one more segfault in handling error messages

We don't want any external requests for an archived timeline. This
includes basebackup requests, i.e. when a compute is being started up.

Therefore, we'd like to forbid such basebackup requests: any attempt to
get a basebackup on an archived timeline (or any getpage request really)
is a cplane bug. Make this a warning for now so that, if there is
potentially a bug, we can detect cases in the wild before they cause
stuck operations, but the intention is to return an error eventually.

Related: #9548

.dockerignore

@@ -24,3 +24,4 @@
 !storage_controller/
 !vendor/postgres-*/
 !workspace_hack/
+!build_tools/patches

.github/ISSUE_TEMPLATE/bug-template.md

@@ -3,6 +3,7 @@ name: Bug Template
 about: Used for describing bugs
 title: ''
 labels: t/bug
+type: Bug
 assignees: ''
 
 ---

.github/ISSUE_TEMPLATE/epic-template.md

@@ -4,6 +4,7 @@ about: A set of related tasks contributing towards specific outcome, comprising
   more than 1 week of work.
 title: 'Epic: '
 labels: t/Epic
+type: Epic
 assignees: ''
 
 ---

.github/actionlint.yml

@@ -4,6 +4,7 @@ self-hosted-runner:
     - large
     - large-arm64
     - small
+    - small-metal
     - small-arm64
     - us-east-2
 config-variables:
@@ -26,3 +27,4 @@ config-variables:
   - SLACK_ON_CALL_QA_STAGING_STREAM
   - DEV_AWS_OIDC_ROLE_MANAGE_BENCHMARK_EC2_VMS_ARN
   - SLACK_ON_CALL_STORAGE_STAGING_STREAM
+  - SLACK_CICD_CHANNEL_ID

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

@@ -41,7 +41,10 @@ inputs:
     description: 'Path to directory containing libpq library - it is caller responsibility to provision the libpq library'
     required: false
     default: '/tmp/neon/pg_install/v16/lib'
-  
+  project_settings:
+    description: 'A JSON object with project settings'
+    required: false
+    default: '{}'
 
 outputs:
   dsn:
@@ -73,7 +76,7 @@ runs:
               \"provisioner\": \"k8s-neonvm\",
               \"autoscaling_limit_min_cu\": ${MIN_CU},
               \"autoscaling_limit_max_cu\": ${MAX_CU},
-              \"settings\": { }
+              \"settings\": ${PROJECT_SETTINGS}
             }
           }")
 
@@ -92,12 +95,12 @@ runs:
         if [ "${SHARD_SPLIT_PROJECT}" = "true" ]; then
           # determine tenant ID
           TENANT_ID=`${PSQL} ${dsn} -t -A -c "SHOW neon.tenant_id"`
-          
+
           echo "Splitting project ${project_id} with tenant_id ${TENANT_ID} into $((SHARD_COUNT)) shards with stripe size $((STRIPE_SIZE))"
 
           echo "Sending PUT request to https://${API_HOST}/regions/${REGION_ID}/api/v1/admin/storage/proxy/control/v1/tenant/${TENANT_ID}/shard_split"
           echo "with body {\"new_shard_count\": $((SHARD_COUNT)), \"new_stripe_size\": $((STRIPE_SIZE))}"
-          
+
           # we need an ADMIN API KEY to invoke storage controller API for shard splitting (bash -u above checks that the variable is set)
           curl -X PUT \
             "https://${API_HOST}/regions/${REGION_ID}/api/v1/admin/storage/proxy/control/v1/tenant/${TENANT_ID}/shard_split" \
@@ -118,3 +121,4 @@ runs:
         STRIPE_SIZE: ${{ inputs.stripe_size }}
         PSQL: ${{ inputs.psql_path }}
         LD_LIBRARY_PATH: ${{ inputs.libpq_lib_path }}
+        PROJECT_SETTINGS: ${{ inputs.project_settings }}

.github/file-filters.yaml

@@ -1,4 +1,5 @@
 rust_code: ['**/*.rs', '**/Cargo.toml', '**/Cargo.lock']
+rust_dependencies: ['**/Cargo.lock']
 
 v14: ['vendor/postgres-v14/**', 'Makefile', 'pgxn/**']
 v15: ['vendor/postgres-v15/**', 'Makefile', 'pgxn/**']

.github/workflows/_benchmarking_preparation.yml

@@ -17,7 +17,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        platform: [ aws-rds-postgres, aws-aurora-serverless-v2-postgres, neon ]
+        platform: [ aws-rds-postgres, aws-aurora-serverless-v2-postgres, neon, neon_pg17 ]
         database: [ clickbench, tpch, userexample ]
 
     env:
@@ -41,6 +41,9 @@ jobs:
           neon)
             CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR }}
             ;;
+          neon_pg17)
+            CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR_PG17 }}
+            ;;
           aws-rds-postgres)
             CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_CONNSTR }}
             ;;

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

@@ -267,6 +267,26 @@ jobs:
           path: /tmp/neon
           aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
 
+      - name: Check diesel schema
+        if: inputs.build-type == 'release' && inputs.arch == 'x64'
+        env:
+          DATABASE_URL: postgresql://localhost:1235/storage_controller
+          POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
+        run: |
+          /tmp/neon/bin/neon_local init
+          /tmp/neon/bin/neon_local storage_controller start
+
+          diesel print-schema > storage_controller/src/schema.rs
+
+          if [ -n "$(git diff storage_controller/src/schema.rs)" ]; then
+            echo >&2 "Uncommitted changes in diesel schema"
+
+            git diff .
+            exit 1
+          fi
+
+          /tmp/neon/bin/neon_local storage_controller stop
+
       # XXX: keep this after the binaries.list is formed, so the coverage can properly work later
       - name: Merge and upload coverage data
         if: inputs.build-type == 'debug'

.github/workflows/_check-codestyle-rust.yml

@@ -16,6 +16,9 @@ defaults:
   run:
     shell: bash -euxo pipefail {0}
 
+# No permission for GITHUB_TOKEN by default; the **minimal required** set of permissions should be granted in each job.
+permissions: {}
+
 jobs:
   check-codestyle-rust:
     strategy:
@@ -84,8 +87,3 @@ jobs:
         run: |
           cargo hakari generate --diff  # workspace-hack Cargo.toml is up-to-date
           cargo hakari manage-deps --dry-run  # all workspace crates depend on workspace-hack
-
-      # https://github.com/EmbarkStudios/cargo-deny
-      - name: Check rust licenses/bans/advisories/sources
-        if: ${{ !cancelled() }}
-        run: cargo deny check --hide-inclusion-graph

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

@@ -94,7 +94,9 @@ jobs:
           echo "LABELS_TO_ADD=${LABELS_TO_ADD}" >> ${GITHUB_OUTPUT}
           echo "LABELS_TO_REMOVE=${LABELS_TO_REMOVE}" >> ${GITHUB_OUTPUT}
 
-      - run: gh pr checkout "${PR_NUMBER}"
+      - uses: actions/checkout@v4
+        with:
+          ref: ${{ github.event.pull_request.head.sha }}
 
       - run: git checkout -b "${BRANCH}"
 

.github/workflows/benchmarking.yml

@@ -63,11 +63,15 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - DEFAULT_PG_VERSION: 16
+          - PG_VERSION: 16
             PLATFORM: "neon-staging"
             region_id: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
             RUNNER: [ self-hosted, us-east-2, x64 ]
-          - DEFAULT_PG_VERSION: 16
+          - PG_VERSION: 17
+            PLATFORM: "neon-staging"
+            region_id: ${{ github.event.inputs.region_id || 'aws-us-east-2' }}
+            RUNNER: [ self-hosted, us-east-2, x64 ]
+          - PG_VERSION: 16
             PLATFORM: "azure-staging"
             region_id: 'azure-eastus2'
             RUNNER: [ self-hosted, eastus2, x64 ]
@@ -75,7 +79,7 @@ jobs:
       TEST_PG_BENCH_DURATIONS_MATRIX: "300"
       TEST_PG_BENCH_SCALES_MATRIX: "10,100"
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: ${{ matrix.DEFAULT_PG_VERSION }}
+      PG_VERSION: ${{ matrix.PG_VERSION }}
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -112,7 +116,7 @@ jobs:
       uses: ./.github/actions/neon-project-create
       with:
         region_id: ${{ matrix.region_id }}
-        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
+        postgres_version: ${{ env.PG_VERSION }}
         api_key: ${{ secrets.NEON_STAGING_API_KEY }}
 
     - name: Run benchmark
@@ -122,7 +126,7 @@ jobs:
         test_selection: performance
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
         # Set --sparse-ordering option of pytest-order plugin
         # to ensure tests are running in order of appears in the file.
@@ -313,7 +317,11 @@ jobs:
                       { "pg_version": 16, "region_id": "azure-eastus2",          "platform": "neonvm-azure-captest-freetier", "db_size": "3gb" ,"runner": '"$runner_azure"',   "image": "neondatabase/build-tools:pinned-bookworm" },
                       { "pg_version": 16, "region_id": "azure-eastus2",          "platform": "neonvm-azure-captest-new",      "db_size": "10gb","runner": '"$runner_azure"',   "image": "neondatabase/build-tools:pinned-bookworm" },
                       { "pg_version": 16, "region_id": "azure-eastus2",          "platform": "neonvm-azure-captest-new",      "db_size": "50gb","runner": '"$runner_azure"',   "image": "neondatabase/build-tools:pinned-bookworm" },
-                      { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-sharding-reuse", "db_size": "50gb","runner": '"$runner_default"', "image": "'"$image_default"'" }]
+                      { "pg_version": 16, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-sharding-reuse", "db_size": "50gb","runner": '"$runner_default"', "image": "'"$image_default"'" },
+                      { "pg_version": 17, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-freetier",       "db_size": "3gb" ,"runner": '"$runner_default"', "image": "'"$image_default"'" },
+                      { "pg_version": 17, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-new",            "db_size": "10gb","runner": '"$runner_default"', "image": "'"$image_default"'" },
+                      { "pg_version": 17, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-new-many-tables","db_size": "10gb","runner": '"$runner_default"', "image": "'"$image_default"'" },
+                      { "pg_version": 17, "region_id": "'"$region_id_default"'", "platform": "neonvm-captest-new",            "db_size": "50gb","runner": '"$runner_default"', "image": "'"$image_default"'" }]
         }'
 
         if [ "$(date +%A)" = "Saturday" ] || [ ${RUN_AWS_RDS_AND_AURORA} = "true" ]; then
@@ -329,12 +337,15 @@ jobs:
         matrix='{
           "platform": [
             "neonvm-captest-reuse"
+          ],
+          "pg_version" : [
+            16,17
           ]
         }'
 
         if [ "$(date +%A)" = "Saturday" ] || [ ${RUN_AWS_RDS_AND_AURORA} = "true" ]; then
-          matrix=$(echo "$matrix" | jq '.include += [{ "platform": "rds-postgres" },
-                                                     { "platform": "rds-aurora"   }]')
+          matrix=$(echo "$matrix" | jq '.include += [{ "pg_version": 16, "platform": "rds-postgres" },
+                                                     { "pg_version": 16, "platform": "rds-aurora"   }]')
         fi
 
         echo "matrix=$(echo "$matrix" | jq --compact-output '.')" >> $GITHUB_OUTPUT
@@ -346,14 +357,14 @@ jobs:
           "platform": [
             "neonvm-captest-reuse"
           ],
-          "scale": [
-            "10"
+          "pg_version" : [
+            16,17
           ]
         }'
 
         if [ "$(date +%A)" = "Saturday" ] || [ ${RUN_AWS_RDS_AND_AURORA} = "true" ]; then
-          matrix=$(echo "$matrix" | jq '.include += [{ "platform": "rds-postgres", "scale": "10" },
-                                                     { "platform": "rds-aurora",   "scale": "10" }]')
+          matrix=$(echo "$matrix" | jq '.include += [{ "pg_version": 16, "platform": "rds-postgres" },
+                                                     { "pg_version": 16, "platform": "rds-aurora"   }]')
         fi
 
         echo "matrix=$(echo "$matrix" | jq --compact-output '.')" >> $GITHUB_OUTPUT
@@ -378,7 +389,7 @@ jobs:
       TEST_PG_BENCH_DURATIONS_MATRIX: "60m"
       TEST_PG_BENCH_SCALES_MATRIX: ${{ matrix.db_size }}
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: ${{ matrix.pg_version }}
+      PG_VERSION: ${{ matrix.pg_version }}
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -416,7 +427,7 @@ jobs:
       uses: ./.github/actions/neon-project-create
       with:
         region_id: ${{ matrix.region_id }}
-        postgres_version: ${{ env.DEFAULT_PG_VERSION }}
+        postgres_version: ${{ env.PG_VERSION }}
         api_key: ${{ secrets.NEON_STAGING_API_KEY }}
         compute_units: ${{ (contains(matrix.platform, 'captest-freetier') && '[0.25, 0.25]') || '[1, 1]' }}
 
@@ -447,7 +458,7 @@ jobs:
 
         echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
 
-    # we want to compare Neon project OLTP throughput and latency at scale factor 10 GB 
+    # we want to compare Neon project OLTP throughput and latency at scale factor 10 GB
     # without (neonvm-captest-new)
     # and with (neonvm-captest-new-many-tables) many relations in the database
     - name: Create many relations before the run
@@ -459,7 +470,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_perf_many_relations
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
@@ -475,7 +486,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_init
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
@@ -490,7 +501,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_simple_update
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
@@ -505,7 +516,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_select_only
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
@@ -549,14 +560,19 @@ jobs:
         include:
           - PLATFORM: "neonvm-captest-pgvector"
             RUNNER: [ self-hosted, us-east-2, x64 ]
+            postgres_version: 16
+          - PLATFORM: "neonvm-captest-pgvector-pg17"
+            RUNNER: [ self-hosted, us-east-2, x64 ]
+            postgres_version: 17
           - PLATFORM: "azure-captest-pgvector"
             RUNNER: [ self-hosted, eastus2, x64 ]
+            postgres_version: 16
 
     env:
       TEST_PG_BENCH_DURATIONS_MATRIX: "15m"
       TEST_PG_BENCH_SCALES_MATRIX: "1"
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      PG_VERSION: ${{ matrix.postgres_version }}
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
 
@@ -574,32 +590,20 @@ jobs:
     steps:
     - uses: actions/checkout@v4
 
-    # until https://github.com/neondatabase/neon/issues/8275 is fixed we temporarily install postgresql-16
-    # instead of using Neon artifacts containing pgbench
-    - name: Install postgresql-16 where pytest expects it
-      run: |
-        # Just to make it easier to test things locally on macOS (with arm64)
-        arch=$(uname -m | sed 's/x86_64/amd64/g' | sed 's/aarch64/arm64/g')
+    - name: Configure AWS credentials
+      uses: aws-actions/configure-aws-credentials@v4
+      with:
+        aws-region: eu-central-1
+        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
+        role-duration-seconds: 18000 # 5 hours
 
-        cd /home/nonroot
-        wget -q "https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-17/libpq5_17.2-1.pgdg120+1_${arch}.deb"
-        wget -q "https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-16/postgresql-client-16_16.6-1.pgdg120+1_${arch}.deb"
-        wget -q "https://apt.postgresql.org/pub/repos/apt/pool/main/p/postgresql-16/postgresql-16_16.6-1.pgdg120+1_${arch}.deb"
-        dpkg -x libpq5_17.2-1.pgdg120+1_${arch}.deb pg
-        dpkg -x postgresql-16_16.6-1.pgdg120+1_${arch}.deb pg
-        dpkg -x postgresql-client-16_16.6-1.pgdg120+1_${arch}.deb pg
-
-        mkdir -p /tmp/neon/pg_install/v16/bin
-        ln -s /home/nonroot/pg/usr/lib/postgresql/16/bin/pgbench /tmp/neon/pg_install/v16/bin/pgbench
-        ln -s /home/nonroot/pg/usr/lib/postgresql/16/bin/psql    /tmp/neon/pg_install/v16/bin/psql
-        ln -s /home/nonroot/pg/usr/lib/$(uname -m)-linux-gnu     /tmp/neon/pg_install/v16/lib
-
-        LD_LIBRARY_PATH="/home/nonroot/pg/usr/lib/$(uname -m)-linux-gnu:${LD_LIBRARY_PATH:-}"
-        export LD_LIBRARY_PATH
-        echo "LD_LIBRARY_PATH=${LD_LIBRARY_PATH}" >> ${GITHUB_ENV}
-
-        /tmp/neon/pg_install/v16/bin/pgbench --version
-        /tmp/neon/pg_install/v16/bin/psql --version
+    - name: Download Neon artifact
+      uses: ./.github/actions/download
+      with:
+        name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
+        path: /tmp/neon/
+        prefix: latest
+        aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
 
     - name: Set up Connection String
       id: set-up-connstr
@@ -608,6 +612,9 @@ jobs:
           neonvm-captest-pgvector)
             CONNSTR=${{ secrets.BENCHMARK_PGVECTOR_CONNSTR }}
             ;;
+          neonvm-captest-pgvector-pg17)
+            CONNSTR=${{ secrets.BENCHMARK_PGVECTOR_CONNSTR_PG17 }}
+            ;;
           azure-captest-pgvector)
             CONNSTR=${{ secrets.BENCHMARK_PGVECTOR_CONNSTR_AZURE }}
             ;;
@@ -619,13 +626,6 @@ jobs:
 
         echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
 
-    - name: Configure AWS credentials
-      uses: aws-actions/configure-aws-credentials@v4
-      with:
-        aws-region: eu-central-1
-        role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
-        role-duration-seconds: 18000 # 5 hours
-
     - name: Benchmark pgvector hnsw indexing
       uses: ./.github/actions/run-python-test-set
       with:
@@ -634,7 +634,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_pgvector_indexing
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
@@ -649,7 +649,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
@@ -696,7 +696,7 @@ jobs:
 
     env:
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      PG_VERSION: ${{ matrix.pg_version }}
       TEST_OUTPUT: /tmp/test_output
       TEST_OLAP_COLLECT_EXPLAIN: ${{ github.event.inputs.collect_olap_explain }}
       TEST_OLAP_COLLECT_PG_STAT_STATEMENTS: ${{ github.event.inputs.collect_pg_stat_statements }}
@@ -739,7 +739,18 @@ jobs:
       run: |
         case "${PLATFORM}" in
           neonvm-captest-reuse)
-            CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CLICKBENCH_10M_CONNSTR }}
+            case "${PG_VERSION}" in
+              16)
+                CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CLICKBENCH_10M_CONNSTR }}
+                ;;
+              17)
+                CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CLICKBENCH_CONNSTR_PG17 }}
+                ;;
+              *)
+                echo >&2 "Unsupported PG_VERSION=${PG_VERSION} for PLATFORM=${PLATFORM}"
+                exit 1
+                ;;
+            esac
             ;;
           rds-aurora)
             CONNSTR=${{ secrets.BENCHMARK_RDS_AURORA_CLICKBENCH_10M_CONNSTR }}
@@ -763,7 +774,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 43200 -k test_clickbench
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
@@ -799,7 +810,7 @@ jobs:
     # We might change it after https://github.com/neondatabase/neon/issues/2900.
     #
     # *_TPCH_S10_CONNSTR: DB generated with scale factor 10 (~10 GB)
-    if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
+    # if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
     permissions:
       contents: write
       statuses: write
@@ -812,12 +823,11 @@ jobs:
 
     env:
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      PG_VERSION: ${{ matrix.pg_version }}
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
       PLATFORM: ${{ matrix.platform }}
-      TEST_OLAP_SCALE: ${{ matrix.scale }}
 
     runs-on: [ self-hosted, us-east-2, x64 ]
     container:
@@ -849,13 +859,24 @@ jobs:
       run: |
         case "${PLATFORM}" in
           neonvm-captest-reuse)
-            ENV_PLATFORM=CAPTEST_TPCH
+            case "${PG_VERSION}" in
+              16)
+                CONNSTR_SECRET_NAME="BENCHMARK_CAPTEST_TPCH_S10_CONNSTR"
+                ;;
+              17)
+                CONNSTR_SECRET_NAME="BENCHMARK_CAPTEST_TPCH_CONNSTR_PG17"
+                ;;
+              *)
+                echo >&2 "Unsupported PG_VERSION=${PG_VERSION} for PLATFORM=${PLATFORM}"
+                exit 1
+                ;;
+            esac
             ;;
           rds-aurora)
-            ENV_PLATFORM=RDS_AURORA_TPCH
+            CONNSTR_SECRET_NAME="BENCHMARK_RDS_AURORA_TPCH_S10_CONNSTR"
             ;;
           rds-postgres)
-            ENV_PLATFORM=RDS_POSTGRES_TPCH
+            CONNSTR_SECRET_NAME="BENCHMARK_RDS_POSTGRES_TPCH_S10_CONNSTR"
             ;;
           *)
             echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neonvm-captest-reuse', 'rds-aurora', or 'rds-postgres'"
@@ -863,7 +884,6 @@ jobs:
             ;;
         esac
 
-        CONNSTR_SECRET_NAME="BENCHMARK_${ENV_PLATFORM}_S${TEST_OLAP_SCALE}_CONNSTR"
         echo "CONNSTR_SECRET_NAME=${CONNSTR_SECRET_NAME}" >> $GITHUB_ENV
 
     - name: Set up Connection String
@@ -881,13 +901,13 @@ 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 }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
         PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
         BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
-        TEST_OLAP_SCALE: ${{ matrix.scale }}
+        TEST_OLAP_SCALE: 10
 
     - name: Create Allure report
       id: create-allure-report
@@ -909,7 +929,7 @@ jobs:
         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) }}
+    # if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
     permissions:
       contents: write
       statuses: write
@@ -922,7 +942,7 @@ jobs:
 
     env:
       POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
+      PG_VERSION: ${{ matrix.pg_version }}
       TEST_OUTPUT: /tmp/test_output
       BUILD_TYPE: remote
       SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
@@ -959,7 +979,18 @@ jobs:
       run: |
         case "${PLATFORM}" in
           neonvm-captest-reuse)
-            CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_CAPTEST_CONNSTR }}
+            case "${PG_VERSION}" in
+              16)
+                CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_CAPTEST_CONNSTR }}
+                ;;
+              17)
+                CONNSTR=${{ secrets.BENCHMARK_CAPTEST_USER_EXAMPLE_CONNSTR_PG17 }}
+                ;;
+              *)
+                echo >&2 "Unsupported PG_VERSION=${PG_VERSION} for PLATFORM=${PLATFORM}"
+                exit 1
+                ;;
+            esac
             ;;
           rds-aurora)
             CONNSTR=${{ secrets.BENCHMARK_USER_EXAMPLE_RDS_AURORA_CONNSTR }}
@@ -983,7 +1014,7 @@ jobs:
         run_in_parallel: false
         save_perf_report: ${{ env.SAVE_PERF_REPORT }}
         extra_params: -m remote_cluster --timeout 21600 -k test_user_examples
-        pg_version: ${{ env.DEFAULT_PG_VERSION }}
+        pg_version: ${{ env.PG_VERSION }}
         aws-oicd-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
       env:
         VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"

.github/workflows/build_and_test.yml

@@ -45,6 +45,26 @@ jobs:
             run cancel-previous-in-concurrency-group.yml \
               --field concurrency_group="${{ env.E2E_CONCURRENCY_GROUP }}"
 
+  files-changed:
+    needs: [ check-permissions ]
+    runs-on: [ self-hosted, small ]
+    timeout-minutes: 3
+    outputs:
+      check-rust-dependencies: ${{ steps.files-changed.outputs.rust_dependencies }}
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          submodules: true
+
+      - name: Check for file changes
+        uses: dorny/paths-filter@de90cc6fb38fc0963ad72b210f1f284cd68cea36  # v3.0.2
+        id: files-changed
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          filters: .github/file-filters.yaml
+
   tag:
     needs: [ check-permissions ]
     runs-on: [ self-hosted, small ]
@@ -170,6 +190,14 @@ jobs:
       archs: '["x64", "arm64"]'
     secrets: inherit
 
