clean up

split out binaries
proxy: experiment with idea to split crates
2026-01-18 02:42:56 +00:00 · 2024-08-13 15:08:57 +01:00 · 2024-08-13 15:08:57 +01:00 · 2024-08-13 15:08:54 +01:00 · 2024-08-13 15:01:48 +01:00 · 2024-08-13 11:08:25 +01:00
410 changed files with 9894 additions and 20585 deletions
--- a/.config/hakari.toml
+++ b/.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
--- a/.github/ISSUE_TEMPLATE/config.yml
+++ b/.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`
--- a/.github/actionlint.yml
+++ b/.github/actionlint.yml
@@ -1,19 +1,13 @@
 self-hosted-runner:
  labels:
    - arm64
+    - gen3
    - large
    - large-arm64
    - small
    - small-arm64
    - us-east-2
 config-variables:
-  - AZURE_DEV_CLIENT_ID
-  - AZURE_DEV_REGISTRY_NAME
-  - AZURE_DEV_SUBSCRIPTION_ID
-  - AZURE_PROD_CLIENT_ID
-  - AZURE_PROD_REGISTRY_NAME
-  - AZURE_PROD_SUBSCRIPTION_ID
-  - AZURE_TENANT_ID
  - BENCHMARK_PROJECT_ID_PUB
  - BENCHMARK_PROJECT_ID_SUB
  - REMOTE_STORAGE_AZURE_CONTAINER
--- a/.github/actions/run-python-test-set/action.yml
+++ b/.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
@@ -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)
@@ -83,6 +83,7 @@ runs:
      uses: actions/checkout@v4
      with:
        submodules: true
+        fetch-depth: 1

    - name: Cache poetry deps
      uses: actions/cache@v4
@@ -169,8 +170,10 @@ 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
@@ -211,13 +214,13 @@ 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() }}
--- a/.github/actions/set-docker-config-dir/action.yml
+++ b/.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
--- a/.github/workflows/_benchmarking_preparation.yml
+++ b/.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;
-        "
--- a/.github/workflows/_build-and-test-locally.yml
+++ b/.github/workflows/_build-and-test-locally.yml
@@ -70,6 +70,7 @@ jobs:
      - uses: actions/checkout@v4
        with:
          submodules: true
+          fetch-depth: 1

      - name: Set pg 14 revision for caching
        id: pg_v14_rev
@@ -94,16 +95,11 @@ jobs:
      # 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
+          if [[ $BUILD_TYPE == "debug" ]]; then
            cov_prefix="scripts/coverage --profraw-prefix=$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"
@@ -163,8 +159,6 @@ jobs:
      # 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/
@@ -179,7 +173,7 @@ jobs:
          done

          # Install test executables and write list of all binaries (for code coverage)
-          if [[ $BUILD_TYPE == "debug" && $ARCH == 'x64' ]]; then
+          if [[ $BUILD_TYPE == "debug" ]]; then
            # Keep bloated coverage data files away from the rest of the artifact
            mkdir -p /tmp/coverage/

@@ -214,16 +208,10 @@ jobs:
          export LD_LIBRARY_PATH

          #nextest does not yet support running doctests
-          ${cov_prefix} cargo test --doc $CARGO_FLAGS $CARGO_FEATURES
+          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 $CARGO_FEATURES  -E 'package(pageserver)'
-            done
+            NEON_PAGESERVER_UNIT_TEST_VIRTUAL_FILE_IOENGINE=$io_engine ${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES
          done

          # Run separate tests for real S3
@@ -256,8 +244,8 @@ jobs:
        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'
+    # Run test on x64 only
+    if: inputs.arch == 'x64'
    needs: [ build-neon ]
    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', inputs.arch == 'arm64' && 'large-arm64' || 'large')) }}
    container:
@@ -275,6 +263,7 @@ jobs:
      - uses: actions/checkout@v4
        with:
          submodules: true
+          fetch-depth: 1

      - name: Pytest regression tests
        uses: ./.github/actions/run-python-test-set
--- a/.github/workflows/_push-to-acr.yml
+++ b/.github/workflows/_push-to-acr.yml
@@ -1,56 +0,0 @@
-name: Push images to ACR
-on:
-  workflow_call:
-    inputs:
-      client_id:
-        description: Client ID of Azure managed identity or Entra app
-        required: true
-        type: string
-      image_tag:
-        description: Tag for the container image
-        required: true
-        type: string
-      images:
-        description: Images to push
-        required: true
-        type: string
-      registry_name:
-        description: Name of the container registry
-        required: true
-        type: string
-      subscription_id:
-        description: Azure subscription ID
-        required: true
-        type: string
-      tenant_id:
-        description: Azure tenant ID
-        required: true
-        type: string
-
-jobs:
-  push-to-acr:
-    runs-on: ubuntu-22.04
-    permissions:
-      contents: read  # This is required for actions/checkout
-      id-token: write # This is required for Azure Login to work.
-
-    steps:
-      - name: Azure login
-        uses: azure/login@6c251865b4e6290e7b78be643ea2d005bc51f69a  # @v2.1.1
-        with:
-          client-id: ${{ inputs.client_id }}
-          subscription-id: ${{ inputs.subscription_id }}
-          tenant-id: ${{ inputs.tenant_id }}
-
-      - name: Login to ACR
-        run: |
-          az acr login --name=${{ inputs.registry_name }}
-
-      - name: Copy docker images to ACR ${{ inputs.registry_name }}
-        run: |
-          images='${{ inputs.images }}'
-          for image in ${images}; do
-            docker buildx imagetools create \
-              -t ${{ inputs.registry_name }}.azurecr.io/neondatabase/${image}:${{ inputs.image_tag }} \
-                                        neondatabase/${image}:${{ inputs.image_tag }}
-          done
--- a/.github/workflows/actionlint.yml
+++ b/.github/workflows/actionlint.yml
@@ -44,7 +44,7 @@ jobs:
            grep -ERl $PAT .github/workflows |\
            while read -r f
            do
-              l=$(grep -nE $PAT $f | awk -F: '{print $1}' | head -1)
+              l=$(grep -nE $PAT .github/workflows/release.yml | awk -F: '{print $1}' | head -1)
              echo "::error file=$f,line=$l::Please use 'ubuntu-22.04' instead of 'ubuntu-latest'"
            done
            exit 1
--- a/.github/workflows/benchmarking.yml
+++ b/.github/workflows/benchmarking.yml
@@ -96,7 +96,7 @@ jobs:
      uses: aws-actions/configure-aws-credentials@v4
      with:
        aws-region: eu-central-1
-        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
+        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}  
        role-duration-seconds: 18000 # 5 hours

    - name: Download Neon artifact
@@ -146,7 +146,6 @@ jobs:
        api_key: ${{ secrets.NEON_STAGING_API_KEY }}

    - name: Create Allure report
-      id: create-allure-report
      if: ${{ !cancelled() }}
      uses: ./.github/actions/allure-report-generate

@@ -155,10 +154,7 @@ jobs:
      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>
+        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 }}

@@ -180,7 +176,7 @@ jobs:
    steps:
    - uses: actions/checkout@v4

-
+    
    - name: Download Neon artifact
      uses: ./.github/actions/download
      with:
@@ -219,23 +215,15 @@ jobs:
        NEON_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 replication testing: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
      env:
        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

@@ -292,9 +280,8 @@ jobs:
                      { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-sharding-reuse", "db_size": "50gb","runner": '"$runner_default"', "image": "'"$image_default"'" }]
        }'

-        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 += [{ "pg_version": 14, "region_id": "'"$region_id_default"'", "platform": "rds-postgres", "db_size": "10gb","runner": '"$runner_default"', "image": "'"$image_default"'" }]')
        fi

        echo "matrix=$(echo "$matrix" | jq --compact-output '.')" >> $GITHUB_OUTPUT
@@ -334,13 +321,9 @@ 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 ]
+    needs: [ generate-matrices ]
    permissions:
      contents: write
      statuses: write
@@ -377,7 +360,7 @@ jobs:
        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:
@@ -472,7 +455,6 @@ jobs:
        api_key: ${{ secrets.NEON_STAGING_API_KEY }}

    - name: Create Allure report
-      id: create-allure-report
      if: ${{ !cancelled() }}
      uses: ./.github/actions/allure-report-generate

@@ -481,10 +463,7 @@ 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>
+        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 }}

@@ -558,7 +537,7 @@ jobs:
        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:
@@ -593,9 +572,8 @@ jobs:
        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

@@ -604,10 +582,7 @@ jobs:
      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 ${PLATFORM}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
      env:
        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

@@ -620,7 +595,7 @@ 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 ]
+    needs: [ generate-matrices, pgbench-compare ]

    strategy:
      fail-fast: false
@@ -628,7 +603,7 @@ 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 }}
@@ -680,7 +655,6 @@ 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 +664,6 @@ jobs:
        TEST_OLAP_SCALE: 10

    - name: Create Allure report
-      id: create-allure-report
      if: ${{ !cancelled() }}
      uses: ./.github/actions/allure-report-generate

@@ -699,10 +672,7 @@ 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 }}

@@ -714,7 +684,7 @@ jobs:
    #
    # *_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 ]
+    needs: [ generate-matrices, clickbench-compare ]

    strategy:
      fail-fast: false
@@ -722,7 +692,7 @@ 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: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -754,7 +724,7 @@ jobs:
            ENV_PLATFORM=RDS_AURORA_TPCH
            ;;
          rds-postgres)
-            ENV_PLATFORM=RDS_POSTGRES_TPCH
+            ENV_PLATFORM=RDS_AURORA_TPCH
            ;;
          *)
            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neonvm-captest-reuse', 'rds-aurora', or 'rds-postgres'"
@@ -780,7 +750,6 @@ 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 +757,6 @@ jobs:
        TEST_OLAP_SCALE: ${{ matrix.scale }}

    - name: Create Allure report
-      id: create-allure-report
      if: ${{ !cancelled() }}
      uses: ./.github/actions/allure-report-generate

@@ -797,16 +765,13 @@ 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 ]
+    needs: [ generate-matrices, tpch-compare ]

    strategy:
      fail-fast: false
@@ -814,7 +779,7 @@ 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' ) }}
@@ -871,7 +836,6 @@ jobs:
        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 +844,6 @@ 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 }}
--- a/.github/workflows/build-build-tools-image.yml
+++ b/.github/workflows/build-build-tools-image.yml
@@ -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", "gen3", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}

    env:
      IMAGE_TAG: ${{ inputs.image-tag }}
@@ -56,7 +56,13 @@ jobs:

      - uses: actions/checkout@v4

-      - uses: ./.github/actions/set-docker-config-dir
+      # 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/setup-buildx-action@v3
        with:
          cache-binary: false
@@ -83,6 +89,11 @@ jobs:
          cache-to: ${{ github.ref_name == 'main' && format('type=registry,ref=cache.neon.build/build-tools:cache-{0},mode=max', matrix.arch) || '' }}
          tags: neondatabase/build-tools:${{ inputs.image-tag }}-${{ matrix.arch }}

+      - name: Remove custom docker config directory
+        if: always()
+        run: |
+          rm -rf /tmp/.docker-custom
+
  merge-images:
    needs: [ build-image ]
    runs-on: ubuntu-22.04
--- a/.github/workflows/build_and_test.yml
+++ b/.github/workflows/build_and_test.yml
@@ -48,7 +48,7 @@ jobs:

  tag:
    needs: [ check-permissions ]
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/base:pinned
    outputs:
      build-tag: ${{steps.build-tag.outputs.tag}}
@@ -90,7 +90,7 @@ jobs:

  check-codestyle-python:
    needs: [ check-permissions, build-build-tools-image ]
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container:
      image: ${{ needs.build-build-tools-image.outputs.image }}
      credentials:
@@ -101,6 +101,9 @@ jobs:
    steps:
      - name: Checkout
        uses: actions/checkout@v4
+        with:
+          submodules: false
+          fetch-depth: 1

      - name: Cache poetry deps
        uses: actions/cache@v4
@@ -139,6 +142,7 @@ jobs:
        uses: actions/checkout@v4
        with:
          submodules: true
+          fetch-depth: 1

 #      Disabled for now
 #      - name: Restore cargo deps cache
@@ -198,9 +202,9 @@ jobs:
    strategy:
      fail-fast: false
      matrix:
-        arch: [ x64, arm64 ]
+        arch: [ x64 ]
        # Do not build or run tests in debug for release branches
-        build-type: ${{ fromJson((startsWith(github.ref_name, 'release') && github.event_name == 'push') && '["release"]' || '["debug", "release"]') }}
+        build-type: ${{ fromJson((startsWith(github.ref_name, 'release' && github.event_name == 'push')) && '["release"]' || '["debug", "release"]') }}
        include:
          - build-type: release
            arch: arm64
@@ -220,7 +224,7 @@ jobs:
    outputs:
      json: ${{ steps.get-benchmark-durations.outputs.json }}
    needs: [ check-permissions, build-build-tools-image ]
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container:
      image: ${{ needs.build-build-tools-image.outputs.image }}
      credentials:
@@ -253,7 +257,7 @@ jobs:
  benchmarks:
    if: github.ref_name == 'main' || contains(github.event.pull_request.labels.*.name, 'run-benchmarks')
    needs: [ check-permissions, build-and-test-locally, build-build-tools-image, get-benchmarks-durations ]
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container:
      image: ${{ needs.build-build-tools-image.outputs.image }}
      credentials:
@@ -280,13 +284,11 @@ jobs:
          save_perf_report: ${{ github.ref_name == 'main' }}
          extra_params: --splits 5 --group ${{ matrix.pytest_split_group }}
          benchmark_durations: ${{ needs.get-benchmarks-durations.outputs.json }}
-          pg_version: v16
        env:
          VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
          PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
          TEST_RESULT_CONNSTR: "${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}"
          PAGESERVER_VIRTUAL_FILE_IO_ENGINE: tokio-epoll-uring
-          SYNC_AFTER_EACH_TEST: true
      # XXX: no coverage data handling here, since benchmarks are run on release builds,
      # while coverage is currently collected for the debug ones

@@ -300,8 +302,9 @@ jobs:
      with:
        channel-id: C060CNA47S9 # on-call-staging-storage-stream
        slack-message: |
-          Benchmarks failed on main <${{ github.event.head_commit.url }}|${{ github.sha }}>
-          <${{ needs.create-test-report.outputs.report-url }}|Allure report>
+          Benchmarks failed on main: ${{ github.event.head_commit.url }}
+
+          Allure report: ${{ needs.create-test-report.outputs.report-url }}
      env:
        SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

@@ -311,7 +314,7 @@ jobs:
    outputs:
      report-url: ${{ steps.create-allure-report.outputs.report-url }}

-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container:
      image: ${{ needs.build-build-tools-image.outputs.image }}
      credentials:
@@ -358,7 +361,7 @@ jobs:

  coverage-report:
    needs: [ check-permissions, build-build-tools-image, build-and-test-locally ]
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container:
      image: ${{ needs.build-build-tools-image.outputs.image }}
      credentials:
@@ -472,7 +475,7 @@ jobs:
      matrix:
        arch: [ x64, arm64 ]

-    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
+    runs-on: ${{ fromJson(format('["self-hosted", "gen3", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}

    steps:
      - name: Checkout
@@ -481,7 +484,12 @@ jobs:
          submodules: true
          fetch-depth: 0

-      - uses: ./.github/actions/set-docker-config-dir
+      # 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 .docker-custom
+          echo DOCKER_CONFIG=$(pwd)/.docker-custom >> $GITHUB_ENV
      - uses: docker/setup-buildx-action@v3
        with:
          cache-binary: false
@@ -500,10 +508,7 @@ jobs:
      - uses: docker/build-push-action@v6
        with:
          context: .
-          # ARM-specific flags are recommended for Graviton ≥ 2, these flags are also supported by Ampere Altra (Azure)
-          # https://github.com/aws/aws-graviton-getting-started/blob/57dc813626d0266f1cc12ef83474745bb1f31fb4/rust.md
          build-args: |
-            ADDITIONAL_RUSTFLAGS=${{ matrix.arch == 'arm64' && '-Ctarget-feature=+lse -Ctarget-cpu=neoverse-n1' || '' }}
            GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
            BUILD_TAG=${{ needs.tag.outputs.build-tag }}
            TAG=${{ needs.build-build-tools-image.outputs.image-tag }}
@@ -516,6 +521,11 @@ jobs:
          tags: |
            neondatabase/neon:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}

+      - name: Remove custom docker config directory
+        if: always()
+        run: |
+          rm -rf .docker-custom
+
  neon-image:
    needs: [ neon-image-arch, tag ]
    runs-on: ubuntu-22.04
@@ -551,7 +561,7 @@ jobs:
        version: [ v14, v15, v16 ]
        arch: [ x64, arm64 ]

-    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
+    runs-on: ${{ fromJson(format('["self-hosted", "gen3", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}

    steps:
      - name: Checkout
@@ -560,7 +570,12 @@ jobs:
          submodules: true
          fetch-depth: 0

-      - uses: ./.github/actions/set-docker-config-dir
+      # 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 .docker-custom
+          echo DOCKER_CONFIG=$(pwd)/.docker-custom >> $GITHUB_ENV
      - uses: docker/setup-buildx-action@v3
        with:
          cache-binary: false
@@ -643,6 +658,11 @@ jobs:
          tags: |
            neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}

+      - name: Remove custom docker config directory
+        if: always()
+        run: |
+          rm -rf .docker-custom
+
  compute-node-image:
    needs: [ compute-node-image-arch, tag ]
    runs-on: ubuntu-22.04
@@ -696,7 +716,7 @@ jobs:

  vm-compute-node-image:
    needs: [ check-permissions, tag, compute-node-image ]
-    runs-on: [ self-hosted, large ]
+    runs-on: [ self-hosted, gen3, large ]
    strategy:
      fail-fast: false
      matrix:
@@ -715,7 +735,13 @@ jobs:
          curl -fL https://github.com/neondatabase/autoscaling/releases/download/$VM_BUILDER_VERSION/vm-builder -o vm-builder
          chmod +x vm-builder

-      - uses: ./.github/actions/set-docker-config-dir
+      # 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 .docker-custom
+          echo DOCKER_CONFIG=$(pwd)/.docker-custom >> $GITHUB_ENV
+
      - uses: docker/login-action@v3
        with:
          username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
@@ -738,6 +764,11 @@ jobs:
        run: |
          docker push neondatabase/vm-compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}

+      - name: Remove custom docker config directory
+        if: always()
+        run: |
+          rm -rf .docker-custom
+
  test-images:
    needs: [ check-permissions, tag, neon-image, compute-node-image ]
    strategy:
@@ -745,7 +776,7 @@ jobs:
      matrix:
        arch: [ x64, arm64 ]

-    runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'small-arm64' || 'small')) }}
+    runs-on: ${{ fromJson(format('["self-hosted", "gen3", "{0}"]', matrix.arch == 'arm64' && 'small-arm64' || 'small')) }}

    steps:
      - name: Checkout
@@ -753,7 +784,13 @@ jobs:
        with:
          fetch-depth: 0

-      - uses: ./.github/actions/set-docker-config-dir
+      # 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 .docker-custom
+          echo DOCKER_CONFIG=$(pwd)/.docker-custom >> $GITHUB_ENV
+
      - uses: docker/login-action@v3
        with:
          username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
@@ -793,7 +830,15 @@ jobs:
          docker compose -f ./docker-compose/docker-compose.yml logs || 0
          docker compose -f ./docker-compose/docker-compose.yml down

+      - name: Remove custom docker config directory
+        if: always()
+        run: |
+          rm -rf .docker-custom
+
  promote-images:
+    permissions:
+      contents: read  # This is required for actions/checkout
+      id-token: write # This is required for Azure Login to work.
    needs: [ check-permissions, tag, test-images, vm-compute-node-image ]
    runs-on: ubuntu-22.04

@@ -820,6 +865,28 @@ jobs:
                                               neondatabase/vm-compute-node-${version}:${{ needs.tag.outputs.build-tag }}
          done

+      - name: Azure login
+        if: github.ref_name == 'main'
+        uses: azure/login@6c251865b4e6290e7b78be643ea2d005bc51f69a  # @v2.1.1
+        with:
+          client-id: ${{ secrets.AZURE_DEV_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          subscription-id: ${{ secrets.AZURE_DEV_SUBSCRIPTION_ID }}
+
+      - name: Login to ACR
+        if: github.ref_name == 'main'
+        run: |
+          az acr login --name=neoneastus2
+
+      - name: Copy docker images to ACR-dev
+        if: github.ref_name == 'main'
+        run: |
+          for image in neon compute-tools {vm-,}compute-node-{v14,v15,v16}; do
+            docker buildx imagetools create \
+              -t neoneastus2.azurecr.io/neondatabase/${image}:${{ needs.tag.outputs.build-tag }} \
+                                        neondatabase/${image}:${{ needs.tag.outputs.build-tag }}
+          done
+
      - name: Add latest tag to images
        if: github.ref_name == 'main'
        run: |
@@ -857,30 +924,6 @@ jobs:
                                               369495373322.dkr.ecr.eu-central-1.amazonaws.com/${image}:${{ needs.tag.outputs.build-tag }}
          done

-  push-to-acr-dev:
-    if: github.ref_name == 'main'
-    needs: [ tag, promote-images ]
-    uses: ./.github/workflows/_push-to-acr.yml
-    with:
-      client_id: ${{ vars.AZURE_DEV_CLIENT_ID }}
-      image_tag: ${{ needs.tag.outputs.build-tag }}
-      images: neon compute-tools vm-compute-node-v14 vm-compute-node-v15 vm-compute-node-v16 compute-node-v14 compute-node-v15 compute-node-v16
-      registry_name: ${{ vars.AZURE_DEV_REGISTRY_NAME }}
-      subscription_id: ${{ vars.AZURE_DEV_SUBSCRIPTION_ID }}
-      tenant_id: ${{ vars.AZURE_TENANT_ID }}
-
-  push-to-acr-prod:
-    if: github.ref_name == 'release'|| github.ref_name == 'release-proxy'
-    needs: [ tag, promote-images ]
-    uses: ./.github/workflows/_push-to-acr.yml
-    with:
-      client_id: ${{ vars.AZURE_PROD_CLIENT_ID }}
-      image_tag: ${{ needs.tag.outputs.build-tag }}
-      images: neon compute-tools vm-compute-node-v14 vm-compute-node-v15 vm-compute-node-v16 compute-node-v14 compute-node-v15 compute-node-v16
-      registry_name: ${{ vars.AZURE_PROD_REGISTRY_NAME }}
-      subscription_id: ${{ vars.AZURE_PROD_SUBSCRIPTION_ID }}
-      tenant_id: ${{ vars.AZURE_TENANT_ID }}
-
  trigger-custom-extensions-build-and-wait:
    needs: [ check-permissions, tag ]
    runs-on: ubuntu-22.04
@@ -956,10 +999,10 @@ jobs:
          exit 1

  deploy:
-    needs: [ check-permissions, promote-images, tag, build-and-test-locally, trigger-custom-extensions-build-and-wait, push-to-acr-dev, push-to-acr-prod ]
-    if: (github.ref_name == 'main' || github.ref_name == 'release' || github.ref_name == 'release-proxy') && !failure() && !cancelled()
+    needs: [ check-permissions, promote-images, tag, build-and-test-locally, trigger-custom-extensions-build-and-wait ]
+    if: github.ref_name == 'main' || github.ref_name == 'release'|| github.ref_name == 'release-proxy'

-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/ansible:latest
    steps:
      - name: Fix git ownership
@@ -979,6 +1022,7 @@ jobs:
      - name: Checkout
        uses: actions/checkout@v4
        with:
+          submodules: false
          fetch-depth: 0

