{
-        "timeline_id"...
-        "lsn"... # Branch point LSN on the parent
-        "requires_data": bool # True if this branch depends on layers in its parent, identify it here
-
-      }>
-    }
-  ]
-}
-```
-
-The information about a timeline in its offload state is intentionally minimal: just enough to decide:
-- Whether it requires [archive optimization](#archive-branch-optimization) by rewriting as a set of image layers: we may infer this
-  by checking if now > last_record_lsn_time - pitr_interval, and pitr_lsn < last_record_lsn.
-- Whether a parent branch should include this offloaded branch in its GC inputs to avoid removing
-  layers that the archived branch depends on
-- Whether requests to delete this `timeline_id` should be executed (i.e. if a deletion request
-  is received for a timeline_id that isn't in the site of live `Timelines` or in the manifest, then
-  we don't need to go to S3 for the deletion.
-- How much archived space to report in consumption metrics
-
-The contents of the manifest's offload list will also be stored as an attribute of `Tenant`, such that the total
-set of timelines may be found by the union of `Tenant::timelines` (non-offloaded timelines) and `Tenant::offloaded`
-(offloaded timelines).
-
-For split-brain protection, the manifest object will be written with a generation suffix, in the same way as
-index_part objects are (see [generation numbers RFC](025-generation-numbers.md)).  This will add some complexity, but
-give us total safety against two pageservers with the same tenant attached fighting over the object.  Existing code
-for finding the latest generation and for cleaning up old generations (in the scrubber) will be generalized to cover
-the manifest file.
-
-### API & Timeline state
-
-Timelines will store a lifecycle state (enum of Active or Archived) in their IndexPart.  This will
-be controlled by a new per-timeline `configure` endpoint.  This is intentionally generic naming, which
-may be used in future to control other per-timeline attributes (e.g. in future we may make PITR interval
-a per-timeline configuration).
-
-`PUT /v1/tenants/{tenant_id}/timelines/{timeline_id}/configure`
-```
-{
-  'state': 'active|archive'
-}
-```
-
-When archiving a timeline, this API will complete as soon as the timeline's state has been set in index_part, and that index has been uploaded.
-
-When activating a timeline, this API will complete as soon as the timeline's state has been set in index_part,
-**and** the `Timeline` object has been instantiated and activated.  This will require reading the timeline's
-index, but not any data: it should be about as fast as a couple of small S3 requests.
-
-The API will be available with identical path via the storage controller: calling this on a sharded tenant
-will simply map the API call to all the shards.
-
-Archived timelines may never have descendent timelines which are active.  This will be enforced at the API level,
-such that activating a timeline requires that all its ancestors are active, and archiving a timeline requires
-that all its descendents are archived.  It is the callers responsibility to walk the hierarchy of timelines
-in the proper order if they would like to archive whole trees of branches.
-
-Because archive timelines will be excluded from the usual timeline listing APIs, a new API specifically
-for archived timelines will be added: this is for use in support/debug:
-
-```
-GET /v1/tenants/{tenant_id}/archived_timelines
-
-{
-  ...same per-timeline content as the tenant manifest...
-}
-
-```
-
-### Tenant attach changes
-
-Currently, during Tenant::spawn we list all the timelines in the S3 bucket, and then for each timeline
-we load their index_part.json.  To avoid the number of GETs scaling linearly with the number of archived
-timelines, we must have a single object that tells us which timelines do not need to be loaded.  The
-number of ListObjects requests while listing timelines will still scale O(N), but this is less problematic
-because each request covers 1000 timelines.
-
-This is **not** literally the same as the set of timelines who have state=archived.  Rather, it is
-the set of timelines which have been offloaded in the background after their state was set to archived.
-
-We may simply skip loading these timelines: there will be no special state of `Timeline`, they just won't
-exist from the perspective of an active `Tenant` apart from in deletion: timeline deletion will need
-to check for offloaded timelines as well as active timelines, to avoid wrongly returning 404 on trying
-to delete an offloaded timeline.
-
-### Warm-up API
-
-`PUT /v1/tenants/{tenant_id}/timelines/{timeline_id}/download?wait_ms=1234`
-
-This API will be similar to the existing `download_remote_layers` API, but smarter:
-- It will not download _all_ remote layers, just the visible set (i.e. layers needed for a read)
-- It will download layers in the visible set until reaching `wait_ms`, then return a struct describing progress
-  of downloads, so that the caller can poll.
-
-The _visible set_ mentioned above will be calculated by the pageserver in the background, by taking the set
-of readable LSNs (i.e. branch points and heads of branches), and walking the layer map to work out which layers
-can possibly be read from these LSNs.  This concept of layer visibility is more generally useful for cache
-eviction and heatmaps, as well as in this specific case of warming up a timeline.
-
-The caller does not have to wait for the warm up API, or call it at all.  But it is strongly advised
-to call it, because otherwise populating local contents for a timeline can take a long time when waiting
-for SQL queries to coincidentally hit all the layers, and during that time query latency remains quite
-volatile.
-
-### Background work
-
-Archived branches are not subject to normal compaction.  Instead, when the compaction loop encounters
-an archived branch, it will consider rewriting the branch to just image layers if the branch has no history
-([archive branch optimization](#archive-branch-optimization)), or offloading the timeline from local disk
-if its state permits that.