+  check-dependencies-rust:
+    needs: [ files-changed, build-build-tools-image ]
+    if: ${{ needs.files-changed.outputs.check-rust-dependencies == 'true' }}
+    uses: ./.github/workflows/cargo-deny.yml
+    with:
+      build-tools-image: ${{ needs.build-build-tools-image.outputs.image }}-bookworm
+    secrets: inherit
+
   build-and-test-locally:
     needs: [ tag, build-build-tools-image ]
     strategy:
@@ -242,7 +270,7 @@ jobs:
       statuses: write
       contents: write
       pull-requests: write
-    runs-on: [ self-hosted, small ]
+    runs-on: [ self-hosted, small-metal ]
     container:
       image: ${{ needs.build-build-tools-image.outputs.image }}-bookworm
       credentials:
@@ -654,7 +682,7 @@ jobs:
           push: true
           pull: true
           file: compute/compute-node.Dockerfile
-          target: neon-pg-ext-test
+          target: extension-tests
           cache-from: type=registry,ref=cache.neon.build/compute-node-${{ matrix.version.pg }}:cache-${{ matrix.version.debian }}-${{ matrix.arch }}
           tags: |
             neondatabase/neon-test-extensions-${{ matrix.version.pg }}:${{needs.tag.outputs.build-tag}}-${{ matrix.version.debian }}-${{ matrix.arch }}
@@ -786,6 +814,17 @@ jobs:
           username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
           password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
 
+      - name: Get the last compute release tag
+        id: get-last-compute-release-tag
+        env:
+          GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
+        run: |
+          tag=$(gh api -q '[.[].tag_name | select(startswith("release-compute"))][0]'\
+            -H "Accept: application/vnd.github+json" \
+            -H "X-GitHub-Api-Version: 2022-11-28" \
+            "/repos/${{ github.repository }}/releases")
+          echo tag=${tag} >> ${GITHUB_OUTPUT}
+
       # `neondatabase/neon` contains multiple binaries, all of them use the same input for the version into the same version formatting library.
       # Pick pageserver as currently the only binary with extra "version" features printed in the string to verify.
       # Regular pageserver version string looks like
@@ -817,6 +856,20 @@ jobs:
           TEST_VERSION_ONLY: ${{ matrix.pg_version }}
         run: ./docker-compose/docker_compose_test.sh
 
+      - name: Print logs and clean up docker-compose test
+        if: always()
+        run: |
+          docker compose --profile test-extensions -f ./docker-compose/docker-compose.yml logs || true
+          docker compose --profile test-extensions -f ./docker-compose/docker-compose.yml down
+
+      - name: Test extension upgrade
+        timeout-minutes: 20
+        if: ${{ needs.tag.outputs.build-tag == github.run_id }}
+        env:
+          NEWTAG: ${{ needs.tag.outputs.build-tag }}
+          OLDTAG: ${{ steps.get-last-compute-release-tag.outputs.tag }}
+        run: ./docker-compose/test_extensions_upgrade.sh
+
       - name: Print logs and clean up
         if: always()
         run: |
@@ -1050,6 +1103,7 @@ jobs:
           retries: 5
           script: |
             const tag = "${{ needs.tag.outputs.build-tag }}";
+            const branch = "${{ github.ref_name }}";
 
             try {
               const existingRef = await github.rest.git.getRef({
@@ -1092,12 +1146,48 @@ jobs:
               }
 
               console.log(`Release for tag ${tag} does not exist. Creating it...`);
+
+              // Find the PR number using the commit SHA
+              const pullRequests = await github.rest.pulls.list({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                state: 'closed',
+                base: branch,
+              });
+
+              const pr = pullRequests.data.find(pr => pr.merge_commit_sha === context.sha);
+              const prNumber = pr ? pr.number : null;
+
+              // Find the previous release on the branch
+              const releases = await github.rest.repos.listReleases({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                per_page: 100,
+              });
+
+              const branchReleases = releases.data
+                .filter((release) => {
+                  const regex = new RegExp(`^${branch}-\\d+$`);
+                  return regex.test(release.tag_name) && !release.draft && !release.prerelease;
+                })
+                .sort((a, b) => new Date(b.created_at) - new Date(a.created_at));
+
+              const previousTag = branchReleases.length > 0 ? branchReleases[0].tag_name : null;
+
+              const releaseNotes = [
+                prNumber
+                  ? `Release PR https://github.com/${context.repo.owner}/${context.repo.repo}/pull/${prNumber}.`
+                  : 'Release PR not found.',
+                previousTag
+                  ? `Diff with the previous release https://github.com/${context.repo.owner}/${context.repo.repo}/compare/${previousTag}...${tag}.`
+                  : `No previous release found on branch ${branch}.`,
+              ].join('\n\n');
+
               await github.rest.repos.createRelease({
                 owner: context.repo.owner,
                 repo: context.repo.repo,
                 tag_name: tag,
-                // TODO: Automate release notes properly
-                generate_release_notes: false,
+                body: releaseNotes,
               });
               console.log(`Release for tag ${tag} created successfully.`);
             }
@@ -1270,6 +1360,8 @@ jobs:
       - build-and-test-locally
       - check-codestyle-python
       - check-codestyle-rust
+      - check-dependencies-rust
+      - files-changed
       - promote-images-dev
       - test-images
       - trigger-custom-extensions-build-and-wait
@@ -1282,4 +1374,11 @@ jobs:
         if: |
           contains(needs.*.result, 'failure')
           || contains(needs.*.result, 'cancelled')
-          || contains(needs.*.result, 'skipped')
+          || (needs.check-dependencies-rust.result == 'skipped' && needs.files-changed.outputs.check-rust-dependencies == 'true')
+          || needs.build-and-test-locally.result == 'skipped'
+          || needs.check-codestyle-python.result == 'skipped'
+          || needs.check-codestyle-rust.result == 'skipped'
+          || needs.files-changed.result == 'skipped'
+          || needs.promote-images-dev.result == 'skipped'
+          || needs.test-images.result == 'skipped'
+          || needs.trigger-custom-extensions-build-and-wait.result == 'skipped'

.github/workflows/cargo-deny.yml

@@ -0,0 +1,57 @@
+name: cargo deny checks
+
+on:
+  workflow_call:
+    inputs:
+      build-tools-image:
+        required: false
+        type: string
+  schedule:
+    - cron: '0 0 * * *'
+
+jobs:
+  cargo-deny:
+    strategy:
+      matrix:
+        ref: >-
+          ${{
+            fromJSON(
+              github.event_name == 'schedule'
+                && '["main","release","release-proxy","release-compute"]'
+                || format('["{0}"]', github.sha)
+            )
+          }}
+
+    runs-on: [self-hosted, small]
+
+    container:
+      image: ${{ inputs.build-tools-image || 'neondatabase/build-tools:pinned' }}
+      credentials:
+        username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
+        password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
+      options: --init
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          ref: ${{ matrix.ref }}
+
+      - name: Check rust licenses/bans/advisories/sources
+        env:
+          CARGO_DENY_TARGET: >-
+            ${{ github.event_name == 'schedule' && 'advisories' || 'all' }}
+        run: cargo deny check --hide-inclusion-graph $CARGO_DENY_TARGET
+
+      - name: Post to a Slack channel
+        if: ${{ github.event_name == 'schedule' && failure() }}
+        uses: slackapi/slack-github-action@v2
+        with:
+          method: chat.postMessage
+          token: ${{ secrets.SLACK_BOT_TOKEN }}
+          payload: |
+            channel: ${{ vars.SLACK_CICD_CHANNEL_ID }}
+            text: |
+              Periodic cargo-deny on ${{ matrix.ref }}: ${{ job.status }}
+              <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
+              Pinging @oncall-devprod.

.github/workflows/pg-clients.yml

@@ -12,8 +12,8 @@ on:
   pull_request:
     paths:
       - '.github/workflows/pg-clients.yml'
-      - 'test_runner/pg_clients/**'
-      - 'test_runner/logical_repl/**'
+      - 'test_runner/pg_clients/**/*.py'
+      - 'test_runner/logical_repl/**/*.py'
       - 'poetry.lock'
   workflow_dispatch:
 
@@ -104,6 +104,8 @@ jobs:
         with:
           api_key: ${{ secrets.NEON_STAGING_API_KEY }}
           postgres_version: ${{ env.DEFAULT_PG_VERSION }}
+          project_settings: >-
+            {"enable_logical_replication": true}
 
       - name: Run tests
         uses: ./.github/actions/run-python-test-set

.github/workflows/pre-merge-checks.yml

@@ -59,7 +59,10 @@ jobs:
           echo "${RUST_CHANGED_FILES}"
 
   build-build-tools-image:
-    if: needs.get-changed-files.outputs.python-changed == 'true'
+    if: |
+      false
+      || needs.get-changed-files.outputs.python-changed == 'true'
+      || needs.get-changed-files.outputs.rust-changed == 'true'
     needs: [ get-changed-files ]
     uses: ./.github/workflows/build-build-tools-image.yml
     with:
@@ -92,7 +95,8 @@ jobs:
   # - conclusion
   # - neon-cloud-e2e
   conclusion:
-    if: always()
+    # Do not run job on Pull Requests as it interferes with the `conclusion` job from the `build_and_test` workflow
+    if: always() && github.event_name == 'merge_group'
     permissions:
       statuses: write # for `github.repos.createCommitStatus(...)`
       contents: write
@@ -124,6 +128,8 @@ jobs:
       - name: Fail the job if any of the dependencies do not succeed or skipped
         run: exit 1
         if: |
-          (contains(needs.check-codestyle-python.result, 'skipped') && needs.get-changed-files.outputs.python-changed == 'true')
+          false
+          || (needs.check-codestyle-python.result == 'skipped' && needs.get-changed-files.outputs.python-changed == 'true')
+          || (needs.check-codestyle-rust.result   == 'skipped' && needs.get-changed-files.outputs.rust-changed   == 'true')
           || contains(needs.*.result, 'failure')
           || contains(needs.*.result, 'cancelled')

Cargo.lock

@@ -206,6 +206,16 @@ dependencies = [
  "syn 2.0.90",
 ]
 
+[[package]]
+name = "assert-json-diff"
+version = "2.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "47e4f2b81832e72834d7518d8487a0396a28cc408186a2e8854c0f98011faf12"
+dependencies = [
+ "serde",
+ "serde_json",
+]
+
 [[package]]
 name = "async-channel"
 version = "1.9.0"
@@ -290,9 +300,9 @@ checksum = "d468802bab17cbc0cc575e9b053f41e72aa36bfa6b7f55e3529ffa43161b97fa"
 
 [[package]]
 name = "aws-config"
-version = "1.5.10"
+version = "1.5.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9b49afaa341e8dd8577e1a2200468f98956d6eda50bcf4a53246cc00174ba924"
+checksum = "dc47e70fc35d054c8fcd296d47a61711f043ac80534a10b4f741904f81e73a90"
 dependencies = [
  "aws-credential-types",
  "aws-runtime",
@@ -301,7 +311,7 @@ dependencies = [
  "aws-sdk-sts",
  "aws-smithy-async",
  "aws-smithy-http",
- "aws-smithy-json 0.60.7",
+ "aws-smithy-json",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
  "aws-smithy-types",
@@ -332,9 +342,9 @@ dependencies = [
 
 [[package]]
 name = "aws-runtime"
-version = "1.4.4"
+version = "1.5.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b5ac934720fbb46206292d2c75b57e67acfc56fe7dfd34fb9a02334af08409ea"
+checksum = "bee7643696e7fdd74c10f9eb42848a87fe469d35eae9c3323f80aa98f350baac"
 dependencies = [
  "aws-credential-types",
  "aws-sigv4",
@@ -366,7 +376,7 @@ dependencies = [
  "aws-runtime",
  "aws-smithy-async",
  "aws-smithy-http",
- "aws-smithy-json 0.61.1",
+ "aws-smithy-json",
  "aws-smithy-query",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
@@ -389,7 +399,7 @@ dependencies = [
  "aws-runtime",
  "aws-smithy-async",
  "aws-smithy-http",
- "aws-smithy-json 0.61.1",
+ "aws-smithy-json",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
  "aws-smithy-types",
@@ -414,7 +424,7 @@ dependencies = [
  "aws-smithy-checksums",
  "aws-smithy-eventstream",
  "aws-smithy-http",
- "aws-smithy-json 0.61.1",
+ "aws-smithy-json",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
  "aws-smithy-types",
@@ -437,15 +447,15 @@ dependencies = [
 
 [[package]]
 name = "aws-sdk-sso"
-version = "1.50.0"
+version = "1.57.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "05ca43a4ef210894f93096039ef1d6fa4ad3edfabb3be92b80908b9f2e4b4eab"
+checksum = "c54bab121fe1881a74c338c5f723d1592bf3b53167f80268a1274f404e1acc38"
 dependencies = [
  "aws-credential-types",
  "aws-runtime",
  "aws-smithy-async",
  "aws-smithy-http",
- "aws-smithy-json 0.61.1",
+ "aws-smithy-json",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
  "aws-smithy-types",
@@ -459,15 +469,15 @@ dependencies = [
 
 [[package]]
 name = "aws-sdk-ssooidc"
-version = "1.51.0"
+version = "1.58.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "abaf490c2e48eed0bb8e2da2fb08405647bd7f253996e0f93b981958ea0f73b0"
+checksum = "8c8234fd024f7ac61c4e44ea008029bde934250f371efe7d4a39708397b1080c"
 dependencies = [
  "aws-credential-types",
  "aws-runtime",
  "aws-smithy-async",
  "aws-smithy-http",
- "aws-smithy-json 0.61.1",
+ "aws-smithy-json",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
  "aws-smithy-types",
@@ -481,15 +491,15 @@ dependencies = [
 
 [[package]]
 name = "aws-sdk-sts"
-version = "1.51.0"
+version = "1.58.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b68fde0d69c8bfdc1060ea7da21df3e39f6014da316783336deff0a9ec28f4bf"
+checksum = "ba60e1d519d6f23a9df712c04fdeadd7872ac911c84b2f62a8bda92e129b7962"
 dependencies = [
  "aws-credential-types",
  "aws-runtime",
  "aws-smithy-async",
  "aws-smithy-http",
- "aws-smithy-json 0.61.1",
+ "aws-smithy-json",
  "aws-smithy-query",
  "aws-smithy-runtime",
  "aws-smithy-runtime-api",
@@ -504,9 +514,9 @@ dependencies = [
 
 [[package]]
 name = "aws-sigv4"
-version = "1.2.6"
+version = "1.2.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7d3820e0c08d0737872ff3c7c1f21ebbb6693d832312d6152bf18ef50a5471c2"
+checksum = "690118821e46967b3c4501d67d7d52dd75106a9c54cf36cefa1985cedbe94e05"
 dependencies = [
  "aws-credential-types",
  "aws-smithy-eventstream",
@@ -533,9 +543,9 @@ dependencies = [
 
 [[package]]
 name = "aws-smithy-async"
-version = "1.2.1"
+version = "1.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "62220bc6e97f946ddd51b5f1361f78996e704677afc518a4ff66b7a72ea1378c"
+checksum = "fa59d1327d8b5053c54bf2eaae63bf629ba9e904434d0835a28ed3c0ed0a614e"
 dependencies = [
  "futures-util",
  "pin-project-lite",
@@ -565,9 +575,9 @@ dependencies = [
 
 [[package]]
 name = "aws-smithy-eventstream"
-version = "0.60.5"
+version = "0.60.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cef7d0a272725f87e51ba2bf89f8c21e4df61b9e49ae1ac367a6d69916ef7c90"
+checksum = "8b18559a41e0c909b77625adf2b8c50de480a8041e5e4a3f5f7d177db70abc5a"
 dependencies = [
  "aws-smithy-types",
  "bytes",
@@ -576,9 +586,9 @@ dependencies = [
 
 [[package]]
 name = "aws-smithy-http"
-version = "0.60.11"
+version = "0.60.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5c8bc3e8fdc6b8d07d976e301c02fe553f72a39b7a9fea820e023268467d7ab6"
+checksum = "7809c27ad8da6a6a68c454e651d4962479e81472aa19ae99e59f9aba1f9713cc"
 dependencies = [
  "aws-smithy-eventstream",
  "aws-smithy-runtime-api",
@@ -597,18 +607,9 @@ dependencies = [
 
 [[package]]
 name = "aws-smithy-json"
-version = "0.60.7"
+version = "0.61.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4683df9469ef09468dad3473d129960119a0d3593617542b7d52086c8486f2d6"
-dependencies = [
- "aws-smithy-types",
-]
-
-[[package]]
-name = "aws-smithy-json"
-version = "0.61.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ee4e69cc50921eb913c6b662f8d909131bb3e6ad6cb6090d3a39b66fc5c52095"
+checksum = "623a51127f24c30776c8b374295f2df78d92517386f77ba30773f15a30ce1422"
 dependencies = [
  "aws-smithy-types",
 ]
@@ -625,9 +626,9 @@ dependencies = [
 
 [[package]]
 name = "aws-smithy-runtime"
-version = "1.7.4"
+version = "1.7.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9f20685047ca9d6f17b994a07f629c813f08b5bce65523e47124879e60103d45"
+checksum = "865f7050bbc7107a6c98a397a9fcd9413690c27fa718446967cf03b2d3ac517e"
 dependencies = [
  "aws-smithy-async",
  "aws-smithy-http",
@@ -669,9 +670,9 @@ dependencies = [
 
 [[package]]
 name = "aws-smithy-types"
-version = "1.2.9"
+version = "1.2.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4fbd94a32b3a7d55d3806fe27d98d3ad393050439dd05eb53ece36ec5e3d3510"
+checksum = "a28f6feb647fb5e0d5b50f0472c19a7db9462b74e2fec01bb0b44eedcc834e97"
 dependencies = [
  "base64-simd",
  "bytes",
@@ -704,9 +705,9 @@ dependencies = [
 
 [[package]]
 name = "aws-types"
-version = "1.3.3"
+version = "1.3.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5221b91b3e441e6675310829fd8984801b772cb1546ef6c0e54dec9f1ac13fef"
+checksum = "b0df5a18c4f951c645300d365fec53a61418bcf4650f604f85fe2a665bfaa0c2"
 dependencies = [
  "aws-credential-types",
  "aws-smithy-async",
@@ -978,7 +979,7 @@ version = "0.70.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f49d8fed880d473ea71efb9bf597651e77201bdd4893efe54c9e5d65ae04ce6f"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "cexpr",
  "clang-sys",
  "itertools 0.12.1",
@@ -1006,9 +1007,9 @@ checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
 
 [[package]]
 name = "bitflags"
-version = "2.4.1"
+version = "2.8.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "327762f6e5a765692301e5bb513e0d9fef63be86bbc14528052b1cd3e6f03e07"
+checksum = "8f68f53c83ab957f72c32642f3868eec03eb974d1fb82e453128456482613d36"
 
 [[package]]
 name = "block-buffer"
@@ -1019,6 +1020,12 @@ dependencies = [
  "generic-array",
 ]
 
+[[package]]
+name = "boxcar"
+version = "0.2.8"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2721c3c5a6f0e7f7e607125d963fedeb765f545f67adc9d71ed934693881eb42"
+
 [[package]]
 name = "bstr"
 version = "1.5.0"
@@ -1225,6 +1232,20 @@ version = "0.7.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "afb84c814227b90d6895e01398aee0d8033c00e7466aca416fb6a8e0eb19d8a7"
 
+[[package]]
+name = "clashmap"
+version = "1.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "93bd59c81e2bd87a775ae2de75f070f7e2bfe97363a6ad652f46824564c23e4d"
+dependencies = [
+ "crossbeam-utils",
+ "hashbrown 0.15.2",
+ "lock_api",
+ "parking_lot_core 0.9.8",
+ "polonius-the-crab",
+ "replace_with",
+]
+
 [[package]]
 name = "colorchoice"
 version = "1.0.0"
@@ -1312,7 +1333,7 @@ dependencies = [
  "tar",
  "thiserror 1.0.69",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-stream",
  "tokio-util",
  "tower 0.5.2",
@@ -1421,7 +1442,7 @@ dependencies = [
  "storage_broker",
  "thiserror 1.0.69",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-util",
  "toml",
  "toml_edit",
@@ -1561,7 +1582,7 @@ version = "0.27.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f476fe445d41c9e991fd07515a6f463074b782242ccf4a5b7b1d1012e70824df"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "crossterm_winapi",
  "libc",
  "parking_lot 0.12.1",
@@ -1792,7 +1813,7 @@ version = "2.2.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ccf1bedf64cdb9643204a36dd15b19a6ce8e7aa7f7b105868e9f1fad5ffa7d12"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "byteorder",
  "chrono",
  "diesel_derives",
@@ -1812,7 +1833,7 @@ dependencies = [
  "futures-util",
  "scoped-futures",
  "tokio",
- "tokio-postgres 0.7.12",
+ "tokio-postgres",
 ]
 
 [[package]]
@@ -2428,6 +2449,16 @@ dependencies = [
  "wasm-bindgen",
 ]
 
+[[package]]
+name = "gettid"
+version = "0.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "397256552fed4a9e577850498071831ec8f18ea83368aecc114cab469dcb43e5"
+dependencies = [
+ "libc",
+ "winapi",
+]
+
 [[package]]
 name = "gimli"
 version = "0.31.1"
@@ -2556,6 +2587,12 @@ dependencies = [
  "allocator-api2",
 ]
 
+[[package]]
+name = "hashbrown"
+version = "0.15.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bf151400ff0baff5465007dd2f3e717f3fe502074ca563069ce3a6629d07b289"
+
 [[package]]
 name = "hashlink"
 version = "0.9.1"
@@ -2606,6 +2643,15 @@ version = "0.4.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6fe2267d4ed49bc07b63801559be28c718ea06c4738b7a03c94df7386d2cde46"
 
+[[package]]
+name = "higher-kinded-types"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "561985554c8b8d4808605c90a5f1979cc6c31a5d20b78465cd59501233c6678e"
+dependencies = [
+ "never-say-never",
+]
+
 [[package]]
 name = "hmac"
 version = "0.12.1"
@@ -3084,11 +3130,11 @@ dependencies = [
 
 [[package]]
 name = "inotify"
-version = "0.9.6"
+version = "0.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f8069d3ec154eb856955c1c0fbffefbf5f3c40a104ec912d4797314c1801abff"
+checksum = "f37dccff2791ab604f9babef0ba14fbe0be30bd368dc541e2b08d07c8aa908f3"
 dependencies = [
- "bitflags 1.3.2",
+ "bitflags 2.8.0",
  "inotify-sys",
  "libc",
 ]
@@ -3265,9 +3311,9 @@ dependencies = [
 
 [[package]]
 name = "kqueue"
-version = "1.0.7"
+version = "1.0.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2c8fc60ba15bf51257aa9807a48a61013db043fcf3a78cb0d916e8e396dcad98"
+checksum = "7447f1ca1b7b563588a205fe93dea8df60fd981423a768bc1c0ded35ed147d0c"
 dependencies = [
  "kqueue-sys",
  "libc",
@@ -3275,9 +3321,9 @@ dependencies = [
 
 [[package]]
 name = "kqueue-sys"
-version = "1.0.3"
+version = "1.0.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8367585489f01bc55dd27404dcf56b95e6da061a256a666ab23be9ba96a2e587"
+checksum = "ed9625ffda8729b85e45cf04090035ac368927b8cebc34898e7c120f52e4838b"
 dependencies = [
  "bitflags 1.3.2",
  "libc",
@@ -3304,9 +3350,9 @@ dependencies = [
 
 [[package]]
 name = "libc"
-version = "0.2.167"
+version = "0.2.169"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "09d6582e104315a817dff97f75133544b2e094ee22447d2acf4a74e189ba06fc"
+checksum = "b5aba8db14291edd000dfcc4d620c7ebfb122c613afb886ca8803fa4e128a20a"
 
 [[package]]
 name = "libloading"
@@ -3553,14 +3599,14 @@ dependencies = [
 
 [[package]]
 name = "mio"
-version = "0.8.11"
+version = "1.0.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a4a650543ca06a924e8b371db273b2756685faae30f8487da1b56505a8f78b0c"
+checksum = "2886843bf800fba2e3377cff24abf6379b4c4d5c6681eaf9ea5b0d15090450bd"
 dependencies = [
  "libc",
  "log",
  "wasi 0.11.0+wasi-snapshot-preview1",
- "windows-sys 0.48.0",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -3569,6 +3615,12 @@ version = "0.8.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e5ce46fe64a9d73be07dcbe690a38ce1b293be448fd8ce1e6c1b8062c9f72c6a"
 