      - name: Trigger deploy workflow
@@ -986,10 +1030,10 @@ jobs:
          GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
        run: |
          if [[ "$GITHUB_REF_NAME" == "main" ]]; then
-            gh workflow --repo neondatabase/infra run deploy-dev.yml --ref main -f branch=main -f dockerTag=${{needs.tag.outputs.build-tag}} -f deployPreprodRegion=false
+            gh workflow --repo neondatabase/aws run deploy-dev.yml --ref main -f branch=main -f dockerTag=${{needs.tag.outputs.build-tag}} -f deployPreprodRegion=false
            gh workflow --repo neondatabase/azure run deploy.yml -f dockerTag=${{needs.tag.outputs.build-tag}}
          elif [[ "$GITHUB_REF_NAME" == "release" ]]; then
-            gh workflow --repo neondatabase/infra run deploy-dev.yml --ref main \
+            gh workflow --repo neondatabase/aws run deploy-dev.yml --ref main \
              -f deployPgSniRouter=false \
              -f deployProxy=false \
              -f deployStorage=true \
@@ -999,14 +1043,14 @@ jobs:
              -f dockerTag=${{needs.tag.outputs.build-tag}} \
              -f deployPreprodRegion=true

-            gh workflow --repo neondatabase/infra run deploy-prod.yml --ref main \
+            gh workflow --repo neondatabase/aws run deploy-prod.yml --ref main \
              -f deployStorage=true \
              -f deployStorageBroker=true \
              -f deployStorageController=true \
              -f branch=main \
              -f dockerTag=${{needs.tag.outputs.build-tag}}
          elif [[ "$GITHUB_REF_NAME" == "release-proxy" ]]; then
-            gh workflow --repo neondatabase/infra run deploy-dev.yml --ref main \
+            gh workflow --repo neondatabase/aws run deploy-dev.yml --ref main \
              -f deployPgSniRouter=true \
              -f deployProxy=true \
              -f deployStorage=false \
@@ -1016,7 +1060,7 @@ jobs:
              -f dockerTag=${{needs.tag.outputs.build-tag}} \
              -f deployPreprodRegion=true

-            gh workflow --repo neondatabase/infra run deploy-proxy-prod.yml --ref main \
+            gh workflow --repo neondatabase/aws run deploy-proxy-prod.yml --ref main \
              -f deployPgSniRouter=true \
              -f deployProxy=true \
              -f branch=main \
@@ -1055,88 +1099,43 @@ jobs:
              generate_release_notes: true,
            })

-  # The job runs on `release` branch and copies compatibility data and Neon artifact from the last *release PR* to the latest directory
  promote-compatibility-data:
-    needs: [ deploy ]
+    needs: [ check-permissions, promote-images, tag, build-and-test-locally ]
    if: github.ref_name == 'release'

-    runs-on: ubuntu-22.04
+    runs-on: [ self-hosted, gen3, small ]
+    container:
+      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/base:pinned
+      options: --init
    steps:
-      - name: Fetch GITHUB_RUN_ID and COMMIT_SHA for the last merged release PR
-        id: fetch-last-release-pr-info
-        env:
-          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-        run: |
-          branch_name_and_pr_number=$(gh pr list \
-            --repo "${GITHUB_REPOSITORY}" \
-            --base release \
-            --state merged \
-            --limit 10 \
-            --json mergeCommit,headRefName,number \
-            --jq ".[] | select(.mergeCommit.oid==\"${GITHUB_SHA}\") | { branch_name: .headRefName, pr_number: .number }")
-          branch_name=$(echo "${branch_name_and_pr_number}" | jq -r '.branch_name')
-          pr_number=$(echo "${branch_name_and_pr_number}" | jq -r '.pr_number')
-
-          run_id=$(gh run list \
-            --repo "${GITHUB_REPOSITORY}" \
-            --workflow build_and_test.yml \
-            --branch "${branch_name}" \
-            --json databaseId \
-            --limit 1 \
-            --jq '.[].databaseId')
-
-          last_commit_sha=$(gh pr view "${pr_number}" \
-            --repo "${GITHUB_REPOSITORY}" \
-            --json commits \
-            --jq '.commits[-1].oid')
-
-          echo "run-id=${run_id}" | tee -a ${GITHUB_OUTPUT}
-          echo "commit-sha=${last_commit_sha}" | tee -a ${GITHUB_OUTPUT}
-
-      - name: Promote compatibility snapshot and Neon artifact
+      - name: Promote compatibility snapshot for the release
        env:
          BUCKET: neon-github-public-dev
-          AWS_REGION: eu-central-1
-          COMMIT_SHA: ${{ steps.fetch-last-release-pr-info.outputs.commit-sha }}
-          RUN_ID: ${{ steps.fetch-last-release-pr-info.outputs.run-id }}
+          PREFIX: artifacts/latest
+          COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
        run: |
-          old_prefix="artifacts/${COMMIT_SHA}/${RUN_ID}"
-          new_prefix="artifacts/latest"
-
-          files_to_promote=()
-          files_on_s3=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${old_prefix} | jq -r '.Contents[]?.Key' || true)
-
-          for arch in X64 ARM64; do
+          # Update compatibility snapshot for the release
+          for pg_version in v14 v15 v16; do
            for build_type in debug release; do
-              neon_artifact_filename="neon-Linux-${arch}-${build_type}-artifact.tar.zst"
-              s3_key=$(echo "${files_on_s3}" | grep ${neon_artifact_filename} | sort --version-sort | tail -1 || true)
-              if [ -z "${s3_key}" ]; then
-                echo >&2 "Neither s3://${BUCKET}/${old_prefix}/${neon_artifact_filename} nor its version from previous attempts exist"
-                exit 1
-              fi
+              OLD_FILENAME=compatibility-snapshot-${build_type}-pg${pg_version}-${GITHUB_RUN_ID}.tar.zst
+              NEW_FILENAME=compatibility-snapshot-${build_type}-pg${pg_version}.tar.zst

-              files_to_promote+=("s3://${BUCKET}/${s3_key}")
-
-              for pg_version in v14 v15 v16; do
-                # We run less tests for debug builds, so we don't need to promote them
-                if [ "${build_type}" == "debug" ] && { [ "${arch}" == "ARM64" ] || [ "${pg_version}" != "v16" ] ; }; then
-                  continue
-                fi
-
-                compatibility_data_filename="compatibility-snapshot-${arch}-${build_type}-pg${pg_version}.tar.zst"
-                s3_key=$(echo "${files_on_s3}" | grep ${compatibility_data_filename} | sort --version-sort | tail -1 || true)
-                if [ -z "${s3_key}" ]; then
-                  echo >&2 "Neither s3://${BUCKET}/${old_prefix}/${compatibility_data_filename} nor its version from previous attempts exist"
-                  exit 1
-                fi
-
-                files_to_promote+=("s3://${BUCKET}/${s3_key}")
-              done
+              time aws s3 mv --only-show-errors s3://${BUCKET}/${PREFIX}/${OLD_FILENAME} s3://${BUCKET}/${PREFIX}/${NEW_FILENAME}
            done
          done

-          for f in "${files_to_promote[@]}"; do
-            time aws s3 cp --only-show-errors ${f} s3://${BUCKET}/${new_prefix}/
+          # Update Neon artifact for the release (reuse already uploaded artifact)
+          for build_type in debug release; do
+            OLD_PREFIX=artifacts/${COMMIT_SHA}/${GITHUB_RUN_ID}
+            FILENAME=neon-${{ runner.os }}-${{ runner.arch }}-${build_type}-artifact.tar.zst
+
+            S3_KEY=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${OLD_PREFIX} | jq -r '.Contents[]?.Key' | grep ${FILENAME} | sort --version-sort | tail -1 || true)
+            if [ -z "${S3_KEY}" ]; then
+              echo >&2 "Neither s3://${BUCKET}/${OLD_PREFIX}/${FILENAME} nor its version from previous attempts exist"
+              exit 1
+            fi
+
+            time aws s3 cp --only-show-errors s3://${BUCKET}/${S3_KEY} s3://${BUCKET}/${PREFIX}/${FILENAME}
          done

  pin-build-tools-image:
@@ -1160,12 +1159,10 @@ jobs:
    # Format `needs` differently to make the list more readable.
    # Usually we do `needs: [...]`
    needs:
-      - build-and-test-locally
      - check-codestyle-python
      - check-codestyle-rust
-      - promote-images
+      - build-and-test-locally
      - test-images
-      - trigger-custom-extensions-build-and-wait
    runs-on: ubuntu-22.04
    steps:
      # The list of possible results:
--- a/.github/workflows/label-for-external-users.yml
+++ b/.github/workflows/label-for-external-users.yml
@@ -1,78 +0,0 @@
-name: Add `external` label to issues and PRs created by external users
-
-on:
-  issues:
-    types:
-      - opened
-  pull_request_target:
-    types:
-      - opened
-  workflow_dispatch:
-    inputs:
-      github-actor:
-        description: 'GitHub username. If empty, the username of the current user will be used'
-        required: false
-
-# 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 }}
-        ACTOR: ${{ inputs.github-actor || github.actor }}
-      run: |
-        expected_error="User does not exist or is not a member of the organization"
-        output_file=output.txt
-
-        for i in $(seq 1 10); do
-          if gh api "/orgs/${GITHUB_REPOSITORY_OWNER}/members/${ACTOR}" \
-              -H "Accept: application/vnd.github+json" \
-              -H "X-GitHub-Api-Version: 2022-11-28" > ${output_file}; then
-
-            is_member=true
-            break
-          elif grep -q "${expected_error}" ${output_file}; then
-            is_member=false
-            break
-          elif [ $i -eq 10 ]; then
-            title="Failed to get memmbership status for ${ACTOR}"
-            message="The latest GitHub API error message: '$(cat ${output_file})'"
-            echo "::error file=.github/workflows/label-for-external-users.yml,title=${title}::${message}"
-
-            exit 1
-          fi
-
-          sleep 1
-        done
-
-        echo "is-member=${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}
--- a/.github/workflows/neon_extra_builds.yml
+++ b/.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
@@ -157,6 +158,7 @@ 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
--- a/.github/workflows/periodic_pagebench.yml
+++ b/.github/workflows/periodic_pagebench.yml
@@ -27,7 +27,7 @@ concurrency:

 jobs:
  trigger_bench_on_ec2_machine_in_eu_central_1:
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, gen3, small ]
    container:
      image: neondatabase/build-tools:pinned
      credentials:
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -915,30 +915,27 @@ dependencies = [

 [[package]]
 name = "bindgen"
-version = "0.70.1"
+version = "0.65.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f49d8fed880d473ea71efb9bf597651e77201bdd4893efe54c9e5d65ae04ce6f"
+checksum = "cfdf7b466f9a4903edc73f95d6d2bcd5baf8ae620638762244d3f60143643cc5"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 1.3.2",
 "cexpr",
 "clang-sys",
- "itertools 0.12.1",
+ "lazy_static",
+ "lazycell",
 "log",
- "prettyplease 0.2.17",
+ "peeking_take_while",
+ "prettyplease 0.2.6",
 "proc-macro2",
 "quote",
 "regex",
 "rustc-hash",
 "shlex",
 "syn 2.0.52",
+ "which",
 ]

-[[package]]
-name = "bit_field"
-version = "0.10.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dc827186963e592360843fb5ba4b973e145841266c1357f7180c43526f2e5b61"
-
 [[package]]
 name = "bitflags"
 version = "1.3.2"
@@ -1211,6 +1208,7 @@ dependencies = [
 "serde_json",
 "serde_with",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -1323,6 +1321,7 @@ dependencies = [
 "serde",
 "serde_with",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -1330,6 +1329,7 @@ name = "control_plane"
 version = "0.1.0"
 dependencies = [
 "anyhow",
+ "async-trait",
 "camino",
 "clap",
 "comfy-table",
@@ -1670,13 +1670,14 @@ dependencies = [
 "smallvec",
 "tracing",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
 name = "diesel"
-version = "2.2.3"
+version = "2.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "65e13bab2796f412722112327f3e575601a3e9cdcbe426f0d30dbf43f3f5dc71"
+checksum = "62d6dcd069e7b5fe49a302411f759d4cf1cf2c27fe798ef46fb8baefc053dd2b"
 dependencies = [
 "bitflags 2.4.1",
 "byteorder",
@@ -2724,12 +2725,6 @@ dependencies = [
 "hashbrown 0.14.5",
 ]

-[[package]]
-name = "indoc"
-version = "2.0.5"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b248f5224d1d606005e02c97f5aa4e88eeb230488bcc03bc9ca4d7991399f2b5"
-
 [[package]]
 name = "infer"
 version = "0.2.3"
@@ -2946,6 +2941,23 @@ dependencies = [
 "spin 0.5.2",
 ]

+[[package]]
+name = "lazycell"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "830d08ce1d1d941e6b30645f1a0eb5643013d835ce3779a5fc208261dbe10f55"
+
+[[package]]
+name = "leaky-bucket"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8eb491abd89e9794d50f93c8db610a29509123e3fbbc9c8c67a528e9391cd853"
+dependencies = [
+ "parking_lot 0.12.1",
+ "tokio",
+ "tracing",
+]
+
 [[package]]
 name = "libc"
 version = "0.2.150"
@@ -3135,6 +3147,7 @@ dependencies = [
 "rand 0.8.5",
 "rand_distr",
 "twox-hash",
+ "workspace_hack",
 ]

 [[package]]
@@ -3674,7 +3687,6 @@ dependencies = [
 "async-compression",
 "async-stream",
 "async-trait",
- "bit_field",
 "byteorder",
 "bytes",
 "camino",
@@ -3698,8 +3710,8 @@ dependencies = [
 "humantime",
 "humantime-serde",
 "hyper 0.14.26",
- "indoc",
 "itertools 0.10.5",
+ "leaky-bucket",
 "md5",
 "metrics",
 "nix 0.27.1",
@@ -3724,7 +3736,6 @@ dependencies = [
 "reqwest 0.12.4",
 "rpds",
 "scopeguard",
- "send-future",
 "serde",
 "serde_json",
 "serde_path_to_error",
@@ -3764,7 +3775,6 @@ dependencies = [
 "bincode",
 "byteorder",
 "bytes",
- "camino",
 "chrono",
 "const_format",
 "enum-map",
@@ -3772,20 +3782,16 @@ dependencies = [
 "humantime",
 "humantime-serde",
 "itertools 0.10.5",
- "nix 0.27.1",
- "postgres_backend",
 "postgres_ffi",
 "rand 0.8.5",
- "remote_storage",
- "reqwest 0.12.4",
 "serde",
 "serde_json",
 "serde_with",
- "storage_broker",
 "strum",
 "strum_macros",
 "thiserror",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -3793,6 +3799,7 @@ name = "pageserver_client"
 version = "0.1.0"
 dependencies = [
 "anyhow",
+ "async-trait",
 "bytes",
 "futures",
 "pageserver_api",
@@ -3968,6 +3975,12 @@ dependencies = [
 "sha2",
 ]

+[[package]]
+name = "peeking_take_while"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "19b17cddbe7ec3f8bc800887bab5e717348c95ea2ca0b1bf0837fb964dc67099"
+
 [[package]]
 name = "pem"
 version = "3.0.3"
@@ -4003,6 +4016,29 @@ dependencies = [
 "indexmap 1.9.3",
 ]

+[[package]]
+name = "pg_sni_router"
+version = "0.1.0"
+dependencies = [
+ "anyhow",
+ "clap",
+ "futures",
+ "git-version",
+ "itertools 0.10.5",
+ "pq_proto",
+ "proxy-core",
+ "proxy-sasl",
+ "rustls 0.22.4",
+ "rustls-pemfile 2.1.1",
+ "socket2 0.5.5",
+ "tokio",
+ "tokio-util",
+ "tracing",
+ "tracing-utils",
+ "utils",
+ "uuid",
+]
+
 [[package]]
 name = "phf"
 version = "0.11.1"
@@ -4180,6 +4216,7 @@ dependencies = [
 "tokio-rustls 0.25.0",
 "tokio-util",
 "tracing",
+ "workspace_hack",
 ]

 [[package]]
@@ -4192,6 +4229,7 @@ dependencies = [
 "postgres",
 "tokio-postgres",
 "url",
+ "workspace_hack",
 ]

 [[package]]
@@ -4214,6 +4252,7 @@ dependencies = [
 "serde",
 "thiserror",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -4251,6 +4290,7 @@ dependencies = [
 "thiserror",
 "tokio",
 "tracing",
+ "workspace_hack",
 ]

 [[package]]
@@ -4265,9 +4305,9 @@ dependencies = [

 [[package]]
 name = "prettyplease"
-version = "0.2.17"
+version = "0.2.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8d3928fb5db768cb86f891ff014f0144589297e3c6a1aba6ed7cecfdace270c7"
+checksum = "3b69d39aab54d069e7f2fe8cb970493e7834601ca2d8c65fd7bbd183578080d1"
 dependencies = [
 "proc-macro2",
 "syn 2.0.52",
@@ -4396,6 +4436,34 @@ dependencies = [
 [[package]]
 name = "proxy"
 version = "0.1.0"
+dependencies = [
+ "anyhow",
+ "aws-config",
+ "clap",
+ "futures",
+ "git-version",
+ "humantime",
+ "itertools 0.10.5",
+ "metrics",
+ "pq_proto",
+ "proxy-core",
+ "proxy-sasl",
+ "remote_storage",
+ "rustls 0.22.4",
+ "rustls-pemfile 2.1.1",
+ "socket2 0.5.5",
+ "tikv-jemallocator",
+ "tokio",
+ "tokio-util",
+ "tracing",
+ "tracing-utils",
+ "utils",
+ "uuid",
+]
+
+[[package]]
+name = "proxy-core"
+version = "0.1.0"
 dependencies = [
 "ahash",
 "anyhow",
@@ -4422,7 +4490,6 @@ dependencies = [
 "fallible-iterator",
 "framed-websockets",
 "futures",
- "git-version",
 "hashbrown 0.14.5",
 "hashlink",
 "hex",
@@ -4445,7 +4512,6 @@ dependencies = [
 "measured",
 "metrics",
 "once_cell",
- "opentelemetry",
 "p256 0.13.2",
 "parking_lot 0.12.1",
 "parquet",
@@ -4455,7 +4521,7 @@ dependencies = [
 "postgres-protocol",
 "postgres_backend",
 "pq_proto",
- "prometheus",
+ "proxy-sasl",
 "rand 0.8.5",
 "rand_distr",
 "rcgen",
@@ -4485,7 +4551,6 @@ dependencies = [
 "task-local-extensions",
 "thiserror",
 "tikv-jemalloc-ctl",
- "tikv-jemallocator",
 "tokio",
 "tokio-postgres",
 "tokio-postgres-rustls",
@@ -4508,6 +4573,35 @@ dependencies = [
 "x509-parser",
 ]

+[[package]]
+name = "proxy-sasl"
+version = "0.1.0"
+dependencies = [
+ "ahash",
+ "anyhow",
+ "base64 0.13.1",
+ "bytes",
+ "crossbeam-deque",
+ "hmac",
+ "itertools 0.10.5",
+ "lasso",
+ "measured",
+ "parking_lot 0.12.1",
+ "pbkdf2",
+ "postgres-protocol",
+ "pq_proto",
+ "rand 0.8.5",
+ "rustls 0.22.4",
+ "sha2",
+ "subtle",
+ "thiserror",
+ "tokio",
+ "tracing",
+ "uuid",
+ "workspace_hack",
+ "x509-parser",
+]
+
 [[package]]
 name = "quick-xml"
 version = "0.31.0"
@@ -4815,6 +4909,7 @@ dependencies = [
 "toml_edit 0.19.10",
 "tracing",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -5339,6 +5434,7 @@ dependencies = [
 "serde",
 "serde_with",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -5447,12 +5543,6 @@ version = "1.0.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bebd363326d05ec3e2f532ab7660680f3b02130d780c299bca73469d521bc0ed"

-[[package]]
-name = "send-future"
-version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "224e328af6e080cddbab3c770b1cf50f0351ba0577091ef2410c3951d835ff87"
-
 [[package]]
 name = "sentry"
 version = "0.32.3"
@@ -5588,12 +5678,11 @@ dependencies = [

 [[package]]
 name = "serde_json"
-version = "1.0.125"
+version = "1.0.96"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "83c8e735a073ccf5be70aa8066aa984eaf2fa000db6c8d0100ae605b366d31ed"
+checksum = "057d394a50403bcac12672b2b18fb387ab6d289d957dab67dd201875391e52f1"
 dependencies = [
 "itoa",
- "memchr",
 "ryu",
 "serde",
 ]
@@ -5948,6 +6037,7 @@ name = "storage_controller_client"
 version = "0.1.0"
 dependencies = [
 "anyhow",
+ "async-trait",
 "bytes",
 "futures",
 "pageserver_api",
@@ -6079,9 +6169,8 @@ checksum = "81cdd64d312baedb58e21336b31bc043b77e01cc99033ce76ef539f78e965ebc"

 [[package]]
 name = "svg_fmt"
-version = "0.4.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "20e16a0f46cf5fd675563ef54f26e83e20f2366bcf027bcb3cc3ed2b98aaf2ca"
+version = "0.4.2"
+source = "git+https://github.com/nical/rust_debug?rev=28a7d96eecff2f28e75b1ea09f2d499a60d0e3b4#28a7d96eecff2f28e75b1ea09f2d499a60d0e3b4"

 [[package]]
 name = "syn"
@@ -6181,6 +6270,7 @@ dependencies = [
 "anyhow",
 "serde",
 "serde_json",
+ "workspace_hack",
 ]

 [[package]]
@@ -6781,6 +6871,7 @@ dependencies = [
 "tracing",
 "tracing-opentelemetry",
 "tracing-subscriber",
+ "workspace_hack",
 ]

 [[package]]
@@ -6951,6 +7042,7 @@ dependencies = [
 "anyhow",
 "arc-swap",
 "async-compression",
+ "async-trait",
 "bincode",
 "byteorder",
 "bytes",
@@ -6966,6 +7058,7 @@ dependencies = [
 "humantime",
 "hyper 0.14.26",
 "jsonwebtoken",
+ "leaky-bucket",
 "metrics",
 "nix 0.27.1",
 "once_cell",
@@ -6996,6 +7089,7 @@ dependencies = [
 "url",
 "uuid",
 "walkdir",
+ "workspace_hack",
 ]

 [[package]]
@@ -7074,6 +7168,7 @@ dependencies = [
 "postgres_ffi",
 "regex",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -7094,6 +7189,7 @@ dependencies = [
 "bindgen",
 "postgres_ffi",
 "utils",
+ "workspace_hack",
 ]

 [[package]]
@@ -7613,7 +7709,6 @@ dependencies = [
 "hyper 0.14.26",
 "indexmap 1.9.3",
 "itertools 0.10.5",
- "itertools 0.12.1",
 "lazy_static",
 "libc",
 "log",
@@ -7651,6 +7746,8 @@ dependencies = [
 "tokio",
 "tokio-rustls 0.24.0",
 "tokio-util",
+ "toml_datetime",
+ "toml_edit 0.19.10",
 "tonic",
 "tower",
 "tracing",
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -9,7 +9,10 @@ members = [
    "pageserver/ctl",
    "pageserver/client",
    "pageserver/pagebench",
-    "proxy",
+    "proxy/core",
+    "proxy/sasl",
+    "proxy/proxy",
+    "proxy/pg_sni_router",
    "safekeeper",
    "storage_broker",
    "storage_controller",
@@ -64,8 +67,7 @@ aws-types = "1.2.0"
 axum = { version = "0.6.20", features = ["ws"] }
 base64 = "0.13.0"
 bincode = "1.3"
-bindgen = "0.70"
-bit_field = "0.10.2"
+bindgen = "0.65"
 bstr = "1.0"
 byteorder = "1.4"
 bytes = "1.0"
@@ -103,18 +105,18 @@ humantime-serde = "1.1.1"
 hyper = "0.14"
 tokio-tungstenite = "0.20.0"
 indexmap = "2"
-indoc = "2"
 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" }
 memoffset = "0.8"
-nix = { version = "0.27", features = ["dir", "fs", "process", "socket", "signal", "poll"] }
+nix = { version = "0.27", features = ["fs", "process", "socket", "signal", "poll"] }
 notify = "6.0.0"
 num_cpus = "1.15"
 num-traits = "0.2.15"
@@ -146,7 +148,6 @@ 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" ] }
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1"
@@ -161,7 +162,8 @@ socket2 = "0.5"
 strum = "0.24"
 strum_macros = "0.24"
 "subtle"  = "2.5.0"
-svg_fmt = "0.4.3"
+# 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" }
 sync_wrapper = "0.1.2"
 tar = "0.4"
 task-local-extensions = "0.1.4"
--- a/4
+++ b/4
@@ -35,9 +35,8 @@ COPY --from=pg-build /home/nonroot/pg_install/v16/include/postgresql/server pg_i
 COPY --from=pg-build /home/nonroot/pg_install/v16/lib                       pg_install/v16/lib
 COPY --chown=nonroot . .