-
-Additionally, the tenant compaction task will walk the state of already offloaded timelines to consider
-optimizing their storage, e.g. if a timeline had some history when offloaded, but since then its PITR
-has elapsed and it can now be rewritten to image layers.
-
-#### Archive branch offload
-
-Recall that when we archive a timeline via the HTTP API, this only sets a state: it doesn't do
-any actual work.
-
-This work is done in the background compaction loop.  It makes sense to tag this work on to the compaction
-loop, because it is spiritually aligned: offloading data for archived branches improves storage efficiency.
-
-The condition for offload is simple:
- - a `Timeline` object exists with state `Archived`
- - the timeline does not have any non-offloaded children.
- 
- Regarding the condition that children must be offloaded, this will always be eventually true, because
- we enforce at the API level that children of archived timelines must themselves be archived, and all
- archived timelines will eventually be offloaded.
-
-Offloading a timeline is simple:
-- Read the timeline's attributes that we will store in its offloaded state (especially its logical size)
-- Call `shutdown()` on the timeline and remove it from the `Tenant` (as if we were about to delete it)
-- Erase all the timeline's content from local storage (`remove_dir_all` on its path)
-- Write the tenant manifest to S3 to prevent this timeline being loaded on next start.
-
-#### Archive branch optimization (flattening)
-
-When we offloaded a branch, it might have had some history that prevented rewriting it to a single
-point in time set of image layers.  For example, a branch might have several days of writes and a 7
-day PITR: when we archive it, it still has those days of history.
-
-Once the PITR has expired, we have an opportunity to reduce the physical footprint of the branch by:
-- Writing compressed image layers within the archived branch, as these are more efficient as a way of storing
-  a point in time compared with delta layers
-- Updating the branch's offload metadata to indicate that this branch no longer depends on its ancestor
-  for data, i.e. the ancestor is free to GC layers files at+below the branch point
-
-Fully compacting an archived branch into image layers at a single LSN may be thought of as *flattening* the
-branch, such that it is now a one-dimensional keyspace rather than a two-dimensional key/lsn space. It becomes
-a true snapshot at that LSN.
-
-It is not always more efficient to flatten a branch than to keep some extra history on the parent: this
-is described in more detail in [optimizations](#delaying-storage-optimization-if-retaining-parent-layers-is-cheaper)
-
-Archive branch optimization should be done _before_ background offloads during compaction, because there may
-be timelines which are ready to be offloaded but also would benefit from the optimization step before
-being offloaded.  For example, a branch which has already fallen out of PITR window and has no history
-of its own may be immediately re-written as a series of image layers before being offloaded.
-
-### Consumption metrics
-
-Archived timelines and offloaded timelines will be excluded from the synthetic size calculation, in anticipating
-that billing structures based on consumption metrics are highly likely to apply different $/GB rates to archived
-vs. ordinary content.
-
-Archived and offloaded timelines' logical size will be reported under the existing `timeline_logical_size`
-variant of `MetricsKey`: receivers are then free to bill on this metric as they please.
-
-### Secondary locations
-
-Archived timelines (including offloaded timelines) will be excluded from heatmaps, and thereby
-when a timeline is archived, after the next cycle of heatmap upload & secondary download, its contents
-will be dropped from secondary locations.
-
-### Sharding
-
-Archiving or activating a timeline will be done symmetrically across all shards in a tenant, in
-the same way that timeline creation and deletion is done.  There are no special rules about ordering:
-the storage controller may dispatch concurrent calls to all shards when archiving or activating a timeline.
-
-Since consumption metrics are only transmitted from shard zero, the state of archival on this shard
-will be authoritative for consumption metrics.
-
-## Error cases
-
-### Errors in sharded tenants
-
-If one shard in a tenant fails an operation but others succeed, the tenant may end up in a mixed
-state, where a timeline is archived on some shards but not on others.  
-
-We will not bother implementing a rollback mechanism for this: errors in archiving/activating a timeline
-are either transient (e.g. S3 unavailable, shutting down), or the fault of the caller (NotFound, BadRequest).
-In the transient case callers are expected to retry until success, or to make appropriate API calls to clear
-up their mistake.  We rely on this good behavior of callers to eventually get timelines into a consistent
-state across all shards.  If callers do leave a timeline in an inconsistent state across shards, this doesn't
-break anything, it's just "weird".
-
-This is similar to the status quo for timeline creation and deletion: callers are expected to retry
-these operations until they succeed.
-
-### Archiving/activating
-
-Archiving/activating a timeline can fail in a limited number of ways:
-1. I/O error storing/reading the timeline's updated index
-    - These errors are always retryable: a fundamental design assumption of the pageserver is that remote
-      storage errors are always transient. 
-2. NotFound if the timeline doesn't exist
-    - Callers of the API are expected to avoid calling deletion and archival APIs concurrently.
-    - The storage controller has runtime locking to prevent races such as deleting a timeline while
-      archiving it.
-3. BadRequest if the rules around ancestors/descendents of archived timelines would be violated
-    - Callers are expected to do their own checks to avoid hitting this case.  If they make
-      a mistake and encounter this error, they should give up.
-
-### Offloading
-
-Offloading can only fail if remote storage is unavailable, which would prevent us from writing the
-tenant manifest.  In such error cases, we give up in the expectation that offloading will be tried 
-again at the next iteration of the compaction loop.