+[[package]]
+name = "never-say-never"
+version = "6.6.666"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cf5a574dadd7941adeaa71823ecba5e28331b8313fb2e1c6a5c7e5981ea53ad6"
+
 [[package]]
 name = "nix"
 version = "0.25.1"
@@ -3600,7 +3652,7 @@ version = "0.27.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2eb04e9c688eff1c89d72b407f168cf79bb9e867a9d3323ed6c01519eb9cc053"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "cfg-if",
  "libc",
  "memoffset 0.9.0",
@@ -3618,12 +3670,11 @@ dependencies = [
 
 [[package]]
 name = "notify"
-version = "6.1.1"
+version = "8.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6205bd8bb1e454ad2e27422015fb5e4f2bcc7e08fa8f27058670d208324a4d2d"
+checksum = "2fee8403b3d66ac7b26aee6e40a897d85dc5ce26f44da36b8b73e987cc52e943"
 dependencies = [
- "bitflags 2.4.1",
- "crossbeam-channel",
+ "bitflags 2.8.0",
  "filetime",
  "fsevent-sys",
  "inotify",
@@ -3631,10 +3682,17 @@ dependencies = [
  "libc",
  "log",
  "mio",
+ "notify-types",
  "walkdir",
- "windows-sys 0.48.0",
+ "windows-sys 0.59.0",
 ]
 
+[[package]]
+name = "notify-types"
+version = "2.0.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5e0826a989adedc2a244799e823aece04662b66609d96af8dff7ac6df9a8925d"
+
 [[package]]
 name = "ntapi"
 version = "0.4.1"
@@ -4060,8 +4118,8 @@ dependencies = [
  "pageserver_compaction",
  "pin-project-lite",
  "postgres",
- "postgres-protocol 0.6.4",
- "postgres-types 0.2.4",
+ "postgres-protocol",
+ "postgres-types",
  "postgres_backend",
  "postgres_connection",
  "postgres_ffi",
@@ -4092,7 +4150,7 @@ dependencies = [
  "tokio",
  "tokio-epoll-uring",
  "tokio-io-timeout",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-stream",
  "tokio-tar",
  "tokio-util",
@@ -4150,7 +4208,7 @@ dependencies = [
  "serde",
  "thiserror 1.0.69",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-stream",
  "tokio-util",
  "utils",
@@ -4180,6 +4238,16 @@ dependencies = [
  "workspace_hack",
 ]
 
+[[package]]
+name = "papaya"
+version = "0.1.8"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dc7c76487f7eaa00a0fc1d7f88dc6b295aec478d11b0fc79f857b62c2874124c"
+dependencies = [
+ "equivalent",
+ "seize",
+]
+
 [[package]]
 name = "parking"
 version = "2.1.1"
@@ -4446,48 +4514,40 @@ dependencies = [
  "plotters-backend",
 ]
 
+[[package]]
+name = "polonius-the-crab"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e97ca2c89572ae41bbec1c99498251f87dd5a94e500c5ec19c382dd593dd5ce9"
+dependencies = [
+ "higher-kinded-types",
+ "never-say-never",
+]
+
 [[package]]
 name = "postgres"
-version = "0.19.4"
-source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#511f998c00148ab7c847bd7e6cfd3a906d0e7473"
+version = "0.19.7"
+source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#1f21e7959a96a34dcfbfce1b14b73286cdadffe9"
 dependencies = [
  "bytes",
  "fallible-iterator",
  "futures-util",
  "log",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
 ]
 
 [[package]]
 name = "postgres-protocol"
-version = "0.6.4"
-source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#511f998c00148ab7c847bd7e6cfd3a906d0e7473"
-dependencies = [
- "base64 0.20.0",
- "byteorder",
- "bytes",
- "fallible-iterator",
- "hmac",
- "lazy_static",
- "md-5",
- "memchr",
- "rand 0.8.5",
- "sha2",
- "stringprep",
-]
-
-[[package]]
-name = "postgres-protocol"
-version = "0.6.7"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "acda0ebdebc28befa84bee35e651e4c5f09073d668c7aed4cf7e23c3cda84b23"
+version = "0.6.6"
+source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#1f21e7959a96a34dcfbfce1b14b73286cdadffe9"
 dependencies = [
  "base64 0.22.1",
  "byteorder",
  "bytes",
  "fallible-iterator",
  "hmac",
+ "lazy_static",
  "md-5",
  "memchr",
  "rand 0.8.5",
@@ -4513,23 +4573,13 @@ dependencies = [
 
 [[package]]
 name = "postgres-types"
-version = "0.2.4"
-source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#511f998c00148ab7c847bd7e6cfd3a906d0e7473"
+version = "0.2.6"
+source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#1f21e7959a96a34dcfbfce1b14b73286cdadffe9"
 dependencies = [
  "bytes",
+ "chrono",
  "fallible-iterator",
- "postgres-protocol 0.6.4",
-]
-
-[[package]]
-name = "postgres-types"
-version = "0.2.8"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f66ea23a2d0e5734297357705193335e0a957696f34bed2f2faefacb2fec336f"
-dependencies = [
- "bytes",
- "fallible-iterator",
- "postgres-protocol 0.6.7",
+ "postgres-protocol",
 ]
 
 [[package]]
@@ -4554,7 +4604,7 @@ dependencies = [
  "serde",
  "thiserror 1.0.69",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-postgres-rustls",
  "tokio-rustls 0.26.0",
  "tokio-util",
@@ -4569,7 +4619,7 @@ dependencies = [
  "itertools 0.10.5",
  "once_cell",
  "postgres",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "url",
 ]
 
@@ -4663,7 +4713,7 @@ dependencies = [
  "byteorder",
  "bytes",
  "itertools 0.10.5",
- "postgres-protocol 0.6.4",
+ "postgres-protocol",
  "rand 0.8.5",
  "serde",
  "thiserror 1.0.69",
@@ -4704,7 +4754,7 @@ version = "0.16.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "731e0d9356b0c25f16f33b5be79b1c57b562f141ebfcdb0ad8ac2c13a24293b4"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "chrono",
  "flate2",
  "hex",
@@ -4719,7 +4769,7 @@ version = "0.16.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "2d3554923a69f4ce04c4a754260c338f505ce22642d3830e049a399fc2059a29"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "chrono",
  "hex",
 ]
@@ -4825,6 +4875,7 @@ dependencies = [
  "ahash",
  "anyhow",
  "arc-swap",
+ "assert-json-diff",
  "async-compression",
  "async-trait",
  "atomic-take",
@@ -4832,15 +4883,16 @@ dependencies = [
  "aws-sdk-iam",
  "aws-sigv4",
  "base64 0.13.1",
+ "boxcar",
  "bstr",
  "bytes",
  "camino",
  "camino-tempfile",
  "chrono",
  "clap",
+ "clashmap",
  "compute_api",
  "consumption_metrics",
- "dashmap 5.5.0",
  "ecdsa 0.16.9",
  "ed25519-dalek",
  "env_logger 0.10.2",
@@ -4848,6 +4900,7 @@ dependencies = [
  "flate2",
  "framed-websockets",
  "futures",
+ "gettid",
  "hashbrown 0.14.5",
  "hashlink",
  "hex",
@@ -4870,7 +4923,9 @@ dependencies = [
  "measured",
  "metrics",
  "once_cell",
+ "opentelemetry",
  "p256 0.13.2",
+ "papaya",
  "parking_lot 0.12.1",
  "parquet",
  "parquet_derive",
@@ -4911,12 +4966,15 @@ dependencies = [
  "tikv-jemalloc-ctl",
  "tikv-jemallocator",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-postgres2",
  "tokio-rustls 0.26.0",
  "tokio-tungstenite 0.21.0",
  "tokio-util",
  "tracing",
+ "tracing-log",
+ "tracing-opentelemetry",
+ "tracing-serde",
  "tracing-subscriber",
  "tracing-utils",
  "try-lock",
@@ -5248,6 +5306,12 @@ dependencies = [
  "utils",
 ]
 
+[[package]]
+name = "replace_with"
+version = "0.1.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e3a8614ee435691de62bcffcf4a66d91b3594bf1428a5722e79103249a095690"
+
 [[package]]
 name = "reqwest"
 version = "0.12.4"
@@ -5527,7 +5591,7 @@ version = "0.38.41"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d7f649912bc1495e167a6edee79151c84b1bad49748cb4f1f1167f459f6224f6"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "errno",
  "libc",
  "linux-raw-sys 0.4.14",
@@ -5699,7 +5763,7 @@ dependencies = [
  "pageserver_api",
  "parking_lot 0.12.1",
  "postgres",
- "postgres-protocol 0.6.4",
+ "postgres-protocol",
  "postgres_backend",
  "postgres_ffi",
  "pprof",
@@ -5723,7 +5787,7 @@ dependencies = [
  "tikv-jemallocator",
  "tokio",
  "tokio-io-timeout",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-stream",
  "tokio-tar",
  "tokio-util",
@@ -5864,6 +5928,16 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "seize"
+version = "0.4.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d84b0c858bdd30cb56f5597f8b3bf702ec23829e652cc636a1e5a7b9de46ae93"
+dependencies = [
+ "libc",
+ "windows-sys 0.52.0",
+]
+
 [[package]]
 name = "semver"
 version = "1.0.17"
@@ -6340,6 +6414,8 @@ dependencies = [
  "rand 0.8.5",
  "reqwest",
  "routerify",
+ "rustls 0.23.18",
+ "rustls-native-certs 0.8.0",
  "scoped-futures",
  "scopeguard",
  "serde",
@@ -6348,6 +6424,8 @@ dependencies = [
  "strum_macros",
  "thiserror 1.0.69",
  "tokio",
+ "tokio-postgres",
+ "tokio-postgres-rustls",
  "tokio-util",
  "tracing",
  "utils",
@@ -6393,7 +6471,7 @@ dependencies = [
  "serde_json",
  "storage_controller_client",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-postgres-rustls",
  "tokio-stream",
  "tokio-util",
@@ -6590,7 +6668,7 @@ dependencies = [
  "fastrand 2.2.0",
  "once_cell",
  "rustix",
- "windows-sys 0.52.0",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -6802,21 +6880,20 @@ checksum = "1f3ccbac311fea05f86f61904b462b55fb3df8837a366dfc601a0161d0532f20"
 
 [[package]]
 name = "tokio"
-version = "1.38.1"
+version = "1.43.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "eb2caba9f80616f438e09748d5acda951967e1ea58508ef53d9c6402485a46df"
+checksum = "3d61fa4ffa3de412bfea335c6ecff681de2b609ba3c77ef3e00e521813a9ed9e"
 dependencies = [
  "backtrace",
  "bytes",
  "libc",
  "mio",
- "num_cpus",
  "parking_lot 0.12.1",
  "pin-project-lite",
  "signal-hook-registry",
  "socket2",
  "tokio-macros",
- "windows-sys 0.48.0",
+ "windows-sys 0.52.0",
 ]
 
 [[package]]
@@ -6847,9 +6924,9 @@ dependencies = [
 
 [[package]]
 name = "tokio-macros"
-version = "2.3.0"
+version = "2.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5f5ae998a069d4b5aba8ee9dad856af7d520c3699e6159b185c2acd48155d39a"
+checksum = "6e06d43f1345a3bcd39f6a56dbb7dcab2ba47e68e8ac134855e7e2bdbaf8cab8"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -6858,8 +6935,8 @@ dependencies = [
 
 [[package]]
 name = "tokio-postgres"
-version = "0.7.7"
-source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#511f998c00148ab7c847bd7e6cfd3a906d0e7473"
+version = "0.7.10"
+source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#1f21e7959a96a34dcfbfce1b14b73286cdadffe9"
 dependencies = [
  "async-trait",
  "byteorder",
@@ -6872,32 +6949,8 @@ dependencies = [
  "percent-encoding",
  "phf",
  "pin-project-lite",
- "postgres-protocol 0.6.4",
- "postgres-types 0.2.4",
- "socket2",
- "tokio",
- "tokio-util",
-]
-
-[[package]]
-name = "tokio-postgres"
-version = "0.7.12"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3b5d3742945bc7d7f210693b0c58ae542c6fd47b17adbbda0885f3dcb34a6bdb"
-dependencies = [
- "async-trait",
- "byteorder",
- "bytes",
- "fallible-iterator",
- "futures-channel",
- "futures-util",
- "log",
- "parking_lot 0.12.1",
- "percent-encoding",
- "phf",
- "pin-project-lite",
- "postgres-protocol 0.6.7",
- "postgres-types 0.2.8",
+ "postgres-protocol",
+ "postgres-types",
  "rand 0.8.5",
  "socket2",
  "tokio",
@@ -6914,7 +6967,7 @@ dependencies = [
  "ring",
  "rustls 0.23.18",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-rustls 0.26.0",
  "x509-certificate",
 ]
@@ -6935,6 +6988,7 @@ dependencies = [
  "pin-project-lite",
  "postgres-protocol2",
  "postgres-types2",
+ "serde",
  "tokio",
  "tokio-util",
 ]
@@ -7157,7 +7211,7 @@ version = "0.6.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "403fa3b783d4b626a8ad51d766ab03cb6d2dbfc46b1c5d4448395e6628dc9697"
 dependencies = [
- "bitflags 2.4.1",
+ "bitflags 2.8.0",
  "bytes",
  "http 1.1.0",
  "http-body 1.0.0",
@@ -7591,7 +7645,7 @@ dependencies = [
  "serde_json",
  "sysinfo",
  "tokio",
- "tokio-postgres 0.7.7",
+ "tokio-postgres",
  "tokio-util",
  "tracing",
  "tracing-subscriber",
@@ -7654,9 +7708,9 @@ dependencies = [
 
 [[package]]
 name = "walkdir"
-version = "2.3.3"
+version = "2.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "36df944cda56c7d8d8b7496af378e6b16de9284591917d307c9b4d313c44e698"
+checksum = "29790946404f91d9c5d06f9874efddea1dc06c5efe94541a7d6863108e3a5e4b"
 dependencies = [
  "same-file",
  "winapi-util",
@@ -7908,6 +7962,15 @@ dependencies = [
  "windows-targets 0.52.6",
 ]
 
+[[package]]
+name = "windows-sys"
+version = "0.59.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1e38bc4d79ed67fd075bcc251a1c39b32a1776bbe92e5bef1f0bf1f8c531853b"
+dependencies = [
+ "windows-targets 0.52.6",
+]
+
 [[package]]
 name = "windows-targets"
 version = "0.48.0"
@@ -8136,6 +8199,7 @@ dependencies = [
  "tower 0.4.13",
  "tracing",
  "tracing-core",
+ "tracing-log",
  "url",
  "zerocopy",
  "zeroize",

Cargo.toml

@@ -54,6 +54,7 @@ async-compression = { version = "0.4.0", features = ["tokio", "gzip", "zstd"] }
 atomic-take = "1.1.0"
 backtrace = "0.3.74"
 flate2 = "1.0.26"
+assert-json-diff = "2"
 async-stream = "0.3"
 async-trait = "0.1"
 aws-config = { version = "1.5", default-features = false, features=["rustls", "sso"] }
@@ -77,10 +78,10 @@ camino = "1.1.6"
 cfg-if = "1.0.0"
 chrono = { version = "0.4", default-features = false, features = ["clock"] }
 clap = { version = "4.0", features = ["derive", "env"] }
+clashmap = { version = "1.0", features = ["raw-api"] }
 comfy-table = "7.1"
 const_format = "0.2"
 crc32c = "0.6"
-dashmap = { version = "5.5.0", features = ["raw-api"] }
 diatomic-waker = { version = "0.2.3" }
 either = "1.8"
 enum-map = "2.4.2"
@@ -123,7 +124,7 @@ measured = { version = "0.0.22", features=["lasso"] }
 measured-process = { version = "0.0.22" }
 memoffset = "0.9"
 nix = { version = "0.27", features = ["dir", "fs", "process", "socket", "signal", "poll"] }
-notify = "6.0.0"
+notify = "8.0.0"
 num_cpus = "1.15"
 num-traits = "0.2.15"
 once_cell = "1.13"
@@ -177,7 +178,7 @@ test-context = "0.3"
 thiserror = "1.0"
 tikv-jemallocator = { version = "0.6", features = ["profiling", "stats", "unprefixed_malloc_on_supported_platforms"] }
 tikv-jemalloc-ctl = { version = "0.6", features = ["stats"] }
-tokio = { version = "1.17", features = ["macros"] }
+tokio = { version = "1.41", features = ["macros"] }
 tokio-epoll-uring = { git = "https://github.com/neondatabase/tokio-epoll-uring.git" , branch = "main" }
 tokio-io-timeout = "1.2.0"
 tokio-postgres-rustls = "0.12.0"
@@ -193,7 +194,9 @@ tower-http = { version = "0.6.2", features = ["request-id", "trace"] }
 tower-service = "0.3.3"
 tracing = "0.1"
 tracing-error = "0.2"
+tracing-log = "0.2"
 tracing-opentelemetry = "0.28"
+tracing-serde = "0.2.0"
 tracing-subscriber = { version = "0.3", default-features = false, features = ["smallvec", "fmt", "tracing-log", "std", "env-filter", "json"] }
 try-lock = "0.2.5"
 twox-hash = { version = "1.6.3", default-features = false }

Dockerfile

@@ -64,6 +64,7 @@ ARG DEFAULT_PG_VERSION
 WORKDIR /data
 
 RUN set -e \
+    && echo 'Acquire::Retries "5";' > /etc/apt/apt.conf.d/80-retries \
     && apt update \
     && apt install -y \
         libreadline-dev \
@@ -72,6 +73,7 @@ RUN set -e \
 	# System postgres for use with client libraries (e.g. in storage controller)
         postgresql-15 \
         openssl \
+    && rm -f /etc/apt/apt.conf.d/80-retries \
     && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* \
     && useradd -d /data neon \
     && chown -R neon:neon /data

build-tools.Dockerfile

@@ -3,6 +3,17 @@ ARG DEBIAN_VERSION=bookworm
 FROM debian:bookworm-slim AS pgcopydb_builder
 ARG DEBIAN_VERSION
 
+# Use strict mode for bash to catch errors early
+SHELL ["/bin/bash", "-euo", "pipefail", "-c"]
+
+# By default, /bin/sh used in debian images will treat '\n' as eol,
+# but as we use bash as SHELL, and built-in echo in bash requires '-e' flag for that.
+RUN echo 'Acquire::Retries "5";' > /etc/apt/apt.conf.d/80-retries && \
+    echo -e "retry_connrefused = on\ntimeout=15\ntries=5\n" > /root/.wgetrc && \
+    echo -e "--retry-connrefused\n--connect-timeout 15\n--retry 5\n--max-time 300\n" > /root/.curlrc
+
+COPY build_tools/patches/pgcopydbv017.patch /pgcopydbv017.patch
+
 RUN if [ "${DEBIAN_VERSION}" = "bookworm" ]; then \
         set -e && \
         apt update && \
@@ -35,6 +46,7 @@ RUN if [ "${DEBIAN_VERSION}" = "bookworm" ]; then \
         mkdir /tmp/pgcopydb && \
         tar -xzf /tmp/pgcopydb.tar.gz -C /tmp/pgcopydb --strip-components=1 && \
         cd /tmp/pgcopydb && \
+        patch -p1 < /pgcopydbv017.patch && \
         make -s clean && \
         make -s -j12 install && \
         libpq_path=$(find /lib /usr/lib -name "libpq.so.5" | head -n 1) && \
@@ -51,7 +63,8 @@ ARG DEBIAN_VERSION
 
 # Add nonroot user
 RUN useradd -ms /bin/bash nonroot -b /home
-SHELL ["/bin/bash", "-c"]
+# Use strict mode for bash to catch errors early
+SHELL ["/bin/bash", "-euo", "pipefail", "-c"]
 
 RUN mkdir -p /pgcopydb/bin && \
     mkdir -p /pgcopydb/lib && \
@@ -61,6 +74,10 @@ RUN mkdir -p /pgcopydb/bin && \
 COPY --from=pgcopydb_builder /usr/lib/postgresql/16/bin/pgcopydb /pgcopydb/bin/pgcopydb
 COPY --from=pgcopydb_builder /pgcopydb/lib/libpq.so.5 /pgcopydb/lib/libpq.so.5
 
+RUN echo 'Acquire::Retries "5";' > /etc/apt/apt.conf.d/80-retries && \
+    echo -e "retry_connrefused = on\ntimeout=15\ntries=5\n" > /root/.wgetrc && \
+    echo -e "--retry-connrefused\n--connect-timeout 15\n--retry 5\n--max-time 300\n" > /root/.curlrc
+
 # System deps
 #
 # 'gdb' is included so that we get backtraces of core dumps produced in
@@ -182,8 +199,14 @@ RUN set -e \
 # It includes several bug fixes on top on v2.0 release (https://github.com/linux-test-project/lcov/compare/v2.0...master)
 # And patches from us:
 # - Generates json file with code coverage summary (https://github.com/neondatabase/lcov/commit/426e7e7a22f669da54278e9b55e6d8caabd00af0.tar.gz)
-RUN for package in Capture::Tiny DateTime Devel::Cover Digest::MD5 File::Spec JSON::XS Memory::Process Time::HiRes JSON; do yes | perl -MCPAN -e "CPAN::Shell->notest('install', '$package')"; done \
-    && wget https://github.com/neondatabase/lcov/archive/426e7e7a22f669da54278e9b55e6d8caabd00af0.tar.gz -O lcov.tar.gz \
+RUN set +o pipefail && \
+	 for package in Capture::Tiny DateTime Devel::Cover Digest::MD5 File::Spec JSON::XS Memory::Process Time::HiRes JSON; do \
+		yes | perl -MCPAN -e "CPAN::Shell->notest('install', '$package')";\
+	 done && \
+	set -o pipefail
+# Split into separate step to debug flaky failures here
+RUN wget https://github.com/neondatabase/lcov/archive/426e7e7a22f669da54278e9b55e6d8caabd00af0.tar.gz -O lcov.tar.gz \
+    && ls -laht lcov.tar.gz && sha256sum lcov.tar.gz \
     && echo "61a22a62e20908b8b9e27d890bd0ea31f567a7b9668065589266371dcbca0992  lcov.tar.gz" | sha256sum --check \
     && mkdir -p lcov && tar -xzf lcov.tar.gz -C lcov --strip-components=1 \
     && cd lcov \
@@ -218,6 +241,8 @@ RUN wget -O /tmp/libicu-${ICU_VERSION}.tgz https://github.com/unicode-org/icu/re
 USER nonroot:nonroot
 WORKDIR /home/nonroot
 