-ARG ADDITIONAL_RUSTFLAGS
 RUN set -e \
-    && PQ_LIB_DIR=$(pwd)/pg_install/v16/lib RUSTFLAGS="-Clinker=clang -Clink-arg=-fuse-ld=mold -Clink-arg=-Wl,--no-rosegment ${ADDITIONAL_RUSTFLAGS}" cargo build \
+    && PQ_LIB_DIR=$(pwd)/pg_install/v16/lib RUSTFLAGS="-Clinker=clang -Clink-arg=-fuse-ld=mold -Clink-arg=-Wl,--no-rosegment" cargo build \
      --bin pg_sni_router  \
      --bin pageserver  \
      --bin pagectl  \
@@ -87,7 +86,6 @@ RUN mkdir -p /data/.neon/ && \
       "pg_distrib_dir='/usr/local/'\n" \
       "listen_pg_addr='0.0.0.0:6400'\n" \
       "listen_http_addr='0.0.0.0:9898'\n" \
-       "availability_zone='local'\n" \
  > /data/.neon/pageserver.toml && \
  chown -R neon:neon /data/.neon

--- a/Dockerfile.build-tools
+++ b/Dockerfile.build-tools
@@ -192,7 +192,7 @@ 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=1.81.0
+ENV RUSTC_VERSION=1.80.1
 ENV RUSTUP_HOME="/home/nonroot/.rustup"
 ENV PATH="/home/nonroot/.cargo/bin:${PATH}"
 ARG RUSTFILT_VERSION=0.2.1
@@ -207,7 +207,7 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
    export PATH="$HOME/.cargo/bin:$PATH" && \
    . "$HOME/.cargo/env" && \
    cargo --version && rustup --version && \
-    rustup component add llvm-tools rustfmt clippy && \
+    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 ${CARGO_DENY_VERSION} && \
--- a/Dockerfile.compute-node
+++ b/Dockerfile.compute-node
@@ -942,7 +942,7 @@ COPY --from=hll-pg-build /hll.tar.gz /ext-src
 COPY --from=plpgsql-check-pg-build /plpgsql_check.tar.gz /ext-src
 #COPY --from=timescaledb-pg-build /timescaledb.tar.gz /ext-src
 COPY --from=pg-hint-plan-pg-build /pg_hint_plan.tar.gz /ext-src
-COPY patches/pg_hint_plan.patch /ext-src
+COPY patches/pg_hintplan.patch /ext-src
 COPY --from=pg-cron-pg-build /pg_cron.tar.gz /ext-src
 COPY patches/pg_cron.patch /ext-src
 #COPY --from=pg-pgx-ulid-build /home/nonroot/pgx_ulid.tar.gz /ext-src
@@ -964,7 +964,7 @@ RUN cd /ext-src/pgvector-src && patch -p1 <../pgvector.patch
 RUN cd /ext-src/rum-src && patch -p1 <../rum.patch
 # cmake is required for the h3 test
 RUN apt-get update && apt-get install -y cmake
-RUN cd /ext-src/pg_hint_plan-src && patch -p1 < /ext-src/pg_hint_plan.patch
+RUN patch -p1 < /ext-src/pg_hintplan.patch
 COPY --chmod=755 docker-compose/run-tests.sh /run-tests.sh
 RUN patch -p1 </ext-src/pg_anon.patch
 RUN patch -p1 </ext-src/pg_cron.patch
--- a/README.md
+++ b/README.md
@@ -64,12 +64,6 @@ brew install protobuf openssl flex bison icu4c pkg-config
 echo 'export PATH="$(brew --prefix openssl)/bin:$PATH"' >> ~/.zshrc
 ```

-If you get errors about missing `m4` you may have to install it manually:
-```
-brew install m4
-brew link --force m4
-```
-
 2. [Install Rust](https://www.rust-lang.org/tools/install)
 ```
 # recommended approach from https://www.rust-lang.org/tools/install
@@ -132,7 +126,7 @@ 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
@@ -268,7 +262,7 @@ 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
--- a/compute_tools/src/bin/compute_ctl.rs
+++ b/compute_tools/src/bin/compute_ctl.rs
@@ -44,7 +44,6 @@ 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};
@@ -367,8 +366,6 @@ fn wait_spec(
        state.start_time = now;
    }

-    launch_lsn_lease_bg_task_for_static(&compute);
-
    Ok(WaitSpecResult {
        compute,
        http_port,
--- a/compute_tools/src/lib.rs
+++ b/compute_tools/src/lib.rs
@@ -11,7 +11,6 @@ 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;
--- a/compute_tools/src/lsn_lease.rs
+++ b/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)
-}
--- a/compute_tools/src/pg_helpers.rs
+++ b/compute_tools/src/pg_helpers.rs
@@ -22,10 +22,9 @@ use compute_api::spec::{Database, GenericOption, GenericOptions, PgIdent, Role};

 const POSTGRES_WAIT_TIMEOUT: Duration = Duration::from_millis(60 * 1000); // milliseconds

-/// Escape a string for including it in a SQL literal.
-///
-/// Wrapping the result with `E'{}'` or `'{}'` is not required,
-/// as it returns a ready-to-use SQL string literal, e.g. `'db'''` or `E'db\\'`.
+/// Escape a string for including it in a SQL literal. Wrapping the result
+/// with `E'{}'` or `'{}'` is not required, as it returns a ready-to-use
+/// SQL string literal, e.g. `'db'''` or `E'db\\'`.
 /// See <https://github.com/postgres/postgres/blob/da98d005cdbcd45af563d0c4ac86d0e9772cd15f/src/backend/utils/adt/quote.c#L47>
 /// for the original implementation.
 pub fn escape_literal(s: &str) -> String {
--- a/control_plane/Cargo.toml
+++ b/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
--- a/control_plane/src/background_process.rs
+++ b/control_plane/src/background_process.rs
@@ -379,7 +379,7 @@ 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),
--- a/control_plane/src/bin/neon_local.rs
+++ b/control_plane/src/bin/neon_local.rs
@@ -15,9 +15,7 @@ use control_plane::local_env::{
 };
 use control_plane::pageserver::PageServerNode;
 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::{
    DEFAULT_HTTP_LISTEN_PORT as DEFAULT_PAGESERVER_HTTP_PORT,
@@ -54,7 +52,7 @@ 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/";

@@ -1054,36 +1052,6 @@ fn get_start_timeout(args: &ArgMatches) -> &Duration {
    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)) => {
@@ -1145,14 +1113,19 @@ async fn handle_storage_controller(
    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 {
+            if let Err(e) = svc.start(get_start_timeout(start_match)).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);
            }
@@ -1255,12 +1228,7 @@ async fn handle_start_all(
    // 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(retry_timeout).await {
            eprintln!("storage_controller start failed: {:#}", e);
            try_stop_all(env, true).await;
            exit(1);
@@ -1390,21 +1358,10 @@ 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:#}");
        }
    }
 }
@@ -1544,18 +1501,6 @@ fn cli() -> Command {
        .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)
@@ -1664,12 +1609,9 @@ fn cli() -> Command {
                .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))
+                            .arg(timeout_arg.clone()))
                .subcommand(Command::new("stop").about("Stop storage controller")
-                            .arg(stop_mode_arg.clone())
-                            .arg(instance_id))
+                            .arg(stop_mode_arg.clone()))
        )
        .subcommand(
            Command::new("safekeeper")
--- a/control_plane/src/endpoint.rs
+++ b/control_plane/src/endpoint.rs
@@ -824,12 +824,11 @@ 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 '{}'",
--- a/control_plane/src/local_env.rs
+++ b/control_plane/src/local_env.rs
@@ -27,7 +27,7 @@ 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
@@ -156,18 +156,10 @@ pub struct NeonStorageControllerConf {
    #[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>,
-
-    #[serde(with = "humantime_serde")]
-    pub heartbeat_interval: Duration,
 }

 impl NeonStorageControllerConf {
@@ -175,9 +167,6 @@ impl NeonStorageControllerConf {
    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);
-
-    // Very tight heartbeat interval to speed up tests
-    const DEFAULT_HEARTBEAT_INTERVAL: std::time::Duration = std::time::Duration::from_millis(100);
 }

 impl Default for NeonStorageControllerConf {
@@ -185,11 +174,8 @@ impl Default for NeonStorageControllerConf {
        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,
-            heartbeat_interval: Self::DEFAULT_HEARTBEAT_INTERVAL,
        }
    }
 }
@@ -406,36 +392,6 @@ 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,
--- a/control_plane/src/pageserver.rs
+++ b/control_plane/src/pageserver.rs
@@ -181,23 +181,6 @@ impl PageServerNode {
        );
        io::stdout().flush()?;

-        // If the config file we got as a CLI argument includes the `availability_zone`
-        // config, then use that to populate the `metadata.json` file for the pageserver.
-        // In production the deployment orchestrator does this for us.
-        let az_id = conf
-            .other
-            .get("availability_zone")
-            .map(|toml| {
-                let az_str = toml.to_string();
-                // Trim the (") chars from the toml representation
-                if az_str.starts_with('"') && az_str.ends_with('"') {
-                    az_str[1..az_str.len() - 1].to_string()
-                } else {
-                    az_str
-                }
-            })
-            .unwrap_or("local".to_string());
-
        let config = self
            .pageserver_init_make_toml(conf)
            .context("make pageserver toml")?;
@@ -233,7 +216,6 @@ impl PageServerNode {
        let (_http_host, http_port) =
            parse_host_port(&self.conf.listen_http_addr).expect("Unable to parse listen_http_addr");
        let http_port = http_port.unwrap_or(9898);
-
        // Intentionally hand-craft JSON: this acts as an implicit format compat test
        // in case the pageserver-side structure is edited, and reflects the real life
        // situation: the metadata is written by some other script.
@@ -244,10 +226,7 @@ impl PageServerNode {
                postgres_port: self.pg_connection_config.port(),
                http_host: "localhost".to_string(),
                http_port,
-                other: HashMap::from([(
-                    "availability_zone_id".to_string(),
-                    serde_json::json!(az_id),
-                )]),
+                other: HashMap::new(),
            })
            .unwrap(),
        )
--- a/control_plane/src/safekeeper.rs
+++ b/control_plane/src/safekeeper.rs
@@ -5,7 +5,6 @@
 //! ```text
 //!   .neon/safekeepers/<safekeeper id>
 //! ```
-use std::future::Future;
 use std::io::Write;
 use std::path::PathBuf;
 use std::time::Duration;
@@ -35,10 +34,12 @@ 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();
--- a/control_plane/src/storage_controller.rs
+++ b/control_plane/src/storage_controller.rs
@@ -3,8 +3,6 @@ use crate::{
    local_env::{LocalEnv, NeonStorageControllerConf},
 };
 use camino::{Utf8Path, Utf8PathBuf};
-use hyper::Uri;
-use nix::unistd::Pid;
 use pageserver_api::{
    controller_api::{
        NodeConfigureRequest, NodeDescribeResponse, NodeRegisterRequest, TenantCreateRequest,
@@ -20,7 +18,7 @@ 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, time::Duration};
 use tokio::process::Command;
 use tracing::instrument;
 use url::Url;
@@ -31,14 +29,12 @@ use utils::{

 pub struct StorageController {
    env: LocalEnv,
+    listen: String,
    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";
@@ -47,36 +43,6 @@ 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,
-        }
-    }
-}
-
 #[derive(Serialize, Deserialize)]
 pub struct AttachHookRequest {
    pub tenant_shard_id: TenantShardId,
@@ -101,6 +67,23 @@ pub struct InspectResponse {

 impl StorageController {
    pub fn from_env(env: &LocalEnv) -> Self {
+        // 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 +126,20 @@ impl StorageController {

        Self {
            env: env.clone(),
+            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
@@ -209,23 +184,23 @@ impl StorageController {
    }

    /// 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> {
+        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 +209,7 @@ impl StorageController {
                "-h",
                "localhost",
                "-p",
-                &format!("{}", postgres_port),
+                &format!("{}", self.postgres_port),
                DB_NAME,
            ])
            .output()
@@ -255,14 +230,13 @@ impl StorageController {

    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)
+            .port(self.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
            //
@@ -278,114 +252,72 @@ impl StorageController {
            .map_err(anyhow::Error::new)
    }

-    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:?}");
-            }
-        }
+    pub async fn start(&self, retry_timeout: &Duration) -> 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_lib_dir = self.get_pg_lib_dir().await?;
+        let pg_log_path = pg_data_path.join("postgres.log");

-        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)
+        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}");
            }
        };

-        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");
+        // 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", self.postgres_port),
+        )
+        .await?;

-        // 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?;
+        println!("Starting storage controller database...");
+        let db_start_args = [
+            "-w",
+            "-D",
+            pg_data_path.as_ref(),
+            "-l",
+            pg_log_path.as_ref(),
+            "start",
+        ];

-        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");
+        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()),
+            retry_timeout,
+            || self.pg_isready(&pg_bin_dir),
+        )
+        .await?;

-            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),
-            )
-            .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);
+        // Run migrations on every startup, in case something changed.
+        let database_url = self.setup_database().await?;

        // We support running a startup SQL script to fiddle with the database before we launch storcon.
        // This is used by the test suite.
@@ -407,7 +339,7 @@ impl StorageController {
                }
            }
        };
-        let (mut client, conn) = self.connect_to_database(postgres_port).await?;
+        let (mut client, conn) = self.connect_to_database().await?;
        let conn = tokio::spawn(conn);
        let tx = client.build_transaction();
        let tx = tx.start().await?;
@@ -416,20 +348,9 @@ impl StorageController {
        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();
-
        let mut args = vec![
            "-l",
-            &listen.to_string(),
+            &self.listen,
            "--dev",
            "--database-url",
            &database_url,
@@ -437,29 +358,15 @@ impl StorageController {
            &humantime::Duration::from(self.config.max_offline).to_string(),
            "--max-warming-up-interval",
            &humantime::Duration::from(self.config.max_warming_up).to_string(),
-            "--heartbeat-interval",
-            &humantime::Duration::from(self.config.heartbeat_interval).to_string(),
-            "--address-for-peers",
-            &address_for_peers.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 {
@@ -487,15 +394,15 @@ impl StorageController {

        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,
+            background_process::InitialPidFile::Create(self.pid_file()),
+            retry_timeout,
            || async {
                match self.ready().await {
                    Ok(_) => Ok(true),
@@ -508,35 +415,8 @@ 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?;
@@ -549,51 +429,27 @@ 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],
@@ -619,31 +475,15 @@ 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 {
--- a/control_plane/storcon_cli/src/main.rs
+++ b/control_plane/storcon_cli/src/main.rs
@@ -4,8 +4,8 @@ use std::{str::FromStr, time::Duration};
 use clap::{Parser, Subcommand};
 use pageserver_api::{
    controller_api::{
-        NodeAvailabilityWrapper, NodeDescribeResponse, NodeShardResponse, ShardSchedulingPolicy,
-        TenantCreateRequest, TenantDescribeResponse, TenantPolicyRequest,
+        NodeAvailabilityWrapper, NodeDescribeResponse, ShardSchedulingPolicy, TenantCreateRequest,
+        TenantDescribeResponse, TenantPolicyRequest,
    },
    models::{
        EvictionPolicy, EvictionPolicyLayerAccessThreshold, LocationConfigSecondary,
@@ -41,8 +41,6 @@ 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
@@ -80,10 +78,7 @@ enum Command {
    /// List nodes known to the storage controller
    Nodes {},
    /// List tenants known to the storage controller
-    Tenants {
-        /// If this field is set, it will list the tenants on a specific node
-        node_id: Option<NodeId>,
-    },
+    Tenants {},
    /// Create a new tenant in the storage controller, and by extension on pageservers.
    TenantCreate {
        #[arg(long)]
@@ -152,9 +147,9 @@ enum Command {
        #[arg(long)]
        threshold: humantime::Duration,
    },
-    // Migrate away from a set of specified pageservers by moving the primary attachments to pageservers
+    // Drain a set of specified pageservers by moving the primary attachments to pageservers
    // outside of the specified set.
-    BulkMigrate {
+    Drain {
        // Set of pageserver node ids to drain.
        #[arg(long)]
        nodes: Vec<NodeId>,
@@ -168,34 +163,6 @@ enum Command {
        #[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)]
@@ -282,34 +249,6 @@ 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?;
-
-            if f(node.scheduling) {
-                return Ok::<NodeSchedulingPolicy, mgmt_api::Error>(node.scheduling);
-            }
-        }
-    });
-
-    Ok(waiter.await??)
-}
-
 #[tokio::main]
 async fn main() -> anyhow::Result<()> {
    let cli = Cli::parse();
@@ -327,7 +266,6 @@ async fn main() -> anyhow::Result<()> {
            listen_pg_port,
            listen_http_addr,
            listen_http_port,
-            availability_zone_id,
        } => {
            storcon_client
                .dispatch::<_, ()>(
@@ -339,7 +277,6 @@ async fn main() -> anyhow::Result<()> {
                        listen_pg_port,
                        listen_http_addr,
                        listen_http_port,
-                        availability_zone_id,
                    }),
                )
                .await?;
@@ -406,41 +343,7 @@ async fn main() -> anyhow::Result<()> {
                )
                .await?;
        }
-        Command::Tenants {
-            node_id: Some(node_id),
-        } => {
-            let describe_response = storcon_client
-                .dispatch::<(), NodeShardResponse>(
-                    Method::GET,
-                    format!("control/v1/node/{node_id}/shards"),
-                    None,
-                )
-                .await?;
-            let shards = describe_response.shards;
-            let mut table = comfy_table::Table::new();
-            table.set_header([
-                "Shard",
-                "Intended Primary/Secondary",
-                "Observed Primary/Secondary",
-            ]);
-            for shard in shards {
-                table.add_row([
-                    format!("{}", shard.tenant_shard_id),
-                    match shard.is_intended_secondary {
-                        None => "".to_string(),
-                        Some(true) => "Secondary".to_string(),
-                        Some(false) => "Primary".to_string(),
-                    },
-                    match shard.is_observed_secondary {
-                        None => "".to_string(),
-                        Some(true) => "Secondary".to_string(),
-                        Some(false) => "Primary".to_string(),
-                    },
-                ]);
-            }
-            println!("{table}");
-        }
-        Command::Tenants { node_id: None } => {
+        Command::Tenants {} => {
            let mut resp = storcon_client
                .dispatch::<(), Vec<TenantDescribeResponse>>(
                    Method::GET,
@@ -719,13 +622,12 @@ async fn main() -> anyhow::Result<()> {
                                threshold: threshold.into(),
                            },
                        )),
-                        heatmap_period: Some("300s".to_string()),
                        ..Default::default()
                    },
                })
                .await?;
        }
-        Command::BulkMigrate {
+        Command::Drain {
            nodes,
            concurrency,
            max_shards,
@@ -754,7 +656,7 @@ async fn main() -> anyhow::Result<()> {
            }

            if nodes.len() != node_to_drain_descs.len() {
-                anyhow::bail!("Bulk migration requested away from node which doesn't exist.")
+                anyhow::bail!("Drain requested for node which doesn't exist.")
            }

            node_to_fill_descs.retain(|desc| {
@@ -766,7 +668,7 @@ async fn main() -> anyhow::Result<()> {
            });

            if node_to_fill_descs.is_empty() {
-                anyhow::bail!("There are no nodes to migrate to")
+                anyhow::bail!("There are no nodes to drain to")
            }

            // Set the node scheduling policy to draining for the nodes which
@@ -787,7 +689,7 @@ async fn main() -> anyhow::Result<()> {
                    .await?;
            }

-            // Perform the migration: move each tenant shard scheduled on a node to
+            // Perform the drain: 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
@@ -800,13 +702,13 @@ async fn main() -> anyhow::Result<()> {

            let mut selected_node_idx = 0;

-            struct MigrationMove {
+            struct DrainMove {
                tenant_shard_id: TenantShardId,
                from: NodeId,
                to: NodeId,
            }

-            let mut moves: Vec<MigrationMove> = Vec::new();
+            let mut moves: Vec<DrainMove> = Vec::new();

            let shards = tenants
                .into_iter()
@@ -836,7 +738,7 @@ async fn main() -> anyhow::Result<()> {
                    continue;
                }

-                moves.push(MigrationMove {
+                moves.push(DrainMove {
                    tenant_shard_id: shard.tenant_shard_id,
                    from: shard
                        .node_attached
@@ -913,67 +815,6 @@ async fn main() -> anyhow::Result<()> {
                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(())
--- a/docker-compose/run-tests.sh
+++ b/docker-compose/run-tests.sh
@@ -3,7 +3,7 @@ set -x

 cd /ext-src || exit 2
 FAILED=
-LIST=$( (echo -e "${SKIP//","/"\n"}"; ls -d -- *-src) | sort | uniq -u)
+LIST=$( (echo "${SKIP//","/"\n"}"; ls -d -- *-src) | sort | uniq -u)
 for d in ${LIST}
 do
       [ -d "${d}" ] || continue
--- a/docs/rfcs/033-storage-controller-drain-and-fill.md
+++ b/docs/rfcs/033-storage-controller-drain-and-fill.md
@@ -14,7 +14,7 @@ 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
+[10 second systemd enforced timeout](https://github.com/neondatabase/aws/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
--- a/docs/rfcs/035-safekeeper-dynamic-membership-change.md
+++ b/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
--- a/docs/rfcs/036-physical-replication.md
+++ b/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).
--- a/docs/rfcs/037-storage-controller-restarts.md
+++ b/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.
--- a/docs/updating-postgres.md
+++ b/docs/updating-postgres.md
@@ -21,21 +21,30 @@ _Example: 15.4 is the new minor version to upgrade to from 15.3._
 1. Create a new branch based on the stable branch you are updating.

    ```shell
-    git checkout -b my-branch-15 REL_15_STABLE_neon
+    git checkout -b my-branch REL_15_STABLE_neon
    ```

-1. Find the upstream release tags you're looking for. They are of the form `REL_X_Y`.
+1. Tag the last commit on the stable branch you are updating.

-1. Merge the upstream tag into the branch you created on the tag and resolve any conflicts.
+    ```shell
+    git tag REL_15_3_neon
+    ```
+
+1. Push the new tag to the Neon Postgres repository.
+
+    ```shell
+    git push origin REL_15_3_neon
+    ```
+
+1. Find the release tags you're looking for. They are of the form `REL_X_Y`.
+
+1. Rebase the branch you created on the tag and resolve any conflicts.

    ```shell
    git fetch upstream REL_15_4
-    git merge REL_15_4
+    git rebase REL_15_4
    ```

-    In the commit message of the merge commit, mention if there were
-    any non-trivial conflicts or other issues.
-
 1. Run the Postgres test suite to make sure our commits have not affected
 Postgres in a negative way.

@@ -48,7 +57,7 @@ Postgres in a negative way.
 1. Push your branch to the Neon Postgres repository.

    ```shell
-    git push origin my-branch-15
+    git push origin my-branch
    ```

 1. Clone the Neon repository if you have not done so already.
@@ -65,7 +74,7 @@ branch.
 1. Update the Git submodule.

    ```shell
-    git submodule set-branch --branch my-branch-15 vendor/postgres-v15
+    git submodule set-branch --branch my-branch vendor/postgres-v15
    git submodule update --remote vendor/postgres-v15
    ```

@@ -80,12 +89,14 @@ minor Postgres release.