-
-### Archive branch optimization
-
-Optimization is a special form of compaction, so can encounter all the same errors as regular compaction
-can: it should return Result<(), CompactionError>, and as with compaction it will be retried on
-the next iteration of the compaction loop.
-
-## Optimizations
-
-### Delaying storage optimization if retaining parent layers is cheaper
-
-Optimizing archived branches to image layers and thereby enabling parent branch GC to progress
-is a safe default: archived branches cannot over-fill a pageserver's local disk, and once they
-are offloaded to S3 they're totally safe, inert things.
-
-However, in some cases it can be advantageous to retain extra history on their parent branch rather
-than flattening the archived branch.  For example, if a 1TB parent branch is rather slow-changing (1GB
-of data per day), and archive branches are being created nightly, then writing out full 1TB image layers
-for each nightly branch is inefficient compared with just keeping more history on the main branch.
-
-Getting this right requires consideration of:
-- Compaction: if keeping more history on the main branch is going to prompt the main branch's compaction to
-  write out extra image layers, then it might make more sense to just write out the image layers on
-  the archived branch.
-- Metadata bloat: keeping extra history on a parent branch doesn't just cost GB of storage, it makes
-  the layer map (and index_part) bigger.  There are practical limits beyond which writing an indefinitely
-  large layer map can cause problems elsewhere.
-
-This optimization can probably be implemented quite cheaply with some basic heuristics like:
-- don't bother doing optimization on an archive branch if the LSN distance between
-  its branch point and the end of the PITR window is <5% of the logical size of the archive branch.
-- ...but, Don't keep more history on the main branch than double the PITR
-
-### Creating a timeline in archived state (a snapshot)
-
-Sometimes, one might want to create a branch with no history, which will not be written to
-before it is archived.  This is a snapshot, although we do not require a special snapshot API,
-since a snapshot can be represented as a timeline with no history.
-
-This can be accomplished by simply creating a timeline and then immediately archiving it, but
-that is somewhat wasteful: this timeline it will spin up various tasks and open a connection to the storage
-broker to try and ingest WAL, before being shutdown in the subsequent archival call.  To explicitly
-support this common special case, we may add a parameter to the timeline creation API which
-creates a timeline directly into the archived state.
-
-Such a timeline creation will do exactly two I/Os at creation time:
-- write the index_part object to record the timeline's existence
-- when the timeline is offloaded in the next iteration of the compaction loop (~20s later),
-  write the tenant manifest.
-
-Later, when the timeline falls off the end of the PITR interval, the usual offload logic will wake
-up the 'snapshot' branch and write out image layers.
-
-## Future Work
-
-### Enabling `fullbackup` dumps from archive branches
-
-It would be useful to be able to export an archive branch to another system, or for use in a local
-postgres database.
-
-This could be implemented as a general capability for all branches, in which case it would "just work"
-for archive branches by activating them.  However, downloading all the layers in a branch just to generate
-a fullbackup is a bit inefficient: we could implement a special case for flattened archived branches
-which streams image layers from S3 and outputs the fullbackup stream without writing the layers out to disk.
-
-Implementing `fullbackup` is a bit more complicated than this because of sharding, but solving that problem
-is unrelated to the topic of archived branches (it probably involves having each shard write out a fullbackup 
-stream to S3 in an intermediate format and, then having one node stitch them together).
-
-### Tagging layers from archived branches
-
-When we know a layer is an image layer written for an archived branch that has fallen off the PITR window,
-we may add tags to the S3 objects to enable writing lifecycle policies that transition such layers to even
-cheaper storage.
-
-This could be done for all archived layers, or it could be driven by the archival API, to give the pageserver
-external hints on which branches are likely to be reactivated, and which branches are good candidates for
-tagging for low performance storage.
-
-Tagging+lifecycles is just one mechanism: one might also directly use S3 storage classes.  Other clouds' object
-stores have similar mechanisms.
-
-### Storing sequences of archive branches as deltas
-
-When archived branches are used as scheduled snapshots, we could store them even more efficiently
-by encoding them as deltas relative to each other (i.e. for nightly snapshots, when we do the
-storage optimization for Tuesday's snapshot, we would read Monday's snapshot and store only the modified
-pages). This is the kind of encoding that many backup storage systems use.
-
-The utility of this depends a lot on the churn rate of the data, and the cost of doing the delta encoding
-vs. just writing out a simple stream of the entire database.  For smaller databases, writing out a full
-copy is pretty trivial (e.g. writing a compressed copy of a 10GiB database to S3 can take under 10 seconds,
-so the complexity tradeoff of diff-encoding it is dubious).
-
-One does not necessarily have to read-back the previous snapshot in order to encoded the next one: if the
-pageserver knows about the schedule, it can intentionally retain extra history on the main branch so that
-we can say: "A branch exists from Monday night.  I have Monday night's data still active in the main branch,
-so now I can read at the Monday LSN and the Tuesday LSN, calculate the delta, and store it as Tuesday's
-delta snapshot".
-
-Clearly this all requires careful housekeeping to retain the relationship between branches that depend on
-each other: perhaps this would be done by making the archive branches have child/parent relationships with
-each other, or perhaps we would permit them to remain children of their original parent, but additionally
-have a relationship with the snapshot they're encoded relative to.