+RUN echo -e "--retry-connrefused\n--connect-timeout 15\n--retry 5\n--max-time 300\n" > /home/nonroot/.curlrc
+
 # Python
 ENV PYTHON_VERSION=3.11.10 \
     PYENV_ROOT=/home/nonroot/.pyenv \
@@ -243,7 +268,7 @@ 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=1.84.0
+ENV RUSTC_VERSION=1.84.1
 ENV RUSTUP_HOME="/home/nonroot/.rustup"
 ENV PATH="/home/nonroot/.cargo/bin:${PATH}"
 ARG RUSTFILT_VERSION=0.2.1
@@ -251,6 +276,7 @@ ARG CARGO_HAKARI_VERSION=0.9.33
 ARG CARGO_DENY_VERSION=0.16.2
 ARG CARGO_HACK_VERSION=0.6.33
 ARG CARGO_NEXTEST_VERSION=0.9.85
+ARG CARGO_DIESEL_CLI_VERSION=2.2.6
 RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux-gnu/rustup-init && whoami && \
 	chmod +x rustup-init && \
 	./rustup-init -y --default-toolchain ${RUSTC_VERSION} && \
@@ -264,6 +290,8 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
     cargo install cargo-deny --locked --version ${CARGO_DENY_VERSION} && \
     cargo install cargo-hack          --version ${CARGO_HACK_VERSION} && \
     cargo install cargo-nextest       --version ${CARGO_NEXTEST_VERSION} && \
+    cargo install diesel_cli          --version ${CARGO_DIESEL_CLI_VERSION} \
+                                      --features postgres-bundled --no-default-features && \
     rm -rf /home/nonroot/.cargo/registry && \
     rm -rf /home/nonroot/.cargo/git
 

build_tools/patches/pgcopydbv017.patch

@@ -0,0 +1,57 @@
+diff --git a/src/bin/pgcopydb/copydb.c b/src/bin/pgcopydb/copydb.c
+index d730b03..69a9be9 100644
+--- a/src/bin/pgcopydb/copydb.c
++++ b/src/bin/pgcopydb/copydb.c
+@@ -44,6 +44,7 @@ GUC dstSettings[] = {
+ 	{ "synchronous_commit", "'off'" },
+ 	{ "statement_timeout", "0" },
+ 	{ "lock_timeout", "0" },
++	{ "idle_in_transaction_session_timeout", "0" },
+ 	{ NULL, NULL },
+ };
+ 
+diff --git a/src/bin/pgcopydb/pgsql.c b/src/bin/pgcopydb/pgsql.c
+index 94f2f46..e051ba8 100644
+--- a/src/bin/pgcopydb/pgsql.c
++++ b/src/bin/pgcopydb/pgsql.c
+@@ -2319,6 +2319,11 @@ pgsql_execute_log_error(PGSQL *pgsql,
+ 
+ 	LinesBuffer lbuf = { 0 };
+ 
++	if (message != NULL){
++		// make sure message is writable by splitLines
++		message = strdup(message);
++	}
++
+ 	if (!splitLines(&lbuf, message))
+ 	{
+ 		/* errors have already been logged */
+@@ -2332,6 +2337,7 @@ pgsql_execute_log_error(PGSQL *pgsql,
+ 				  PQbackendPID(pgsql->connection),
+ 				  lbuf.lines[lineNumber]);
+ 	}
++        free(message); // free copy of message we created above
+ 
+ 	if (pgsql->logSQL)
+ 	{
+@@ -3174,11 +3180,18 @@ pgcopy_log_error(PGSQL *pgsql, PGresult *res, const char *context)
+ 		/* errors have already been logged */
+ 		return;
+ 	}
+-
+ 	if (res != NULL)
+ 	{
+ 		char *sqlstate = PQresultErrorField(res, PG_DIAG_SQLSTATE);
+-		strlcpy(pgsql->sqlstate, sqlstate, sizeof(pgsql->sqlstate));
++		if (sqlstate == NULL)
++		{
++			// PQresultErrorField returned NULL!
++			pgsql->sqlstate[0] = '\0';  // Set to an empty string to avoid segfault
++		}
++		else
++		{
++			strlcpy(pgsql->sqlstate, sqlstate, sizeof(pgsql->sqlstate));
++		}
+ 	}
+ 
+ 	char *endpoint =

compute/patches/contrib_pg16.patch

@@ -0,0 +1,242 @@
+diff --git a/contrib/amcheck/expected/check_heap.out b/contrib/amcheck/expected/check_heap.out
+index 979e5e8..2375b45 100644
+--- a/contrib/amcheck/expected/check_heap.out
++++ b/contrib/amcheck/expected/check_heap.out
+@@ -80,12 +80,9 @@ INSERT INTO heaptest (a, b)
+ -- same transaction.  The heaptest table is smaller than the default
+ -- wal_skip_threshold, so a wal_level=minimal commit reads the table into
+ -- shared_buffers.  A transaction delays that and excludes any autovacuum.
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_test_stats_tblspc LOCATION '';
+ SELECT sum(reads) AS stats_bulkreads_before
+   FROM pg_stat_io WHERE context = 'bulkread' \gset
+ BEGIN;
+-ALTER TABLE heaptest SET TABLESPACE regress_test_stats_tblspc;
+ -- Check that valid options are not rejected nor corruption reported
+ -- for a non-empty table
+ SELECT * FROM verify_heapam(relation := 'heaptest', skip := 'none');
+@@ -118,14 +115,6 @@ SELECT pg_stat_force_next_flush();
+  
+ (1 row)
+ 
+-SELECT sum(reads) AS stats_bulkreads_after
+-  FROM pg_stat_io WHERE context = 'bulkread' \gset
+-SELECT :stats_bulkreads_after > :stats_bulkreads_before;
+- ?column? 
+-----------
+- t
+-(1 row)
+-
+ CREATE ROLE regress_heaptest_role;
+ -- verify permissions are checked (error due to function not callable)
+ SET ROLE regress_heaptest_role;
+@@ -233,7 +222,6 @@ ERROR:  cannot check relation "test_foreign_table"
+ DETAIL:  This operation is not supported for foreign tables.
+ -- cleanup
+ DROP TABLE heaptest;
+-DROP TABLESPACE regress_test_stats_tblspc;
+ DROP TABLE test_partition;
+ DROP TABLE test_partitioned;
+ DROP OWNED BY regress_heaptest_role; -- permissions
+diff --git a/contrib/amcheck/sql/check_heap.sql b/contrib/amcheck/sql/check_heap.sql
+index 1745bae..3b429c3 100644
+--- a/contrib/amcheck/sql/check_heap.sql
++++ b/contrib/amcheck/sql/check_heap.sql
+@@ -40,12 +40,9 @@ INSERT INTO heaptest (a, b)
+ -- same transaction.  The heaptest table is smaller than the default
+ -- wal_skip_threshold, so a wal_level=minimal commit reads the table into
+ -- shared_buffers.  A transaction delays that and excludes any autovacuum.
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_test_stats_tblspc LOCATION '';
+ SELECT sum(reads) AS stats_bulkreads_before
+   FROM pg_stat_io WHERE context = 'bulkread' \gset
+ BEGIN;
+-ALTER TABLE heaptest SET TABLESPACE regress_test_stats_tblspc;
+ -- Check that valid options are not rejected nor corruption reported
+ -- for a non-empty table
+ SELECT * FROM verify_heapam(relation := 'heaptest', skip := 'none');
+@@ -58,9 +55,6 @@ COMMIT;
+ --   ALTER TABLE ... SET TABLESPACE ...
+ -- causing an additional bulkread, which should be reflected in pg_stat_io.
+ SELECT pg_stat_force_next_flush();
+-SELECT sum(reads) AS stats_bulkreads_after
+-  FROM pg_stat_io WHERE context = 'bulkread' \gset
+-SELECT :stats_bulkreads_after > :stats_bulkreads_before;
+ 
+ CREATE ROLE regress_heaptest_role;
+ 
+@@ -140,7 +134,6 @@ SELECT * FROM verify_heapam('test_foreign_table',
+ 
+ -- cleanup
+ DROP TABLE heaptest;
+-DROP TABLESPACE regress_test_stats_tblspc;
+ DROP TABLE test_partition;
+ DROP TABLE test_partitioned;
+ DROP OWNED BY regress_heaptest_role; -- permissions
+diff --git a/contrib/citext/expected/create_index_acl.out b/contrib/citext/expected/create_index_acl.out
+index 33be13a..70a406c 100644
+--- a/contrib/citext/expected/create_index_acl.out
++++ b/contrib/citext/expected/create_index_acl.out
+@@ -5,9 +5,6 @@
+ -- owner having as few applicable privileges as possible.  (The privileges.sql
+ -- regress_sro_user tests look for the opposite defect; they confirm that
+ -- DefineIndex() uses the table owner userid where necessary.)
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_create_idx_tblspace LOCATION '';
+-RESET allow_in_place_tablespaces;
+ BEGIN;
+ CREATE ROLE regress_minimal;
+ CREATE SCHEMA s;
+@@ -49,11 +46,9 @@ ALTER TABLE s.x OWNER TO regress_minimal;
+ -- Empty-table DefineIndex()
+ CREATE UNIQUE INDEX u0rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e0rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Make the table nonempty.
+ INSERT INTO s.x VALUES ('foo'), ('bar');
+@@ -66,11 +61,9 @@ RESET search_path;
+ GRANT EXECUTE ON FUNCTION s.index_this_expr TO regress_minimal;
+ CREATE UNIQUE INDEX u2rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Shall not find s.coll via search_path, despite the s.const->public.setter
+ -- call having set search_path=s during expression planning.  Suppress the
+@@ -78,9 +71,7 @@ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+ \set VERBOSITY sqlstate
+ ALTER TABLE s.x ADD CONSTRAINT underqualified EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ ERROR:  42704
+ \set VERBOSITY default
+ ROLLBACK;
+-DROP TABLESPACE regress_create_idx_tblspace;
+diff --git a/contrib/citext/sql/create_index_acl.sql b/contrib/citext/sql/create_index_acl.sql
+index 10b5225..ae442e1 100644
+--- a/contrib/citext/sql/create_index_acl.sql
++++ b/contrib/citext/sql/create_index_acl.sql
+@@ -6,10 +6,6 @@
+ -- regress_sro_user tests look for the opposite defect; they confirm that
+ -- DefineIndex() uses the table owner userid where necessary.)
+ 
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_create_idx_tblspace LOCATION '';
+-RESET allow_in_place_tablespaces;
+-
+ BEGIN;
+ CREATE ROLE regress_minimal;
+ CREATE SCHEMA s;
+@@ -51,11 +47,9 @@ ALTER TABLE s.x OWNER TO regress_minimal;
+ -- Empty-table DefineIndex()
+ CREATE UNIQUE INDEX u0rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e0rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Make the table nonempty.
+ INSERT INTO s.x VALUES ('foo'), ('bar');
+@@ -68,11 +62,9 @@ RESET search_path;
+ GRANT EXECUTE ON FUNCTION s.index_this_expr TO regress_minimal;
+ CREATE UNIQUE INDEX u2rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Shall not find s.coll via search_path, despite the s.const->public.setter
+ -- call having set search_path=s during expression planning.  Suppress the
+@@ -80,9 +72,7 @@ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+ \set VERBOSITY sqlstate
+ ALTER TABLE s.x ADD CONSTRAINT underqualified EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ \set VERBOSITY default
+ ROLLBACK;
+ 
+-DROP TABLESPACE regress_create_idx_tblspace;
+diff --git a/contrib/file_fdw/expected/file_fdw.out b/contrib/file_fdw/expected/file_fdw.out
+index 72304e0..ebe131b 100644
+--- a/contrib/file_fdw/expected/file_fdw.out
++++ b/contrib/file_fdw/expected/file_fdw.out
+@@ -4,6 +4,7 @@
+ -- directory paths are passed to us in environment variables
+ \getenv abs_srcdir PG_ABS_SRCDIR
+ -- Clean up in case a prior regression run failed
++SET compute_query_id TO 'off';
+ SET client_min_messages TO 'warning';
+ DROP ROLE IF EXISTS regress_file_fdw_superuser, regress_file_fdw_user, regress_no_priv_user;
+ RESET client_min_messages;
+diff --git a/contrib/file_fdw/sql/file_fdw.sql b/contrib/file_fdw/sql/file_fdw.sql
+index f0548e1..848a08c 100644
+--- a/contrib/file_fdw/sql/file_fdw.sql
++++ b/contrib/file_fdw/sql/file_fdw.sql
+@@ -6,6 +6,7 @@
+ \getenv abs_srcdir PG_ABS_SRCDIR
+ 
+ -- Clean up in case a prior regression run failed
++SET compute_query_id TO 'off';
+ SET client_min_messages TO 'warning';
+ DROP ROLE IF EXISTS regress_file_fdw_superuser, regress_file_fdw_user, regress_no_priv_user;
+ RESET client_min_messages;
+diff --git a/contrib/pageinspect/expected/gist.out b/contrib/pageinspect/expected/gist.out
+index d1adbab..38b52ac 100644
+--- a/contrib/pageinspect/expected/gist.out
++++ b/contrib/pageinspect/expected/gist.out
+@@ -10,25 +10,6 @@ BEGIN;
+ CREATE TABLE test_gist AS SELECT point(i,i) p, i::text t FROM
+     generate_series(1,1000) i;
+ CREATE INDEX test_gist_idx ON test_gist USING gist (p);
+--- Page 0 is the root, the rest are leaf pages
+-SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 0));
+- lsn | nsn | rightlink  | flags 
+------+-----+------------+-------
+- 0/1 | 0/0 | 4294967295 | {}
+-(1 row)
+-
+-SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 1));
+- lsn | nsn | rightlink  | flags  
+------+-----+------------+--------
+- 0/1 | 0/0 | 4294967295 | {leaf}
+-(1 row)
+-
+-SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 2));
+- lsn | nsn | rightlink | flags  
+------+-----+-----------+--------
+- 0/1 | 0/0 |         1 | {leaf}
+-(1 row)
+-
+ COMMIT;
+ SELECT * FROM gist_page_items(get_raw_page('test_gist_idx', 0), 'test_gist_idx');
+  itemoffset |   ctid    | itemlen | dead |             keys              
+diff --git a/contrib/pageinspect/sql/gist.sql b/contrib/pageinspect/sql/gist.sql
+index d263542..607992f 100644
+--- a/contrib/pageinspect/sql/gist.sql
++++ b/contrib/pageinspect/sql/gist.sql
+@@ -12,11 +12,6 @@ CREATE TABLE test_gist AS SELECT point(i,i) p, i::text t FROM
+     generate_series(1,1000) i;
+ CREATE INDEX test_gist_idx ON test_gist USING gist (p);
+ 
+--- Page 0 is the root, the rest are leaf pages
+-SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 0));
+-SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 1));
+-SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 2));
+-
+ COMMIT;
+ 
+ SELECT * FROM gist_page_items(get_raw_page('test_gist_idx', 0), 'test_gist_idx');

compute/patches/contrib_pg17.patch

@@ -0,0 +1,196 @@
+diff --git a/contrib/amcheck/expected/check_heap.out b/contrib/amcheck/expected/check_heap.out
+index 979e5e8..2375b45 100644
+--- a/contrib/amcheck/expected/check_heap.out
++++ b/contrib/amcheck/expected/check_heap.out
+@@ -80,12 +80,9 @@ INSERT INTO heaptest (a, b)
+ -- same transaction.  The heaptest table is smaller than the default
+ -- wal_skip_threshold, so a wal_level=minimal commit reads the table into
+ -- shared_buffers.  A transaction delays that and excludes any autovacuum.
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_test_stats_tblspc LOCATION '';
+ SELECT sum(reads) AS stats_bulkreads_before
+   FROM pg_stat_io WHERE context = 'bulkread' \gset
+ BEGIN;
+-ALTER TABLE heaptest SET TABLESPACE regress_test_stats_tblspc;
+ -- Check that valid options are not rejected nor corruption reported
+ -- for a non-empty table
+ SELECT * FROM verify_heapam(relation := 'heaptest', skip := 'none');
+@@ -118,14 +115,6 @@ SELECT pg_stat_force_next_flush();
+  
+ (1 row)
+ 
+-SELECT sum(reads) AS stats_bulkreads_after
+-  FROM pg_stat_io WHERE context = 'bulkread' \gset
+-SELECT :stats_bulkreads_after > :stats_bulkreads_before;
+- ?column? 
+-----------
+- t
+-(1 row)
+-
+ CREATE ROLE regress_heaptest_role;
+ -- verify permissions are checked (error due to function not callable)
+ SET ROLE regress_heaptest_role;
+@@ -233,7 +222,6 @@ ERROR:  cannot check relation "test_foreign_table"
+ DETAIL:  This operation is not supported for foreign tables.
+ -- cleanup
+ DROP TABLE heaptest;
+-DROP TABLESPACE regress_test_stats_tblspc;
+ DROP TABLE test_partition;
+ DROP TABLE test_partitioned;
+ DROP OWNED BY regress_heaptest_role; -- permissions
+diff --git a/contrib/amcheck/sql/check_heap.sql b/contrib/amcheck/sql/check_heap.sql
+index 1745bae..3b429c3 100644
+--- a/contrib/amcheck/sql/check_heap.sql
++++ b/contrib/amcheck/sql/check_heap.sql
+@@ -40,12 +40,9 @@ INSERT INTO heaptest (a, b)
+ -- same transaction.  The heaptest table is smaller than the default
+ -- wal_skip_threshold, so a wal_level=minimal commit reads the table into
+ -- shared_buffers.  A transaction delays that and excludes any autovacuum.
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_test_stats_tblspc LOCATION '';
+ SELECT sum(reads) AS stats_bulkreads_before
+   FROM pg_stat_io WHERE context = 'bulkread' \gset
+ BEGIN;
+-ALTER TABLE heaptest SET TABLESPACE regress_test_stats_tblspc;
+ -- Check that valid options are not rejected nor corruption reported
+ -- for a non-empty table
+ SELECT * FROM verify_heapam(relation := 'heaptest', skip := 'none');
+@@ -58,9 +55,6 @@ COMMIT;
+ --   ALTER TABLE ... SET TABLESPACE ...
+ -- causing an additional bulkread, which should be reflected in pg_stat_io.
+ SELECT pg_stat_force_next_flush();
+-SELECT sum(reads) AS stats_bulkreads_after
+-  FROM pg_stat_io WHERE context = 'bulkread' \gset
+-SELECT :stats_bulkreads_after > :stats_bulkreads_before;
+ 
+ CREATE ROLE regress_heaptest_role;
+ 
+@@ -140,7 +134,6 @@ SELECT * FROM verify_heapam('test_foreign_table',
+ 
+ -- cleanup
+ DROP TABLE heaptest;
+-DROP TABLESPACE regress_test_stats_tblspc;
+ DROP TABLE test_partition;
+ DROP TABLE test_partitioned;
+ DROP OWNED BY regress_heaptest_role; -- permissions
+diff --git a/contrib/citext/expected/create_index_acl.out b/contrib/citext/expected/create_index_acl.out
+index 33be13a..70a406c 100644
+--- a/contrib/citext/expected/create_index_acl.out
++++ b/contrib/citext/expected/create_index_acl.out
+@@ -5,9 +5,6 @@
+ -- owner having as few applicable privileges as possible.  (The privileges.sql
+ -- regress_sro_user tests look for the opposite defect; they confirm that
+ -- DefineIndex() uses the table owner userid where necessary.)
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_create_idx_tblspace LOCATION '';
+-RESET allow_in_place_tablespaces;
+ BEGIN;
+ CREATE ROLE regress_minimal;
+ CREATE SCHEMA s;
+@@ -49,11 +46,9 @@ ALTER TABLE s.x OWNER TO regress_minimal;
+ -- Empty-table DefineIndex()
+ CREATE UNIQUE INDEX u0rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e0rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Make the table nonempty.
+ INSERT INTO s.x VALUES ('foo'), ('bar');
+@@ -66,11 +61,9 @@ RESET search_path;
+ GRANT EXECUTE ON FUNCTION s.index_this_expr TO regress_minimal;
+ CREATE UNIQUE INDEX u2rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Shall not find s.coll via search_path, despite the s.const->public.setter
+ -- call having set search_path=s during expression planning.  Suppress the
+@@ -78,9 +71,7 @@ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+ \set VERBOSITY sqlstate
+ ALTER TABLE s.x ADD CONSTRAINT underqualified EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ ERROR:  42704
+ \set VERBOSITY default
+ ROLLBACK;
+-DROP TABLESPACE regress_create_idx_tblspace;
+diff --git a/contrib/citext/sql/create_index_acl.sql b/contrib/citext/sql/create_index_acl.sql
+index 10b5225..ae442e1 100644
+--- a/contrib/citext/sql/create_index_acl.sql
++++ b/contrib/citext/sql/create_index_acl.sql
+@@ -6,10 +6,6 @@
+ -- regress_sro_user tests look for the opposite defect; they confirm that
+ -- DefineIndex() uses the table owner userid where necessary.)
+ 
+-SET allow_in_place_tablespaces = true;
+-CREATE TABLESPACE regress_create_idx_tblspace LOCATION '';
+-RESET allow_in_place_tablespaces;
+-
+ BEGIN;
+ CREATE ROLE regress_minimal;
+ CREATE SCHEMA s;
+@@ -51,11 +47,9 @@ ALTER TABLE s.x OWNER TO regress_minimal;
+ -- Empty-table DefineIndex()
+ CREATE UNIQUE INDEX u0rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e0rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Make the table nonempty.
+ INSERT INTO s.x VALUES ('foo'), ('bar');
+@@ -68,11 +62,9 @@ RESET search_path;
+ GRANT EXECUTE ON FUNCTION s.index_this_expr TO regress_minimal;
+ CREATE UNIQUE INDEX u2rows ON s.x USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll s.citext_pattern_ops)
+-  TABLESPACE regress_create_idx_tblspace
+   WHERE s.index_row_if(y);
+ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE s.coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ -- Shall not find s.coll via search_path, despite the s.const->public.setter
+ -- call having set search_path=s during expression planning.  Suppress the
+@@ -80,9 +72,7 @@ ALTER TABLE s.x ADD CONSTRAINT e2rows EXCLUDE USING btree
+ \set VERBOSITY sqlstate
+ ALTER TABLE s.x ADD CONSTRAINT underqualified EXCLUDE USING btree
+   ((s.index_this_expr(y, s.const())) COLLATE coll WITH s.=)
+-  USING INDEX TABLESPACE regress_create_idx_tblspace
+   WHERE (s.index_row_if(y));
+ \set VERBOSITY default
+ ROLLBACK;
+ 
+-DROP TABLESPACE regress_create_idx_tblspace;
+diff --git a/contrib/file_fdw/expected/file_fdw.out b/contrib/file_fdw/expected/file_fdw.out
+index 86c148a..81bdb2c 100644
+--- a/contrib/file_fdw/expected/file_fdw.out
++++ b/contrib/file_fdw/expected/file_fdw.out
+@@ -4,6 +4,7 @@
+ -- directory paths are passed to us in environment variables
+ \getenv abs_srcdir PG_ABS_SRCDIR
+ -- Clean up in case a prior regression run failed
++SET compute_query_id TO 'off';
+ SET client_min_messages TO 'warning';
+ DROP ROLE IF EXISTS regress_file_fdw_superuser, regress_file_fdw_user, regress_no_priv_user;
+ RESET client_min_messages;
+diff --git a/contrib/file_fdw/sql/file_fdw.sql b/contrib/file_fdw/sql/file_fdw.sql
+index f0548e1..848a08c 100644
+--- a/contrib/file_fdw/sql/file_fdw.sql
++++ b/contrib/file_fdw/sql/file_fdw.sql
+@@ -6,6 +6,7 @@
+ \getenv abs_srcdir PG_ABS_SRCDIR
+ 
+ -- Clean up in case a prior regression run failed
++SET compute_query_id TO 'off';
+ SET client_min_messages TO 'warning';
+ DROP ROLE IF EXISTS regress_file_fdw_superuser, regress_file_fdw_user, regress_no_priv_user;
+ RESET client_min_messages;

compute/patches/pg_graphql.patch

@@ -0,0 +1,19 @@
+commit ec6a491d126882966a696f9ad5d3698935361d55
+Author: Alexey Masterov <alexeymasterov@neon.tech>
+Date:   Tue Dec 17 10:25:00 2024 +0100
+
+    Changes required to run tests on Neon
+
+diff --git a/test/expected/permissions_functions.out b/test/expected/permissions_functions.out
+index 1e9fbc2..94cbe25 100644
+--- a/test/expected/permissions_functions.out
++++ b/test/expected/permissions_functions.out
+@@ -64,7 +64,7 @@ begin;
+     select current_user;
+  current_user 
+ --------------
+- postgres
++ cloud_admin
+ (1 row)
+ 
+     -- revoke default access from the public role for new functions

compute_tools/src/bin/compute_ctl.rs

@@ -34,6 +34,7 @@
 //!             -r http://pg-ext-s3-gateway \
 //! ```
 use std::collections::HashMap;
+use std::ffi::OsString;
 use std::fs::File;
 use std::path::Path;
 use std::process::exit;
@@ -44,7 +45,7 @@ use std::{thread, time::Duration};
 
 use anyhow::{Context, Result};
 use chrono::Utc;
-use clap::Arg;
+use clap::Parser;
 use compute_tools::disk_quota::set_disk_quota;
 use compute_tools::lsn_lease::launch_lsn_lease_bg_task_for_static;
 use signal_hook::consts::{SIGQUIT, SIGTERM};
@@ -73,10 +74,75 @@ use utils::failpoint_support;
 // in-case of not-set environment var
 const BUILD_TAG_DEFAULT: &str = "latest";
 
-fn main() -> Result<()> {
-    let scenario = failpoint_support::init();
+// Compatibility hack: if the control plane specified any remote-ext-config
+// use the default value for extension storage proxy gateway.
+// Remove this once the control plane is updated to pass the gateway URL
+fn parse_remote_ext_config(arg: &str) -> Result<String> {
+    if arg.starts_with("http") {
+        Ok(arg.trim_end_matches('/').to_string())
+    } else {
+        Ok("http://pg-ext-s3-gateway".to_string())
+    }
+}
 
-    let (build_tag, clap_args) = init()?;
+#[derive(Parser)]
+#[command(rename_all = "kebab-case")]
+struct Cli {
+    #[arg(short = 'b', long, default_value = "postgres", env = "POSTGRES_PATH")]
+    pub pgbin: String,
+
+    #[arg(short = 'r', long, value_parser = parse_remote_ext_config)]
+    pub remote_ext_config: Option<String>,
+
+    #[arg(long, default_value_t = 3080)]
+    pub http_port: u16,
+
+    #[arg(short = 'D', long, value_name = "DATADIR")]
+    pub pgdata: String,
+
+    #[arg(short = 'C', long, value_name = "DATABASE_URL")]
+    pub connstr: String,
+
+    #[cfg(target_os = "linux")]
+    #[arg(long, default_value = "neon-postgres")]
+    pub cgroup: String,
+
+    #[cfg(target_os = "linux")]
+    #[arg(
+        long,
+        default_value = "host=localhost port=5432 dbname=postgres user=cloud_admin sslmode=disable application_name=vm-monitor"
+    )]
+    pub filecache_connstr: String,
+
+    #[cfg(target_os = "linux")]
+    #[arg(long, default_value = "0.0.0.0:10301")]
+    pub vm_monitor_addr: String,
+
+    #[arg(long, action = clap::ArgAction::SetTrue)]
+    pub resize_swap_on_bind: bool,
+
+    #[arg(long)]
+    pub set_disk_quota_for_fs: Option<String>,
+
+    #[arg(short = 's', long = "spec", group = "spec")]
+    pub spec_json: Option<String>,
+
+    #[arg(short = 'S', long, group = "spec-path")]
+    pub spec_path: Option<OsString>,
+
+    #[arg(short = 'i', long, group = "compute-id", conflicts_with_all = ["spec", "spec-path"])]
+    pub compute_id: Option<String>,
+
+    #[arg(short = 'p', long, conflicts_with_all = ["spec", "spec-path"], requires = "compute-id", value_name = "CONTROL_PLANE_API_BASE_URL")]
+    pub control_plane_uri: Option<String>,
+}
+
+fn main() -> Result<()> {
+    let cli = Cli::parse();
+
+    let build_tag = init()?;
+
+    let scenario = failpoint_support::init();
 
     // enable core dumping for all child processes
     setrlimit(Resource::CORE, rlimit::INFINITY, rlimit::INFINITY)?;
@@ -85,13 +151,11 @@ fn main() -> Result<()> {
         // Enter startup tracing context
         let _startup_context_guard = startup_context_from_env();
 
-        let cli_args = process_cli(&clap_args)?;
+        let cli_spec = try_spec_from_cli(&cli)?;
 
-        let cli_spec = try_spec_from_cli(&clap_args, &cli_args)?;
+        let compute = wait_spec(build_tag, &cli, cli_spec)?;
 
-        let wait_spec_result = wait_spec(build_tag, cli_args, cli_spec)?;
-
-        start_postgres(&clap_args, wait_spec_result)?
+        start_postgres(&cli, compute)?
 