 1. Create a pull request, and wait for CI to go green.

-1. Push the Postgres branches with the merge commits into the Neon Postgres repository.
+1. Force push the rebased Postgres branches into the Neon Postgres repository.

    ```shell
-    git push origin my-branch-15:REL_15_STABLE_neon
+    git push --force origin my-branch:REL_15_STABLE_neon
    ```

+    It may require disabling various branch protections.
+
 1. Update your Neon PR to point at the branches.

    ```shell
--- a/libs/compute_api/Cargo.toml
+++ b/libs/compute_api/Cargo.toml
@@ -14,3 +14,5 @@ regex.workspace = true

 utils = { path = "../utils" }
 remote_storage = { version = "0.1", path = "../remote_storage/" }
+
+workspace_hack.workspace = true
--- a/libs/consumption_metrics/Cargo.toml
+++ b/libs/consumption_metrics/Cargo.toml
@@ -6,8 +6,10 @@ license = "Apache-2.0"

 [dependencies]
 anyhow.workspace = true
-chrono = { workspace = true, features = ["serde"] }
+chrono.workspace = true
 rand.workspace = true
 serde.workspace = true
 serde_with.workspace = true
 utils.workspace = true
+
+workspace_hack.workspace = true
--- a/libs/desim/Cargo.toml
+++ b/libs/desim/Cargo.toml
@@ -14,3 +14,5 @@ parking_lot.workspace = true
 hex.workspace = true
 scopeguard.workspace = true
 smallvec = { workspace = true, features = ["write"] }
+
+workspace_hack.workspace = true
--- a/libs/metrics/Cargo.toml
+++ b/libs/metrics/Cargo.toml
@@ -12,6 +12,8 @@ chrono.workspace = true
 twox-hash.workspace = true
 measured.workspace = true

+workspace_hack.workspace = true
+
 [target.'cfg(target_os = "linux")'.dependencies]
 procfs.workspace = true
 measured-process.workspace = true
--- a/libs/metrics/src/lib.rs
+++ b/libs/metrics/src/lib.rs
@@ -68,7 +68,6 @@ macro_rules! register_uint_gauge {
 static INTERNAL_REGISTRY: Lazy<Registry> = Lazy::new(Registry::new);

 /// Register a collector in the internal registry. MUST be called before the first call to `gather()`.
-///
 /// Otherwise, we can have a deadlock in the `gather()` call, trying to register a new collector
 /// while holding the lock.
 pub fn register_internal(c: Box<dyn Collector>) -> prometheus::Result<()> {
--- a/libs/pageserver_api/Cargo.toml
+++ b/libs/pageserver_api/Cargo.toml
@@ -4,10 +4,6 @@ version = "0.1.0"
 edition.workspace = true
 license.workspace = true

-[features]
-# See pageserver/Cargo.toml
-testing = ["dep:nix"]
-
 [dependencies]
 serde.workspace = true
 serde_with.workspace = true
@@ -25,14 +21,10 @@ hex.workspace = true
 humantime.workspace = true
 thiserror.workspace = true
 humantime-serde.workspace = true
-chrono = { workspace = true, features = ["serde"] }
+chrono.workspace = true
 itertools.workspace = true
-storage_broker.workspace = true
-camino = {workspace = true, features = ["serde1"]}
-remote_storage.workspace = true
-postgres_backend.workspace = true
-nix = {workspace = true, optional = true}
-reqwest.workspace = true
+
+workspace_hack.workspace = true

 [dev-dependencies]
 bincode.workspace = true
--- a/libs/pageserver_api/src/config.rs
+++ b/libs/pageserver_api/src/config.rs
@@ -1,28 +1,15 @@
-use camino::Utf8PathBuf;
+use std::collections::HashMap;
+
+use const_format::formatcp;

 #[cfg(test)]
 mod tests;

-use const_format::formatcp;
 pub const DEFAULT_PG_LISTEN_PORT: u16 = 64000;
 pub const DEFAULT_PG_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_PG_LISTEN_PORT}");
 pub const DEFAULT_HTTP_LISTEN_PORT: u16 = 9898;
 pub const DEFAULT_HTTP_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_HTTP_LISTEN_PORT}");

-use postgres_backend::AuthType;
-use remote_storage::RemoteStorageConfig;
-use serde_with::serde_as;
-use std::{
-    collections::HashMap,
-    num::{NonZeroU64, NonZeroUsize},
-    str::FromStr,
-    time::Duration,
-};
-use utils::logging::LogFormat;
-
-use crate::models::ImageCompressionAlgorithm;
-use crate::models::LsnLease;
-
 // Certain metadata (e.g. externally-addressable name, AZ) is delivered
 // as a separate structure.  This information is not neeed by the pageserver
 // itself, it is only used for registering the pageserver with the control
@@ -42,476 +29,3 @@ pub struct NodeMetadata {
    #[serde(flatten)]
    pub other: HashMap<String, serde_json::Value>,
 }
-
-/// `pageserver.toml`
-///
-/// We use serde derive with `#[serde(default)]` to generate a deserializer
-/// that fills in the default values for each config field.
-///
-/// If there cannot be a static default value because we need to make runtime
-/// checks to determine the default, make it an `Option` (which defaults to None).
-/// The runtime check should be done in the consuming crate, i.e., `pageserver`.
-#[serde_as]
-#[derive(Clone, Debug, serde::Deserialize, serde::Serialize)]
-#[serde(default, deny_unknown_fields)]
-pub struct ConfigToml {
-    // types mapped 1:1 into the runtime PageServerConfig type
-    pub listen_pg_addr: String,
-    pub listen_http_addr: String,
-    pub availability_zone: Option<String>,
-    #[serde(with = "humantime_serde")]
-    pub wait_lsn_timeout: Duration,
-    #[serde(with = "humantime_serde")]
-    pub wal_redo_timeout: Duration,
-    pub superuser: String,
-    pub page_cache_size: usize,
-    pub max_file_descriptors: usize,
-    pub pg_distrib_dir: Option<Utf8PathBuf>,
-    #[serde_as(as = "serde_with::DisplayFromStr")]
-    pub http_auth_type: AuthType,
-    #[serde_as(as = "serde_with::DisplayFromStr")]
-    pub pg_auth_type: AuthType,
-    pub auth_validation_public_key_path: Option<Utf8PathBuf>,
-    pub remote_storage: Option<RemoteStorageConfig>,
-    pub tenant_config: TenantConfigToml,
-    #[serde_as(as = "serde_with::DisplayFromStr")]
-    pub broker_endpoint: storage_broker::Uri,
-    #[serde(with = "humantime_serde")]
-    pub broker_keepalive_interval: Duration,
-    #[serde_as(as = "serde_with::DisplayFromStr")]
-    pub log_format: LogFormat,
-    pub concurrent_tenant_warmup: NonZeroUsize,
-    pub concurrent_tenant_size_logical_size_queries: NonZeroUsize,
-    #[serde(with = "humantime_serde")]
-    pub metric_collection_interval: Duration,
-    pub metric_collection_endpoint: Option<reqwest::Url>,
-    pub metric_collection_bucket: Option<RemoteStorageConfig>,
-    #[serde(with = "humantime_serde")]
-    pub synthetic_size_calculation_interval: Duration,
-    pub disk_usage_based_eviction: Option<DiskUsageEvictionTaskConfig>,
-    pub test_remote_failures: u64,
-    pub ondemand_download_behavior_treat_error_as_warn: bool,
-    #[serde(with = "humantime_serde")]
-    pub background_task_maximum_delay: Duration,
-    pub control_plane_api: Option<reqwest::Url>,
-    pub control_plane_api_token: Option<String>,
-    pub control_plane_emergency_mode: bool,
-    pub heatmap_upload_concurrency: usize,
-    pub secondary_download_concurrency: usize,
-    pub virtual_file_io_engine: Option<crate::models::virtual_file::IoEngineKind>,
-    pub ingest_batch_size: u64,
-    pub max_vectored_read_bytes: MaxVectoredReadBytes,
-    pub image_compression: ImageCompressionAlgorithm,
-    pub ephemeral_bytes_per_memory_kb: usize,
-    pub l0_flush: Option<crate::models::L0FlushConfig>,
-    #[serde(skip_serializing)]
-    // TODO(https://github.com/neondatabase/neon/issues/8184): remove after this field is removed from all pageserver.toml's
-    pub compact_level0_phase1_value_access: serde::de::IgnoredAny,
-    pub virtual_file_direct_io: crate::models::virtual_file::DirectIoMode,
-    pub io_buffer_alignment: usize,
-}
-
-#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-#[serde(deny_unknown_fields)]
-pub struct DiskUsageEvictionTaskConfig {
-    pub max_usage_pct: utils::serde_percent::Percent,
-    pub min_avail_bytes: u64,
-    #[serde(with = "humantime_serde")]
-    pub period: Duration,
-    #[cfg(feature = "testing")]
-    pub mock_statvfs: Option<statvfs::mock::Behavior>,
-    /// Select sorting for evicted layers
-    #[serde(default)]
-    pub eviction_order: EvictionOrder,
-}
-
-pub mod statvfs {
-    pub mod mock {
-        #[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-        #[serde(tag = "type")]
-        pub enum Behavior {
-            Success {
-                blocksize: u64,
-                total_blocks: u64,
-                name_filter: Option<utils::serde_regex::Regex>,
-            },
-            #[cfg(feature = "testing")]
-            Failure { mocked_error: MockedError },
-        }
-
-        #[cfg(feature = "testing")]
-        #[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-        #[allow(clippy::upper_case_acronyms)]
-        pub enum MockedError {
-            EIO,
-        }
-
-        #[cfg(feature = "testing")]
-        impl From<MockedError> for nix::Error {
-            fn from(e: MockedError) -> Self {
-                match e {
-                    MockedError::EIO => nix::Error::EIO,
-                }
-            }
-        }
-    }
-}
-
-#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-#[serde(tag = "type", content = "args")]
-pub enum EvictionOrder {
-    RelativeAccessed {
-        highest_layer_count_loses_first: bool,
-    },
-}
-
-impl Default for EvictionOrder {
-    fn default() -> Self {
-        Self::RelativeAccessed {
-            highest_layer_count_loses_first: true,
-        }
-    }
-}
-
-#[derive(
-    Eq,
-    PartialEq,
-    Debug,
-    Copy,
-    Clone,
-    strum_macros::EnumString,
-    strum_macros::Display,
-    serde_with::DeserializeFromStr,
-    serde_with::SerializeDisplay,
-)]
-#[strum(serialize_all = "kebab-case")]
-pub enum GetVectoredImpl {
-    Sequential,
-    Vectored,
-}
-
-#[derive(
-    Eq,
-    PartialEq,
-    Debug,
-    Copy,
-    Clone,
-    strum_macros::EnumString,
-    strum_macros::Display,
-    serde_with::DeserializeFromStr,
-    serde_with::SerializeDisplay,
-)]
-#[strum(serialize_all = "kebab-case")]
-pub enum GetImpl {
-    Legacy,
-    Vectored,
-}
-
-#[derive(Copy, Clone, Debug, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-#[serde(transparent)]
-pub struct MaxVectoredReadBytes(pub NonZeroUsize);
-
-/// A tenant's calcuated configuration, which is the result of merging a
-/// tenant's TenantConfOpt with the global TenantConf from PageServerConf.
-///
-/// For storing and transmitting individual tenant's configuration, see
-/// TenantConfOpt.
-#[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
-#[serde(deny_unknown_fields, default)]
-pub struct TenantConfigToml {
-    // Flush out an inmemory layer, if it's holding WAL older than this
-    // This puts a backstop on how much WAL needs to be re-digested if the
-    // page server crashes.
-    // This parameter actually determines L0 layer file size.
-    pub checkpoint_distance: u64,
-    // Inmemory layer is also flushed at least once in checkpoint_timeout to
-    // eventually upload WAL after activity is stopped.
-    #[serde(with = "humantime_serde")]
-    pub checkpoint_timeout: Duration,
-    // Target file size, when creating image and delta layers.
-    // This parameter determines L1 layer file size.
-    pub compaction_target_size: u64,
-    // How often to check if there's compaction work to be done.
-    // Duration::ZERO means automatic compaction is disabled.
-    #[serde(with = "humantime_serde")]
-    pub compaction_period: Duration,
-    // Level0 delta layer threshold for compaction.
-    pub compaction_threshold: usize,
-    pub compaction_algorithm: crate::models::CompactionAlgorithmSettings,
-    // Determines how much history is retained, to allow
-    // branching and read replicas at an older point in time.
-    // The unit is #of bytes of WAL.
-    // Page versions older than this are garbage collected away.
-    pub gc_horizon: u64,
-    // Interval at which garbage collection is triggered.
-    // Duration::ZERO means automatic GC is disabled
-    #[serde(with = "humantime_serde")]
-    pub gc_period: Duration,
-    // Delta layer churn threshold to create L1 image layers.
-    pub image_creation_threshold: usize,
-    // Determines how much history is retained, to allow
-    // branching and read replicas at an older point in time.
-    // The unit is time.
-    // Page versions older than this are garbage collected away.
-    #[serde(with = "humantime_serde")]
-    pub pitr_interval: Duration,
-    /// Maximum amount of time to wait while opening a connection to receive wal, before erroring.
-    #[serde(with = "humantime_serde")]
-    pub walreceiver_connect_timeout: Duration,
-    /// Considers safekeepers stalled after no WAL updates were received longer than this threshold.
-    /// A stalled safekeeper will be changed to a newer one when it appears.
-    #[serde(with = "humantime_serde")]
-    pub lagging_wal_timeout: Duration,
-    /// Considers safekeepers lagging when their WAL is behind another safekeeper for more than this threshold.
-    /// A lagging safekeeper will be changed after `lagging_wal_timeout` time elapses since the last WAL update,
-    /// to avoid eager reconnects.
-    pub max_lsn_wal_lag: NonZeroU64,
-    pub eviction_policy: crate::models::EvictionPolicy,
-    pub min_resident_size_override: Option<u64>,
-    // See the corresponding metric's help string.
-    #[serde(with = "humantime_serde")]
-    pub evictions_low_residence_duration_metric_threshold: Duration,
-
-    /// If non-zero, the period between uploads of a heatmap from attached tenants.  This
-    /// may be disabled if a Tenant will not have secondary locations: only secondary
-    /// locations will use the heatmap uploaded by attached locations.
-    #[serde(with = "humantime_serde")]
-    pub heatmap_period: Duration,
-
-    /// If true then SLRU segments are dowloaded on demand, if false SLRU segments are included in basebackup
-    pub lazy_slru_download: bool,
-
-    pub timeline_get_throttle: crate::models::ThrottleConfig,
-
-    // How much WAL must be ingested before checking again whether a new image layer is required.
-    // Expresed in multiples of checkpoint distance.
-    pub image_layer_creation_check_threshold: u8,
-
-    /// Switch to a new aux file policy. Switching this flag requires the user has not written any aux file into
-    /// the storage before, and this flag cannot be switched back. Otherwise there will be data corruptions.
-    /// There is a `last_aux_file_policy` flag which gets persisted in `index_part.json` once the first aux
-    /// file is written.
-    pub switch_aux_file_policy: crate::models::AuxFilePolicy,
-
-    /// The length for an explicit LSN lease request.
-    /// Layers needed to reconstruct pages at LSN will not be GC-ed during this interval.
-    #[serde(with = "humantime_serde")]
-    pub lsn_lease_length: Duration,
-
-    /// The length for an implicit LSN lease granted as part of `get_lsn_by_timestamp` request.
-    /// Layers needed to reconstruct pages at LSN will not be GC-ed during this interval.
-    #[serde(with = "humantime_serde")]
-    pub lsn_lease_length_for_ts: Duration,
-}
-
-pub mod defaults {
-    use crate::models::ImageCompressionAlgorithm;
-
-    pub use storage_broker::DEFAULT_ENDPOINT as BROKER_DEFAULT_ENDPOINT;
-
-    pub const DEFAULT_WAIT_LSN_TIMEOUT: &str = "300 s";
-    pub const DEFAULT_WAL_REDO_TIMEOUT: &str = "60 s";
-
-    pub const DEFAULT_SUPERUSER: &str = "cloud_admin";
-
-    pub const DEFAULT_PAGE_CACHE_SIZE: usize = 8192;
-    pub const DEFAULT_MAX_FILE_DESCRIPTORS: usize = 100;
-
-    pub const DEFAULT_LOG_FORMAT: &str = "plain";
-
-    pub const DEFAULT_CONCURRENT_TENANT_WARMUP: usize = 8;
-
-    pub const DEFAULT_CONCURRENT_TENANT_SIZE_LOGICAL_SIZE_QUERIES: usize = 1;
-
-    pub const DEFAULT_METRIC_COLLECTION_INTERVAL: &str = "10 min";
-    pub const DEFAULT_METRIC_COLLECTION_ENDPOINT: Option<reqwest::Url> = None;
-    pub const DEFAULT_SYNTHETIC_SIZE_CALCULATION_INTERVAL: &str = "10 min";
-    pub const DEFAULT_BACKGROUND_TASK_MAXIMUM_DELAY: &str = "10s";
-
-    pub const DEFAULT_HEATMAP_UPLOAD_CONCURRENCY: usize = 8;
-    pub const DEFAULT_SECONDARY_DOWNLOAD_CONCURRENCY: usize = 1;
-
-    pub const DEFAULT_INGEST_BATCH_SIZE: u64 = 100;
-
-    pub const DEFAULT_MAX_VECTORED_READ_BYTES: usize = 128 * 1024; // 128 KiB
-
-    pub const DEFAULT_IMAGE_COMPRESSION: ImageCompressionAlgorithm =
-        ImageCompressionAlgorithm::Zstd { level: Some(1) };
-
-    pub const DEFAULT_VALIDATE_VECTORED_GET: bool = false;
-
-    pub const DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB: usize = 0;
-
-    pub const DEFAULT_IO_BUFFER_ALIGNMENT: usize = 512;
-}
-
-impl Default for ConfigToml {
-    fn default() -> Self {
-        use defaults::*;
-
-        Self {
-            listen_pg_addr: (DEFAULT_PG_LISTEN_ADDR.to_string()),
-            listen_http_addr: (DEFAULT_HTTP_LISTEN_ADDR.to_string()),
-            availability_zone: (None),
-            wait_lsn_timeout: (humantime::parse_duration(DEFAULT_WAIT_LSN_TIMEOUT)
-                .expect("cannot parse default wait lsn timeout")),
-            wal_redo_timeout: (humantime::parse_duration(DEFAULT_WAL_REDO_TIMEOUT)
-                .expect("cannot parse default wal redo timeout")),
-            superuser: (DEFAULT_SUPERUSER.to_string()),
-            page_cache_size: (DEFAULT_PAGE_CACHE_SIZE),
-            max_file_descriptors: (DEFAULT_MAX_FILE_DESCRIPTORS),
-            pg_distrib_dir: None, // Utf8PathBuf::from("./pg_install"), // TODO: formely, this was std::env::current_dir()
-            http_auth_type: (AuthType::Trust),
-            pg_auth_type: (AuthType::Trust),
-            auth_validation_public_key_path: (None),
-            remote_storage: None,
-            broker_endpoint: (storage_broker::DEFAULT_ENDPOINT
-                .parse()
-                .expect("failed to parse default broker endpoint")),
-            broker_keepalive_interval: (humantime::parse_duration(
-                storage_broker::DEFAULT_KEEPALIVE_INTERVAL,
-            )
-            .expect("cannot parse default keepalive interval")),
-            log_format: (LogFormat::from_str(DEFAULT_LOG_FORMAT).unwrap()),
-
-            concurrent_tenant_warmup: (NonZeroUsize::new(DEFAULT_CONCURRENT_TENANT_WARMUP)
-                .expect("Invalid default constant")),
-            concurrent_tenant_size_logical_size_queries: NonZeroUsize::new(1).unwrap(),
-            metric_collection_interval: (humantime::parse_duration(
-                DEFAULT_METRIC_COLLECTION_INTERVAL,
-            )
-            .expect("cannot parse default metric collection interval")),
-            synthetic_size_calculation_interval: (humantime::parse_duration(
-                DEFAULT_SYNTHETIC_SIZE_CALCULATION_INTERVAL,
-            )
-            .expect("cannot parse default synthetic size calculation interval")),
-            metric_collection_endpoint: (DEFAULT_METRIC_COLLECTION_ENDPOINT),
-
-            metric_collection_bucket: (None),
-
-            disk_usage_based_eviction: (None),
-
-            test_remote_failures: (0),
-
-            ondemand_download_behavior_treat_error_as_warn: (false),
-
-            background_task_maximum_delay: (humantime::parse_duration(
-                DEFAULT_BACKGROUND_TASK_MAXIMUM_DELAY,
-            )
-            .unwrap()),
-
-            control_plane_api: (None),
-            control_plane_api_token: (None),
-            control_plane_emergency_mode: (false),
-
-            heatmap_upload_concurrency: (DEFAULT_HEATMAP_UPLOAD_CONCURRENCY),
-            secondary_download_concurrency: (DEFAULT_SECONDARY_DOWNLOAD_CONCURRENCY),
-
-            ingest_batch_size: (DEFAULT_INGEST_BATCH_SIZE),
-
-            virtual_file_io_engine: None,
-
-            max_vectored_read_bytes: (MaxVectoredReadBytes(
-                NonZeroUsize::new(DEFAULT_MAX_VECTORED_READ_BYTES).unwrap(),
-            )),
-            image_compression: (DEFAULT_IMAGE_COMPRESSION),
-            ephemeral_bytes_per_memory_kb: (DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB),
-            l0_flush: None,
-            compact_level0_phase1_value_access: Default::default(),
-            virtual_file_direct_io: crate::models::virtual_file::DirectIoMode::default(),
-
-            io_buffer_alignment: DEFAULT_IO_BUFFER_ALIGNMENT,
-
-            tenant_config: TenantConfigToml::default(),
-        }
-    }
-}
-
-pub mod tenant_conf_defaults {
-
-    // FIXME: This current value is very low. I would imagine something like 1 GB or 10 GB
-    // would be more appropriate. But a low value forces the code to be exercised more,
-    // which is good for now to trigger bugs.
-    // This parameter actually determines L0 layer file size.
-    pub const DEFAULT_CHECKPOINT_DISTANCE: u64 = 256 * 1024 * 1024;
-    pub const DEFAULT_CHECKPOINT_TIMEOUT: &str = "10 m";
-
-    // FIXME the below configs are only used by legacy algorithm. The new algorithm
-    // has different parameters.
-
-    // Target file size, when creating image and delta layers.
-    // This parameter determines L1 layer file size.
-    pub const DEFAULT_COMPACTION_TARGET_SIZE: u64 = 128 * 1024 * 1024;
-
-    pub const DEFAULT_COMPACTION_PERIOD: &str = "20 s";
-    pub const DEFAULT_COMPACTION_THRESHOLD: usize = 10;
-    pub const DEFAULT_COMPACTION_ALGORITHM: crate::models::CompactionAlgorithm =
-        crate::models::CompactionAlgorithm::Legacy;
-
-    pub const DEFAULT_GC_HORIZON: u64 = 64 * 1024 * 1024;
-
-    // Large DEFAULT_GC_PERIOD is fine as long as PITR_INTERVAL is larger.
-    // If there's a need to decrease this value, first make sure that GC
-    // doesn't hold a layer map write lock for non-trivial operations.
-    // Relevant: https://github.com/neondatabase/neon/issues/3394
-    pub const DEFAULT_GC_PERIOD: &str = "1 hr";
-    pub const DEFAULT_IMAGE_CREATION_THRESHOLD: usize = 3;
-    pub const DEFAULT_PITR_INTERVAL: &str = "7 days";
-    pub const DEFAULT_WALRECEIVER_CONNECT_TIMEOUT: &str = "10 seconds";
-    pub const DEFAULT_WALRECEIVER_LAGGING_WAL_TIMEOUT: &str = "10 seconds";
-    // The default limit on WAL lag should be set to avoid causing disconnects under high throughput
-    // scenarios: since the broker stats are updated ~1/s, a value of 1GiB should be sufficient for
-    // throughputs up to 1GiB/s per timeline.
-    pub const DEFAULT_MAX_WALRECEIVER_LSN_WAL_LAG: u64 = 1024 * 1024 * 1024;
-    pub const DEFAULT_EVICTIONS_LOW_RESIDENCE_DURATION_METRIC_THRESHOLD: &str = "24 hour";
-    // By default ingest enough WAL for two new L0 layers before checking if new image
-    // image layers should be created.
-    pub const DEFAULT_IMAGE_LAYER_CREATION_CHECK_THRESHOLD: u8 = 2;
-
-    pub const DEFAULT_INGEST_BATCH_SIZE: u64 = 100;
-}
-
-impl Default for TenantConfigToml {
-    fn default() -> Self {
-        use tenant_conf_defaults::*;
-        Self {
-            checkpoint_distance: DEFAULT_CHECKPOINT_DISTANCE,
-            checkpoint_timeout: humantime::parse_duration(DEFAULT_CHECKPOINT_TIMEOUT)
-                .expect("cannot parse default checkpoint timeout"),
-            compaction_target_size: DEFAULT_COMPACTION_TARGET_SIZE,
-            compaction_period: humantime::parse_duration(DEFAULT_COMPACTION_PERIOD)
-                .expect("cannot parse default compaction period"),
-            compaction_threshold: DEFAULT_COMPACTION_THRESHOLD,
-            compaction_algorithm: crate::models::CompactionAlgorithmSettings {
-                kind: DEFAULT_COMPACTION_ALGORITHM,
-            },
-            gc_horizon: DEFAULT_GC_HORIZON,
-            gc_period: humantime::parse_duration(DEFAULT_GC_PERIOD)
-                .expect("cannot parse default gc period"),
-            image_creation_threshold: DEFAULT_IMAGE_CREATION_THRESHOLD,
-            pitr_interval: humantime::parse_duration(DEFAULT_PITR_INTERVAL)
-                .expect("cannot parse default PITR interval"),
-            walreceiver_connect_timeout: humantime::parse_duration(
-                DEFAULT_WALRECEIVER_CONNECT_TIMEOUT,
-            )
-            .expect("cannot parse default walreceiver connect timeout"),
-            lagging_wal_timeout: humantime::parse_duration(DEFAULT_WALRECEIVER_LAGGING_WAL_TIMEOUT)
-                .expect("cannot parse default walreceiver lagging wal timeout"),
-            max_lsn_wal_lag: NonZeroU64::new(DEFAULT_MAX_WALRECEIVER_LSN_WAL_LAG)
-                .expect("cannot parse default max walreceiver Lsn wal lag"),
-            eviction_policy: crate::models::EvictionPolicy::NoEviction,
-            min_resident_size_override: None,
-            evictions_low_residence_duration_metric_threshold: humantime::parse_duration(
-                DEFAULT_EVICTIONS_LOW_RESIDENCE_DURATION_METRIC_THRESHOLD,
-            )
-            .expect("cannot parse default evictions_low_residence_duration_metric_threshold"),
-            heatmap_period: Duration::ZERO,
-            lazy_slru_download: false,
-            timeline_get_throttle: crate::models::ThrottleConfig::disabled(),
-            image_layer_creation_check_threshold: DEFAULT_IMAGE_LAYER_CREATION_CHECK_THRESHOLD,
-            switch_aux_file_policy: crate::models::AuxFilePolicy::default_tenant_config(),
-            lsn_lease_length: LsnLease::DEFAULT_LENGTH,
-            lsn_lease_length_for_ts: LsnLease::DEFAULT_LENGTH_FOR_TS,
-        }
-    }
-}
--- a/libs/pageserver_api/src/controller_api.rs
+++ b/libs/pageserver_api/src/controller_api.rs
@@ -1,4 +1,4 @@
-use std::collections::{HashMap, HashSet};
+use std::collections::HashSet;
 use std::str::FromStr;
 use std::time::{Duration, Instant};