-
-Activating a branch that is diff-encoded may require activating several earlier branches too, so figuring
-out how frequently to write a full copy is important.  This is essentially a zoomed-out version of what
-we do with delta layers and image layers within a timeline, except each "layer" is a whole timeline.
-
-
-## FAQ/Alternatives
-
-### Store all timelines in the tenant manifest
-
-Rather than special-casing offloaded timelines in the offload manifest, we could store a total
-manifest of all timelines, eliminating the need for the pageserver to list timelines in S3 on
-startup.
-
-That would be a more invasive change (require hooking in to timeline creation), and would
-generate much more I/O to this manifest for tenants that had many branches _and_ frequent
-create/delete cycles for short lived branches.  Restricting the manifest to offloaded timelines
-means that we only have to cope with the rate at which long-lived timelines are archived, rather
-than the rate at which sort lived timelines are created & destroyed.
-
-### Automatically archiving/activating timelines without external API calls
-
-We could implement TTL driven offload of timelines, waking them up when a page request
-arrives.
-
-This has downsides:
-- Opacity: if we do TTL-driven offload inside the pageserver, then the end user doesn't
-  know which of their branches are in this state, and might get a surprise when they try
-  to use such a branch.
-- Price fluctuation: if the archival of a branch is used in end user pricing, then users
-  prefer clarity & consistency.  Ideally a branch's storage should cost the same from the moment it
-  is created, rather than having a usage-dependency storage price.
-- Complexity: enabling the page service to call up into the Tenant to activate a timeline
-  would be awkward, compared with an external entry point.
-
-### Make offloaded a state of Timeline
-
-To reduce the operator-facing complexity of having some timelines APIs that only return
-non-offloaded timelines, we could build the offloaded state into the Timeline type.
-
-`timeline.rs` is already one of the most egregiously long source files in the tree, so
-this is rejected on the basis that we need to avoid making that complexity worse.

docs/rfcs/036-physical-replication.md

@@ -1,265 +0,0 @@
-# Physical Replication
-
-This RFC is a bit special in that we have already implemented physical
-replication a long time ago. However, we never properly wrote down all
-the decisions and assumptions, and in the last months when more users
-have started to use the feature, numerous issues have surfaced.
-
-This RFC documents the design decisions that have been made.
-
-## Summary
-
-PostgreSQL has a feature called streaming replication, where a replica
-streams WAL from the primary and continuously applies it. It is also
-known as "physical replication", to distinguish it from logical
-replication.  In PostgreSQL, a replica is initialized by taking a
-physical backup of the primary. In Neon, the replica is initialized
-from a slim "base backup" from the pageserver, just like a primary,
-and the primary and the replicas connect to the same pageserver,
-sharing the storage.
-
-There are two kinds of read-only replicas in Neon:
-- replicas that follow the primary, and
-- "static" replicas that are pinned at a particular LSN.
-
-A static replica is useful e.g. for performing time-travel queries and
-running one-off slow queries without affecting the primary. A replica
-that follows the primary can be used e.g. to scale out read-only
-workloads.
-
-## Motivation
-
-Read-only replicas allow offloading read-only queries. It's useful for
-isolation, if you want to make sure that read-only queries don't
-affect the primary, and it's also an easy way to provide guaranteed
-read-only access to an application, without having to mess with access
-controls.
-
-## Non Goals (if relevant)
-
-This RFC is all about WAL-based *physical* replication. Logical
-replication is a different feature.
-
-Neon also has the capability to launch "static" read-only nodes which
-do not follow the primary, but are pinned to a particular LSN. They
-can be used for long-running one-off queries, or for Point-in-time
-queries. They work similarly to read replicas that follow the primary,
-but some things are simpler: there are no concerns about cache
-invalidation when the data changes on the primary, or worrying about
-transactions that are in-progress on the primary.
-
-## Impacted components (e.g. pageserver, safekeeper, console, etc)
-
-- Control plane launches the replica
-- Replica Postgres instance connects to the safekeepers, to stream the WAL
-- The primary does not know about the standby, except for the hot standby feedback
-- The primary and replicas all connect to the same pageservers
-
-
-# Context
-
-Some useful things to know about hot standby and replicas in
-PostgreSQL.
-
-## PostgreSQL startup sequence
-
-"Running" and "start up" terms are little imprecise. PostgreSQL
-replica startup goes through several stages:
-
-1. First, the process is started up, and various initialization steps
-   are performed, like initializing shared memory. If you try to
-   connect to the server in this stage, you get an error: ERROR: the
-   database system is starting up. This stage happens very quickly, no
-
-2. Then the server reads the checpoint record from the WAL and starts
-   the WAL replay starting from the checkpoint. This works differently
-   in Neon: we start the WAL replay at the basebackup LSN, not from a
-   checkpoint! If you connect to the server in this state, you get an
-   error: ERROR: the database system is not yet accepting
-   connections. We proceed to the next stage, when the WAL replay sees
-   a running-xacts record. Or in Neon, the "CLOG scanning" mechanism
-   can allow us to move directly to next stage, with all the caveats
-   listed in this RFC.
-
-3. When the running-xacts information is established, the server
-   starts to accept connections normally.
-
-From PostgreSQL's point of view, the server is already running in
-stage 2, even though it's not accepting connections yet. Our
-`compute_ctl` does not consider it as running until stage 3. If the
-transition from stage 2 to 3 doesn't happen fast enough, the control
-plane will mark the start operation as failed.