         // Startup is finished, exit the startup tracing span
     };
@@ -108,7 +172,7 @@ fn main() -> Result<()> {
     deinit_and_exit(wait_pg_result);
 }
 
-fn init() -> Result<(String, clap::ArgMatches)> {
+fn init() -> Result<String> {
     init_tracing_and_logging(DEFAULT_LOG_LEVEL)?;
 
     let mut signals = Signals::new([SIGINT, SIGTERM, SIGQUIT])?;
@@ -123,66 +187,7 @@ fn init() -> Result<(String, clap::ArgMatches)> {
         .to_string();
     info!("build_tag: {build_tag}");
 
-    Ok((build_tag, cli().get_matches()))
-}
-
-fn process_cli(matches: &clap::ArgMatches) -> Result<ProcessCliResult> {
-    let pgbin_default = "postgres";
-    let pgbin = matches
-        .get_one::<String>("pgbin")
-        .map(|s| s.as_str())
-        .unwrap_or(pgbin_default);
-
-    let ext_remote_storage = matches
-        .get_one::<String>("remote-ext-config")
-        // Compatibility hack: if the control plane specified any remote-ext-config
-        // use the default value for extension storage proxy gateway.
-        // Remove this once the control plane is updated to pass the gateway URL
-        .map(|conf| {
-            if conf.starts_with("http") {
-                conf.trim_end_matches('/')
-            } else {
-                "http://pg-ext-s3-gateway"
-            }
-        });
-
-    let http_port = *matches
-        .get_one::<u16>("http-port")
-        .expect("http-port is required");
-    let pgdata = matches
-        .get_one::<String>("pgdata")
-        .expect("PGDATA path is required");
-    let connstr = matches
-        .get_one::<String>("connstr")
-        .expect("Postgres connection string is required");
-    let spec_json = matches.get_one::<String>("spec");
-    let spec_path = matches.get_one::<String>("spec-path");
-    let resize_swap_on_bind = matches.get_flag("resize-swap-on-bind");
-    let set_disk_quota_for_fs = matches.get_one::<String>("set-disk-quota-for-fs");
-
-    Ok(ProcessCliResult {
-        connstr,
-        pgdata,
-        pgbin,
-        ext_remote_storage,
-        http_port,
-        spec_json,
-        spec_path,
-        resize_swap_on_bind,
-        set_disk_quota_for_fs,
-    })
-}
-
-struct ProcessCliResult<'clap> {
-    connstr: &'clap str,
-    pgdata: &'clap str,
-    pgbin: &'clap str,
-    ext_remote_storage: Option<&'clap str>,
-    http_port: u16,
-    spec_json: Option<&'clap String>,
-    spec_path: Option<&'clap String>,
-    resize_swap_on_bind: bool,
-    set_disk_quota_for_fs: Option<&'clap String>,
+    Ok(build_tag)
 }
 
 fn startup_context_from_env() -> Option<opentelemetry::ContextGuard> {
@@ -235,19 +240,9 @@ fn startup_context_from_env() -> Option<opentelemetry::ContextGuard> {
     }
 }
 
-fn try_spec_from_cli(
-    matches: &clap::ArgMatches,
-    ProcessCliResult {
-        spec_json,
-        spec_path,
-        ..
-    }: &ProcessCliResult,
-) -> Result<CliSpecParams> {
-    let compute_id = matches.get_one::<String>("compute-id");
-    let control_plane_uri = matches.get_one::<String>("control-plane-uri");
-
+fn try_spec_from_cli(cli: &Cli) -> Result<CliSpecParams> {
     // First, try to get cluster spec from the cli argument
-    if let Some(spec_json) = spec_json {
+    if let Some(ref spec_json) = cli.spec_json {
         info!("got spec from cli argument {}", spec_json);
         return Ok(CliSpecParams {
             spec: Some(serde_json::from_str(spec_json)?),
@@ -256,7 +251,7 @@ fn try_spec_from_cli(
     }
 
     // Second, try to read it from the file if path is provided
-    if let Some(spec_path) = spec_path {
+    if let Some(ref spec_path) = cli.spec_path {
         let file = File::open(Path::new(spec_path))?;
         return Ok(CliSpecParams {
             spec: Some(serde_json::from_reader(file)?),
@@ -264,17 +259,20 @@ fn try_spec_from_cli(
         });
     }
 
-    let Some(compute_id) = compute_id else {
+    if cli.compute_id.is_none() {
         panic!(
             "compute spec should be provided by one of the following ways: \
                 --spec OR --spec-path OR --control-plane-uri and --compute-id"
         );
     };
-    let Some(control_plane_uri) = control_plane_uri else {
+    if cli.control_plane_uri.is_none() {
         panic!("must specify both --control-plane-uri and --compute-id or none");
     };
 
-    match get_spec_from_control_plane(control_plane_uri, compute_id) {
+    match get_spec_from_control_plane(
+        cli.control_plane_uri.as_ref().unwrap(),
+        cli.compute_id.as_ref().unwrap(),
+    ) {
         Ok(spec) => Ok(CliSpecParams {
             spec,
             live_config_allowed: true,
@@ -298,21 +296,12 @@ struct CliSpecParams {
 
 fn wait_spec(
     build_tag: String,
-    ProcessCliResult {
-        connstr,
-        pgdata,
-        pgbin,
-        ext_remote_storage,
-        resize_swap_on_bind,
-        set_disk_quota_for_fs,
-        http_port,
-        ..
-    }: ProcessCliResult,
+    cli: &Cli,
     CliSpecParams {
         spec,
         live_config_allowed,
     }: CliSpecParams,
-) -> Result<WaitSpecResult> {
+) -> Result<Arc<ComputeNode>> {
     let mut new_state = ComputeState::new();
     let spec_set;
 
@@ -324,7 +313,7 @@ fn wait_spec(
     } else {
         spec_set = false;
     }
-    let connstr = Url::parse(connstr).context("cannot parse connstr as a URL")?;
+    let connstr = Url::parse(&cli.connstr).context("cannot parse connstr as a URL")?;
     let conn_conf = postgres::config::Config::from_str(connstr.as_str())
         .context("cannot build postgres config from connstr")?;
     let tokio_conn_conf = tokio_postgres::config::Config::from_str(connstr.as_str())
@@ -333,14 +322,14 @@ fn wait_spec(
         connstr,
         conn_conf,
         tokio_conn_conf,
-        pgdata: pgdata.to_string(),
-        pgbin: pgbin.to_string(),
-        pgversion: get_pg_version_string(pgbin),
-        http_port,
+        pgdata: cli.pgdata.clone(),
+        pgbin: cli.pgbin.clone(),
+        pgversion: get_pg_version_string(&cli.pgbin),
+        http_port: cli.http_port,
         live_config_allowed,
         state: Mutex::new(new_state),
         state_changed: Condvar::new(),
-        ext_remote_storage: ext_remote_storage.map(|s| s.to_string()),
+        ext_remote_storage: cli.remote_ext_config.clone(),
         ext_download_progress: RwLock::new(HashMap::new()),
         build_tag,
     };
@@ -357,7 +346,7 @@ fn wait_spec(
     // Launch http service first, so that we can serve control-plane requests
     // while configuration is still in progress.
     let _http_handle =
-        launch_http_server(http_port, &compute).expect("cannot launch http endpoint thread");
+        launch_http_server(cli.http_port, &compute).expect("cannot launch http endpoint thread");
 
     if !spec_set {
         // No spec provided, hang waiting for it.
@@ -389,27 +378,12 @@ fn wait_spec(
 
     launch_lsn_lease_bg_task_for_static(&compute);
 
-    Ok(WaitSpecResult {
-        compute,
-        resize_swap_on_bind,
-        set_disk_quota_for_fs: set_disk_quota_for_fs.cloned(),
-    })
-}
-
-struct WaitSpecResult {
-    compute: Arc<ComputeNode>,
-    resize_swap_on_bind: bool,
-    set_disk_quota_for_fs: Option<String>,
+    Ok(compute)
 }
 
 fn start_postgres(
-    // need to allow unused because `matches` is only used if target_os = "linux"
-    #[allow(unused_variables)] matches: &clap::ArgMatches,
-    WaitSpecResult {
-        compute,
-        resize_swap_on_bind,
-        set_disk_quota_for_fs,
-    }: WaitSpecResult,
+    cli: &Cli,
+    compute: Arc<ComputeNode>,
 ) -> Result<(Option<PostgresHandle>, StartPostgresResult)> {
     // We got all we need, update the state.
     let mut state = compute.state.lock().unwrap();
@@ -437,7 +411,7 @@ fn start_postgres(
     let mut delay_exit = false;
 
     // Resize swap to the desired size if the compute spec says so
-    if let (Some(size_bytes), true) = (swap_size_bytes, resize_swap_on_bind) {
+    if let (Some(size_bytes), true) = (swap_size_bytes, cli.resize_swap_on_bind) {
         // To avoid 'swapoff' hitting postgres startup, we need to run resize-swap to completion
         // *before* starting postgres.
         //
@@ -464,9 +438,9 @@ fn start_postgres(
 
     // Set disk quota if the compute spec says so
     if let (Some(disk_quota_bytes), Some(disk_quota_fs_mountpoint)) =
-        (disk_quota_bytes, set_disk_quota_for_fs)
+        (disk_quota_bytes, cli.set_disk_quota_for_fs.as_ref())
     {
-        match set_disk_quota(disk_quota_bytes, &disk_quota_fs_mountpoint) {
+        match set_disk_quota(disk_quota_bytes, disk_quota_fs_mountpoint) {
             Ok(()) => {
                 let size_mib = disk_quota_bytes as f32 / (1 << 20) as f32; // just for more coherent display.
                 info!(%disk_quota_bytes, %size_mib, "set disk quota");
@@ -509,13 +483,7 @@ fn start_postgres(
         if #[cfg(target_os = "linux")] {
             use std::env;
             use tokio_util::sync::CancellationToken;
-            let vm_monitor_addr = matches
-                .get_one::<String>("vm-monitor-addr")
-                .expect("--vm-monitor-addr should always be set because it has a default arg");
-            let file_cache_connstr = matches.get_one::<String>("filecache-connstr");
-            let cgroup = matches.get_one::<String>("cgroup");
 
-            // Only make a runtime if we need to.
             // Note: it seems like you can make a runtime in an inner scope and
             // if you start a task in it it won't be dropped. However, make it
             // in the outermost scope just to be safe.
@@ -538,15 +506,15 @@ fn start_postgres(
             let pgconnstr = if disable_lfc_resizing.unwrap_or(false) {
                 None
             } else {
-                file_cache_connstr.cloned()
+                Some(cli.filecache_connstr.clone())
             };
 
             let vm_monitor = rt.as_ref().map(|rt| {
                 rt.spawn(vm_monitor::start(
                     Box::leak(Box::new(vm_monitor::Args {
-                        cgroup: cgroup.cloned(),
+                        cgroup: Some(cli.cgroup.clone()),
                         pgconnstr,
-                        addr: vm_monitor_addr.clone(),
+                        addr: cli.vm_monitor_addr.clone(),
                     })),
                     token.clone(),
                 ))
@@ -702,105 +670,6 @@ fn deinit_and_exit(WaitPostgresResult { exit_code }: WaitPostgresResult) -> ! {
     exit(exit_code.unwrap_or(1))
 }
 
-fn cli() -> clap::Command {
-    // Env variable is set by `cargo`
-    let version = option_env!("CARGO_PKG_VERSION").unwrap_or("unknown");
-    clap::Command::new("compute_ctl")
-        .version(version)
-        .arg(
-            Arg::new("http-port")
-                .long("http-port")
-                .value_name("HTTP_PORT")
-                .default_value("3080")
-                .value_parser(clap::value_parser!(u16))
-                .required(false),
-        )
-        .arg(
-            Arg::new("connstr")
-                .short('C')
-                .long("connstr")
-                .value_name("DATABASE_URL")
-                .required(true),
-        )
-        .arg(
-            Arg::new("pgdata")
-                .short('D')
-                .long("pgdata")
-                .value_name("DATADIR")
-                .required(true),
-        )
-        .arg(
-            Arg::new("pgbin")
-                .short('b')
-                .long("pgbin")
-                .default_value("postgres")
-                .value_name("POSTGRES_PATH"),
-        )
-        .arg(
-            Arg::new("spec")
-                .short('s')
-                .long("spec")
-                .value_name("SPEC_JSON"),
-        )
-        .arg(
-            Arg::new("spec-path")
-                .short('S')
-                .long("spec-path")
-                .value_name("SPEC_PATH"),
-        )
-        .arg(
-            Arg::new("compute-id")
-                .short('i')
-                .long("compute-id")
-                .value_name("COMPUTE_ID"),
-        )
-        .arg(
-            Arg::new("control-plane-uri")
-                .short('p')
-                .long("control-plane-uri")
-                .value_name("CONTROL_PLANE_API_BASE_URI"),
-        )
-        .arg(
-            Arg::new("remote-ext-config")
-                .short('r')
-                .long("remote-ext-config")
-                .value_name("REMOTE_EXT_CONFIG"),
-        )
-        // TODO(fprasx): we currently have default arguments because the cloud PR
-        // to pass them in hasn't been merged yet. We should get rid of them once
-        // the PR is merged.
-        .arg(
-            Arg::new("vm-monitor-addr")
-                .long("vm-monitor-addr")
-                .default_value("0.0.0.0:10301")
-                .value_name("VM_MONITOR_ADDR"),
-        )
-        .arg(
-            Arg::new("cgroup")
-                .long("cgroup")
-                .default_value("neon-postgres")
-                .value_name("CGROUP"),
-        )
-        .arg(
-            Arg::new("filecache-connstr")
-                .long("filecache-connstr")
-                .default_value(
-                    "host=localhost port=5432 dbname=postgres user=cloud_admin sslmode=disable application_name=vm-monitor",
-                )
-                .value_name("FILECACHE_CONNSTR"),
-        )
-        .arg(
-            Arg::new("resize-swap-on-bind")
-                .long("resize-swap-on-bind")
-                .action(clap::ArgAction::SetTrue),
-        )
-        .arg(
-            Arg::new("set-disk-quota-for-fs")
-                .long("set-disk-quota-for-fs")
-                .value_name("SET_DISK_QUOTA_FOR_FS")
-        )
-}
-
 /// When compute_ctl is killed, send also termination signal to sync-safekeepers
 /// to prevent leakage. TODO: it is better to convert compute_ctl to async and
 /// wait for termination which would be easy then.
@@ -810,7 +679,14 @@ fn handle_exit_signal(sig: i32) {
     exit(1);
 }
 
-#[test]
-fn verify_cli() {
-    cli().debug_assert()
+#[cfg(test)]
+mod test {
+    use clap::CommandFactory;
+
+    use super::Cli;
+
+    #[test]
+    fn verify_cli() {
+        Cli::command().debug_assert()
+    }
 }

compute_tools/src/extension_server.rs

@@ -85,6 +85,8 @@ use tracing::info;
 use tracing::log::warn;
 use zstd::stream::read::Decoder;
 
+use crate::metrics::{REMOTE_EXT_REQUESTS_TOTAL, UNKNOWN_HTTP_STATUS};
+
 fn get_pg_config(argument: &str, pgbin: &str) -> String {
     // gives the result of `pg_config [argument]`
     // where argument is a flag like `--version` or `--sharedir`
@@ -256,23 +258,60 @@ pub fn create_control_files(remote_extensions: &RemoteExtSpec, pgbin: &str) {
 async fn download_extension_tar(ext_remote_storage: &str, ext_path: &str) -> Result<Bytes> {
     let uri = format!("{}/{}", ext_remote_storage, ext_path);
 
-    info!("Download extension {:?} from uri {:?}", ext_path, uri);
+    info!("Download extension {} from uri {}", ext_path, uri);
 
-    let resp = reqwest::get(uri).await?;
+    match do_extension_server_request(&uri).await {
+        Ok(resp) => {
+            info!("Successfully downloaded remote extension data {}", ext_path);
+            REMOTE_EXT_REQUESTS_TOTAL
+                .with_label_values(&[&StatusCode::OK.to_string()])
+                .inc();
+            Ok(resp)
+        }
+        Err((msg, status)) => {
+            REMOTE_EXT_REQUESTS_TOTAL
+                .with_label_values(&[&status])
+                .inc();
+            bail!(msg);
+        }
+    }
+}
 
-    match resp.status() {
+// Do a single remote extensions server request.
+// Return result or (error message + stringified status code) in case of any failures.
+async fn do_extension_server_request(uri: &str) -> Result<Bytes, (String, String)> {
+    let resp = reqwest::get(uri).await.map_err(|e| {
+        (
+            format!(
+                "could not perform remote extensions server request: {:?}",
+                e
+            ),
+            UNKNOWN_HTTP_STATUS.to_string(),
+        )
+    })?;
+    let status = resp.status();
+
+    match status {
         StatusCode::OK => match resp.bytes().await {
-            Ok(resp) => {
-                info!("Download extension {:?} completed successfully", ext_path);
-                Ok(resp)
-            }
-            Err(e) => bail!("could not deserialize remote extension response: {}", e),
+            Ok(resp) => Ok(resp),
+            Err(e) => Err((
+                format!("could not read remote extensions server response: {:?}", e),
+                // It's fine to return and report error with status as 200 OK,
+                // because we still failed to read the response.
+                status.to_string(),
+            )),
         },
-        StatusCode::SERVICE_UNAVAILABLE => bail!("remote extension is temporarily unavailable"),
-        _ => bail!(
-            "unexpected remote extension response status code: {}",
-            resp.status()
-        ),
+        StatusCode::SERVICE_UNAVAILABLE => Err((
+            "remote extensions server is temporarily unavailable".to_string(),
+            status.to_string(),
+        )),
+        _ => Err((
+            format!(
+                "unexpected remote extensions server response status code: {}",
+                status
+            ),
+            status.to_string(),
+        )),
     }
 }
 

compute_tools/src/http/openapi_spec.yaml

@@ -68,35 +68,6 @@ paths:
               schema:
                 $ref: "#/components/schemas/ComputeInsights"
 
-  /installed_extensions:
-    get:
-      tags:
-      - Info
-      summary: Get installed extensions.
-      description: ""
-      operationId: getInstalledExtensions
-      responses:
-        200:
-          description: List of installed extensions
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/InstalledExtensions"
-  /info:
-    get:
-      tags:
-      - Info
-      summary: Get info about the compute pod / VM.
-      description: ""
-      operationId: getInfo
-      responses:
-        200:
-          description: Info
-          content:
-            application/json:
-              schema:
-                $ref: "#/components/schemas/Info"
-
   /dbs_and_roles:
     get:
       tags:

compute_tools/src/http/routes/info.rs

@@ -1,11 +0,0 @@
-use axum::response::Response;
-use compute_api::responses::InfoResponse;
-use http::StatusCode;
-
-use crate::http::JsonResponse;
-
-/// Get information about the physical characteristics about the compute.
-pub(in crate::http) async fn get_info() -> Response {
-    let num_cpus = num_cpus::get_physical();
-    JsonResponse::success(StatusCode::OK, &InfoResponse { num_cpus })
-}

compute_tools/src/http/routes/installed_extensions.rs

@@ -1,33 +0,0 @@
-use std::sync::Arc;
-
-use axum::{extract::State, response::Response};
-use compute_api::responses::ComputeStatus;
-use http::StatusCode;
-use tokio::task;
-
-use crate::{compute::ComputeNode, http::JsonResponse, installed_extensions};
-
-/// Get a list of installed extensions.
-pub(in crate::http) async fn get_installed_extensions(
-    State(compute): State<Arc<ComputeNode>>,
-) -> Response {
-    let status = compute.get_status();
-    if status != ComputeStatus::Running {
-        return JsonResponse::invalid_status(status);
-    }
-
-    let conf = compute.get_conn_conf(None);
-    let res = task::spawn_blocking(move || installed_extensions::get_installed_extensions(conf))
-        .await
-        .unwrap();
-
-    match res {
-        Ok(installed_extensions) => {
-            JsonResponse::success(StatusCode::OK, Some(installed_extensions))
-        }
-        Err(e) => JsonResponse::error(
-            StatusCode::INTERNAL_SERVER_ERROR,
-            format!("failed to get list of installed extensions: {e}"),
-        ),
-    }
-}

compute_tools/src/http/routes/metrics.rs

@@ -2,17 +2,16 @@ use axum::{body::Body, response::Response};
 use http::header::CONTENT_TYPE;
 use http::StatusCode;
 use metrics::proto::MetricFamily;
-use metrics::Encoder;
-use metrics::TextEncoder;
+use metrics::{Encoder, TextEncoder};
 