@@ -8,7 +8,6 @@ use std::time::{Duration, Instant};
 use serde::{Deserialize, Serialize};
 use utils::id::{NodeId, TenantId};

-use crate::models::PageserverUtilization;
 use crate::{
    models::{ShardParameters, TenantConfig},
    shard::{ShardStripeSize, TenantShardId},
@@ -56,8 +55,6 @@ pub struct NodeRegisterRequest {

    pub listen_http_addr: String,
    pub listen_http_port: u16,
-
-    pub availability_zone_id: String,
 }

 #[derive(Serialize, Deserialize)]
@@ -74,17 +71,6 @@ pub struct TenantPolicyRequest {
    pub scheduling: Option<ShardSchedulingPolicy>,
 }

-#[derive(Serialize, Deserialize)]
-pub struct ShardsPreferredAzsRequest {
-    #[serde(flatten)]
-    pub preferred_az_ids: HashMap<TenantShardId, String>,
-}
-
-#[derive(Serialize, Deserialize)]
-pub struct ShardsPreferredAzsResponse {
-    pub updated: Vec<TenantShardId>,
-}
-
 #[derive(Serialize, Deserialize, Debug)]
 pub struct TenantLocateResponseShard {
    pub shard_id: TenantShardId,
@@ -112,21 +98,6 @@ pub struct TenantDescribeResponse {
    pub config: TenantConfig,
 }

-#[derive(Serialize, Deserialize, Debug)]
-pub struct NodeShardResponse {
-    pub node_id: NodeId,
-    pub shards: Vec<NodeShard>,
-}
-
-#[derive(Serialize, Deserialize, Debug)]
-pub struct NodeShard {
-    pub tenant_shard_id: TenantShardId,
-    /// Whether the shard is observed secondary on a specific node. True = yes, False = no, None = not on this node.
-    pub is_observed_secondary: Option<bool>,
-    /// Whether the shard is intended to be a secondary on a specific node. True = yes, False = no, None = not on this node.
-    pub is_intended_secondary: Option<bool>,
-}
-
 #[derive(Serialize, Deserialize)]
 pub struct NodeDescribeResponse {
    pub id: NodeId,
@@ -158,12 +129,8 @@ pub struct TenantDescribeResponseShard {
    pub is_splitting: bool,

    pub scheduling_policy: ShardSchedulingPolicy,
-
-    pub preferred_az_id: Option<String>,
 }

-/// Migration request for a given tenant shard to a given node.
-///
 /// Explicitly migrating a particular shard is a low level operation
 /// TODO: higher level "Reschedule tenant" operation where the request
 /// specifies some constraints, e.g. asking it to get off particular node(s)
@@ -173,11 +140,23 @@ pub struct TenantShardMigrateRequest {
    pub node_id: NodeId,
 }

-#[derive(Serialize, Clone, Debug)]
+/// Utilisation score indicating how good a candidate a pageserver
+/// is for scheduling the next tenant. See [`crate::models::PageserverUtilization`].
+/// Lower values are better.
+#[derive(Serialize, Deserialize, Clone, Copy, Eq, PartialEq, PartialOrd, Ord, Debug)]
+pub struct UtilizationScore(pub u64);
+
+impl UtilizationScore {
+    pub fn worst() -> Self {
+        UtilizationScore(u64::MAX)
+    }
+}
+
+#[derive(Serialize, Clone, Copy, Debug)]
 #[serde(into = "NodeAvailabilityWrapper")]
 pub enum NodeAvailability {
    // Normal, happy state
-    Active(PageserverUtilization),
+    Active(UtilizationScore),
    // Node is warming up, but we expect it to become available soon. Covers
    // the time span between the re-attach response being composed on the storage controller
    // and the first successful heartbeat after the processing of the re-attach response
@@ -216,9 +195,7 @@ impl From<NodeAvailabilityWrapper> for NodeAvailability {
        match val {
            // Assume the worst utilisation score to begin with. It will later be updated by
            // the heartbeats.
-            NodeAvailabilityWrapper::Active => {
-                NodeAvailability::Active(PageserverUtilization::full())
-            }
+            NodeAvailabilityWrapper::Active => NodeAvailability::Active(UtilizationScore::worst()),
            NodeAvailabilityWrapper::WarmingUp => NodeAvailability::WarmingUp(Instant::now()),
            NodeAvailabilityWrapper::Offline => NodeAvailability::Offline,
        }
@@ -336,17 +313,20 @@ pub struct MetadataHealthUpdateRequest {
 pub struct MetadataHealthUpdateResponse {}

 #[derive(Serialize, Deserialize, Debug)]
+
 pub struct MetadataHealthListUnhealthyResponse {
    pub unhealthy_tenant_shards: Vec<TenantShardId>,
 }

 #[derive(Serialize, Deserialize, Debug)]
+
 pub struct MetadataHealthListOutdatedRequest {
    #[serde(with = "humantime_serde")]
    pub not_scrubbed_for: Duration,
 }

 #[derive(Serialize, Deserialize, Debug)]
+
 pub struct MetadataHealthListOutdatedResponse {
    pub health_records: Vec<MetadataHealthRecord>,
 }
--- a/libs/pageserver_api/src/key.rs
+++ b/libs/pageserver_api/src/key.rs
@@ -22,11 +22,6 @@ pub struct Key {
    pub field6: u32,
 }

-/// When working with large numbers of Keys in-memory, it is more efficient to handle them as i128 than as
-/// a struct of fields.
-#[derive(Clone, Copy, Hash, PartialEq, Eq, Ord, PartialOrd)]
-pub struct CompactKey(i128);
-
 /// The storage key size.
 pub const KEY_SIZE: usize = 18;

@@ -108,41 +103,14 @@ impl Key {
        }
    }

-    /// This function checks more extensively what keys we can take on the write path.
-    /// If a key beginning with 00 does not have a global/default tablespace OID, it
-    /// will be rejected on the write path.
-    #[allow(dead_code)]
-    pub fn is_valid_key_on_write_path_strong(&self) -> bool {
-        use postgres_ffi::pg_constants::{DEFAULTTABLESPACE_OID, GLOBALTABLESPACE_OID};
-        if !self.is_i128_representable() {
-            return false;
-        }
-        if self.field1 == 0
-            && !(self.field2 == GLOBALTABLESPACE_OID
-                || self.field2 == DEFAULTTABLESPACE_OID
-                || self.field2 == 0)
-        {
-            return false; // User defined tablespaces are not supported
-        }
-        true
-    }
-
-    /// This is a weaker version of `is_valid_key_on_write_path_strong` that simply
-    /// checks if the key is i128 representable. Note that some keys can be successfully
-    /// ingested into the pageserver, but will cause errors on generating basebackup.
-    pub fn is_valid_key_on_write_path(&self) -> bool {
-        self.is_i128_representable()
-    }
-
-    pub fn is_i128_representable(&self) -> bool {
-        self.field2 <= 0xFFFF || self.field2 == 0xFFFFFFFF || self.field2 == 0x22222222
-    }
-
    /// 'field2' is used to store tablespaceid for relations and small enum numbers for other relish.
    /// As long as Neon does not support tablespace (because of lack of access to local file system),
    /// we can assume that only some predefined namespace OIDs are used which can fit in u16
    pub fn to_i128(&self) -> i128 {
-        assert!(self.is_i128_representable(), "invalid key: {self}");
+        assert!(
+            self.field2 <= 0xFFFF || self.field2 == 0xFFFFFFFF || self.field2 == 0x22222222,
+            "invalid key: {self}",
+        );
        (((self.field1 & 0x7F) as i128) << 120)
            | (((self.field2 & 0xFFFF) as i128) << 104)
            | ((self.field3 as i128) << 72)
@@ -162,14 +130,6 @@ impl Key {
        }
    }

-    pub fn to_compact(&self) -> CompactKey {
-        CompactKey(self.to_i128())
-    }
-
-    pub fn from_compact(k: CompactKey) -> Self {
-        Self::from_i128(k.0)
-    }
-
    pub const fn next(&self) -> Key {
        self.add(1)
    }
@@ -239,13 +199,6 @@ impl fmt::Display for Key {
    }
 }

-impl fmt::Display for CompactKey {
-    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
-        let k = Key::from_compact(*self);
-        k.fmt(f)
-    }
-}
-
 impl Key {
    pub const MIN: Key = Key {
        field1: u8::MIN,
@@ -263,15 +216,6 @@ impl Key {
        field5: u8::MAX,
        field6: u32::MAX,
    };
-    /// A key slightly smaller than [`Key::MAX`] for use in layer key ranges to avoid them to be confused with L0 layers
-    pub const NON_L0_MAX: Key = Key {
-        field1: u8::MAX,
-        field2: u32::MAX,
-        field3: u32::MAX,
-        field4: u32::MAX,
-        field5: u8::MAX,
-        field6: u32::MAX - 1,
-    };

    pub fn from_hex(s: &str) -> Result<Self> {
        if s.len() != 36 {
--- a/libs/pageserver_api/src/models.rs
+++ b/libs/pageserver_api/src/models.rs
@@ -6,9 +6,8 @@ pub use utilization::PageserverUtilization;

 use std::{
    collections::HashMap,
-    fmt::Display,
    io::{BufRead, Read},
-    num::{NonZeroU32, NonZeroU64, NonZeroUsize},
+    num::{NonZeroU64, NonZeroUsize},
    str::FromStr,
    sync::atomic::AtomicUsize,
    time::{Duration, SystemTime},
@@ -305,10 +304,8 @@ pub struct TenantConfig {
    pub lsn_lease_length_for_ts: Option<String>,
 }

-/// The policy for the aux file storage.
-///
-/// It can be switched through `switch_aux_file_policy` tenant config.
-/// When the first aux file written, the policy will be persisted in the
+/// The policy for the aux file storage. It can be switched through `switch_aux_file_policy`
+/// tenant config. When the first aux file written, the policy will be persisted in the
 /// `index_part.json` file and has a limited migration path.
 ///
 /// Currently, we only allow the following migration path:
@@ -351,7 +348,7 @@ impl AuxFilePolicy {

    /// If a tenant writes aux files without setting `switch_aux_policy`, this value will be used.
    pub fn default_tenant_config() -> Self {
-        Self::V2
+        Self::V1
    }
 }

@@ -438,9 +435,7 @@ pub enum CompactionAlgorithm {
    Tiered,
 }

-#[derive(
-    Debug, Clone, Copy, PartialEq, Eq, serde_with::DeserializeFromStr, serde_with::SerializeDisplay,
-)]
+#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
 pub enum ImageCompressionAlgorithm {
    // Disabled for writes, support decompressing during read path
    Disabled,
@@ -475,33 +470,11 @@ impl FromStr for ImageCompressionAlgorithm {
    }
 }

-impl Display for ImageCompressionAlgorithm {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        match self {
-            ImageCompressionAlgorithm::Disabled => write!(f, "disabled"),
-            ImageCompressionAlgorithm::Zstd { level } => {
-                if let Some(level) = level {
-                    write!(f, "zstd({})", level)
-                } else {
-                    write!(f, "zstd")
-                }
-            }
-        }
-    }
-}
-
 #[derive(Eq, PartialEq, Debug, Clone, Serialize, Deserialize)]
 pub struct CompactionAlgorithmSettings {
    pub kind: CompactionAlgorithm,
 }

-#[derive(Debug, PartialEq, Eq, Clone, serde::Deserialize, serde::Serialize)]
-#[serde(tag = "mode", rename_all = "kebab-case", deny_unknown_fields)]
-pub enum L0FlushConfig {
-    #[serde(rename_all = "snake_case")]
-    Direct { max_concurrency: NonZeroUsize },
-}
-
 #[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
 pub struct EvictionPolicyLayerAccessThreshold {
    #[serde(with = "humantime_serde")]
@@ -513,11 +486,12 @@ pub struct EvictionPolicyLayerAccessThreshold {
 #[derive(Debug, Serialize, Deserialize, Clone, PartialEq, Eq)]
 pub struct ThrottleConfig {
    pub task_kinds: Vec<String>, // TaskKind
-    pub initial: u32,
+    pub initial: usize,
    #[serde(with = "humantime_serde")]
    pub refill_interval: Duration,
-    pub refill_amount: NonZeroU32,
-    pub max: u32,
+    pub refill_amount: NonZeroUsize,
+    pub max: usize,
+    pub fair: bool,
 }

 impl ThrottleConfig {
@@ -527,8 +501,9 @@ impl ThrottleConfig {
            // other values don't matter with emtpy `task_kinds`.
            initial: 0,
            refill_interval: Duration::from_millis(1),
-            refill_amount: NonZeroU32::new(1).unwrap(),
+            refill_amount: NonZeroUsize::new(1).unwrap(),
            max: 1,
+            fair: true,
        }
    }
    /// The requests per second allowed  by the given config.
@@ -746,14 +721,8 @@ pub struct TimelineInfo {

    pub walreceiver_status: String,

-    // ALWAYS add new fields at the end of the struct with `Option` to ensure forward/backward compatibility.
-    // Backward compatibility: you will get a JSON not containing the newly-added field.
-    // Forward compatibility: a previous version of the pageserver will receive a JSON. serde::Deserialize does
-    // not deny unknown fields by default so it's safe to set the field to some value, though it won't be
-    // read.
    /// The last aux file policy being used on this timeline
    pub last_aux_file_policy: Option<AuxFilePolicy>,
-    pub is_archived: Option<bool>,
 }

 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -898,9 +867,7 @@ pub struct WalRedoManagerStatus {
    pub process: Option<WalRedoManagerProcessStatus>,
 }

-/// The progress of a secondary tenant.
-///
-/// It is mostly useful when doing a long running download: e.g. initiating
+/// The progress of a secondary tenant is mostly useful when doing a long running download: e.g. initiating
 /// a download job, timing out while waiting for it to run, and then inspecting this status to understand
 /// what's happening.
 #[derive(Default, Debug, Serialize, Deserialize, Clone)]
@@ -1095,7 +1062,7 @@ impl TryFrom<u8> for PagestreamBeMessageTag {
    }
 }

-// A GetPage request contains two LSN values:
+// In the V2 protocol version, a GetPage request contains two LSN values:
 //
 // request_lsn: Get the page version at this point in time.  Lsn::Max is a special value that means
 // "get the latest version present". It's used by the primary server, which knows that no one else
@@ -1108,7 +1075,7 @@ impl TryFrom<u8> for PagestreamBeMessageTag {
 // passing an earlier LSN can speed up the request, by allowing the pageserver to process the
 // request without waiting for 'request_lsn' to arrive.
 //
-// The now-defunct V1 interface contained only one LSN, and a boolean 'latest' flag. The V1 interface was
+// The legacy V1 interface contained only one LSN, and a boolean 'latest' flag. The V1 interface was
 // sufficient for the primary; the 'lsn' was equivalent to the 'not_modified_since' value, and
 // 'latest' was set to true. The V2 interface was added because there was no correct way for a
 // standby to request a page at a particular non-latest LSN, and also include the
@@ -1116,11 +1083,15 @@ impl TryFrom<u8> for PagestreamBeMessageTag {
 // request, if the standby knows that the page hasn't been modified since, and risk getting an error
 // if that LSN has fallen behind the GC horizon, or requesting the current replay LSN, which could
 // require the pageserver unnecessarily to wait for the WAL to arrive up to that point. The new V2
-// interface allows sending both LSNs, and let the pageserver do the right thing. There was no
+// interface allows sending both LSNs, and let the pageserver do the right thing. There is no
 // difference in the responses between V1 and V2.
 //
+// The Request structs below reflect the V2 interface. If V1 is used, the parse function
+// maps the old format requests to the new format.
+//
 #[derive(Clone, Copy)]
 pub enum PagestreamProtocolVersion {
+    V1,
    V2,
 }

@@ -1259,17 +1230,36 @@ impl PagestreamFeMessage {
        bytes.into()
    }

-    pub fn parse<R: std::io::Read>(body: &mut R) -> anyhow::Result<PagestreamFeMessage> {
+    pub fn parse<R: std::io::Read>(
+        body: &mut R,
+        protocol_version: PagestreamProtocolVersion,
+    ) -> anyhow::Result<PagestreamFeMessage> {
        // these correspond to the NeonMessageTag enum in pagestore_client.h
        //
        // TODO: consider using protobuf or serde bincode for less error prone
        // serialization.
        let msg_tag = body.read_u8()?;

-        // these two fields are the same for every request type
-        let request_lsn = Lsn::from(body.read_u64::<BigEndian>()?);
-        let not_modified_since = Lsn::from(body.read_u64::<BigEndian>()?);
+        let (request_lsn, not_modified_since) = match protocol_version {
+            PagestreamProtocolVersion::V2 => (
+                Lsn::from(body.read_u64::<BigEndian>()?),
+                Lsn::from(body.read_u64::<BigEndian>()?),
+            ),
+            PagestreamProtocolVersion::V1 => {
+                // In the old protocol, each message starts with a boolean 'latest' flag,
+                // followed by 'lsn'. Convert that to the two LSNs, 'request_lsn' and
+                // 'not_modified_since', used in the new protocol version.
+                let latest = body.read_u8()? != 0;
+                let request_lsn = Lsn::from(body.read_u64::<BigEndian>()?);
+                if latest {
+                    (Lsn::MAX, request_lsn) // get latest version
+                } else {
+                    (request_lsn, request_lsn) // get version at specified LSN
+                }
+            }
+        };

+        // The rest of the messages are the same between V1 and V2
        match msg_tag {
            0 => Ok(PagestreamFeMessage::Exists(PagestreamExistsRequest {
                request_lsn,
@@ -1477,7 +1467,9 @@ mod tests {
        ];
        for msg in messages {
            let bytes = msg.serialize();
-            let reconstructed = PagestreamFeMessage::parse(&mut bytes.reader()).unwrap();
+            let reconstructed =
+                PagestreamFeMessage::parse(&mut bytes.reader(), PagestreamProtocolVersion::V2)
+                    .unwrap();
            assert!(msg == reconstructed);
        }
    }
@@ -1685,33 +1677,21 @@ mod tests {
    #[test]
    fn test_image_compression_algorithm_parsing() {
        use ImageCompressionAlgorithm::*;
-        let cases = [
-            ("disabled", Disabled),
-            ("zstd", Zstd { level: None }),
-            ("zstd(18)", Zstd { level: Some(18) }),
-            ("zstd(-3)", Zstd { level: Some(-3) }),
-        ];
-
-        for (display, expected) in cases {
-            assert_eq!(
-                ImageCompressionAlgorithm::from_str(display).unwrap(),
-                expected,
-                "parsing works"
-            );
-            assert_eq!(format!("{expected}"), display, "Display FromStr roundtrip");
-
-            let ser = serde_json::to_string(&expected).expect("serialization");
-            assert_eq!(
-                serde_json::from_str::<ImageCompressionAlgorithm>(&ser).unwrap(),
-                expected,
-                "serde roundtrip"
-            );
-
-            assert_eq!(
-                serde_json::Value::String(display.to_string()),
-                serde_json::to_value(expected).unwrap(),
-                "Display is the serde serialization"
-            );
-        }
+        assert_eq!(
+            ImageCompressionAlgorithm::from_str("disabled").unwrap(),
+            Disabled
+        );
+        assert_eq!(
+            ImageCompressionAlgorithm::from_str("zstd").unwrap(),
+            Zstd { level: None }
+        );
+        assert_eq!(
+            ImageCompressionAlgorithm::from_str("zstd(18)").unwrap(),
+            Zstd { level: Some(18) }
+        );
+        assert_eq!(
+            ImageCompressionAlgorithm::from_str("zstd(-3)").unwrap(),
+            Zstd { level: Some(-3) }
+        );
    }
 }
--- a/libs/pageserver_api/src/models/utilization.rs
+++ b/libs/pageserver_api/src/models/utilization.rs
@@ -1,5 +1,4 @@
-use std::time::SystemTime;
-use utils::{serde_percent::Percent, serde_system_time};
+use utils::serde_system_time::SystemTime;