-
-
-## Decisions, Issues
-
-### Cache invalidation in replica
-
-When a read replica follows the primary in PostgreSQL, it needs to
-stream all the WAL from the primary and apply all the records, to keep
-the local copy of the data consistent with the primary. In Neon, the
-replica can fetch the updated page versions from the pageserver, so
-it's not necessary to apply all the WAL. However, it needs to ensure
-that any pages that are currently in the Postgres buffer cache, or the
-Local File Cache, are either updated, or thrown away so that the next
-read of the page will fetch the latest version.
-
-We choose to apply the WAL records for pages that are already in the
-buffer cache, and skip records for other pages. Somewhat arbitrarily,
-we also apply records affecting catalog relations, fetching the old
-page version from the pageserver if necessary first. See
-`neon_redo_read_buffer_filter()` function.
-
-The replica wouldn't necessarily need to see all the WAL records, only
-the records that apply to cached pages. For simplicity, we do stream
-all the WAL to the replica, and the replica simply ignores WAL records
-that require no action.
-
-Like in PostgreSQL, the read replica maintains a "replay LSN", which
-is the LSN up to which the replica has received and replayed the
-WAL. The replica can lag behind the primary, if it cannot quite keep
-up with the primary, or if a long-running query conflicts with changes
-that are about to be applied, or even intentionally if the user wishes
-to see delayed data (see recovery_min_apply_delay). It's important
-that the replica sees a consistent view of the whole cluster at the
-replay LSN, when it's lagging behind.
-
-In Neon, the replica connects to a safekeeper to get the WAL
-stream. That means that the safekeepers must be able to regurgitate
-the original WAL as far back as the replay LSN of any running read
-replica. (A static read-only node that does not follow the primary
-does not require a WAL stream however). The primary does not need to
-be running, and when it is, the replicas don't incur any extra
-overhead to the primary (see hot standby feedback though).
-
-### In-progress transactions
-
-In PostgreSQL, when a hot standby server starts up, it cannot
-immediately open up for queries (see [PostgreSQL startup
-sequence]). It first needs to establish a complete list of in-progress
-transactions, including subtransactions, that are running at the
-primary, at the current replay LSN. Normally that happens quickly,
-when the replica sees a "running-xacts" WAL record, because the
-primary writes a running-xacts WAL record at every checkpoint, and in
-PostgreSQL the replica always starts the WAL replay from a checkpoint
-REDO point. (A shutdown checkpoint WAL record also implies that all
-the non-prepared transactions have ended.) If there are a lot of
-subtransactions in progress, however, the standby might need to wait
-for old transactions to complete before it can open up for queries.
-
-In Neon that problem is worse: a replica can start at any LSN, so
-there's no guarantee that it will see a running-xacts record any time
-soon. In particular, if the primary is not running when the replica is
-started, it might never see a running-xacts record.
-
-To make things worse, we initially missed this issue, and always
-started accepting queries at replica startup, even if it didn't have
-the transaction information. That could lead to incorrect query
-results and data corruption later. However, as we fixed that, we
-introduced a new problem compared to what we had before: previously
-the replica would always start up, but after fixing that bug, it might
-not. In a superficial way, the old behavior was better (but could lead
-to serious issues later!). That made fixing that bug was very hard,
-because as we fixed it, we made things (superficially) worse for
-others.
-
-See https://github.com/neondatabase/neon/pull/7288 which fixed the
-bug, and follow-up PRs https://github.com/neondatabase/neon/pull/8323
-and https://github.com/neondatabase/neon/pull/8484 to try to claw back
-the cases that started to cause trouble as fixing it. As of this
-writing, there are still cases where a replica might not immediately
-start up, causing the control plane operation to fail, the remaining
-issues are tracked in https://github.com/neondatabase/neon/issues/6211.
-
-One long-term fix for this is to switch to using so-called CSN
-snapshots in read replica. That would make it unnecessary to have the
-full in-progress transaction list in the replica at startup time. See
-https://commitfest.postgresql.org/48/4912/ for a work-in-progress
-patch to upstream to implement that.
-
-Another thing we could do is to teach the control plane about that
-distinction between "starting up" and "running but haven't received
-running-xacts information yet", so that we could keep the replica
-waiting longer in that stage, and also give any client connections the
-same `ERROR: the database system is not yet accepting connections`
-error that you get in standalone PostgreSQL in that state.
-
-
-### Recovery conflicts and Hot standby feedback
-
-It's possible that a tuple version is vacuumed away in the primary,
-even though it is still needed by a running transactions in the
-replica. This is called a "recovery conflict", and PostgreSQL provides
-various options for dealing with it. By default, the WAL replay will
-wait up to 30 s for the conflicting query to finish. After that, it
-will kill the running query, so that the WAL replay can proceed.
-
-Another way to avoid the situation is to enable the
-[`hot_standby_feedback`](https://www.postgresql.org/docs/current/runtime-config-replication.html#GUC-HOT-STANDBY-FEEDBACK)
-option. When it is enabled, the primary will refrain from vacuuming
-tuples that are still needed in the primary. That means potentially
-bloating the primary, which violates the usual rule that read replicas
-don't affect the operations on the primary, which is why it's off by
-default. We leave it to users to decide if they want to turn it on,
-same as PostgreSQL.