-use crate::{http::JsonResponse, installed_extensions};
+use crate::{http::JsonResponse, metrics::collect};
 
 /// Expose Prometheus metrics.
 pub(in crate::http) async fn get_metrics() -> Response {
     // When we call TextEncoder::encode() below, it will immediately return an
     // error if a metric family has no metrics, so we need to preemptively
     // filter out metric families with no metrics.
-    let metrics = installed_extensions::collect()
+    let metrics = collect()
         .into_iter()
         .filter(|m| !m.get_metric().is_empty())
         .collect::<Vec<MetricFamily>>();

compute_tools/src/http/routes/mod.rs

@@ -10,9 +10,7 @@ pub(in crate::http) mod extension_server;
 pub(in crate::http) mod extensions;
 pub(in crate::http) mod failpoints;
 pub(in crate::http) mod grants;
-pub(in crate::http) mod info;
 pub(in crate::http) mod insights;
-pub(in crate::http) mod installed_extensions;
 pub(in crate::http) mod metrics;
 pub(in crate::http) mod metrics_json;
 pub(in crate::http) mod status;

compute_tools/src/http/server.rs

@@ -22,8 +22,7 @@ use uuid::Uuid;
 
 use super::routes::{
     check_writability, configure, database_schema, dbs_and_roles, extension_server, extensions,
-    grants, info as info_route, insights, installed_extensions, metrics, metrics_json, status,
-    terminate,
+    grants, insights, metrics, metrics_json, status, terminate,
 };
 use crate::compute::ComputeNode;
 
@@ -60,12 +59,7 @@ async fn serve(port: u16, compute: Arc<ComputeNode>) {
         )
         .route("/extensions", post(extensions::install_extension))
         .route("/grants", post(grants::add_grant))
-        .route("/info", get(info_route::get_info))
         .route("/insights", get(insights::get_insights))
-        .route(
-            "/installed_extensions",
-            get(installed_extensions::get_installed_extensions),
-        )
         .route("/metrics", get(metrics::get_metrics))
         .route("/metrics.json", get(metrics_json::get_metrics))
         .route("/status", get(status::get_status))

compute_tools/src/installed_extensions.rs

@@ -1,13 +1,10 @@
 use compute_api::responses::{InstalledExtension, InstalledExtensions};
-use metrics::proto::MetricFamily;
 use std::collections::HashMap;
 
 use anyhow::Result;
 use postgres::{Client, NoTls};
 
-use metrics::core::Collector;
-use metrics::{register_uint_gauge_vec, UIntGaugeVec};
-use once_cell::sync::Lazy;
+use crate::metrics::INSTALLED_EXTENSIONS;
 
 /// We don't reuse get_existing_dbs() just for code clarity
 /// and to make database listing query here more explicit.
@@ -102,16 +99,3 @@ pub fn get_installed_extensions(mut conf: postgres::config::Config) -> Result<In
         extensions: extensions_map.into_values().collect(),
     })
 }
-
-static INSTALLED_EXTENSIONS: Lazy<UIntGaugeVec> = Lazy::new(|| {
-    register_uint_gauge_vec!(
-        "compute_installed_extensions",
-        "Number of databases where the version of extension is installed",
-        &["extension_name", "version", "owned_by_superuser"]
-    )
-    .expect("failed to define a metric")
-});
-
-pub fn collect() -> Vec<MetricFamily> {
-    INSTALLED_EXTENSIONS.collect()
-}

compute_tools/src/lib.rs

@@ -16,6 +16,7 @@ pub mod extension_server;
 pub mod installed_extensions;
 pub mod local_proxy;
 pub mod lsn_lease;
+pub mod metrics;
 mod migration;
 pub mod monitor;
 pub mod params;

compute_tools/src/metrics.rs

@@ -0,0 +1,70 @@
+use metrics::core::Collector;
+use metrics::proto::MetricFamily;
+use metrics::{register_int_counter_vec, register_uint_gauge_vec, IntCounterVec, UIntGaugeVec};
+use once_cell::sync::Lazy;
+
+pub(crate) static INSTALLED_EXTENSIONS: Lazy<UIntGaugeVec> = Lazy::new(|| {
+    register_uint_gauge_vec!(
+        "compute_installed_extensions",
+        "Number of databases where the version of extension is installed",
+        &["extension_name", "version", "owned_by_superuser"]
+    )
+    .expect("failed to define a metric")
+});
+
+// Normally, any HTTP API request is described by METHOD (e.g. GET, POST, etc.) + PATH,
+// but for all our APIs we defined a 'slug'/method/operationId in the OpenAPI spec.
+// And it's fair to call it a 'RPC' (Remote Procedure Call).
+pub enum CPlaneRequestRPC {
+    GetSpec,
+}
+
+impl CPlaneRequestRPC {
+    pub fn as_str(&self) -> &str {
+        match self {
+            CPlaneRequestRPC::GetSpec => "GetSpec",
+        }
+    }
+}
+
+pub const UNKNOWN_HTTP_STATUS: &str = "unknown";
+
+pub(crate) static CPLANE_REQUESTS_TOTAL: Lazy<IntCounterVec> = Lazy::new(|| {
+    register_int_counter_vec!(
+        "compute_ctl_cplane_requests_total",
+        "Total number of control plane requests made by compute_ctl by status",
+        &["rpc", "http_status"]
+    )
+    .expect("failed to define a metric")
+});
+
+/// Total number of failed database migrations. Per-compute, this is actually a boolean metric,
+/// either empty or with a single value (1, migration_id) because we stop at the first failure.
+/// Yet, the sum over the fleet will provide the total number of failures.
+pub(crate) static DB_MIGRATION_FAILED: Lazy<IntCounterVec> = Lazy::new(|| {
+    register_int_counter_vec!(
+        "compute_ctl_db_migration_failed_total",
+        "Total number of failed database migrations",
+        &["migration_id"]
+    )
+    .expect("failed to define a metric")
+});
+
+pub(crate) static REMOTE_EXT_REQUESTS_TOTAL: Lazy<IntCounterVec> = Lazy::new(|| {
+    register_int_counter_vec!(
+        "compute_ctl_remote_ext_requests_total",
+        "Total number of requests made by compute_ctl to download extensions from S3 proxy by status",
+        // Do not use any labels like extension name yet.
+        // We can add them later if needed.
+        &["http_status"]
+    )
+    .expect("failed to define a metric")
+});
+
+pub fn collect() -> Vec<MetricFamily> {
+    let mut metrics = INSTALLED_EXTENSIONS.collect();
+    metrics.extend(CPLANE_REQUESTS_TOTAL.collect());
+    metrics.extend(REMOTE_EXT_REQUESTS_TOTAL.collect());
+    metrics.extend(DB_MIGRATION_FAILED.collect());
+    metrics
+}

compute_tools/src/migration.rs

@@ -1,7 +1,9 @@
 use anyhow::{Context, Result};
 use fail::fail_point;
 use postgres::{Client, Transaction};
-use tracing::info;
+use tracing::{error, info};
+
+use crate::metrics::DB_MIGRATION_FAILED;
 
 /// Runs a series of migrations on a target database
 pub(crate) struct MigrationRunner<'m> {
@@ -78,24 +80,31 @@ impl<'m> MigrationRunner<'m> {
         Ok(())
     }
 
-    /// Run an individual migration
-    fn run_migration(txn: &mut Transaction, migration_id: i64, migration: &str) -> Result<()> {
+    /// Run an individual migration in a separate transaction block.
+    fn run_migration(client: &mut Client, migration_id: i64, migration: &str) -> Result<()> {
+        let mut txn = client
+            .transaction()
+            .with_context(|| format!("begin transaction for migration {migration_id}"))?;
+
         if migration.starts_with("-- SKIP") {
             info!("Skipping migration id={}", migration_id);
 
             // Even though we are skipping the migration, updating the
             // migration ID should help keep logic easy to understand when
             // trying to understand the state of a cluster.
-            Self::update_migration_id(txn, migration_id)?;
+            Self::update_migration_id(&mut txn, migration_id)?;
         } else {
             info!("Running migration id={}:\n{}\n", migration_id, migration);
 
             txn.simple_query(migration)
                 .with_context(|| format!("apply migration {migration_id}"))?;
 
-            Self::update_migration_id(txn, migration_id)?;
+            Self::update_migration_id(&mut txn, migration_id)?;
         }
 
+        txn.commit()
+            .with_context(|| format!("commit transaction for migration {migration_id}"))?;
+
         Ok(())
     }
 
@@ -109,19 +118,20 @@ impl<'m> MigrationRunner<'m> {
             // The index lags the migration ID by 1, so the current migration
             // ID is also the next index
             let migration_id = (current_migration + 1) as i64;
+            let migration = self.migrations[current_migration];
 
-            let mut txn = self
-                .client
-                .transaction()
-                .with_context(|| format!("begin transaction for migration {migration_id}"))?;
-
-            Self::run_migration(&mut txn, migration_id, self.migrations[current_migration])
-                .with_context(|| format!("running migration {migration_id}"))?;
-
-            txn.commit()
-                .with_context(|| format!("commit transaction for migration {migration_id}"))?;
-
-            info!("Finished migration id={}", migration_id);
+            match Self::run_migration(self.client, migration_id, migration) {
+                Ok(_) => {
+                    info!("Finished migration id={}", migration_id);
+                }
+                Err(e) => {
+                    error!("Failed to run migration id={}: {:?}", migration_id, e);
+                    DB_MIGRATION_FAILED
+                        .with_label_values(&[migration_id.to_string().as_str()])
+                        .inc();
+                    return Err(e);
+                }
+            }
 
             current_migration += 1;
         }

compute_tools/src/spec.rs

@@ -6,6 +6,7 @@ use std::path::Path;
 use tracing::{error, info, instrument, warn};
 
 use crate::config;
+use crate::metrics::{CPlaneRequestRPC, CPLANE_REQUESTS_TOTAL, UNKNOWN_HTTP_STATUS};
 use crate::migration::MigrationRunner;
 use crate::params::PG_HBA_ALL_MD5;
 use crate::pg_helpers::*;
@@ -19,7 +20,7 @@ use compute_api::spec::ComputeSpec;
 fn do_control_plane_request(
     uri: &str,
     jwt: &str,
-) -> Result<ControlPlaneSpecResponse, (bool, String)> {
+) -> Result<ControlPlaneSpecResponse, (bool, String, String)> {
     let resp = reqwest::blocking::Client::new()
         .get(uri)
         .header("Authorization", format!("Bearer {}", jwt))
@@ -27,35 +28,42 @@ fn do_control_plane_request(
         .map_err(|e| {
             (
                 true,
-                format!("could not perform spec request to control plane: {}", e),
+                format!("could not perform spec request to control plane: {:?}", e),
+                UNKNOWN_HTTP_STATUS.to_string(),
             )
         })?;
 
-    match resp.status() {
+    let status = resp.status();
+    match status {
         StatusCode::OK => match resp.json::<ControlPlaneSpecResponse>() {
             Ok(spec_resp) => Ok(spec_resp),
             Err(e) => Err((
                 true,
-                format!("could not deserialize control plane response: {}", e),
+                format!("could not deserialize control plane response: {:?}", e),
+                status.to_string(),
             )),
         },
-        StatusCode::SERVICE_UNAVAILABLE => {
-            Err((true, "control plane is temporarily unavailable".to_string()))
-        }
+        StatusCode::SERVICE_UNAVAILABLE => Err((
+            true,
+            "control plane is temporarily unavailable".to_string(),
+            status.to_string(),
+        )),
         StatusCode::BAD_GATEWAY => {
             // We have a problem with intermittent 502 errors now
             // https://github.com/neondatabase/cloud/issues/2353
             // It's fine to retry GET request in this case.
-            Err((true, "control plane request failed with 502".to_string()))
+            Err((
+                true,
+                "control plane request failed with 502".to_string(),
+                status.to_string(),
+            ))
         }
         // Another code, likely 500 or 404, means that compute is unknown to the control plane
         // or some internal failure happened. Doesn't make much sense to retry in this case.
         _ => Err((
             false,
-            format!(
-                "unexpected control plane response status code: {}",
-                resp.status()
-            ),
+            format!("unexpected control plane response status code: {}", status),
+            status.to_string(),
         )),
     }
 }
@@ -83,17 +91,28 @@ pub fn get_spec_from_control_plane(
     // - got spec -> return Ok(Some(spec))
     while attempt < 4 {
         spec = match do_control_plane_request(&cp_uri, &jwt) {
-            Ok(spec_resp) => match spec_resp.status {
-                ControlPlaneComputeStatus::Empty => Ok(None),
-                ControlPlaneComputeStatus::Attached => {
-                    if let Some(spec) = spec_resp.spec {
-                        Ok(Some(spec))
-                    } else {
-                        bail!("compute is attached, but spec is empty")
+            Ok(spec_resp) => {
+                CPLANE_REQUESTS_TOTAL
+                    .with_label_values(&[
+                        CPlaneRequestRPC::GetSpec.as_str(),
+                        &StatusCode::OK.to_string(),
+                    ])
+                    .inc();
+                match spec_resp.status {
+                    ControlPlaneComputeStatus::Empty => Ok(None),
+                    ControlPlaneComputeStatus::Attached => {
+                        if let Some(spec) = spec_resp.spec {
+                            Ok(Some(spec))
+                        } else {
+                            bail!("compute is attached, but spec is empty")
+                        }
                     }
                 }
-            },
-            Err((retry, msg)) => {
+            }
+            Err((retry, msg, status)) => {
+                CPLANE_REQUESTS_TOTAL
+                    .with_label_values(&[CPlaneRequestRPC::GetSpec.as_str(), &status])
+                    .inc();
                 if retry {
                     Err(anyhow!(msg))
                 } else {

control_plane/src/pageserver.rs

@@ -347,6 +347,11 @@ impl PageServerNode {
                 .map(|x| x.parse::<usize>())
                 .transpose()
                 .context("Failed to parse 'compaction_threshold' as an integer")?,
+            compaction_upper_limit: settings
+                .remove("compaction_upper_limit")
+                .map(|x| x.parse::<usize>())
+                .transpose()
+                .context("Failed to parse 'compaction_upper_limit' as an integer")?,
             compaction_algorithm: settings
                 .remove("compaction_algorithm")
                 .map(serde_json::from_str)
@@ -383,6 +388,11 @@ impl PageServerNode {
                 .map(|x| x.parse::<u8>())
                 .transpose()
                 .context("Failed to parse 'image_creation_check_threshold' as integer")?,
+            image_creation_preempt_threshold: settings
+                .remove("image_creation_preempt_threshold")
+                .map(|x| x.parse::<usize>())
+                .transpose()
+                .context("Failed to parse 'image_creation_preempt_threshold' as integer")?,
             pitr_interval: settings.remove("pitr_interval").map(|x| x.to_string()),
             walreceiver_connect_timeout: settings
                 .remove("walreceiver_connect_timeout")

control_plane/storcon_cli/src/main.rs

@@ -10,8 +10,8 @@ use pageserver_api::{
     controller_api::{
         AvailabilityZone, NodeAvailabilityWrapper, NodeDescribeResponse, NodeShardResponse,
         SafekeeperDescribeResponse, SafekeeperSchedulingPolicyRequest, ShardSchedulingPolicy,
-        ShardsPreferredAzsRequest, SkSchedulingPolicy, TenantCreateRequest, TenantDescribeResponse,
-        TenantPolicyRequest,
+        ShardsPreferredAzsRequest, ShardsPreferredAzsResponse, SkSchedulingPolicy,
+        TenantCreateRequest, TenantDescribeResponse, TenantPolicyRequest,
     },
     models::{
         EvictionPolicy, EvictionPolicyLayerAccessThreshold, LocationConfigSecondary,
@@ -800,7 +800,7 @@ async fn main() -> anyhow::Result<()> {
                     .collect(),
             };
             storcon_client
-                .dispatch::<ShardsPreferredAzsRequest, ()>(
+                .dispatch::<ShardsPreferredAzsRequest, ShardsPreferredAzsResponse>(
                     Method::PUT,
                     "control/v1/preferred_azs".to_string(),
                     Some(req),

deny.toml

@@ -32,6 +32,7 @@ reason = "the marvin attack only affects private key decryption, not public key
 # https://embarkstudios.github.io/cargo-deny/checks/licenses/cfg.html
 [licenses]
 allow = [
+    "0BSD",
     "Apache-2.0",
     "BSD-2-Clause",
     "BSD-3-Clause",

docker-compose/compute_wrapper/Dockerfile

@@ -7,11 +7,12 @@ FROM $REPOSITORY/${COMPUTE_IMAGE}:$TAG
 ARG COMPUTE_IMAGE
 
 USER root
-RUN apt-get update &&       \
+RUN echo 'Acquire::Retries "5";' > /etc/apt/apt.conf.d/80-retries && \
+    apt-get update &&       \
     apt-get install -y curl \
                        jq   \
                        netcat-openbsd
 #This is required for the pg_hintplan test
-RUN mkdir -p /ext-src/pg_hint_plan-src && chown postgres /ext-src/pg_hint_plan-src 
+RUN mkdir -p /ext-src/pg_hint_plan-src /postgres/contrib/file_fdw && chown postgres /ext-src/pg_hint_plan-src /postgres/contrib/file_fdw
 
-USER postgres
+USER postgres

docker-compose/compute_wrapper/shell/compute.sh

@@ -20,30 +20,55 @@ while ! nc -z pageserver 6400; do
 done
 echo "Page server is ready."
 
-echo "Create a tenant and timeline"
-generate_id tenant_id
-PARAMS=(
-     -X PUT
-     -H "Content-Type: application/json"
-     -d "{\"mode\": \"AttachedSingle\", \"generation\": 1, \"tenant_conf\": {}}"
-     "http://pageserver:9898/v1/tenant/${tenant_id}/location_config"
-)
-result=$(curl "${PARAMS[@]}")
-echo $result | jq .
+cp ${SPEC_FILE_ORG} ${SPEC_FILE}
 
-generate_id timeline_id
-PARAMS=(
-     -sbf
-     -X POST
-     -H "Content-Type: application/json"
-     -d "{\"new_timeline_id\": \"${timeline_id}\", \"pg_version\": ${PG_VERSION}}"
-     "http://pageserver:9898/v1/tenant/${tenant_id}/timeline/"
-)
-result=$(curl "${PARAMS[@]}")
-echo $result | jq .
+ if [ -n "${TENANT_ID:-}" ] && [ -n "${TIMELINE_ID:-}" ]; then
+   tenant_id=${TENANT_ID}
+   timeline_id=${TIMELINE_ID}
+else
+  echo "Check if a tenant present"
+  PARAMS=(
+       -X GET
+       -H "Content-Type: application/json"
+       "http://pageserver:9898/v1/tenant"
+  )
+  tenant_id=$(curl "${PARAMS[@]}" | jq -r .[0].id)
+  if [ -z "${tenant_id}" ] || [ "${tenant_id}" = null ]; then
+    echo "Create a tenant"
+    generate_id tenant_id
+    PARAMS=(
+         -X PUT
+         -H "Content-Type: application/json"
+         -d "{\"mode\": \"AttachedSingle\", \"generation\": 1, \"tenant_conf\": {}}"
+        "http://pageserver:9898/v1/tenant/${tenant_id}/location_config"
+    )
+    result=$(curl "${PARAMS[@]}")
+    echo $result | jq .
+  fi
+
+  echo "Check if a timeline present"
+  PARAMS=(
+       -X GET
+       -H "Content-Type: application/json"
+       "http://pageserver:9898/v1/tenant/${tenant_id}/timeline"
+  )
+  timeline_id=$(curl "${PARAMS[@]}" | jq -r .[0].timeline_id)
+  if [ -z "${timeline_id}" ] || [ "${timeline_id}" = null ]; then
+    generate_id timeline_id
+    PARAMS=(
+        -sbf
+        -X POST
+        -H "Content-Type: application/json"
+        -d "{\"new_timeline_id\": \"${timeline_id}\", \"pg_version\": ${PG_VERSION}}"
+        "http://pageserver:9898/v1/tenant/${tenant_id}/timeline/"
+    )
+    result=$(curl "${PARAMS[@]}")
+    echo $result | jq .
+  fi
+fi
 
 echo "Overwrite tenant id and timeline id in spec file"
-sed "s/TENANT_ID/${tenant_id}/" ${SPEC_FILE_ORG} > ${SPEC_FILE}
+sed -i "s/TENANT_ID/${tenant_id}/" ${SPEC_FILE}
 sed -i "s/TIMELINE_ID/${timeline_id}/" ${SPEC_FILE}
 
 cat ${SPEC_FILE}

docker-compose/docker-compose.yml

@@ -149,11 +149,13 @@ services:
       args:
         - REPOSITORY=${REPOSITORY:-neondatabase}
         - COMPUTE_IMAGE=compute-node-v${PG_VERSION:-16}
-        - TAG=${TAG:-latest}
+        - TAG=${COMPUTE_TAG:-${TAG:-latest}}
         - http_proxy=${http_proxy:-}
         - https_proxy=${https_proxy:-}
     environment:
       - PG_VERSION=${PG_VERSION:-16}
+      - TENANT_ID=${TENANT_ID:-}
+      - TIMELINE_ID=${TIMELINE_ID:-}
       #- RUST_BACKTRACE=1
     # Mount the test files directly, for faster editing cycle.
     volumes:

docker-compose/docker_compose_test.sh

@@ -31,7 +31,7 @@ for pg_version in ${TEST_VERSION_ONLY-14 15 16 17}; do
     echo "clean up containers if exists"
     cleanup
     PG_TEST_VERSION=$((pg_version < 16 ? 16 : pg_version))
-    PG_VERSION=$pg_version PG_TEST_VERSION=$PG_TEST_VERSION docker compose --profile test-extensions -f $COMPOSE_FILE up --build -d
+    PG_VERSION=$pg_version PG_TEST_VERSION=$PG_TEST_VERSION docker compose --profile test-extensions -f $COMPOSE_FILE up --quiet-pull --build -d
 
     echo "wait until the compute is ready. timeout after 60s. "
     cnt=0
@@ -51,6 +51,8 @@ for pg_version in ${TEST_VERSION_ONLY-14 15 16 17}; do
     done
 
     if [ $pg_version -ge 16 ]; then
+        docker cp ext-src $TEST_CONTAINER_NAME:/
+        docker exec $TEST_CONTAINER_NAME bash -c "apt update && apt install -y libtap-parser-sourcehandler-pgtap-perl"
         # This is required for the pg_hint_plan test, to prevent flaky log message causing the test to fail
         # It cannot be moved to Dockerfile now because the database directory is created after the start of the container
         echo Adding dummy config
@@ -60,17 +62,32 @@ for pg_version in ${TEST_VERSION_ONLY-14 15 16 17}; do
         docker cp $TEST_CONTAINER_NAME:/ext-src/pg_hint_plan-src/data $TMPDIR/data
         docker cp $TMPDIR/data $COMPUTE_CONTAINER_NAME:/ext-src/pg_hint_plan-src/
         rm -rf $TMPDIR
+        # The following block does the same for the contrib/file_fdw test
+        TMPDIR=$(mktemp -d)
+        docker cp $TEST_CONTAINER_NAME:/postgres/contrib/file_fdw/data $TMPDIR/data
+        docker cp $TMPDIR/data $COMPUTE_CONTAINER_NAME:/postgres/contrib/file_fdw/data
+        rm -rf $TMPDIR
+        # Apply patches
+        cat ../compute/patches/contrib_pg${pg_version}.patch | docker exec -i $TEST_CONTAINER_NAME bash -c "(cd /postgres && patch -p1)"
         # We are running tests now
-        if ! docker exec -e SKIP=timescaledb-src,rdkit-src,postgis-src,pgx_ulid-src,pgtap-src,pg_tiktoken-src,pg_jsonschema-src,pg_graphql-src,kq_imcx-src,wal2json_2_5-src \
-            $TEST_CONTAINER_NAME /run-tests.sh | tee testout.txt
-        then
-            FAILED=$(tail -1 testout.txt)
-            for d in $FAILED
-            do
-                mkdir $d
-                docker cp $TEST_CONTAINER_NAME:/ext-src/$d/regression.diffs $d || true
-                docker cp $TEST_CONTAINER_NAME:/ext-src/$d/regression.out $d || true
-                cat $d/regression.out $d/regression.diffs || true
+        rm -f testout.txt testout_contrib.txt
+        docker exec -e USE_PGXS=1 -e SKIP=timescaledb-src,rdkit-src,postgis-src,pgx_ulid-src,pgtap-src,pg_tiktoken-src,pg_jsonschema-src,kq_imcx-src,wal2json_2_5-src \
+        $TEST_CONTAINER_NAME /run-tests.sh /ext-src | tee testout.txt && EXT_SUCCESS=1 || EXT_SUCCESS=0
+        docker exec -e SKIP=start-scripts,postgres_fdw,ltree_plpython,jsonb_plpython,jsonb_plperl,hstore_plpython,hstore_plperl,dblink,bool_plperl \
+        $TEST_CONTAINER_NAME /run-tests.sh /postgres/contrib | tee testout_contrib.txt && CONTRIB_SUCCESS=1 || CONTRIB_SUCCESS=0
+        if [ $EXT_SUCCESS -eq 0 ] || [ $CONTRIB_SUCCESS -eq 0 ]; then
+            CONTRIB_FAILED=
+            FAILED=
+            [ $EXT_SUCCESS -eq 0 ] && FAILED=$(tail -1 testout.txt | awk '{for(i=1;i<=NF;i++){print "/ext-src/"$i;}}')
+            [ $CONTRIB_SUCCESS -eq 0 ] && CONTRIB_FAILED=$(tail -1 testout_contrib.txt | awk '{for(i=0;i<=NF;i++){print "/postgres/contrib/"$i;}}')
+            for d in $FAILED $CONTRIB_FAILED; do
+                dn="$(basename $d)"
+                rm -rf $dn
+                mkdir $dn
+                docker cp $TEST_CONTAINER_NAME:$d/regression.diffs $dn || [ $? -eq 1 ]
+                docker cp $TEST_CONTAINER_NAME:$d/regression.out $dn || [ $? -eq 1 ]
+                cat $dn/regression.out $dn/regression.diffs || true
+                rm -rf $dn
             done
         rm -rf $FAILED
         exit 1