 /// Pageserver current utilization and scoring for how good candidate the pageserver would be for
 /// the next tenant.
@@ -10,143 +9,19 @@ use utils::{serde_percent::Percent, serde_system_time};
 /// not handle full u64 values properly.
 #[derive(serde::Serialize, serde::Deserialize, Debug, Clone)]
 pub struct PageserverUtilization {
-    /// Used disk space (physical, ground truth from statfs())
+    /// Used disk space
    #[serde(serialize_with = "ser_saturating_u63")]
    pub disk_usage_bytes: u64,
    /// Free disk space
    #[serde(serialize_with = "ser_saturating_u63")]
    pub free_space_bytes: u64,
-
-    /// Wanted disk space, based on the tenant shards currently present on this pageserver: this
-    /// is like disk_usage_bytes, but it is stable and does not change with the cache state of
-    /// tenants, whereas disk_usage_bytes may reach the disk eviction `max_usage_pct` and stay
-    /// there, or may be unrealistically low if the pageserver has attached tenants which haven't
-    /// downloaded layers yet.
-    #[serde(serialize_with = "ser_saturating_u63", default)]
-    pub disk_wanted_bytes: u64,
-
-    // What proportion of total disk space will this pageserver use before it starts evicting data?
-    #[serde(default = "unity_percent")]
-    pub disk_usable_pct: Percent,
-
-    // How many shards are currently on this node?
-    #[serde(default)]
-    pub shard_count: u32,
-
-    // How many shards should this node be able to handle at most?
-    #[serde(default)]
-    pub max_shard_count: u32,
-
-    /// Cached result of [`Self::score`]
-    pub utilization_score: Option<u64>,
-
+    /// Lower is better score for how good candidate for a next tenant would this pageserver be.
+    #[serde(serialize_with = "ser_saturating_u63")]
+    pub utilization_score: u64,
    /// When was this snapshot captured, pageserver local time.
    ///
    /// Use millis to give confidence that the value is regenerated often enough.
-    pub captured_at: serde_system_time::SystemTime,
-}
-
-fn unity_percent() -> Percent {
-    Percent::new(0).unwrap()
-}
-
-pub type RawScore = u64;
-
-impl PageserverUtilization {
-    const UTILIZATION_FULL: u64 = 1000000;
-
-    /// Calculate a utilization score.  The result is to be inrepreted as a fraction of
-    /// Self::UTILIZATION_FULL.
-    ///
-    /// Lower values are more affine to scheduling more work on this node.
-    /// - UTILIZATION_FULL represents an ideal node which is fully utilized but should not receive any more work.
-    /// - 0.0 represents an empty node.
-    /// - Negative values are forbidden
-    /// - Values over UTILIZATION_FULL indicate an overloaded node, which may show degraded performance due to
-    ///   layer eviction.
-    pub fn score(&self) -> RawScore {
-        let disk_usable_capacity = ((self.disk_usage_bytes + self.free_space_bytes)
-            * self.disk_usable_pct.get() as u64)
-            / 100;
-        let disk_utilization_score =
-            self.disk_wanted_bytes * Self::UTILIZATION_FULL / disk_usable_capacity;
-
-        let shard_utilization_score =
-            self.shard_count as u64 * Self::UTILIZATION_FULL / self.max_shard_count as u64;
-        std::cmp::max(disk_utilization_score, shard_utilization_score)
-    }
-
-    pub fn cached_score(&mut self) -> RawScore {
-        match self.utilization_score {
-            None => {
-                let s = self.score();
-                self.utilization_score = Some(s);
-                s
-            }
-            Some(s) => s,
-        }
-    }
-
-    /// If a node is currently hosting more work than it can comfortably handle.  This does not indicate that
-    /// it will fail, but it is a strong signal that more work should not be added unless there is no alternative.
-    pub fn is_overloaded(score: RawScore) -> bool {
-        score >= Self::UTILIZATION_FULL
-    }
-
-    pub fn adjust_shard_count_max(&mut self, shard_count: u32) {
-        if self.shard_count < shard_count {
-            self.shard_count = shard_count;
-
-            // Dirty cache: this will be calculated next time someone retrives the score
-            self.utilization_score = None;
-        }
-    }
-
-    /// A utilization structure that has a full utilization score: use this as a placeholder when
-    /// you need a utilization but don't have real values yet.
-    pub fn full() -> Self {
-        Self {
-            disk_usage_bytes: 1,
-            free_space_bytes: 0,
-            disk_wanted_bytes: 1,
-            disk_usable_pct: Percent::new(100).unwrap(),
-            shard_count: 1,
-            max_shard_count: 1,
-            utilization_score: Some(Self::UTILIZATION_FULL),
-            captured_at: serde_system_time::SystemTime(SystemTime::now()),
-        }
-    }
-}
-
-/// Test helper
-pub mod test_utilization {
-    use super::PageserverUtilization;
-    use std::time::SystemTime;
-    use utils::{
-        serde_percent::Percent,
-        serde_system_time::{self},
-    };
-
-    // Parameters of the imaginary node used for test utilization instances
-    const TEST_DISK_SIZE: u64 = 1024 * 1024 * 1024 * 1024;
-    const TEST_SHARDS_MAX: u32 = 1000;
-
-    /// Unit test helper.  Unconditionally compiled because cfg(test) doesn't carry across crates.  Do
-    /// not abuse this function from non-test code.
-    ///
-    /// Emulates a node with a 1000 shard limit and a 1TB disk.
-    pub fn simple(shard_count: u32, disk_wanted_bytes: u64) -> PageserverUtilization {
-        PageserverUtilization {
-            disk_usage_bytes: disk_wanted_bytes,
-            free_space_bytes: TEST_DISK_SIZE - std::cmp::min(disk_wanted_bytes, TEST_DISK_SIZE),
-            disk_wanted_bytes,
-            disk_usable_pct: Percent::new(100).unwrap(),
-            shard_count,
-            max_shard_count: TEST_SHARDS_MAX,
-            utilization_score: None,
-            captured_at: serde_system_time::SystemTime(SystemTime::now()),
-        }
-    }
+    pub captured_at: SystemTime,
 }

 /// openapi knows only `format: int64`, so avoid outputting a non-parseable value by generated clients.
@@ -174,19 +49,15 @@ mod tests {
        let doc = PageserverUtilization {
            disk_usage_bytes: u64::MAX,
            free_space_bytes: 0,
-            disk_wanted_bytes: u64::MAX,
-            utilization_score: Some(13),
-            disk_usable_pct: Percent::new(90).unwrap(),
-            shard_count: 100,
-            max_shard_count: 200,
-            captured_at: serde_system_time::SystemTime(
+            utilization_score: u64::MAX,
+            captured_at: SystemTime(
                std::time::SystemTime::UNIX_EPOCH + Duration::from_secs(1708509779),
            ),
        };

        let s = serde_json::to_string(&doc).unwrap();

-        let expected = "{\"disk_usage_bytes\":9223372036854775807,\"free_space_bytes\":0,\"disk_wanted_bytes\":9223372036854775807,\"disk_usable_pct\":90,\"shard_count\":100,\"max_shard_count\":200,\"utilization_score\":13,\"captured_at\":\"2024-02-21T10:02:59.000Z\"}";
+        let expected = r#"{"disk_usage_bytes":9223372036854775807,"free_space_bytes":0,"utilization_score":9223372036854775807,"captured_at":"2024-02-21T10:02:59.000Z"}"#;

        assert_eq!(s, expected);
    }
--- a/libs/postgres_backend/Cargo.toml
+++ b/libs/postgres_backend/Cargo.toml
@@ -18,6 +18,7 @@ tokio-rustls.workspace = true
 tracing.workspace = true

 pq_proto.workspace = true
+workspace_hack.workspace = true

 [dev-dependencies]
 once_cell.workspace = true
--- a/libs/postgres_backend/src/lib.rs
+++ b/libs/postgres_backend/src/lib.rs
@@ -69,10 +69,8 @@ impl QueryError {
 }

 /// Returns true if the given error is a normal consequence of a network issue,
-/// or the client closing the connection.
-///
-/// These errors can happen during normal operations,
-/// and don't indicate a bug in our code.
+/// or the client closing the connection. These errors can happen during normal
+/// operations, and don't indicate a bug in our code.
 pub fn is_expected_io_error(e: &io::Error) -> bool {
    use io::ErrorKind::*;
    matches!(
--- a/libs/postgres_connection/Cargo.toml
+++ b/libs/postgres_connection/Cargo.toml
@@ -11,5 +11,7 @@ postgres.workspace = true
 tokio-postgres.workspace = true
 url.workspace = true

+workspace_hack.workspace = true
+
 [dev-dependencies]
 once_cell.workspace = true
--- a/libs/postgres_connection/src/lib.rs
+++ b/libs/postgres_connection/src/lib.rs
@@ -7,7 +7,6 @@ use std::fmt;
 use url::Host;

 /// Parses a string of format either `host:port` or `host` into a corresponding pair.
-///
 /// The `host` part should be a correct `url::Host`, while `port` (if present) should be
 /// a valid decimal u16 of digits only.
 pub fn parse_host_port<S: AsRef<str>>(host_port: S) -> Result<(Host, Option<u16>), anyhow::Error> {
--- a/libs/postgres_ffi/Cargo.toml
+++ b/libs/postgres_ffi/Cargo.toml
@@ -19,6 +19,8 @@ thiserror.workspace = true
 serde.workspace = true
 utils.workspace = true

+workspace_hack.workspace = true
+
 [dev-dependencies]
 env_logger.workspace = true
 postgres.workspace = true
--- a/libs/postgres_ffi/build.rs
+++ b/libs/postgres_ffi/build.rs
@@ -14,7 +14,7 @@ impl ParseCallbacks for PostgresFfiCallbacks {
    fn include_file(&self, filename: &str) {
        // This does the equivalent of passing bindgen::CargoCallbacks
        // to the builder .parse_callbacks() method.
-        let cargo_callbacks = bindgen::CargoCallbacks::new();
+        let cargo_callbacks = bindgen::CargoCallbacks;
        cargo_callbacks.include_file(filename)
    }

--- a/libs/postgres_ffi/src/lib.rs
+++ b/libs/postgres_ffi/src/lib.rs
@@ -136,15 +136,15 @@ pub const MAX_SEND_SIZE: usize = XLOG_BLCKSZ * 16;

 // Export some version independent functions that are used outside of this mod
 pub use v14::xlog_utils::encode_logical_message;
+pub use v14::xlog_utils::from_pg_timestamp;
 pub use v14::xlog_utils::get_current_timestamp;
 pub use v14::xlog_utils::to_pg_timestamp;
-pub use v14::xlog_utils::try_from_pg_timestamp;
 pub use v14::xlog_utils::XLogFileName;

 pub use v14::bindings::DBState_DB_SHUTDOWNED;

-pub fn bkpimage_is_compressed(bimg_info: u8, version: u32) -> bool {
-    dispatch_pgversion!(version, pgv::bindings::bkpimg_is_compressed(bimg_info))
+pub fn bkpimage_is_compressed(bimg_info: u8, version: u32) -> anyhow::Result<bool> {
+    dispatch_pgversion!(version, Ok(pgv::bindings::bkpimg_is_compressed(bimg_info)))
 }

 pub fn generate_wal_segment(
--- a/libs/postgres_ffi/src/xlog_utils.rs
+++ b/libs/postgres_ffi/src/xlog_utils.rs
@@ -135,8 +135,6 @@ pub fn get_current_timestamp() -> TimestampTz {
 mod timestamp_conversions {
    use std::time::Duration;

-    use anyhow::Context;
-
    use super::*;

    const UNIX_EPOCH_JDATE: u64 = 2440588; // == date2j(1970, 1, 1)
@@ -156,18 +154,18 @@ mod timestamp_conversions {
        }
    }

-    pub fn try_from_pg_timestamp(time: TimestampTz) -> anyhow::Result<SystemTime> {
+    pub fn from_pg_timestamp(time: TimestampTz) -> SystemTime {
        let time: u64 = time
            .try_into()
-            .context("timestamp before millenium (postgres epoch)")?;
+            .expect("timestamp before millenium (postgres epoch)");
        let since_unix_epoch = time + SECS_DIFF_UNIX_TO_POSTGRES_EPOCH * USECS_PER_SEC;
        SystemTime::UNIX_EPOCH
            .checked_add(Duration::from_micros(since_unix_epoch))
-            .context("SystemTime overflow")
+            .expect("SystemTime overflow")
    }
 }

-pub use timestamp_conversions::{to_pg_timestamp, try_from_pg_timestamp};
+pub use timestamp_conversions::{from_pg_timestamp, to_pg_timestamp};

 // Returns (aligned) end_lsn of the last record in data_dir with WAL segments.
 // start_lsn must point to some previously known record boundary (beginning of
@@ -547,14 +545,14 @@ mod tests {
    #[test]
    fn test_ts_conversion() {
        let now = SystemTime::now();
-        let round_trip = try_from_pg_timestamp(to_pg_timestamp(now)).unwrap();
+        let round_trip = from_pg_timestamp(to_pg_timestamp(now));

        let now_since = now.duration_since(SystemTime::UNIX_EPOCH).unwrap();
        let round_trip_since = round_trip.duration_since(SystemTime::UNIX_EPOCH).unwrap();
        assert_eq!(now_since.as_micros(), round_trip_since.as_micros());

        let now_pg = get_current_timestamp();
-        let round_trip_pg = to_pg_timestamp(try_from_pg_timestamp(now_pg).unwrap());
+        let round_trip_pg = to_pg_timestamp(from_pg_timestamp(now_pg));

        assert_eq!(now_pg, round_trip_pg);
    }
--- a/libs/postgres_ffi/wal_craft/Cargo.toml
+++ b/libs/postgres_ffi/wal_craft/Cargo.toml
@@ -14,6 +14,8 @@ postgres.workspace = true
 postgres_ffi.workspace = true
 camino-tempfile.workspace = true

+workspace_hack.workspace = true
+
 [dev-dependencies]
 regex.workspace = true
 utils.workspace = true
--- a/libs/pq_proto/Cargo.toml
+++ b/libs/pq_proto/Cargo.toml
@@ -11,7 +11,9 @@ itertools.workspace = true
 pin-project-lite.workspace = true
 postgres-protocol.workspace = true
 rand.workspace = true
-tokio = { workspace = true, features = ["io-util"] }
+tokio.workspace = true
 tracing.workspace = true
 thiserror.workspace = true
 serde.workspace = true
+
+workspace_hack.workspace = true
--- a/libs/remote_storage/Cargo.toml
+++ b/libs/remote_storage/Cargo.toml
@@ -32,7 +32,7 @@ scopeguard.workspace = true
 metrics.workspace = true
 utils.workspace = true
 pin-project-lite.workspace = true
-
+workspace_hack.workspace = true
 azure_core.workspace = true
 azure_identity.workspace = true
 azure_storage.workspace = true
@@ -46,4 +46,3 @@ sync_wrapper = { workspace = true, features = ["futures"] }
 camino-tempfile.workspace = true
 test-context.workspace = true
 rand.workspace = true
-tokio = { workspace = true, features = ["test-util"] }
--- a/libs/remote_storage/src/azure_blob.rs
+++ b/libs/remote_storage/src/azure_blob.rs
@@ -383,48 +383,6 @@ impl RemoteStorage for AzureBlobStorage {
        }
    }

-    async fn head_object(
-        &self,
-        key: &RemotePath,
-        cancel: &CancellationToken,
-    ) -> Result<ListingObject, DownloadError> {
-        let kind = RequestKind::Head;
-        let _permit = self.permit(kind, cancel).await?;
-
-        let started_at = start_measuring_requests(kind);
-
-        let blob_client = self.client.blob_client(self.relative_path_to_name(key));
-        let properties_future = blob_client.get_properties().into_future();
-
-        let properties_future = tokio::time::timeout(self.timeout, properties_future);
-
-        let res = tokio::select! {
-            res = properties_future => res,
-            _ = cancel.cancelled() => return Err(TimeoutOrCancel::Cancel.into()),
-        };
-
-        if let Ok(inner) = &res {
-            // do not incl. timeouts as errors in metrics but cancellations
-            let started_at = ScopeGuard::into_inner(started_at);
-            crate::metrics::BUCKET_METRICS
-                .req_seconds
-                .observe_elapsed(kind, inner, started_at);
-        }
-
-        let data = match res {
-            Ok(Ok(data)) => Ok(data),
-            Ok(Err(sdk)) => Err(to_download_error(sdk)),
-            Err(_timeout) => Err(DownloadError::Timeout),
-        }?;
-
-        let properties = data.blob.properties;
-        Ok(ListingObject {
-            key: key.to_owned(),
-            last_modified: SystemTime::from(properties.last_modified),
-            size: properties.content_length,
-        })
-    }
-
    async fn upload(
        &self,
        from: impl Stream<Item = std::io::Result<Bytes>> + Send + Sync + 'static,
--- a/libs/remote_storage/src/config.rs
+++ b/libs/remote_storage/src/config.rs
@@ -235,31 +235,6 @@ timeout = '5s'";
        );
    }

-    #[test]
-    fn test_storage_class_serde_roundtrip() {
-        let classes = [
-            None,
-            Some(StorageClass::Standard),
-            Some(StorageClass::IntelligentTiering),
-        ];
-        for class in classes {
-            #[derive(Serialize, Deserialize)]
-            struct Wrapper {
-                #[serde(
-                    deserialize_with = "deserialize_storage_class",
-                    serialize_with = "serialize_storage_class"
-                )]
-                class: Option<StorageClass>,
-            }
-            let wrapped = Wrapper {
-                class: class.clone(),
-            };
-            let serialized = serde_json::to_string(&wrapped).unwrap();
-            let deserialized: Wrapper = serde_json::from_str(&serialized).unwrap();
-            assert_eq!(class, deserialized.class);
-        }
-    }
-
    #[test]
    fn test_azure_parsing() {
        let toml = "\
--- a/libs/remote_storage/src/error.rs
+++ b/libs/remote_storage/src/error.rs
@@ -42,10 +42,6 @@ impl DownloadError {
            Timeout | Other(_) => false,
        }
    }
-
-    pub fn is_cancelled(&self) -> bool {
-        matches!(self, DownloadError::Cancelled)
-    }
 }

 impl From<std::io::Error> for DownloadError {
--- a/libs/remote_storage/src/lib.rs
+++ b/libs/remote_storage/src/lib.rs
@@ -45,8 +45,6 @@ pub use azure_core::Etag;

 pub use error::{DownloadError, TimeTravelError, TimeoutOrCancel};

-/// Default concurrency limit for S3 operations
-///
 /// Currently, sync happens with AWS S3, that has two limits on requests per second:
 /// ~200 RPS for IAM services
 /// <https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide/UsingWithRDS.IAMDBAuth.html>
@@ -152,7 +150,7 @@ pub enum ListingMode {
    NoDelimiter,
 }

-#[derive(PartialEq, Eq, Debug, Clone)]
+#[derive(PartialEq, Eq, Debug)]
 pub struct ListingObject {
    pub key: RemotePath,
    pub last_modified: SystemTime,
@@ -217,13 +215,6 @@ pub trait RemoteStorage: Send + Sync + 'static {
        Ok(combined)
    }

-    /// Obtain metadata information about an object.
-    async fn head_object(
-        &self,
-        key: &RemotePath,
-        cancel: &CancellationToken,
-    ) -> Result<ListingObject, DownloadError>;
-
    /// Streams the local file contents into remote into the remote storage entry.
    ///
    /// If the operation fails because of timeout or cancellation, the root cause of the error will be
@@ -302,9 +293,7 @@ pub trait RemoteStorage: Send + Sync + 'static {
    ) -> Result<(), TimeTravelError>;
 }

-/// Data part of an ongoing [`Download`].
-///
-/// `DownloadStream` is sensitive to the timeout and cancellation used with the original
+/// DownloadStream is sensitive to the timeout and cancellation used with the original
 /// [`RemoteStorage::download`] request. The type yields `std::io::Result<Bytes>` to be compatible
 /// with `tokio::io::copy_buf`.
 // This has 'static because safekeepers do not use cancellation tokens (yet)
@@ -374,20 +363,6 @@ impl<Other: RemoteStorage> GenericRemoteStorage<Arc<Other>> {
        }
    }

-    // See [`RemoteStorage::head_object`].
-    pub async fn head_object(
-        &self,
-        key: &RemotePath,
-        cancel: &CancellationToken,
-    ) -> Result<ListingObject, DownloadError> {
-        match self {
-            Self::LocalFs(s) => s.head_object(key, cancel).await,
-            Self::AwsS3(s) => s.head_object(key, cancel).await,
-            Self::AzureBlob(s) => s.head_object(key, cancel).await,
-            Self::Unreliable(s) => s.head_object(key, cancel).await,
-        }
-    }
-
    /// See [`RemoteStorage::upload`]
    pub async fn upload(
        &self,
@@ -623,7 +598,6 @@ impl ConcurrencyLimiter {
            RequestKind::Delete => &self.write,
            RequestKind::Copy => &self.write,
            RequestKind::TimeTravel => &self.write,
-            RequestKind::Head => &self.read,
        }
    }

--- a/libs/remote_storage/src/local_fs.rs
+++ b/libs/remote_storage/src/local_fs.rs
@@ -445,20 +445,6 @@ impl RemoteStorage for LocalFs {
        }
    }

-    async fn head_object(
-        &self,
-        key: &RemotePath,
-        _cancel: &CancellationToken,
-    ) -> Result<ListingObject, DownloadError> {
-        let target_file_path = key.with_base(&self.storage_root);
-        let metadata = file_metadata(&target_file_path).await?;
-        Ok(ListingObject {
-            key: key.clone(),
-            last_modified: metadata.modified()?,
-            size: metadata.len(),
-        })
-    }
-
    async fn upload(
        &self,
        data: impl Stream<Item = std::io::Result<Bytes>> + Send + Sync,
--- a/libs/remote_storage/src/metrics.rs
+++ b/libs/remote_storage/src/metrics.rs
@@ -13,7 +13,6 @@ pub(crate) enum RequestKind {
    List = 3,
    Copy = 4,
    TimeTravel = 5,
-    Head = 6,
 }

 use scopeguard::ScopeGuard;
@@ -28,7 +27,6 @@ impl RequestKind {
            List => "list_objects",
            Copy => "copy_object",
            TimeTravel => "time_travel_recover",
-            Head => "head_object",
        }
    }
    const fn as_index(&self) -> usize {
@@ -36,8 +34,7 @@ impl RequestKind {
    }
 }

-const REQUEST_KIND_COUNT: usize = 7;
-pub(crate) struct RequestTyped<C>([C; REQUEST_KIND_COUNT]);
+pub(crate) struct RequestTyped<C>([C; 6]);

 impl<C> RequestTyped<C> {
    pub(crate) fn get(&self, kind: RequestKind) -> &C {
@@ -46,8 +43,8 @@ impl<C> RequestTyped<C> {

    fn build_with(mut f: impl FnMut(RequestKind) -> C) -> Self {
        use RequestKind::*;
-        let mut it = [Get, Put, Delete, List, Copy, TimeTravel, Head].into_iter();
-        let arr = std::array::from_fn::<C, REQUEST_KIND_COUNT, _>(|index| {
+        let mut it = [Get, Put, Delete, List, Copy, TimeTravel].into_iter();
+        let arr = std::array::from_fn::<C, 6, _>(|index| {
            let next = it.next().unwrap();
            assert_eq!(index, next.as_index());
            f(next)
--- a/libs/remote_storage/src/s3_bucket.rs
+++ b/libs/remote_storage/src/s3_bucket.rs
@@ -23,7 +23,7 @@ use aws_config::{
 use aws_sdk_s3::{
    config::{AsyncSleep, IdentityCache, Region, SharedAsyncSleep},
    error::SdkError,
-    operation::{get_object::GetObjectError, head_object::HeadObjectError},
+    operation::get_object::GetObjectError,
    types::{Delete, DeleteMarkerEntry, ObjectIdentifier, ObjectVersion, StorageClass},
    Client,
 };
@@ -604,78 +604,6 @@ impl RemoteStorage for S3Bucket {
        }
    }