-
-Neon supports `hot_standby_feedback` by passing the feedback messages
-from the replica to the safekeepers, and from safekeepers to the
-primary.
-
-### Relationship of settings between primary and replica
-
-In order to enter hot standby mode, some configuration options need to
-be set to the same or larger values in the standby, compared to the
-primary.  See [explanation in the PostgreSQL
-docs](https://www.postgresql.org/docs/current/hot-standby.html#HOT-STANDBY-ADMIN)
-
-In Neon, we have this problem too. To prevent customers from hitting
-it, the control plane automatically adjusts the settings of a replica,
-so that they match or exceed the primary's settings (see
-https://github.com/neondatabase/cloud/issues/14903). However, you
-can still hit the issue if the primary is restarted with larger
-settings, while the replica is running.
-
-
-### Interaction with Pageserver GC
-
-The read replica can lag behind the primary. If there are recovery
-conflicts or the replica cannot keep up for some reason, the lag can
-in principle grow indefinitely. The replica will issue all GetPage
-requests to the pageservers at the current replay LSN, and needs to
-see the old page versions.
-
-If the retention period in the pageserver is set to be small, it may
-have already garbage collected away the old page versions. That will
-cause read errors in the compute, and can mean that the replica cannot
-make progress with the replication anymore.
-
-There is a mechanism for replica to pass information about its replay
-LSN to the pageserver, so that the pageserver refrains from GC'ing
-data that is still needed by the standby. It's called
-'standby_horizon' in the pageserver code, see
-https://github.com/neondatabase/neon/pull/7368. A separate "lease"
-mechanism also is in the works, where the replica could hold a lease
-on the old LSN, preventing the pageserver from advancing the GC
-horizon past that point. The difference is that the standby_horizon
-mechanism relies on a feedback message from replica to safekeeper,
-while the least API is exposed directly from the pageserver. A static
-read-only node is not connected to safekeepers, so it cannot use the
-standby_horizon mechanism.
-
-
-### Synchronous replication
-
-We haven't put any effort into synchronous replication yet.
-
-PostgreSQL provides multiple levels of synchronicity. In the weaker
-levels, a transaction is not acknowledged as committed to the client
-in the primary until the WAL has been streamed to a replica or flushed
-to disk there. Those modes don't make senses in Neon, because the
-safekeepers handle durability.
-
-`synchronous_commit=remote_apply` mode would make sense. In that mode,
-the commit is not acknowledged to the client until it has been
-replayed in the replica. That ensures that after commit, you can see
-the commit in the replica too (aka. read-your-write consistency).

docs/rfcs/037-storage-controller-restarts.md

@@ -1,259 +0,0 @@
-# Rolling Storage Controller Restarts
-
-## Summary
-
-This RFC describes the issues around the current storage controller restart procedure
-and describes an implementation which reduces downtime to a few milliseconds on the happy path.
-
-## Motivation
-
-Storage controller upgrades (restarts, more generally) can cause multi-second availability gaps.
-While the storage controller does not sit on the main data path, it's generally not acceptable
-to block management requests for extended periods of time (e.g. https://github.com/neondatabase/neon/issues/8034).
-
-### Current Implementation
-
-The storage controller runs in a Kubernetes Deployment configured for one replica and strategy set to [Recreate](https://kubernetes.io/docs/concepts/workloads/controllers/deployment/#recreate-deployment).
-In non Kubernetes terms, during an upgrade, the currently running storage controller is stopped and, only after,
-a new instance is created.
-
-At start-up, the storage controller calls into all the pageservers it manages (retrieved from DB) to learn the
-latest locations of all tenant shards present on them. This is usually fast, but can push into tens of seconds
-under unfavourable circumstances: pageservers are heavily loaded or unavailable.
-
-## Prior Art
-
-There's probably as many ways of handling restarts gracefully as there are distributed systems. Some examples include:
-* Active/Standby architectures: Two or more instance of the same service run, but traffic is only routed to one of them.
-For fail-over, traffic is routed to one of the standbys (which becomes active).
-* Consensus Algorithms (Raft, Paxos and friends): The part of consensus we care about here is leader election: peers communicate to each other
-and use a voting scheme that ensures the existence of a single leader (e.g. Raft epochs).
-
-## Requirements
-
-* Reduce storage controller unavailability during upgrades to milliseconds
-* Minimize the interval in which it's possible for more than one storage controller
-to issue reconciles.
-* Have one uniform implementation for restarts and upgrades
-* Fit in with the current Kubernetes deployment scheme
-
-## Non Goals
-
-* Implement our own consensus algorithm from scratch
-* Completely eliminate downtime storage controller downtime. Instead we aim to reduce it to the point where it looks
-like a transient error to the control plane
-
-## Impacted Components
-
-* storage controller
-* deployment orchestration (i.e. Ansible)
-* helm charts
-
-## Terminology
-
-* Observed State: in-memory mapping between tenant shards and their current pageserver locations - currently built up
-at start-up by quering pageservers
-* Deployment: Kubernetes [primitive](https://kubernetes.io/docs/concepts/workloads/controllers/deployment/) that models
-a set of replicas
-
-## Implementation
-
-### High Level Flow
-
-At a very high level the proposed idea is to start a new storage controller instance while
-the previous one is still running and cut-over to it when it becomes ready. The new instance,
-should coordinate with the existing one and transition responsibility gracefully. While the controller
-has built in safety against split-brain situations (via generation numbers), we'd like to avoid such
-scenarios since they can lead to availability issues for tenants that underwent changes while two controllers
-were operating at the same time and require operator intervention to remedy.