docker-compose/ext-src/hll-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --dbname=contrib_regression add_agg agg_oob auto_sparse card_op cast_shape copy_binary cumulative_add_cardinality_correction cumulative_add_comprehensive_promotion cumulative_add_sparse_edge cumulative_add_sparse_random cumulative_add_sparse_step cumulative_union_comprehensive cumulative_union_explicit_explicit cumulative_union_explicit_promotion cumulative_union_probabilistic_probabilistic cumulative_union_sparse_full_representation cumulative_union_sparse_promotion cumulative_union_sparse_sparse disable_hashagg equal explicit_thresh hash hash_any meta_func murmur_bigint murmur_bytea nosparse notequal scalar_oob storedproc transaction typmod typmod_insert union_op

docker-compose/ext-src/hypopg-src/test-upgrade.patch

@@ -0,0 +1,27 @@
+diff --git a/expected/hypopg.out b/expected/hypopg.out
+index 90121d0..859260b 100644
+--- a/expected/hypopg.out
++++ b/expected/hypopg.out
+@@ -11,7 +11,8 @@ BEGIN
+ END;
+ $_$
+ LANGUAGE plpgsql;
+-CREATE EXTENSION hypopg;
++CREATE EXTENSION IF NOT EXISTS hypopg;
++NOTICE:  extension "hypopg" already exists, skipping
+ CREATE TABLE hypo (id integer, val text, "Id2" bigint);
+ INSERT INTO hypo SELECT i, 'line ' || i
+ FROM generate_series(1,100000) f(i);
+diff --git a/test/sql/hypopg.sql b/test/sql/hypopg.sql
+index 99722b0..8d6bacb 100644
+--- a/test/sql/hypopg.sql
++++ b/test/sql/hypopg.sql
+@@ -12,7 +12,7 @@ END;
+ $_$
+ LANGUAGE plpgsql;
+
+-CREATE EXTENSION hypopg;
++CREATE EXTENSION IF NOT EXISTS hypopg;
+
+ CREATE TABLE hypo (id integer, val text, "Id2" bigint);
+

docker-compose/ext-src/hypopg-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin' --use-existing --inputdir=test --dbname=contrib_regression hypopg hypo_brin hypo_index_part hypo_include hypo_hash hypo_hide_index

docker-compose/ext-src/ip4r-src/test-upgrade.patch

@@ -0,0 +1,23 @@
+diff --git a/expected/ip4r.out b/expected/ip4r.out
+index 7527af3..b38ed29 100644
+--- a/expected/ip4r.out
++++ b/expected/ip4r.out
+@@ -1,6 +1,5 @@
+ --
+ /*CUT-HERE*/
+-CREATE EXTENSION ip4r;
+ -- Check whether any of our opclasses fail amvalidate
+ DO $d$
+   DECLARE
+diff --git a/sql/ip4r.sql b/sql/ip4r.sql
+index 65c49ec..24ade09 100644
+--- a/sql/ip4r.sql
++++ b/sql/ip4r.sql
+@@ -1,7 +1,6 @@
+ --
+
+ /*CUT-HERE*/
+-CREATE EXTENSION ip4r;
+
+ -- Check whether any of our opclasses fail amvalidate
+

docker-compose/ext-src/ip4r-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --dbname=contrib_regression ip4r ip4r-softerr ip4r-v11

docker-compose/ext-src/pg_cron-src/test-upgrade.patch

@@ -0,0 +1,75 @@
+diff --git a/expected/pg_cron-test.out b/expected/pg_cron-test.out
+index d79d542..1663886 100644
+--- a/expected/pg_cron-test.out
++++ b/expected/pg_cron-test.out
+@@ -1,30 +1,3 @@
+-CREATE EXTENSION pg_cron VERSION '1.0';
+-SELECT extversion FROM pg_extension WHERE extname='pg_cron';
+- extversion 
+-------------
+- 1.0
+-(1 row)
+-
+--- Test binary compatibility with v1.4 function signature.
+-ALTER EXTENSION pg_cron UPDATE TO '1.4';
+-SELECT cron.unschedule(job_name := 'no_such_job');
+-ERROR:  could not find valid entry for job 'no_such_job'
+-SELECT cron.schedule('testjob', '* * * * *', 'SELECT 1');
+- schedule 
+-----------
+-        1
+-(1 row)
+-
+-SELECT cron.unschedule('testjob');
+- unschedule 
+-------------
+- t
+-(1 row)
+-
+--- Test cache invalidation
+-DROP EXTENSION pg_cron;
+-CREATE EXTENSION pg_cron VERSION '1.4';
+-ALTER EXTENSION pg_cron UPDATE;
+ -- Vacuum every day at 10:00am (GMT)
+ SELECT cron.schedule('0 10 * * *', 'VACUUM');
+  schedule 
+@@ -300,8 +273,3 @@ SELECT jobid, jobname, schedule, command FROM cron.job ORDER BY jobid;
+ SELECT cron.schedule('bad-last-dom-job1', '0 11 $foo * *', 'VACUUM FULL');
+ ERROR:  invalid schedule: 0 11 $foo * *
+ HINT:  Use cron format (e.g. 5 4 * * *), or interval format '[1-59] seconds'
+--- cleaning
+-DROP EXTENSION pg_cron;
+-drop user pgcron_cront;
+-drop database pgcron_dbno;
+-drop database pgcron_dbyes;
+diff --git a/sql/pg_cron-test.sql b/sql/pg_cron-test.sql
+index 45f94d9..241cf73 100644
+--- a/sql/pg_cron-test.sql
++++ b/sql/pg_cron-test.sql
+@@ -1,17 +1,3 @@
+-CREATE EXTENSION pg_cron VERSION '1.0';
+-SELECT extversion FROM pg_extension WHERE extname='pg_cron';
+--- Test binary compatibility with v1.4 function signature.
+-ALTER EXTENSION pg_cron UPDATE TO '1.4';
+-SELECT cron.unschedule(job_name := 'no_such_job');
+-SELECT cron.schedule('testjob', '* * * * *', 'SELECT 1');
+-SELECT cron.unschedule('testjob');
+-
+--- Test cache invalidation
+-DROP EXTENSION pg_cron;
+-CREATE EXTENSION pg_cron VERSION '1.4';
+-
+-ALTER EXTENSION pg_cron UPDATE;
+-
+ -- Vacuum every day at 10:00am (GMT)
+ SELECT cron.schedule('0 10 * * *', 'VACUUM');
+ 
+@@ -156,8 +142,3 @@ SELECT jobid, jobname, schedule, command FROM cron.job ORDER BY jobid;
+ -- invalid last of day job
+ SELECT cron.schedule('bad-last-dom-job1', '0 11 $foo * *', 'VACUUM FULL');
+ 
+--- cleaning
+-DROP EXTENSION pg_cron;
+-drop user pgcron_cront;
+-drop database pgcron_dbno;
+-drop database pgcron_dbyes;

docker-compose/ext-src/pg_cron-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --dbname=contrib_regression pg_cron-test

docker-compose/ext-src/pg_graphql-src/neon-test.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+set -ex
+cd "$(dirname "${0}")"
+dropdb --if-exists contrib_regression
+createdb contrib_regression
+PGXS="$(dirname "$(pg_config --pgxs)" )"
+REGRESS="${PGXS}/../test/regress/pg_regress"
+TESTDIR="test"
+TESTS=$(ls "${TESTDIR}/sql" | sort )
+TESTS=${TESTS//\.sql/}
+psql -v ON_ERROR_STOP=1 -f test/fixtures.sql -d contrib_regression
+${REGRESS} --use-existing --dbname=contrib_regression --inputdir=${TESTDIR} ${TESTS}
+

docker-compose/ext-src/pg_ivm-src/test-upgrade.patch

@@ -0,0 +1,18 @@
+diff --git a/expected/pg_ivm.out b/expected/pg_ivm.out
+index e8798ee..cca58d0 100644
+--- a/expected/pg_ivm.out
++++ b/expected/pg_ivm.out
+@@ -1,4 +1,3 @@
+-CREATE EXTENSION pg_ivm;
+ GRANT ALL ON SCHEMA public TO public;
+ -- create a table to use as a basis for views and materialized views in various combinations
+ CREATE TABLE mv_base_a (i int, j int);
+diff --git a/sql/pg_ivm.sql b/sql/pg_ivm.sql
+index d3c1a01..9382d7f 100644
+--- a/sql/pg_ivm.sql
++++ b/sql/pg_ivm.sql
+@@ -1,4 +1,3 @@
+-CREATE EXTENSION pg_ivm;
+ GRANT ALL ON SCHEMA public TO public;
+ 
+ -- create a table to use as a basis for views and materialized views in various combinations

docker-compose/ext-src/pg_ivm-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin' --dbname=contrib_regression pg_ivm create_immv refresh_immv

docker-compose/ext-src/pg_roaringbitmap-src/test-upgrade.patch

@@ -0,0 +1,25 @@
+diff --git a/expected/roaringbitmap.out b/expected/roaringbitmap.out
+index de70531..a5f7c15 100644
+--- a/expected/roaringbitmap.out
++++ b/expected/roaringbitmap.out
+@@ -1,7 +1,6 @@
+ --
+ --  Test roaringbitmap extension
+ --
+-CREATE EXTENSION if not exists roaringbitmap;
+ -- Test input and output
+ set roaringbitmap.output_format='array';
+ set extra_float_digits = 0;
+diff --git a/sql/roaringbitmap.sql b/sql/roaringbitmap.sql
+index a0e9c74..84bc966 100644
+--- a/sql/roaringbitmap.sql
++++ b/sql/roaringbitmap.sql
+@@ -2,8 +2,6 @@
+ --  Test roaringbitmap extension
+ --
+ 
+-CREATE EXTENSION if not exists roaringbitmap;
+-
+ -- Test input and output
+ 
+ set roaringbitmap.output_format='array';

docker-compose/ext-src/pg_roaringbitmap-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --dbname=contrib_regression roaringbitmap

docker-compose/ext-src/pg_semver-src/test-upgrade.patch

@@ -0,0 +1,24 @@
+diff --git a/test/sql/base.sql b/test/sql/base.sql
+index af599d8..2eed91b 100644
+--- a/test/sql/base.sql
++++ b/test/sql/base.sql
+@@ -2,7 +2,6 @@
+ BEGIN;
+ 
+ \i test/pgtap-core.sql
+-\i sql/semver.sql
+ 
+ SELECT plan(334);
+ --SELECT * FROM no_plan();
+diff --git a/test/sql/corpus.sql b/test/sql/corpus.sql
+index 1f5f637..a519905 100644
+--- a/test/sql/corpus.sql
++++ b/test/sql/corpus.sql
+@@ -4,7 +4,6 @@ BEGIN;
+ -- Test the SemVer corpus from https://regex101.com/r/Ly7O1x/3/.
+ 
+ \i test/pgtap-core.sql
+-\i sql/semver.sql
+ 
+ SELECT plan(71);
+ --SELECT * FROM no_plan();

docker-compose/ext-src/pg_semver-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --inputdir=test --dbname=contrib_regression base corpus

docker-compose/ext-src/pg_uuidv7-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --inputdir=test --dbname=contrib_regression  002_uuid_generate_v7 003_uuid_v7_to_timestamptz 004_uuid_timestamptz_to_v7 005_uuid_v7_to_timestamp 006_uuid_timestamp_to_v7

docker-compose/ext-src/pgjwt-src/neon-test.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+set -ex
+cd "$(dirname "${0}")"
+pg_prove test.sql

docker-compose/ext-src/pgjwt-src/test-upgrade.patch

@@ -0,0 +1,15 @@
+diff --git a/test.sql b/test.sql
+index d7a0ca8..f15bc76 100644
+--- a/test.sql
++++ b/test.sql
+@@ -9,9 +9,7 @@
+ \set ON_ERROR_STOP true
+ \set QUIET 1
+ 
+-CREATE EXTENSION pgcrypto;
+-CREATE EXTENSION pgtap;
+-CREATE EXTENSION pgjwt;
++CREATE EXTENSION IF NOT EXISTS pgtap;
+ 
+ BEGIN;
+ SELECT plan(23);

docker-compose/ext-src/pgjwt-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+pg_prove test.sql

docker-compose/ext-src/pgvector-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin' --inputdir=test --use-existing --dbname=contrib_regression bit btree cast copy halfvec hnsw_bit hnsw_halfvec hnsw_sparsevec hnsw_vector ivfflat_bit ivfflat_halfvec ivfflat_vector sparsevec vector_type

docker-compose/ext-src/plv8-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin'  --use-existing --dbname=contrib_regression plv8 plv8-errors scalar_args inline json startup_pre startup varparam json_conv jsonb_conv window guc es6 arraybuffer composites currentresource startup_perms bytea find_function_perms memory_limits reset show array_spread regression dialect bigint procedure

docker-compose/ext-src/postgresql-unit-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin' --use-existing --dbname=contrib_regression extension tables unit binary unicode prefix units time temperature functions language_functions round derived compare aggregate iec custom crosstab convert

docker-compose/ext-src/prefix-src/test-upgrade.sh

@@ -0,0 +1,5 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --dbname=contrib_regression prefix falcon explain queries

docker-compose/ext-src/rum-src/test-upgrade.patch

@@ -0,0 +1,19 @@
+diff --git a/expected/rum.out b/expected/rum.out
+index 5966d19..8860b79 100644
+--- a/expected/rum.out
++++ b/expected/rum.out
+@@ -1,4 +1,3 @@
+-CREATE EXTENSION rum;
+ CREATE TABLE test_rum( t text, a tsvector );
+ CREATE TRIGGER tsvectorupdate
+ BEFORE UPDATE OR INSERT ON test_rum
+diff --git a/sql/rum.sql b/sql/rum.sql
+index 8414bb9..898e6ab 100644
+--- a/sql/rum.sql
++++ b/sql/rum.sql
+@@ -1,5 +1,3 @@
+-CREATE EXTENSION rum;
+-
+ CREATE TABLE test_rum( t text, a tsvector );
+
+ CREATE TRIGGER tsvectorupdate

docker-compose/ext-src/rum-src/test-upgrade.sh

@@ -0,0 +1,6 @@
+#!/bin/sh
+set -ex
+cd "$(dirname ${0})"
+patch -p1 <test-upgrade.patch
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin' --use-existing --dbname=contrib_regression rum rum_validate rum_hash ruminv timestamp orderby orderby_hash altorder altorder_hash limits int2 int4 int8 float4 float8 money oid time timetz date interval macaddr inet cidr text varchar char bytea bit varbit numeric rum_weight expr array

docker-compose/run-tests.sh

@@ -1,17 +1,22 @@
 #!/bin/bash
 set -x
 
-cd /ext-src || exit 2
+extdir=${1}
+
+cd "${extdir}" || exit 2
 FAILED=
-LIST=$( (echo -e "${SKIP//","/"\n"}"; ls -d -- *-src) | sort | uniq -u)
-for d in ${LIST}
-do
-       [ -d "${d}" ] || continue
-       if ! psql -w -c "select 1" >/dev/null; then
-          FAILED="${d} ${FAILED}"
-          break
-       fi
+LIST=$( (echo -e "${SKIP//","/"\n"}"; ls) | sort | uniq -u)
+for d in ${LIST}; do
+    [ -d "${d}" ] || continue
+    if ! psql -w -c "select 1" >/dev/null; then
+      FAILED="${d} ${FAILED}"
+      break
+    fi
+    if [ -f "${d}/neon-test.sh" ]; then
+       "${d}/neon-test.sh" || FAILED="${d} ${FAILED}"
+    else
        USE_PGXS=1 make -C "${d}" installcheck || FAILED="${d} ${FAILED}"
+    fi
 done
 [ -z "${FAILED}" ] && exit 0
 echo "${FAILED}"

docker-compose/test_extensions_upgrade.sh

@@ -0,0 +1,94 @@
+#!/bin/bash
+set -eux -o pipefail
+cd "$(dirname "${0}")"
+# Takes a variable name as argument. The result is stored in that variable.
+generate_id() {
+    local -n resvar=$1
+    printf -v resvar '%08x%08x%08x%08x' $SRANDOM $SRANDOM $SRANDOM $SRANDOM
+}
+if [ -z ${OLDTAG+x} ] || [ -z ${NEWTAG+x} ] || [ -z "${OLDTAG}" ] || [ -z "${NEWTAG}" ]; then
+  echo OLDTAG and NEWTAG must be defined
+  exit 1
+fi
+export PG_VERSION=${PG_VERSION:-16}
+function wait_for_ready {
+  TIME=0
+  while ! docker compose logs compute_is_ready | grep -q "accepting connections" && [ ${TIME} -le 300 ] ; do
+    ((TIME += 1 ))
+    sleep 1
+  done
+  if [ ${TIME} -gt 300 ]; then
+    echo Time is out.
+    exit 2
+  fi
+}
+function create_extensions() {
+  for ext in ${1}; do
+    docker compose exec neon-test-extensions psql -X -v ON_ERROR_STOP=1 -d contrib_regression -c "CREATE EXTENSION IF NOT EXISTS ${ext} CASCADE"
+  done
+}
+EXTENSIONS='[
+{"extname": "plv8", "extdir": "plv8-src"},
+{"extname": "vector", "extdir": "pgvector-src"},
+{"extname": "unit", "extdir": "postgresql-unit-src"},
+{"extname": "hypopg", "extdir": "hypopg-src"},
+{"extname": "rum", "extdir": "rum-src"},
+{"extname": "ip4r", "extdir": "ip4r-src"},
+{"extname": "prefix", "extdir": "prefix-src"},
+{"extname": "hll", "extdir": "hll-src"},
+{"extname": "pg_cron", "extdir": "pg_cron-src"},
+{"extname": "pg_uuidv7", "extdir": "pg_uuidv7-src"},
+{"extname": "roaringbitmap", "extdir": "pg_roaringbitmap-src"},
+{"extname": "semver", "extdir": "pg_semver-src"},
+{"extname": "pg_ivm", "extdir": "pg_ivm-src"},
+{"extname": "pgjwt", "extdir": "pgjwt-src"}
+]'
+EXTNAMES=$(echo ${EXTENSIONS} | jq -r '.[].extname' | paste -sd ' ' -)
+TAG=${NEWTAG} docker compose --profile test-extensions up --quiet-pull --build -d
+wait_for_ready
+docker compose exec neon-test-extensions psql -c "DROP DATABASE IF EXISTS contrib_regression"
+docker compose exec neon-test-extensions psql -c "CREATE DATABASE contrib_regression"
+create_extensions "${EXTNAMES}"
+query="select json_object_agg(extname,extversion) from pg_extension where extname in ('${EXTNAMES// /\',\'}')"
+new_vers=$(docker compose exec neon-test-extensions psql -Aqt -d contrib_regression -c "$query")
+docker compose --profile test-extensions down
+TAG=${OLDTAG} docker compose --profile test-extensions up --quiet-pull --build -d --force-recreate
+wait_for_ready
+docker compose cp  ext-src neon-test-extensions:/
+docker compose exec neon-test-extensions psql -c "DROP DATABASE IF EXISTS contrib_regression"
+docker compose exec neon-test-extensions psql -c "CREATE DATABASE contrib_regression"
+create_extensions "${EXTNAMES}"
+query="select pge.extname from pg_extension pge join (select key as extname, value as extversion from json_each_text('${new_vers}')) x on pge.extname=x.extname and pge.extversion <> x.extversion"
+exts=$(docker compose exec neon-test-extensions psql -Aqt -d contrib_regression -c "$query")
+if [ -z "${exts}" ]; then
+  echo "No extensions were upgraded"
+else
+  tenant_id=$(docker compose exec neon-test-extensions psql -Aqt -c "SHOW neon.tenant_id")
+  timeline_id=$(docker compose exec neon-test-extensions psql -Aqt -c "SHOW neon.timeline_id")
+  for ext in ${exts}; do
+    echo Testing ${ext}...
+    EXTDIR=$(echo ${EXTENSIONS} | jq -r '.[] | select(.extname=="'${ext}'") | .extdir')
+    generate_id new_timeline_id
+    PARAMS=(
+        -sbf
+        -X POST
+        -H "Content-Type: application/json"
+        -d "{\"new_timeline_id\": \"${new_timeline_id}\", \"pg_version\": ${PG_VERSION}, \"ancestor_timeline_id\": \"${timeline_id}\"}"
+        "http://127.0.0.1:9898/v1/tenant/${tenant_id}/timeline/"
+    )
+    result=$(curl "${PARAMS[@]}")
+    echo $result | jq .
+    TENANT_ID=${tenant_id} TIMELINE_ID=${new_timeline_id} TAG=${OLDTAG} docker compose down compute compute_is_ready
+    COMPUTE_TAG=${NEWTAG} TAG=${OLDTAG} TENANT_ID=${tenant_id} TIMELINE_ID=${new_timeline_id} docker compose up --quiet-pull -d --build compute compute_is_ready
+    wait_for_ready
+    TID=$(docker compose exec neon-test-extensions psql -Aqt -c "SHOW neon.timeline_id")
+    if [ ${TID} != ${new_timeline_id} ]; then
+      echo Timeline mismatch
+      exit 1
+    fi
+    docker compose exec neon-test-extensions psql -d contrib_regression -c "\dx ${ext}"
+    docker compose exec neon-test-extensions sh -c /ext-src/${EXTDIR}/test-upgrade.sh
+    docker compose exec neon-test-extensions psql -d contrib_regression -c "alter extension ${ext} update"
+    docker compose exec neon-test-extensions psql -d contrib_regression -c "\dx ${ext}"
+  done
+fi

libs/compute_api/src/responses.rs

@@ -15,11 +15,6 @@ pub struct GenericAPIError {
     pub error: String,
 }
 
-#[derive(Debug, Clone, Serialize)]
-pub struct InfoResponse {
-    pub num_cpus: usize,
-}
-
 #[derive(Debug, Clone, Serialize)]
 pub struct ExtensionInstallResponse {
     pub extension: PgIdent,

libs/compute_api/src/spec.rs

@@ -204,14 +204,16 @@ impl RemoteExtSpec {
 
         // Check if extension is present in public or custom.
         // If not, then it is not allowed to be used by this compute.
-        if let Some(public_extensions) = &self.public_extensions {
-            if !public_extensions.contains(&real_ext_name.to_string()) {
-                if let Some(custom_extensions) = &self.custom_extensions {
-                    if !custom_extensions.contains(&real_ext_name.to_string()) {
-                        return Err(anyhow::anyhow!("extension {} is not found", real_ext_name));
-                    }
-                }
-            }
+        if !self
+            .public_extensions
+            .as_ref()
+            .is_some_and(|exts| exts.iter().any(|e| e == ext_name))
+            && !self
+                .custom_extensions
+                .as_ref()
+                .is_some_and(|exts| exts.iter().any(|e| e == ext_name))
+        {
+            return Err(anyhow::anyhow!("extension {} is not found", real_ext_name));
         }
 
         match self.extension_data.get(real_ext_name) {
@@ -340,6 +342,96 @@ mod tests {
     use super::*;
     use std::fs::File;
 