-    async fn head_object(
-        &self,
-        key: &RemotePath,
-        cancel: &CancellationToken,
-    ) -> Result<ListingObject, DownloadError> {
-        let kind = RequestKind::Head;
-        let _permit = self.permit(kind, cancel).await?;
-
-        let started_at = start_measuring_requests(kind);
-
-        let head_future = self
-            .client
-            .head_object()
-            .bucket(self.bucket_name())
-            .key(self.relative_path_to_s3_object(key))
-            .send();
-
-        let head_future = tokio::time::timeout(self.timeout, head_future);
-
-        let res = tokio::select! {
-            res = head_future => res,
-            _ = cancel.cancelled() => return Err(TimeoutOrCancel::Cancel.into()),
-        };
-
-        let res = res.map_err(|_e| DownloadError::Timeout)?;
-
-        // do not incl. timeouts as errors in metrics but cancellations
-        let started_at = ScopeGuard::into_inner(started_at);
-        crate::metrics::BUCKET_METRICS
-            .req_seconds
-            .observe_elapsed(kind, &res, started_at);
-
-        let data = match res {
-            Ok(object_output) => object_output,
-            Err(SdkError::ServiceError(e)) if matches!(e.err(), HeadObjectError::NotFound(_)) => {
-                // Count this in the AttemptOutcome::Ok bucket, because 404 is not
-                // an error: we expect to sometimes fetch an object and find it missing,
-                // e.g. when probing for timeline indices.
-                crate::metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
-                    kind,
-                    AttemptOutcome::Ok,
-                    started_at,
-                );
-                return Err(DownloadError::NotFound);
-            }
-            Err(e) => {
-                crate::metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
-                    kind,
-                    AttemptOutcome::Err,
-                    started_at,
-                );
-
-                return Err(DownloadError::Other(
-                    anyhow::Error::new(e).context("s3 head object"),
-                ));
-            }
-        };
-
-        let (Some(last_modified), Some(size)) = (data.last_modified, data.content_length) else {
-            return Err(DownloadError::Other(anyhow!(
-                "head_object doesn't contain last_modified or content_length"
-            )))?;
-        };
-        Ok(ListingObject {
-            key: key.to_owned(),
-            last_modified: SystemTime::try_from(last_modified).map_err(|e| {
-                DownloadError::Other(anyhow!("can't convert time '{last_modified}': {e}"))
-            })?,
-            size: size as u64,
-        })
-    }
-
    async fn upload(
        &self,
        from: impl Stream<Item = std::io::Result<Bytes>> + Send + Sync + 'static,
--- a/libs/remote_storage/src/simulate_failures.rs
+++ b/libs/remote_storage/src/simulate_failures.rs
@@ -30,7 +30,6 @@ pub struct UnreliableWrapper {
 #[derive(Debug, Hash, Eq, PartialEq)]
 enum RemoteOp {
    ListPrefixes(Option<RemotePath>),
-    HeadObject(RemotePath),
    Upload(RemotePath),
    Download(RemotePath),
    Delete(RemotePath),
@@ -138,16 +137,6 @@ impl RemoteStorage for UnreliableWrapper {
        self.inner.list(prefix, mode, max_keys, cancel).await
    }

-    async fn head_object(
-        &self,
-        key: &RemotePath,
-        cancel: &CancellationToken,
-    ) -> Result<crate::ListingObject, DownloadError> {
-        self.attempt(RemoteOp::HeadObject(key.clone()))
-            .map_err(DownloadError::Other)?;
-        self.inner.head_object(key, cancel).await
-    }
-
    async fn upload(
        &self,
        data: impl Stream<Item = std::io::Result<Bytes>> + Send + Sync + 'static,
--- a/libs/safekeeper_api/Cargo.toml
+++ b/libs/safekeeper_api/Cargo.toml
@@ -9,3 +9,5 @@ serde.workspace = true
 serde_with.workspace = true
 const_format.workspace = true
 utils.workspace = true
+
+workspace_hack.workspace = true
--- a/libs/safekeeper_api/src/models.rs
+++ b/libs/safekeeper_api/src/models.rs
@@ -60,16 +60,3 @@ pub struct TimelineCopyRequest {
    pub target_timeline_id: TimelineId,
    pub until_lsn: Lsn,
 }
-
-#[derive(Debug, Clone, Deserialize, Serialize)]
-pub struct TimelineTermBumpRequest {
-    /// bump to
-    pub term: Option<u64>,
-}
-
-#[derive(Debug, Clone, Deserialize, Serialize)]
-pub struct TimelineTermBumpResponse {
-    // before the request
-    pub previous_term: u64,
-    pub current_term: u64,
-}
--- a/libs/tenant_size_model/Cargo.toml
+++ b/libs/tenant_size_model/Cargo.toml
@@ -9,3 +9,5 @@ license.workspace = true
 anyhow.workspace = true
 serde.workspace = true
 serde_json.workspace = true
+
+workspace_hack.workspace = true
--- a/libs/tenant_size_model/src/lib.rs
+++ b/libs/tenant_size_model/src/lib.rs
@@ -5,10 +5,9 @@
 mod calculation;
 pub mod svg;

-/// StorageModel is the input to the synthetic size calculation.
-///
-/// It represents a tree of timelines, with just the information that's needed
-/// for the calculation. This doesn't track timeline names or where each timeline
+/// StorageModel is the input to the synthetic size calculation. It represents
+/// a tree of timelines, with just the information that's needed for the
+/// calculation. This doesn't track timeline names or where each timeline
 /// begins and ends, for example. Instead, it consists of "points of interest"
 /// on the timelines. A point of interest could be the timeline start or end point,
 /// the oldest point on a timeline that needs to be retained because of PITR
--- a/libs/tracing-utils/Cargo.toml
+++ b/libs/tracing-utils/Cargo.toml
@@ -14,3 +14,5 @@ tokio = { workspace = true, features = ["rt", "rt-multi-thread"] }
 tracing.workspace = true
 tracing-opentelemetry.workspace = true
 tracing-subscriber.workspace = true
+
+workspace_hack.workspace = true
--- a/libs/utils/Cargo.toml
+++ b/libs/utils/Cargo.toml
@@ -14,6 +14,7 @@ testing = ["fail/failpoints"]
 arc-swap.workspace = true
 sentry.workspace = true
 async-compression.workspace = true
+async-trait.workspace = true
 anyhow.workspace = true
 bincode.workspace = true
 bytes.workspace = true
@@ -25,6 +26,7 @@ hyper = { workspace = true, features = ["full"] }
 fail.workspace = true
 futures = { workspace = true}
 jsonwebtoken.workspace = true
+leaky-bucket.workspace = true
 nix.workspace = true
 once_cell.workspace = true
 pin-project-lite.workspace = true
@@ -37,7 +39,7 @@ thiserror.workspace = true
 tokio.workspace = true
 tokio-tar.workspace = true
 tokio-util.workspace = true
-toml_edit = { workspace = true, features = ["serde"] }
+toml_edit.workspace = true
 tracing.workspace = true
 tracing-error.workspace = true
 tracing-subscriber = { workspace = true, features = ["json", "registry"] }
@@ -52,6 +54,7 @@ walkdir.workspace = true
 pq_proto.workspace = true
 postgres_connection.workspace = true
 metrics.workspace = true
+workspace_hack.workspace = true

 const_format.workspace = true

@@ -68,7 +71,6 @@ criterion.workspace = true
 hex-literal.workspace = true
 camino-tempfile.workspace = true
 serde_assert.workspace = true
-tokio = { workspace = true, features = ["test-util"] }

 [[bench]]
 name = "benchmarks"
--- a/libs/utils/src/circuit_breaker.rs
+++ b/libs/utils/src/circuit_breaker.rs
@@ -5,10 +5,8 @@ use std::{

 use metrics::IntCounter;

-/// Circuit breakers are for operations that are expensive and fallible.
-///
-/// If a circuit breaker fails repeatedly, we will stop attempting it for some
-/// period of time, to avoid denial-of-service from retries, and
+/// Circuit breakers are for operations that are expensive and fallible: if they fail repeatedly,
+/// we will stop attempting them for some period of time, to avoid denial-of-service from retries, and
 /// to mitigate the log spam from repeated failures.
 pub struct CircuitBreaker {
    /// An identifier that enables us to log useful errors when a circuit is broken
--- a/libs/utils/src/completion.rs
+++ b/libs/utils/src/completion.rs
@@ -5,40 +5,13 @@ use tokio_util::task::{task_tracker::TaskTrackerToken, TaskTracker};
 /// Can be cloned, moved and kept around in futures as "guard objects".
 #[derive(Clone)]
 pub struct Completion {
-    token: TaskTrackerToken,
-}
-
-impl std::fmt::Debug for Completion {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("Completion")
-            .field("siblings", &self.token.task_tracker().len())
-            .finish()
-    }
-}
-
-impl Completion {
-    /// Returns true if this completion is associated with the given barrier.
-    pub fn blocks(&self, barrier: &Barrier) -> bool {
-        TaskTracker::ptr_eq(self.token.task_tracker(), &barrier.0)
-    }
-
-    pub fn barrier(&self) -> Barrier {
-        Barrier(self.token.task_tracker().clone())
-    }
+    _token: TaskTrackerToken,
 }

 /// Barrier will wait until all clones of [`Completion`] have been dropped.
 #[derive(Clone)]
 pub struct Barrier(TaskTracker);

-impl std::fmt::Debug for Barrier {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        f.debug_struct("Barrier")
-            .field("remaining", &self.0.len())
-            .finish()
-    }
-}
-
 impl Default for Barrier {
    fn default() -> Self {
        let (_, rx) = channel();
@@ -78,5 +51,5 @@ pub fn channel() -> (Completion, Barrier) {
    tracker.close();

    let token = tracker.token();
-    (Completion { token }, Barrier(tracker))
+    (Completion { _token: token }, Barrier(tracker))
 }
--- a/libs/utils/src/crashsafe.rs
+++ b/libs/utils/src/crashsafe.rs
@@ -1,4 +1,3 @@
-use std::os::fd::AsRawFd;
 use std::{
    borrow::Cow,
    fs::{self, File},
@@ -204,27 +203,6 @@ pub fn overwrite(
    Ok(())
 }

-/// Syncs the filesystem for the given file descriptor.
-#[cfg_attr(target_os = "macos", allow(unused_variables))]
-pub fn syncfs(fd: impl AsRawFd) -> anyhow::Result<()> {
-    // Linux guarantees durability for syncfs.
-    // POSIX doesn't have syncfs, and further does not actually guarantee durability of sync().
-    #[cfg(target_os = "linux")]
-    {
-        use anyhow::Context;
-        nix::unistd::syncfs(fd.as_raw_fd()).context("syncfs")?;
-    }
-    #[cfg(target_os = "macos")]
-    {
-        // macOS is not a production platform for Neon, don't even bother.
-    }
-    #[cfg(not(any(target_os = "linux", target_os = "macos")))]
-    {
-        compile_error!("Unsupported OS");
-    }
-    Ok(())
-}
-
 #[cfg(test)]
 mod tests {

--- a/libs/utils/src/id.rs
+++ b/libs/utils/src/id.rs
@@ -249,10 +249,8 @@ macro_rules! id_newtype {
    };
 }

-/// Neon timeline ID.
-///
-/// They are different from PostgreSQL timeline
-/// IDs, but serve a similar purpose: they differentiate
+/// Neon timeline IDs are different from PostgreSQL timeline
+/// IDs. They serve a similar purpose though: they differentiate
 /// between different "histories" of the same cluster.  However,
 /// PostgreSQL timeline IDs are a bit cumbersome, because they are only
 /// 32-bits wide, and they must be in ascending order in any given
--- a/libs/utils/src/leaky_bucket.rs
+++ b/libs/utils/src/leaky_bucket.rs
@@ -1,280 +0,0 @@
-//! This module implements the Generic Cell Rate Algorithm for a simplified
-//! version of the Leaky Bucket rate limiting system.
-//!
-//! # Leaky Bucket
-//!
-//! If the bucket is full, no new requests are allowed and are throttled/errored.
-//! If the bucket is partially full/empty, new requests are added to the bucket in
-//! terms of "tokens".
-//!
-//! Over time, tokens are removed from the bucket, naturally allowing new requests at a steady rate.
-//!
-//! The bucket size tunes the burst support. The drain rate tunes the steady-rate requests per second.
-//!
-//! # [GCRA](https://en.wikipedia.org/wiki/Generic_cell_rate_algorithm)
-//!
-//! GCRA is a continuous rate leaky-bucket impl that stores minimal state and requires
-//! no background jobs to drain tokens, as the design utilises timestamps to drain automatically over time.
-//!
-//! We store an "empty_at" timestamp as the only state. As time progresses, we will naturally approach
-//! the empty state. The full-bucket state is calculated from `empty_at - config.bucket_width`.
-//!
-//! Another explaination can be found here: <https://brandur.org/rate-limiting>
-
-use std::{sync::Mutex, time::Duration};
-
-use tokio::{sync::Notify, time::Instant};
-
-pub struct LeakyBucketConfig {
-    /// This is the "time cost" of a single request unit.
-    /// Should loosely represent how long it takes to handle a request unit in active resource time.
-    /// Loosely speaking this is the inverse of the steady-rate requests-per-second
-    pub cost: Duration,
-
-    /// total size of the bucket
-    pub bucket_width: Duration,
-}
-
-impl LeakyBucketConfig {
-    pub fn new(rps: f64, bucket_size: f64) -> Self {
-        let cost = Duration::from_secs_f64(rps.recip());
-        let bucket_width = cost.mul_f64(bucket_size);
-        Self { cost, bucket_width }
-    }
-}
-
-pub struct LeakyBucketState {
-    /// Bucket is represented by `allow_at..empty_at` where `allow_at = empty_at - config.bucket_width`.
-    ///
-    /// At any given time, `empty_at - now` represents the number of tokens in the bucket, multiplied by the "time_cost".
-    /// Adding `n` tokens to the bucket is done by moving `empty_at` forward by `n * config.time_cost`.
-    /// If `now < allow_at`, the bucket is considered filled and cannot accept any more tokens.
-    /// Draining the bucket will happen naturally as `now` moves forward.
-    ///
-    /// Let `n` be some "time cost" for the request,
-    /// If now is after empty_at, the bucket is empty and the empty_at is reset to now,
-    /// If now is within the `bucket window + n`, we are within time budget.
-    /// If now is before the `bucket window + n`, we have run out of budget.
-    ///
-    /// This is inspired by the generic cell rate algorithm (GCRA) and works
-    /// exactly the same as a leaky-bucket.
-    pub empty_at: Instant,
-}
-
-impl LeakyBucketState {
-    pub fn with_initial_tokens(config: &LeakyBucketConfig, initial_tokens: f64) -> Self {
-        LeakyBucketState {
-            empty_at: Instant::now() + config.cost.mul_f64(initial_tokens),
-        }
-    }
-
-    pub fn bucket_is_empty(&self, now: Instant) -> bool {
-        // if self.end is after now, the bucket is not empty
-        self.empty_at <= now
-    }
-
-    /// Immediately adds tokens to the bucket, if there is space.
-    ///
-    /// In a scenario where you are waiting for available rate,
-    /// rather than just erroring immediately, `started` corresponds to when this waiting started.
-    ///
-    /// `n` is the number of tokens that will be filled in the bucket.
-    ///
-    /// # Errors
-    ///
-    /// If there is not enough space, no tokens are added. Instead, an error is returned with the time when
-    /// there will be space again.
-    pub fn add_tokens(
-        &mut self,
-        config: &LeakyBucketConfig,
-        started: Instant,
-        n: f64,
-    ) -> Result<(), Instant> {
-        let now = Instant::now();
-
-        // invariant: started <= now
-        debug_assert!(started <= now);
-
-        // If the bucket was empty when we started our search,
-        // we should update the `empty_at` value accordingly.
-        // this prevents us from having negative tokens in the bucket.
-        let mut empty_at = self.empty_at;
-        if empty_at < started {
-            empty_at = started;
-        }
-
-        let n = config.cost.mul_f64(n);
-        let new_empty_at = empty_at + n;
-        let allow_at = new_empty_at.checked_sub(config.bucket_width);
-
-        //                     empty_at
-        //          allow_at    |   new_empty_at
-        //           /          |   /
-        // -------o-[---------o-|--]---------
-        //   now1 ^      now2 ^
-        //
-        // at now1, the bucket would be completely filled if we add n tokens.
-        // at now2, the bucket would be partially filled if we add n tokens.
-
-        match allow_at {
-            Some(allow_at) if now < allow_at => Err(allow_at),
-            _ => {
-                self.empty_at = new_empty_at;
-                Ok(())
-            }
-        }
-    }
-}
-
-pub struct RateLimiter {
-    pub config: LeakyBucketConfig,
-    pub state: Mutex<LeakyBucketState>,
-    /// a queue to provide this fair ordering.
-    pub queue: Notify,
-}
-
-struct Requeue<'a>(&'a Notify);
-
-impl Drop for Requeue<'_> {
-    fn drop(&mut self) {
-        self.0.notify_one();
-    }
-}
-
-impl RateLimiter {
-    pub fn with_initial_tokens(config: LeakyBucketConfig, initial_tokens: f64) -> Self {
-        RateLimiter {
-            state: Mutex::new(LeakyBucketState::with_initial_tokens(
-                &config,
-                initial_tokens,
-            )),
-            config,
-            queue: {
-                let queue = Notify::new();
-                queue.notify_one();
-                queue
-            },
-        }
-    }
-
-    pub fn steady_rps(&self) -> f64 {
-        self.config.cost.as_secs_f64().recip()
-    }
-
-    /// returns true if we did throttle
-    pub async fn acquire(&self, count: usize) -> bool {
-        let mut throttled = false;
-
-        let start = tokio::time::Instant::now();
-
-        // wait until we are the first in the queue
-        let mut notified = std::pin::pin!(self.queue.notified());
-        if !notified.as_mut().enable() {
-            throttled = true;
-            notified.await;
-        }
-
-        // notify the next waiter in the queue when we are done.
-        let _guard = Requeue(&self.queue);
-
-        loop {
-            let res = self
-                .state
-                .lock()
-                .unwrap()
-                .add_tokens(&self.config, start, count as f64);
-            match res {
-                Ok(()) => return throttled,
-                Err(ready_at) => {
-                    throttled = true;
-                    tokio::time::sleep_until(ready_at).await;
-                }
-            }
-        }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use std::time::Duration;
-
-    use tokio::time::Instant;
-
-    use super::{LeakyBucketConfig, LeakyBucketState};
-
-    #[tokio::test(start_paused = true)]
-    async fn check() {
-        let config = LeakyBucketConfig {
-            // average 100rps
-            cost: Duration::from_millis(10),
-            // burst up to 100 requests
-            bucket_width: Duration::from_millis(1000),
-        };
-
-        let mut state = LeakyBucketState {
-            empty_at: Instant::now(),
-        };
-
-        // supports burst
-        {
-            // should work for 100 requests this instant
-            for _ in 0..100 {
-                state.add_tokens(&config, Instant::now(), 1.0).unwrap();
-            }
-            let ready = state.add_tokens(&config, Instant::now(), 1.0).unwrap_err();
-            assert_eq!(ready - Instant::now(), Duration::from_millis(10));
-        }
-
-        // doesn't overfill
-        {
-            // after 1s we should have an empty bucket again.
-            tokio::time::advance(Duration::from_secs(1)).await;
-            assert!(state.bucket_is_empty(Instant::now()));
-
-            // after 1s more, we should not over count the tokens and allow more than 200 requests.
-            tokio::time::advance(Duration::from_secs(1)).await;
-            for _ in 0..100 {
-                state.add_tokens(&config, Instant::now(), 1.0).unwrap();
-            }
-            let ready = state.add_tokens(&config, Instant::now(), 1.0).unwrap_err();
-            assert_eq!(ready - Instant::now(), Duration::from_millis(10));
-        }
-
-        // supports sustained rate over a long period
-        {
-            tokio::time::advance(Duration::from_secs(1)).await;
-
-            // should sustain 100rps
-            for _ in 0..2000 {
-                tokio::time::advance(Duration::from_millis(10)).await;
-                state.add_tokens(&config, Instant::now(), 1.0).unwrap();
-            }
-        }
-
-        // supports requesting more tokens than can be stored in the bucket
-        // we just wait a little bit longer upfront.
-        {
-            // start the bucket completely empty
-            tokio::time::advance(Duration::from_secs(5)).await;
-            assert!(state.bucket_is_empty(Instant::now()));
-
-            // requesting 200 tokens of space should take 200*cost = 2s
-            // but we already have 1s available, so we wait 1s from start.
-            let start = Instant::now();
-
-            let ready = state.add_tokens(&config, start, 200.0).unwrap_err();
-            assert_eq!(ready - Instant::now(), Duration::from_secs(1));
-
-            tokio::time::advance(Duration::from_millis(500)).await;
-            let ready = state.add_tokens(&config, start, 200.0).unwrap_err();
-            assert_eq!(ready - Instant::now(), Duration::from_millis(500));
-
-            tokio::time::advance(Duration::from_millis(500)).await;
-            state.add_tokens(&config, start, 200.0).unwrap();
-
-            // bucket should be completely full now
-            let ready = state.add_tokens(&config, Instant::now(), 1.0).unwrap_err();
-            assert_eq!(ready - Instant::now(), Duration::from_millis(10));
-        }
-    }
-}
--- a/libs/utils/src/lib.rs
+++ b/libs/utils/src/lib.rs
@@ -71,7 +71,6 @@ pub mod postgres_client;

 pub mod tracing_span_assert;

-pub mod leaky_bucket;
 pub mod rate_limit;

 /// Simple once-barrier and a guard which keeps barrier awaiting.
--- a/libs/utils/src/lock_file.rs
+++ b/libs/utils/src/lock_file.rs
@@ -100,9 +100,7 @@ pub enum LockFileRead {
 }

 /// Open & try to lock the lock file at the given `path`, returning a [handle][`LockFileRead`] to
-/// inspect its content.
-///
-/// It is not an `Err(...)` if the file does not exist or is already locked.
+/// inspect its content. It is not an `Err(...)` if the file does not exist or is already locked.
 /// Check the [`LockFileRead`] variants for details.
 pub fn read_and_hold_lock_file(path: &Utf8Path) -> anyhow::Result<LockFileRead> {
    let res = fs::OpenOptions::new().read(true).open(path);
--- a/libs/utils/src/logging.rs
+++ b/libs/utils/src/logging.rs
@@ -5,9 +5,7 @@ use metrics::{IntCounter, IntCounterVec};
 use once_cell::sync::Lazy;
 use strum_macros::{EnumString, EnumVariantNames};

-#[derive(
-    EnumString, strum_macros::Display, EnumVariantNames, Eq, PartialEq, Debug, Clone, Copy,
-)]
+#[derive(EnumString, EnumVariantNames, Eq, PartialEq, Debug, Clone, Copy)]
 #[strum(serialize_all = "snake_case")]
 pub enum LogFormat {
    Plain,
@@ -190,7 +188,7 @@ impl Drop for TracingPanicHookGuard {
 }

 /// Named symbol for our panic hook, which logs the panic.
-fn tracing_panic_hook(info: &std::panic::PanicHookInfo) {
+fn tracing_panic_hook(info: &std::panic::PanicInfo) {
    // following rust 1.66.1 std implementation:
    // https://github.com/rust-lang/rust/blob/90743e7298aca107ddaa0c202a4d3604e29bfeb6/library/std/src/panicking.rs#L235-L288
    let location = info.location();
@@ -276,14 +274,6 @@ impl From<String> for SecretString {
    }
 }

-impl FromStr for SecretString {
-    type Err = std::convert::Infallible;
-
-    fn from_str(s: &str) -> Result<Self, Self::Err> {
-        Ok(Self(s.to_string()))
-    }
-}
-
 impl std::fmt::Debug for SecretString {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        write!(f, "[SECRET]")
--- a/libs/utils/src/pageserver_feedback.rs
+++ b/libs/utils/src/pageserver_feedback.rs
@@ -8,7 +8,6 @@ use tracing::{trace, warn};
 use crate::lsn::Lsn;