-
-### Kubernetes Deployment Configuration
-
-On the Kubernetes configuration side, the proposal is to update the storage controller `Deployment`
-to use `spec.strategy.type = RollingUpdate`, `spec.strategy.rollingUpdate.maxSurge=1` and `spec.strategy.maxUnavailable=0`.
-Under the hood, Kubernetes creates a new replica set and adds one pod to it (`maxSurge=1`). The old replica set does not
-scale down until the new replica set has one replica in the ready state (`maxUnavailable=0`).
-
-The various possible failure scenarios are investigated in the [Handling Failures](#handling-failures) section.
-
-### Storage Controller Start-Up
-
-This section describes the primitives required on the storage controller side and the flow of the happy path.
-
-#### Database Table For Leader Synchronization
-
-A new table should be added to the storage controller database for leader synchronization during startup.
-This table will always contain at most one row. The proposed name for the table is `leader` and the schema
-contains two elements:
-* `hostname`: represents the hostname for the current storage controller leader - should be addressible
-from other pods in the deployment
-* `start_timestamp`: holds the start timestamp for the current storage controller leader (UTC timezone) - only required
-for failure case handling: see [Previous Leader Crashes Before New Leader Readiness](#previous-leader-crashes-before-new-leader-readiness)
-
-Storage controllers will read the leader row at start-up and then update it to mark themselves as the leader
-at the end of the start-up sequence. We want compare-and-exchange semantics for the update: avoid the
-situation where two concurrent updates succeed and overwrite each other. The default Postgres isolation
-level is `READ COMMITTED`, which isn't strict enough here. This update transaction should use at least `REPEATABLE
-READ` isolation level in order to [prevent lost updates](https://www.interdb.jp/pg/pgsql05/08.html). Currently,
-the storage controller uses the stricter `SERIALIZABLE` isolation level for all transactions. This more than suits
-our needs here.
-
-```
-START TRANSACTION ISOLATION LEVEL REPEATABLE READ
-UPDATE leader SET hostname=<new_hostname>, start_timestamp=<new_start_ts>
-WHERE hostname=<old_hostname>, start_timestampt=<old_start_ts>;
-```
-
-If the transaction fails or if no rows have been updated, then the compare-and-exchange is regarded as a failure.
-
-#### Step Down API
-
-A new HTTP endpoint should be added to the storage controller: `POST /control/v1/step_down`. Upon receiving this
-request the leader cancels any pending reconciles and goes into a mode where it replies with 503 to all other APIs
-and does not issue any location configurations to its pageservers. The successful HTTP response will return a serialized
-snapshot of the observed state.
-
-If other step down requests come in after the initial one, the request is handled and the observed state is returned (required
-for failure scenario handling - see [Handling Failures](#handling-failures)).
-
-#### Graceful Restart Happy Path
-
-At start-up, the first thing the storage controller does is retrieve the sole row from the new
-`leader` table. If such an entry exists, send a `/step_down` PUT API call to the current leader.
-This should be retried a few times with a short backoff (see [1]). The aspiring leader loads the
-observed state into memory and the start-up sequence proceeds as usual, but *without* querying the
-pageservers in order to build up the observed state.
-
-Before doing any reconciliations or persistence change, update the `leader` database table as described in the [Database Table For Leader Synchronization](database-table-for-leader-synchronization)
-section. If this step fails, the storage controller process exits.
-
-Note that no row will exist in the `leaders` table for the first graceful restart. In that case, force update the `leader` table
-(without the WHERE clause) and perform with the pre-existing start-up procedure (i.e. build observed state by querying pageservers).
-
-Summary of proposed new start-up sequence:
-1. Call `/step_down`
-2. Perform any pending database migrations
-3. Load state from database
-4. Load observed state returned in step (1) into memory
-5. Do initial heartbeat round (may be moved after 5)
-7. Mark self as leader by updating the database
-8. Reschedule and reconcile everything
-
-Some things to note from the steps above:
-* The storage controller makes no changes to the cluster state before step (5) (i.e. no location config
-calls to the pageserver and no compute notifications)
-* Ask the current leader to step down before loading state from database so we don't get a lost update
-if the transactions overlap.
-* Before loading the observed state at step (3), cross-validate against the database. If validation fails,
-fall back to asking the pageservers about their current locations.
-* Database migrations should only run **after** the previous instance steps down (or the step down times out).
-
-
-[1] The API call might fail because there's no storage controller running (i.e. [restart](#storage-controller-crash-or-restart)),
-so we don't want to extend the unavailability period by much. We still want to retry since that's not the common case.
-
-### Handling Failures
-
-#### Storage Controller Crash Or Restart
-
-The storage controller may crash or be restarted outside of roll-outs. When a new pod is created, its call to
-`/step_down` will fail since the previous leader is no longer reachable. In this case perform the pre-existing
-start-up procedure and update the leader table (with the WHERE clause). If the update fails, the storage controller
-exists and consistency is maintained.