 /// Feedback pageserver sends to safekeeper and safekeeper resends to compute.
-///
 /// Serialized in custom flexible key/value format. In replication protocol, it
 /// is marked with NEON_STATUS_UPDATE_TAG_BYTE to differentiate from postgres
 /// Standby status update / Hot standby feedback messages.
--- a/libs/utils/src/poison.rs
+++ b/libs/utils/src/poison.rs
@@ -65,8 +65,6 @@ impl<T> Poison<T> {
    }
 }

-/// Armed pointer to a [`Poison`].
-///
 /// Use [`Self::data`] and [`Self::data_mut`] to access the wrapped state.
 /// Once modifications are done, use [`Self::disarm`].
 /// If [`Guard`] gets dropped instead of calling [`Self::disarm`], the state is poisoned
--- a/libs/utils/src/rate_limit.rs
+++ b/libs/utils/src/rate_limit.rs
@@ -5,15 +5,6 @@ use std::time::{Duration, Instant};
 pub struct RateLimit {
    last: Option<Instant>,
    interval: Duration,
-    dropped: u64,
-}
-
-pub struct RateLimitStats(u64);
-
-impl std::fmt::Display for RateLimitStats {
-    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
-        write!(f, "{} dropped calls", self.0)
-    }
 }

 impl RateLimit {
@@ -21,27 +12,20 @@ impl RateLimit {
        Self {
            last: None,
            interval,
-            dropped: 0,
        }
    }

    /// Call `f` if the rate limit allows.
    /// Don't call it otherwise.
    pub fn call<F: FnOnce()>(&mut self, f: F) {
-        self.call2(|_| f())
-    }
-
-    pub fn call2<F: FnOnce(RateLimitStats)>(&mut self, f: F) {
        let now = Instant::now();
        match self.last {
            Some(last) if now - last <= self.interval => {
                // ratelimit
-                self.dropped += 1;
            }
            _ => {
                self.last = Some(now);
-                f(RateLimitStats(self.dropped));
-                self.dropped = 0;
+                f();
            }
        }
    }
--- a/libs/utils/src/shard.rs
+++ b/libs/utils/src/shard.rs
@@ -13,11 +13,10 @@ pub struct ShardNumber(pub u8);
 #[derive(Ord, PartialOrd, Eq, PartialEq, Clone, Copy, Serialize, Deserialize, Debug, Hash)]
 pub struct ShardCount(pub u8);

-/// Combination of ShardNumber and ShardCount.
-///
-/// For use within the context of a particular tenant, when we need to know which shard we're
-/// dealing with, but do not need to know the full ShardIdentity (because we won't be doing
-/// any page->shard mapping), and do not need to know the fully qualified TenantShardId.
+/// Combination of ShardNumber and ShardCount.  For use within the context of a particular tenant,
+/// when we need to know which shard we're dealing with, but do not need to know the full
+/// ShardIdentity (because we won't be doing any page->shard mapping), and do not need to know
+/// the fully qualified TenantShardId.
 #[derive(Eq, PartialEq, PartialOrd, Ord, Clone, Copy, Hash)]
 pub struct ShardIndex {
    pub shard_number: ShardNumber,
--- a/libs/utils/src/simple_rcu.rs
+++ b/libs/utils/src/simple_rcu.rs
@@ -49,11 +49,12 @@ use std::sync::{RwLock, RwLockWriteGuard};

 use tokio::sync::watch;

-/// Rcu allows multiple readers to read and hold onto a value without blocking
-/// (for very long).
 ///
-/// Storing to the Rcu updates the value, making new readers immediately see
-/// the new value, but it also waits for all current readers to finish.
+/// Rcu allows multiple readers to read and hold onto a value without blocking
+/// (for very long).  Storing to the Rcu updates the value, making new readers
+/// immediately see the new value, but it also waits for all current readers to
+/// finish.
+///
 pub struct Rcu<V> {
    inner: RwLock<RcuInner<V>>,
 }
--- a/libs/utils/src/sync/heavier_once_cell.rs
+++ b/libs/utils/src/sync/heavier_once_cell.rs
@@ -5,9 +5,7 @@ use std::sync::{
 use tokio::sync::Semaphore;

 /// Custom design like [`tokio::sync::OnceCell`] but using [`OwnedSemaphorePermit`] instead of
-/// `SemaphorePermit`.
-///
-/// Allows use of `take` which does not require holding an outer mutex guard
+/// `SemaphorePermit`, allowing use of `take` which does not require holding an outer mutex guard
 /// for the duration of initialization.
 ///
 /// Has no unsafe, builds upon [`tokio::sync::Semaphore`] and [`std::sync::Mutex`].
--- a/libs/utils/src/vec_map.rs
+++ b/libs/utils/src/vec_map.rs
@@ -7,7 +7,6 @@ pub enum VecMapOrdering {
 }

 /// Ordered map datastructure implemented in a Vec.
-///
 /// Append only - can only add keys that are larger than the
 /// current max key.
 /// Ordering can be adjusted using [`VecMapOrdering`]
--- a/libs/utils/src/yielding_loop.rs
+++ b/libs/utils/src/yielding_loop.rs
@@ -6,10 +6,9 @@ pub enum YieldingLoopError {
    Cancelled,
 }

-/// Helper for long synchronous loops, e.g. over all tenants in the system.
-///
-/// Periodically yields to avoid blocking the executor, and after resuming
-/// checks the provided cancellation token to drop out promptly on shutdown.
+/// Helper for long synchronous loops, e.g. over all tenants in the system.  Periodically
+/// yields to avoid blocking the executor, and after resuming checks the provided
+/// cancellation token to drop out promptly on shutdown.
 #[inline(always)]
 pub async fn yielding_loop<I, T, F>(
    interval: usize,
@@ -24,7 +23,7 @@ where
    for (i, item) in iter.enumerate() {
        visitor(item);

-        if (i + 1) % interval == 0 {
+        if i + 1 % interval == 0 {
            tokio::task::yield_now().await;
            if cancel.is_cancelled() {
                return Err(YieldingLoopError::Cancelled);
--- a/libs/walproposer/Cargo.toml
+++ b/libs/walproposer/Cargo.toml
@@ -9,6 +9,8 @@ anyhow.workspace = true
 utils.workspace = true
 postgres_ffi.workspace = true

+workspace_hack.workspace = true
+
 [build-dependencies]
 anyhow.workspace = true
 bindgen.workspace = true
--- a/libs/walproposer/build.rs
+++ b/libs/walproposer/build.rs
@@ -4,6 +4,7 @@
 use std::{env, path::PathBuf, process::Command};

 use anyhow::{anyhow, Context};
+use bindgen::CargoCallbacks;

 fn main() -> anyhow::Result<()> {
    // Tell cargo to invalidate the built crate whenever the wrapper changes
@@ -63,25 +64,16 @@ fn main() -> anyhow::Result<()> {
            .map_err(|s| anyhow!("Bad postgres server path {s:?}"))?
    };

-    let unwind_abi_functions = [
-        "log_internal",
-        "recovery_download",
-        "start_streaming",
-        "finish_sync_safekeepers",
-        "wait_event_set",
-        "WalProposerStart",
-    ];
-
    // The bindgen::Builder is the main entry point
    // to bindgen, and lets you build up options for
    // the resulting bindings.
-    let mut builder = bindgen::Builder::default()
+    let bindings = bindgen::Builder::default()
        // The input header we would like to generate
        // bindings for.
        .header("bindgen_deps.h")
        // Tell cargo to invalidate the built crate whenever any of the
        // included header files changed.
-        .parse_callbacks(Box::new(bindgen::CargoCallbacks::new()))
+        .parse_callbacks(Box::new(CargoCallbacks))
        .allowlist_type("WalProposer")
        .allowlist_type("WalProposerConfig")
        .allowlist_type("walproposer_api")
@@ -103,7 +95,6 @@ fn main() -> anyhow::Result<()> {
        .allowlist_var("ERROR")
        .allowlist_var("FATAL")
        .allowlist_var("PANIC")
-        .allowlist_var("PG_VERSION_NUM")
        .allowlist_var("WPEVENT")
        .allowlist_var("WL_LATCH_SET")
        .allowlist_var("WL_SOCKET_READABLE")
@@ -113,12 +104,7 @@ fn main() -> anyhow::Result<()> {
        .allowlist_var("WL_SOCKET_MASK")
        .clang_arg("-DWALPROPOSER_LIB")
        .clang_arg(format!("-I{pgxn_neon}"))
-        .clang_arg(format!("-I{inc_server_path}"));
-
-    for name in unwind_abi_functions {
-        builder = builder.override_abi(bindgen::Abi::CUnwind, name);
-    }
-    let bindings = builder
+        .clang_arg(format!("-I{inc_server_path}"))
        // Finish the builder and generate the bindings.
        .generate()
        // Unwrap the Result and panic on failure.
--- a/libs/walproposer/src/api_bindings.rs
+++ b/libs/walproposer/src/api_bindings.rs
@@ -33,7 +33,7 @@ extern "C" fn get_shmem_state(wp: *mut WalProposer) -> *mut WalproposerShmemStat
    }
 }

-extern "C-unwind" fn start_streaming(wp: *mut WalProposer, startpos: XLogRecPtr) {
+extern "C" fn start_streaming(wp: *mut WalProposer, startpos: XLogRecPtr) {
    unsafe {
        let callback_data = (*(*wp).config).callback_data;
        let api = callback_data as *mut Box<dyn ApiImpl>;
@@ -187,7 +187,7 @@ extern "C" fn conn_blocking_write(
    }
 }

-extern "C-unwind" fn recovery_download(wp: *mut WalProposer, sk: *mut Safekeeper) -> bool {
+extern "C" fn recovery_download(wp: *mut WalProposer, sk: *mut Safekeeper) -> bool {
    unsafe {
        let callback_data = (*(*(*sk).wp).config).callback_data;
        let api = callback_data as *mut Box<dyn ApiImpl>;
@@ -272,7 +272,7 @@ extern "C" fn rm_safekeeper_event_set(sk: *mut Safekeeper) {
    }
 }

-extern "C-unwind" fn wait_event_set(
+extern "C" fn wait_event_set(
    wp: *mut WalProposer,
    timeout: ::std::os::raw::c_long,
    event_sk: *mut *mut Safekeeper,
@@ -324,7 +324,7 @@ extern "C" fn get_redo_start_lsn(wp: *mut WalProposer) -> XLogRecPtr {
    }
 }

-extern "C-unwind" fn finish_sync_safekeepers(wp: *mut WalProposer, lsn: XLogRecPtr) {
+extern "C" fn finish_sync_safekeepers(wp: *mut WalProposer, lsn: XLogRecPtr) {
    unsafe {
        let callback_data = (*(*wp).config).callback_data;
        let api = callback_data as *mut Box<dyn ApiImpl>;
@@ -340,7 +340,7 @@ extern "C" fn process_safekeeper_feedback(wp: *mut WalProposer, sk: *mut Safekee
    }
 }

-extern "C-unwind" fn log_internal(
+extern "C" fn log_internal(
    wp: *mut WalProposer,
    level: ::std::os::raw::c_int,
    line: *const ::std::os::raw::c_char,
--- a/libs/walproposer/src/walproposer.rs
+++ b/libs/walproposer/src/walproposer.rs
@@ -282,11 +282,7 @@ mod tests {
    use std::cell::UnsafeCell;
    use utils::id::TenantTimelineId;

-    use crate::{
-        api_bindings::Level,
-        bindings::{NeonWALReadResult, PG_VERSION_NUM},
-        walproposer::Wrapper,
-    };
+    use crate::{api_bindings::Level, bindings::NeonWALReadResult, walproposer::Wrapper};

    use super::ApiImpl;

@@ -493,79 +489,41 @@ mod tests {

        let (sender, receiver) = sync_channel(1);

-        // Messages definitions are at walproposer.h
-        // xxx: it would be better to extract them from safekeeper crate and
-        // use serialization/deserialization here.
-        let greeting_tag = (b'g' as u64).to_ne_bytes();
-        let proto_version = 2_u32.to_ne_bytes();
-        let pg_version: [u8; 4] = PG_VERSION_NUM.to_ne_bytes();
-        let proposer_id = [0; 16];
-        let system_id = 0_u64.to_ne_bytes();
-        let tenant_id = ttid.tenant_id.as_arr();
-        let timeline_id = ttid.timeline_id.as_arr();
-        let pg_tli = 1_u32.to_ne_bytes();
-        let wal_seg_size = 16777216_u32.to_ne_bytes();
-        let proposer_greeting = [
-            greeting_tag.as_slice(),
-            proto_version.as_slice(),
-            pg_version.as_slice(),
-            proposer_id.as_slice(),
-            system_id.as_slice(),
-            tenant_id.as_slice(),
-            timeline_id.as_slice(),
-            pg_tli.as_slice(),
-            wal_seg_size.as_slice(),
-        ]
-        .concat();
-
-        let voting_tag = (b'v' as u64).to_ne_bytes();
-        let vote_request_term = 3_u64.to_ne_bytes();
-        let proposer_id = [0; 16];
-        let vote_request = [
-            voting_tag.as_slice(),
-            vote_request_term.as_slice(),
-            proposer_id.as_slice(),
-        ]
-        .concat();
-
-        let acceptor_greeting_term = 2_u64.to_ne_bytes();
-        let acceptor_greeting_node_id = 1_u64.to_ne_bytes();
-        let acceptor_greeting = [
-            greeting_tag.as_slice(),
-            acceptor_greeting_term.as_slice(),
-            acceptor_greeting_node_id.as_slice(),
-        ]
-        .concat();
-
-        let vote_response_term = 3_u64.to_ne_bytes();
-        let vote_given = 1_u64.to_ne_bytes();
-        let flush_lsn = 0x539_u64.to_ne_bytes();
-        let truncate_lsn = 0x539_u64.to_ne_bytes();
-        let th_len = 1_u32.to_ne_bytes();
-        let th_term = 2_u64.to_ne_bytes();
-        let th_lsn = 0x539_u64.to_ne_bytes();
-        let timeline_start_lsn = 0x539_u64.to_ne_bytes();
-        let vote_response = [
-            voting_tag.as_slice(),
-            vote_response_term.as_slice(),
-            vote_given.as_slice(),
-            flush_lsn.as_slice(),
-            truncate_lsn.as_slice(),
-            th_len.as_slice(),
-            th_term.as_slice(),
-            th_lsn.as_slice(),
-            timeline_start_lsn.as_slice(),
-        ]
-        .concat();
-
        let my_impl: Box<dyn ApiImpl> = Box::new(MockImpl {
            wait_events: Cell::new(WaitEventsData {
                sk: std::ptr::null_mut(),
                event_mask: 0,
            }),
-            expected_messages: vec![proposer_greeting, vote_request],
+            expected_messages: vec![
+                // TODO: When updating Postgres versions, this test will cause
+                // problems. Postgres version in message needs updating.
+                //
+                // Greeting(ProposerGreeting { protocol_version: 2, pg_version: 160003, proposer_id: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], system_id: 0, timeline_id: 9e4c8f36063c6c6e93bc20d65a820f3d, tenant_id: 9e4c8f36063c6c6e93bc20d65a820f3d, tli: 1, wal_seg_size: 16777216 })
+                vec![
+                    103, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 3, 113, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 158, 76, 143, 54, 6, 60, 108, 110,
+                    147, 188, 32, 214, 90, 130, 15, 61, 158, 76, 143, 54, 6, 60, 108, 110, 147,
+                    188, 32, 214, 90, 130, 15, 61, 1, 0, 0, 0, 0, 0, 0, 1,
+                ],
+                // VoteRequest(VoteRequest { term: 3 })
+                vec![
+                    118, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                    0, 0, 0, 0, 0, 0,
+                ],
+            ],
            expected_ptr: AtomicUsize::new(0),
-            safekeeper_replies: vec![acceptor_greeting, vote_response],
+            safekeeper_replies: vec![
+                // Greeting(AcceptorGreeting { term: 2, node_id: NodeId(1) })
+                vec![
+                    103, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0,
+                ],
+                // VoteResponse(VoteResponse { term: 3, vote_given: 1, flush_lsn: 0/539, truncate_lsn: 0/539, term_history: [(2, 0/539)], timeline_start_lsn: 0/539 })
+                vec![
+                    118, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 57,
+                    5, 0, 0, 0, 0, 0, 0, 57, 5, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0,
+                    0, 57, 5, 0, 0, 0, 0, 0, 0, 57, 5, 0, 0, 0, 0, 0, 0,
+                ],
+            ],
            replies_ptr: AtomicUsize::new(0),
            sync_channel: sender,
            shmem: UnsafeCell::new(crate::api_bindings::empty_shmem()),
--- a/pageserver/Cargo.toml
+++ b/pageserver/Cargo.toml
@@ -8,7 +8,7 @@ license.workspace = true
 default = []
 # Enables test-only APIs, incuding failpoints. In particular, enables the `fail_point!` macro,
 # which adds some runtime cost to run tests on outage conditions
-testing = ["fail/failpoints", "pageserver_api/testing" ]
+testing = ["fail/failpoints"]

 [dependencies]
 anyhow.workspace = true
@@ -16,7 +16,6 @@ arc-swap.workspace = true
 async-compression.workspace = true
 async-stream.workspace = true
 async-trait.workspace = true
-bit_field.workspace = true
 byteorder.workspace = true
 bytes.workspace = true
 camino.workspace = true
@@ -37,6 +36,7 @@ humantime.workspace = true
 humantime-serde.workspace = true
 hyper.workspace = true
 itertools.workspace = true
+leaky-bucket.workspace = true
 md5.workspace = true
 nix.workspace = true
 # hack to get the number of worker threads tokio uses
@@ -52,7 +52,6 @@ rand.workspace = true
 range-set-blaze = { version = "0.1.16", features = ["alloc"] }
 regex.workspace = true
 scopeguard.workspace = true
-send-future.workspace = true
 serde.workspace = true
 serde_json = { workspace = true, features = ["raw_value"] }
 serde_path_to_error.workspace = true
@@ -101,7 +100,6 @@ procfs.workspace = true
 criterion.workspace = true
 hex-literal.workspace = true
 tokio = { workspace = true, features = ["process", "sync", "fs", "rt", "io-util", "time", "test-util"] }
-indoc.workspace = true

 [[bench]]
 name = "bench_layer_map"
--- a/pageserver/benches/bench_ingest.rs
+++ b/pageserver/benches/bench_ingest.rs
@@ -10,7 +10,6 @@ use pageserver::{
    page_cache,
    repository::Value,
    task_mgr::TaskKind,
-    tenant::storage_layer::inmemory_layer::SerializedBatch,
    tenant::storage_layer::InMemoryLayer,
    virtual_file,
 };
@@ -68,16 +67,12 @@ async fn ingest(
    let layer =
        InMemoryLayer::create(conf, timeline_id, tenant_shard_id, lsn, entered, &ctx).await?;

-    let data = Value::Image(Bytes::from(vec![0u8; put_size]));
-    let data_ser_size = data.serialized_size().unwrap() as usize;
+    let data = Value::Image(Bytes::from(vec![0u8; put_size])).ser()?;
    let ctx = RequestContext::new(
        pageserver::task_mgr::TaskKind::WalReceiverConnectionHandler,
        pageserver::context::DownloadBehavior::Download,
    );

-    const BATCH_SIZE: usize = 16;
-    let mut batch = Vec::new();
-
    for i in 0..put_count {
        lsn += put_size as u64;

@@ -100,17 +95,7 @@ async fn ingest(
            }
        }

-        batch.push((key.to_compact(), lsn, data_ser_size, data.clone()));
-        if batch.len() >= BATCH_SIZE {
-            let this_batch = std::mem::take(&mut batch);
-            let serialized = SerializedBatch::from_values(this_batch).unwrap();
-            layer.put_batch(serialized, &ctx).await?;
-        }
-    }
-    if !batch.is_empty() {
-        let this_batch = std::mem::take(&mut batch);
-        let serialized = SerializedBatch::from_values(this_batch).unwrap();
-        layer.put_batch(serialized, &ctx).await?;
+        layer.put_value(key, lsn, &data, &ctx).await?;
    }
    layer.freeze(lsn + 1).await;

@@ -164,11 +149,7 @@ fn criterion_benchmark(c: &mut Criterion) {
    let conf: &'static PageServerConf = Box::leak(Box::new(
        pageserver::config::PageServerConf::dummy_conf(temp_dir.path().to_path_buf()),
    ));
-    virtual_file::init(
-        16384,
-        virtual_file::io_engine_for_bench(),
-        pageserver_api::config::defaults::DEFAULT_IO_BUFFER_ALIGNMENT,
-    );
+    virtual_file::init(16384, virtual_file::io_engine_for_bench());
    page_cache::init(conf.page_cache_size);

    {
--- a/pageserver/client/Cargo.toml
+++ b/pageserver/client/Cargo.toml
@@ -7,6 +7,7 @@ license.workspace = true
 [dependencies]
 pageserver_api.workspace = true
 thiserror.workspace = true
+async-trait.workspace = true
 reqwest = { workspace = true, features = [ "stream" ] }
 utils.workspace = true
 serde.workspace = true
--- a/pageserver/client/src/lib.rs
+++ b/pageserver/client/src/lib.rs
@@ -1,20 +1,2 @@
 pub mod mgmt_api;
 pub mod page_service;
-
-/// For timeline_block_unblock_gc, distinguish the two different operations. This could be a bool.
-// If file structure is per-kind not per-feature then where to put this?
-#[derive(Clone, Copy)]
-pub enum BlockUnblock {
-    Block,
-    Unblock,
-}
-
-impl std::fmt::Display for BlockUnblock {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        let s = match self {
-            BlockUnblock::Block => "block",
-            BlockUnblock::Unblock => "unblock",
-        };
-        f.write_str(s)
-    }
-}
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Conrad Ludgate	18303e4d68	clean up	2024-08-13 15:08:57 +01:00
Conrad Ludgate	3df6d368e3	split out binaries	2024-08-13 15:08:57 +01:00
Conrad Ludgate	b62e7c0138	proxy: experiment with idea to split crates	2024-08-13 15:08:54 +01:00
Conrad Ludgate	a2968c6cf8	move proxy to proxy/code	2024-08-13 15:01:48 +01:00
Conrad Ludgate	bae1288671	make jwk renewal permits a bit more type safe	2024-08-13 11:08:25 +01:00
Conrad Ludgate	1254d8f56e	address some comments	2024-08-13 10:24:14 +01:00
Conrad Ludgate	073508493c	remove async_trait for FetchAuthRules	2024-08-12 16:14:53 +01:00
Conrad Ludgate	7cb2349296	add jwks size limiter	2024-08-12 11:48:57 +01:00
Conrad Ludgate	87151f9efd	ignore marvin vuln	2024-08-12 09:01:30 +01:00
Conrad Ludgate	96fe084c57	compact mock server	2024-08-12 09:01:04 +01:00
Conrad Ludgate	20fdf3e19f	extract fetch/update routine	2024-08-12 09:01:04 +01:00
Conrad Ludgate	c6b36d8171	fix lints	2024-08-12 09:01:04 +01:00
Conrad Ludgate	0e8a848937	finish happy path test	2024-08-12 09:01:04 +01:00
Conrad Ludgate	db4085fe22	mock tests for jwk renewal	2024-08-12 09:01:04 +01:00
Conrad Ludgate	0d895ba002	strip down supported algorithms to just RS256 and ES256	2024-08-12 09:01:04 +01:00
Conrad Ludgate	103f34e954	flesh out JWKs cache	2024-08-12 09:01:04 +01:00
Conrad Ludgate	262378e561	flesh out jwt code	2024-08-12 09:01:04 +01:00
Conrad Ludgate	9f38ab39c6	stash jwts	2024-08-12 09:01:04 +01:00
Conrad Ludgate	fa92328423	start stubbing jwt	2024-08-12 09:01:04 +01:00