-
-#### Previous Leader Crashes Before New Leader Readiness
-
-When the previous leader (P1) crashes before the new leader (P2) passses the readiness check, Kubernetes will
-reconcile the old replica set and create a new pod for it (P1'). The `/step_down` API call will fail for P1'
-(see [2]).
-
-Now we have two cases to consider:
-* P2 updates the `leader` table first: The database update from P1' will fail and P1' will exit, or be terminated
-by Kubernetes depending on timings.
-* P1' updates the `leader` table first: The `hostname` field of the `leader` row stays the same, but the `start_timestamp` field changes.
-The database update from P2 will fail (since `start_timestamp` does not match). P2 will exit and Kubernetes will
-create a new replacement pod for it (P2'). Now the entire dance starts again, but with P1' as the leader and P2' as the incumbent.
-
-[2] P1 and P1' may (more likely than not) be the same pod and have the same hostname. The implementation
-should avoid this self reference and fail the API call at the client if the persisted hostname matches
-the current one.
-
-#### Previous Leader Crashes After New Leader Readiness
-
-The deployment's replica sets already satisfy the deployment's replica count requirements and the
-Kubernetes deployment rollout will just clean up the dead pod.
-
-#### New Leader Crashes Before Pasing Readiness Check
-
-The deployment controller scales up the new replica sets by creating a new pod. The entire procedure is repeated
-with the new pod.
-
-#### Network Partition Between New Pod and Previous Leader
-
-This feels very unlikely, but should be considered in any case. P2 (the new aspiring leader) fails the `/step_down`
-API call into P1 (the current leader). P2 proceeds with the pre-existing startup procedure and updates the `leader` table.
-Kubernetes will terminate P1, but there may be a brief period where both storage controller can drive reconciles.
-
-### Dealing With Split Brain Scenarios
-
-As we've seen in the previous section, we can end up with two storage controller running at the same time. The split brain
-duration is not bounded since the Kubernetes controller might become partitioned from the pods (unlikely though). While these
-scenarios are not fatal, they can cause tenant unavailability, so we'd like to reduce the chances of this happening.
-The rest of this section sketches some safety measure. It's likely overkill to implement all of them however.
-
-### Ensure Leadership Before Producing Side Effects
-
-The storage controller has two types of side effects: location config requests into pageservers and compute notifications into the control plane.
-Before issuing either, the storage controller could check that it is indeed still the leader by querying the database. Side effects might still be
-applied if they race with the database updatem, but the situation will eventually be detected. The storage controller process should terminate in these cases.
-
-### Leadership Lease
-
-Up until now, the leadership defined by this RFC is static. In order to bound the length of the split brain scenario, we could require the leadership
-to be renewed periodically. Two new columns would be added to the leaders table:
-1. `last_renewed` - timestamp indicating when the lease was last renewed
-2. `lease_duration` - duration indicating the amount of time after which the lease expires
-
-The leader periodically attempts to renew the lease by checking that it is in fact still the legitimate leader and updating `last_renewed` in the
-same transaction. If the update fails, the process exits. New storage controller instances wishing to become leaders must wait for the current lease
-to expire before acquiring leadership if they have not succesfully received a response to the `/step_down` request.
-
-### Notify Pageserver Of Storage Controller Term
-
-Each time that leadership changes, we can bump a `term` integer column in the `leader` table. This term uniquely identifies a leader.
-Location config requests and re-attach responses can include this term. On the pageserver side, keep the latest term in memory and refuse
-anything which contains a stale term (i.e. smaller than the current one).
-
-### Observability
-
-* The storage controller should expose a metric which describes it's state (`Active | WarmingUp | SteppedDown`).
-Per region alerts should be added on this metric which triggers when:
-  + no storage controller has been in the `Active` state for an extended period of time
-  + more than one storage controllers are in the `Active` state
-
-* An alert that periodically verifies that the `leader` table is in sync with the metric above would be very useful.
-We'd have to expose the storage controller read only database to Grafana (perhaps it is already done).
-
-## Alternatives
-
-### Kubernetes Leases
-
-Kubernetes has a [lease primitive](https://kubernetes.io/docs/concepts/architecture/leases/) which can be used to implement leader election.
-Only one instance may hold a lease at any given time. This lease needs to be periodically renewed and has an expiration period.
-
-In our case, it would work something like this:
-* `/step_down` deletes the lease or stops it from renewing
-* lease acquisition becomes part of the start-up procedure
-
-The kubert crate implements a [lightweight lease API](https://docs.rs/kubert/latest/kubert/lease/struct.LeaseManager.html), but it's still
-not exactly trivial to implement.
-
-This approach has the benefit of baked in observability (`kubectl describe lease`), but:
-* We offload the responsibility to Kubernetes which makes it harder to debug when things go wrong.
-* More code surface than the simple "row in database" approach. Also, most of this code would be in
-a dependency not subject to code review, etc.
-* Hard to test. Our testing infra does not run the storage controller in Kubernetes and changing it do
-so is not simple and complictes and the test set-up.
-
-To my mind, the "row in database" approach is straightforward enough that we don't have to offload this
-to something external.

docs/settings.md

@@ -134,7 +134,7 @@ depends on that, so if you change it, bad things will happen.
 
 #### page_cache_size
 
-Size of the page cache. Unit is
+Size of the page cache, to hold materialized page versions. Unit is
 number of 8 kB blocks. The default is 8192, which means 64 MB.
 
 #### max_file_descriptors