+    #[test]
+    fn allow_installing_remote_extensions() {
+        let rspec: RemoteExtSpec = serde_json::from_value(serde_json::json!({
+            "public_extensions": null,
+            "custom_extensions": null,
+            "library_index": {},
+            "extension_data": {},
+        }))
+        .unwrap();
+
+        rspec
+            .get_ext("ext", false, "latest", "v17")
+            .expect_err("Extension should not be found");
+
+        let rspec: RemoteExtSpec = serde_json::from_value(serde_json::json!({
+            "public_extensions": [],
+            "custom_extensions": null,
+            "library_index": {},
+            "extension_data": {},
+        }))
+        .unwrap();
+
+        rspec
+            .get_ext("ext", false, "latest", "v17")
+            .expect_err("Extension should not be found");
+
+        let rspec: RemoteExtSpec = serde_json::from_value(serde_json::json!({
+            "public_extensions": [],
+            "custom_extensions": [],
+            "library_index": {
+                "ext": "ext"
+            },
+            "extension_data": {
+                "ext": {
+                    "control_data": {
+                        "ext.control": ""
+                    },
+                    "archive_path": ""
+                }
+            },
+        }))
+        .unwrap();
+
+        rspec
+            .get_ext("ext", false, "latest", "v17")
+            .expect_err("Extension should not be found");
+
+        let rspec: RemoteExtSpec = serde_json::from_value(serde_json::json!({
+            "public_extensions": [],
+            "custom_extensions": ["ext"],
+            "library_index": {
+                "ext": "ext"
+            },
+            "extension_data": {
+                "ext": {
+                    "control_data": {
+                        "ext.control": ""
+                    },
+                    "archive_path": ""
+                }
+            },
+        }))
+        .unwrap();
+
+        rspec
+            .get_ext("ext", false, "latest", "v17")
+            .expect("Extension should be found");
+
+        let rspec: RemoteExtSpec = serde_json::from_value(serde_json::json!({
+            "public_extensions": ["ext"],
+            "custom_extensions": [],
+            "library_index": {
+                "ext": "ext"
+            },
+            "extension_data": {
+                "ext": {
+                    "control_data": {
+                        "ext.control": ""
+                    },
+                    "archive_path": ""
+                }
+            },
+        }))
+        .unwrap();
+
+        rspec
+            .get_ext("ext", false, "latest", "v17")
+            .expect("Extension should be found");
+    }
+
     #[test]
     fn parse_spec_file() {
         let file = File::open("tests/cluster_spec.json").unwrap();

libs/pageserver_api/src/config.rs

@@ -256,6 +256,11 @@ pub struct TenantConfigToml {
     pub compaction_period: Duration,
     /// Level0 delta layer threshold for compaction.
     pub compaction_threshold: usize,
+    /// Controls the amount of L0 included in a single compaction iteration.
+    /// The unit is `checkpoint_distance`, i.e., a size.
+    /// We add L0s to the set of layers to compact until their cumulative
+    /// size exceeds `compaction_upper_limit * checkpoint_distance`.
+    pub compaction_upper_limit: usize,
     pub compaction_algorithm: crate::models::CompactionAlgorithmSettings,
     /// Level0 delta layer threshold at which to delay layer flushes for compaction backpressure,
     /// such that they take 2x as long, and start waiting for layer flushes during ephemeral layer
@@ -318,6 +323,10 @@ pub struct TenantConfigToml {
     // Expresed in multiples of checkpoint distance.
     pub image_layer_creation_check_threshold: u8,
 
+    // How many multiples of L0 `compaction_threshold` will preempt image layer creation and do L0 compaction.
+    // Set to 0 to disable preemption.
+    pub image_creation_preempt_threshold: usize,
+
     /// 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")]
@@ -523,6 +532,12 @@ pub mod tenant_conf_defaults {
 
     pub const DEFAULT_COMPACTION_PERIOD: &str = "20 s";
     pub const DEFAULT_COMPACTION_THRESHOLD: usize = 10;
+
+    // This value needs to be tuned to avoid OOM. We have 3/4 of the total CPU threads to do background works, that's 16*3/4=9 on
+    // most of our pageservers. Compaction ~50 layers requires about 2GB memory (could be reduced later by optimizing L0 hole
+    // calculation to avoid loading all keys into the memory). So with this config, we can get a maximum peak compaction usage of 18GB.
+    pub const DEFAULT_COMPACTION_UPPER_LIMIT: usize = 50;
+
     pub const DEFAULT_COMPACTION_ALGORITHM: crate::models::CompactionAlgorithm =
         crate::models::CompactionAlgorithm::Legacy;
 
@@ -536,6 +551,10 @@ pub mod tenant_conf_defaults {
     // Relevant: https://github.com/neondatabase/neon/issues/3394
     pub const DEFAULT_GC_PERIOD: &str = "1 hr";
     pub const DEFAULT_IMAGE_CREATION_THRESHOLD: usize = 3;
+    // If there are more than threshold * compaction_threshold (that is 3 * 10 in the default config) L0 layers, image
+    // layer creation will end immediately. Set to 0 to disable. The target default will be 3 once we
+    // want to enable this feature.
+    pub const DEFAULT_IMAGE_CREATION_PREEMPT_THRESHOLD: usize = 0;
     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";
@@ -563,6 +582,7 @@ impl Default for TenantConfigToml {
             compaction_period: humantime::parse_duration(DEFAULT_COMPACTION_PERIOD)
                 .expect("cannot parse default compaction period"),
             compaction_threshold: DEFAULT_COMPACTION_THRESHOLD,
+            compaction_upper_limit: DEFAULT_COMPACTION_UPPER_LIMIT,
             compaction_algorithm: crate::models::CompactionAlgorithmSettings {
                 kind: DEFAULT_COMPACTION_ALGORITHM,
             },
@@ -593,6 +613,7 @@ impl Default for TenantConfigToml {
             lazy_slru_download: false,
             timeline_get_throttle: crate::models::ThrottleConfig::disabled(),
             image_layer_creation_check_threshold: DEFAULT_IMAGE_LAYER_CREATION_CHECK_THRESHOLD,
+            image_creation_preempt_threshold: DEFAULT_IMAGE_CREATION_PREEMPT_THRESHOLD,
             lsn_lease_length: LsnLease::DEFAULT_LENGTH,
             lsn_lease_length_for_ts: LsnLease::DEFAULT_LENGTH_FOR_TS,
             timeline_offloading: false,

libs/pageserver_api/src/models.rs

@@ -458,6 +458,8 @@ pub struct TenantConfigPatch {
     pub compaction_period: FieldPatch<String>,
     #[serde(skip_serializing_if = "FieldPatch::is_noop")]
     pub compaction_threshold: FieldPatch<usize>,
+    #[serde(skip_serializing_if = "FieldPatch::is_noop")]
+    pub compaction_upper_limit: FieldPatch<usize>,
     // defer parsing compaction_algorithm, like eviction_policy
     #[serde(skip_serializing_if = "FieldPatch::is_noop")]
     pub compaction_algorithm: FieldPatch<CompactionAlgorithmSettings>,
@@ -496,6 +498,8 @@ pub struct TenantConfigPatch {
     #[serde(skip_serializing_if = "FieldPatch::is_noop")]
     pub image_layer_creation_check_threshold: FieldPatch<u8>,
     #[serde(skip_serializing_if = "FieldPatch::is_noop")]
+    pub image_creation_preempt_threshold: FieldPatch<usize>,
+    #[serde(skip_serializing_if = "FieldPatch::is_noop")]
     pub lsn_lease_length: FieldPatch<String>,
     #[serde(skip_serializing_if = "FieldPatch::is_noop")]
     pub lsn_lease_length_for_ts: FieldPatch<String>,
@@ -522,6 +526,7 @@ pub struct TenantConfig {
     pub compaction_target_size: Option<u64>,
     pub compaction_period: Option<String>,
     pub compaction_threshold: Option<usize>,
+    pub compaction_upper_limit: Option<usize>,
     // defer parsing compaction_algorithm, like eviction_policy
     pub compaction_algorithm: Option<CompactionAlgorithmSettings>,
     pub l0_flush_delay_threshold: Option<usize>,
@@ -541,6 +546,7 @@ pub struct TenantConfig {
     pub lazy_slru_download: Option<bool>,
     pub timeline_get_throttle: Option<ThrottleConfig>,
     pub image_layer_creation_check_threshold: Option<u8>,
+    pub image_creation_preempt_threshold: Option<usize>,
     pub lsn_lease_length: Option<String>,
     pub lsn_lease_length_for_ts: Option<String>,
     pub timeline_offloading: Option<bool>,
@@ -559,6 +565,7 @@ impl TenantConfig {
             mut compaction_target_size,
             mut compaction_period,
             mut compaction_threshold,
+            mut compaction_upper_limit,
             mut compaction_algorithm,
             mut l0_flush_delay_threshold,
             mut l0_flush_stall_threshold,
@@ -577,6 +584,7 @@ impl TenantConfig {
             mut lazy_slru_download,
             mut timeline_get_throttle,
             mut image_layer_creation_check_threshold,
+            mut image_creation_preempt_threshold,
             mut lsn_lease_length,
             mut lsn_lease_length_for_ts,
             mut timeline_offloading,
@@ -594,6 +602,9 @@ impl TenantConfig {
             .apply(&mut compaction_target_size);
         patch.compaction_period.apply(&mut compaction_period);
         patch.compaction_threshold.apply(&mut compaction_threshold);
+        patch
+            .compaction_upper_limit
+            .apply(&mut compaction_upper_limit);
         patch.compaction_algorithm.apply(&mut compaction_algorithm);
         patch
             .l0_flush_delay_threshold
@@ -628,6 +639,9 @@ impl TenantConfig {
         patch
             .image_layer_creation_check_threshold
             .apply(&mut image_layer_creation_check_threshold);
+        patch
+            .image_creation_preempt_threshold
+            .apply(&mut image_creation_preempt_threshold);
         patch.lsn_lease_length.apply(&mut lsn_lease_length);
         patch
             .lsn_lease_length_for_ts
@@ -653,6 +667,7 @@ impl TenantConfig {
             compaction_target_size,
             compaction_period,
             compaction_threshold,
+            compaction_upper_limit,
             compaction_algorithm,
             l0_flush_delay_threshold,
             l0_flush_stall_threshold,
@@ -671,6 +686,7 @@ impl TenantConfig {
             lazy_slru_download,
             timeline_get_throttle,
             image_layer_creation_check_threshold,
+            image_creation_preempt_threshold,
             lsn_lease_length,
             lsn_lease_length_for_ts,
             timeline_offloading,

libs/pq_proto/src/lib.rs

@@ -182,6 +182,13 @@ pub struct CancelKeyData {
     pub cancel_key: i32,
 }
 
+pub fn id_to_cancel_key(id: u64) -> CancelKeyData {
+    CancelKeyData {
+        backend_pid: (id >> 32) as i32,
+        cancel_key: (id & 0xffffffff) as i32,
+    }
+}
+
 impl fmt::Display for CancelKeyData {
     fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         let hi = (self.backend_pid as u64) << 32;

libs/proxy/tokio-postgres2/Cargo.toml

@@ -19,3 +19,4 @@ postgres-protocol2 = { path = "../postgres-protocol2" }
 postgres-types2 = { path = "../postgres-types2" }
 tokio = { workspace = true, features = ["io-util", "time", "net"] }
 tokio-util = { workspace = true, features = ["codec"] }
+serde = { workspace = true, features = ["derive"] }

libs/proxy/tokio-postgres2/src/cancel_token.rs

@@ -3,12 +3,13 @@ use crate::tls::TlsConnect;
 
 use crate::{cancel_query, client::SocketConfig, tls::MakeTlsConnect};
 use crate::{cancel_query_raw, Error};
+use serde::{Deserialize, Serialize};
 use tokio::io::{AsyncRead, AsyncWrite};
 use tokio::net::TcpStream;
 
 /// The capability to request cancellation of in-progress queries on a
 /// connection.
-#[derive(Clone)]
+#[derive(Clone, Serialize, Deserialize)]
 pub struct CancelToken {
     pub socket_config: Option<SocketConfig>,
     pub ssl_mode: SslMode,

libs/proxy/tokio-postgres2/src/client.rs

@@ -18,6 +18,7 @@ use fallible_iterator::FallibleIterator;
 use futures_util::{future, ready, TryStreamExt};
 use parking_lot::Mutex;
 use postgres_protocol2::message::{backend::Message, frontend};
+use serde::{Deserialize, Serialize};
 use std::collections::HashMap;
 use std::fmt;
 use std::sync::Arc;
@@ -137,7 +138,7 @@ impl InnerClient {
     }
 }
 
-#[derive(Clone)]
+#[derive(Clone, Serialize, Deserialize)]
 pub struct SocketConfig {
     pub host: Host,
     pub port: u16,

libs/proxy/tokio-postgres2/src/config.rs

@@ -7,6 +7,7 @@ use crate::tls::MakeTlsConnect;
 use crate::tls::TlsConnect;
 use crate::{Client, Connection, Error};
 use postgres_protocol2::message::frontend::StartupMessageParams;
+use serde::{Deserialize, Serialize};
 use std::fmt;
 use std::str;
 use std::time::Duration;
@@ -16,7 +17,7 @@ pub use postgres_protocol2::authentication::sasl::ScramKeys;
 use tokio::net::TcpStream;
 
 /// TLS configuration.
-#[derive(Debug, Copy, Clone, PartialEq, Eq)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Serialize, Deserialize)]
 #[non_exhaustive]
 pub enum SslMode {
     /// Do not use TLS.
@@ -50,7 +51,7 @@ pub enum ReplicationMode {
 }
 
 /// A host specification.
-#[derive(Debug, Clone, PartialEq, Eq)]
+#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
 pub enum Host {
     /// A TCP hostname.
     Tcp(String),

libs/remote_storage/src/config.rs

@@ -7,7 +7,7 @@ use serde::{Deserialize, Serialize};
 
 use crate::{
     DEFAULT_MAX_KEYS_PER_LIST_RESPONSE, DEFAULT_REMOTE_STORAGE_AZURE_CONCURRENCY_LIMIT,
-    DEFAULT_REMOTE_STORAGE_S3_CONCURRENCY_LIMIT,
+    DEFAULT_REMOTE_STORAGE_LOCALFS_CONCURRENCY_LIMIT, DEFAULT_REMOTE_STORAGE_S3_CONCURRENCY_LIMIT,
 };
 
 /// External backup storage configuration, enough for creating a client for that storage.
@@ -45,11 +45,11 @@ impl RemoteStorageKind {
 
 impl RemoteStorageConfig {
     /// Helper to fetch the configured concurrency limit.
-    pub fn concurrency_limit(&self) -> Option<usize> {
+    pub fn concurrency_limit(&self) -> usize {
         match &self.storage {
-            RemoteStorageKind::LocalFs { .. } => None,
-            RemoteStorageKind::AwsS3(c) => Some(c.concurrency_limit.into()),
-            RemoteStorageKind::AzureContainer(c) => Some(c.concurrency_limit.into()),
+            RemoteStorageKind::LocalFs { .. } => DEFAULT_REMOTE_STORAGE_LOCALFS_CONCURRENCY_LIMIT,
+            RemoteStorageKind::AwsS3(c) => c.concurrency_limit.into(),
+            RemoteStorageKind::AzureContainer(c) => c.concurrency_limit.into(),
         }
     }
 }

libs/remote_storage/src/lib.rs

@@ -65,6 +65,12 @@ pub const DEFAULT_REMOTE_STORAGE_S3_CONCURRENCY_LIMIT: usize = 100;
 /// Here, a limit of max 20k concurrent connections was noted.
 /// <https://learn.microsoft.com/en-us/answers/questions/1301863/is-there-any-limitation-to-concurrent-connections>
 pub const DEFAULT_REMOTE_STORAGE_AZURE_CONCURRENCY_LIMIT: usize = 100;
+/// Set this limit analogously to the S3 limit.
+///
+/// The local filesystem backend doesn't enforce a concurrency limit itself, but this also bounds
+/// the upload queue concurrency. Some tests create thousands of uploads, which slows down the
+/// quadratic scheduling of the upload queue, and there is no point spawning so many Tokio tasks.
+pub const DEFAULT_REMOTE_STORAGE_LOCALFS_CONCURRENCY_LIMIT: usize = 100;
 /// No limits on the client side, which currenltly means 1000 for AWS S3.
 /// <https://docs.aws.amazon.com/AmazonS3/latest/API/API_ListObjectsV2.html#API_ListObjectsV2_RequestSyntax>
 pub const DEFAULT_MAX_KEYS_PER_LIST_RESPONSE: Option<i32> = None;

libs/utils/src/logging.rs

@@ -5,6 +5,24 @@ use metrics::{IntCounter, IntCounterVec};
 use once_cell::sync::Lazy;
 use strum_macros::{EnumString, VariantNames};
 
+/// Logs a critical error, similarly to `tracing::error!`. This will:
+///
+/// * Emit an ERROR log message with prefix "CRITICAL:" and a backtrace.
+/// * Increment libmetrics_tracing_event_count{level="critical"}, and indirectly level="error".
+/// * Trigger a pageable alert (via the metric above).
+/// * In debug builds, panic the process.
+#[macro_export]
+macro_rules! critical {
+    ($($arg:tt)*) => {
+        if cfg!(debug_assertions) {
+            panic!($($arg)*);
+        }
+        $crate::logging::TRACING_EVENT_COUNT_METRIC.inc_critical();
+        let backtrace = std::backtrace::Backtrace::capture();
+        tracing::error!("CRITICAL: {}\n{backtrace}", format!($($arg)*));
+    };
+}
+
 #[derive(EnumString, strum_macros::Display, VariantNames, Eq, PartialEq, Debug, Clone, Copy)]
 #[strum(serialize_all = "snake_case")]
 pub enum LogFormat {
@@ -25,7 +43,10 @@ impl LogFormat {
     }
 }
 
-struct TracingEventCountMetric {
+pub struct TracingEventCountMetric {
+    /// CRITICAL is not a `tracing` log level. Instead, we increment it in the `critical!` macro,
+    /// and also emit it as a regular error. These are thus double-counted, but that seems fine.
+    critical: IntCounter,
     error: IntCounter,
     warn: IntCounter,
     info: IntCounter,
@@ -33,7 +54,7 @@ struct TracingEventCountMetric {
     trace: IntCounter,
 }
 
-static TRACING_EVENT_COUNT_METRIC: Lazy<TracingEventCountMetric> = Lazy::new(|| {
+pub static TRACING_EVENT_COUNT_METRIC: Lazy<TracingEventCountMetric> = Lazy::new(|| {
     let vec = metrics::register_int_counter_vec!(
         "libmetrics_tracing_event_count",
         "Number of tracing events, by level",
@@ -46,6 +67,7 @@ static TRACING_EVENT_COUNT_METRIC: Lazy<TracingEventCountMetric> = Lazy::new(||
 impl TracingEventCountMetric {
     fn new(vec: IntCounterVec) -> Self {
         Self {
+            critical: vec.with_label_values(&["critical"]),
             error: vec.with_label_values(&["error"]),
             warn: vec.with_label_values(&["warn"]),
             info: vec.with_label_values(&["info"]),
@@ -54,6 +76,11 @@ impl TracingEventCountMetric {
         }
     }
 
+    // Allow public access from `critical!` macro.
+    pub fn inc_critical(&self) {
+        self.critical.inc();
+    }
+
     fn inc_for_level(&self, level: tracing::Level) {
         let counter = match level {
             tracing::Level::ERROR => &self.error,

libs/vm_monitor/src/filecache.rs

@@ -177,8 +177,8 @@ impl FileCacheState {
         crate::spawn_with_cancel(
             token,
             |res| {
-                if let Err(error) = res {
-                    error!(%error, "postgres error")
+                if let Err(e) = res {
+                    error!(error = format_args!("{e:#}"), "postgres error");
                 }
             },
             conn,
@@ -205,7 +205,7 @@ impl FileCacheState {
         {
             Ok(rows) => Ok(rows),
             Err(e) => {
-                error!(error = ?e, "postgres error: {e} -> retrying");
+                error!(error = format_args!("{e:#}"), "postgres error -> retrying");
 
                 let client = FileCacheState::connect(&self.conn_str, self.token.clone())
                     .await

libs/vm_monitor/src/lib.rs

@@ -191,15 +191,12 @@ async fn start_monitor(
     .await;
     let mut monitor = match monitor {
         Ok(Ok(monitor)) => monitor,
-        Ok(Err(error)) => {
-            error!(?error, "failed to create monitor");
+        Ok(Err(e)) => {
+            error!(error = format_args!("{e:#}"), "failed to create monitor");
             return;
         }
         Err(_) => {
-            error!(
-                ?timeout,
-                "creating monitor timed out (probably waiting to receive protocol range)"
-            );
+            error!(?timeout, "creating monitor timed out");
             return;
         }
     };
@@ -207,6 +204,9 @@ async fn start_monitor(
 
     match monitor.run().await {
         Ok(()) => info!("monitor was killed due to new connection"),
-        Err(e) => error!(error = ?e, "monitor terminated unexpectedly"),
+        Err(e) => error!(
+            error = format_args!("{e:#}"),
+            "monitor terminated unexpectedly"
+        ),
     }
 }

libs/vm_monitor/src/runner.rs

@@ -370,12 +370,16 @@ impl Runner {
                 }),
             InboundMsgKind::InvalidMessage { error } => {
                 warn!(
-                    %error, id, "received notification of an invalid message we sent"
+                    error = format_args!("{error:#}"),
+                    id, "received notification of an invalid message we sent"
                 );
                 Ok(None)
             }
             InboundMsgKind::InternalError { error } => {
-                warn!(error, id, "agent experienced an internal error");
+                warn!(
+                    error = format_args!("{error:#}"),
+                    id, "agent experienced an internal error"
+                );
                 Ok(None)
             }
             InboundMsgKind::HealthCheck {} => {
@@ -476,7 +480,7 @@ impl Runner {
                                         // gives the outermost cause, and the debug impl
                                         // pretty-prints the error, whereas {:#} contains all the
                                         // causes, but is compact (no newlines).
-                                        warn!(error = format!("{e:#}"), "error handling message");
+                                        warn!(error = format_args!("{e:#}"), "error handling message");
                                         OutboundMsg::new(
                                             OutboundMsgKind::InternalError {
                                                 error: e.to_string(),
@@ -492,7 +496,7 @@ impl Runner {
                                     .context("failed to send message")?;
                             }
                             Err(e) => warn!(
-                                error = format!("{e}"),
+                                error = format_args!("{e:#}"),
                                 msg = ?msg,
                                 "received error message"
                             ),

pageserver/Cargo.toml

@@ -36,7 +36,7 @@ itertools.workspace = true
 md5.workspace = true
 nix.workspace = true
 # hack to get the number of worker threads tokio uses
-num_cpus = { version = "1.15" }
+num_cpus.workspace = true
 num-traits.workspace = true
 once_cell.workspace = true
 pin-project-lite.workspace = true

pageserver/src/http/openapi_spec.yml

@@ -984,6 +984,8 @@ components:
           type: string
         compaction_threshold:
           type: string
+        compaction_upper_limit:
+          type: string
         image_creation_threshold:
           type: integer
         walreceiver_connect_timeout:

pageserver/src/http/routes.rs

@@ -1472,7 +1472,13 @@ async fn layer_download_handler(
     let downloaded = timeline
         .download_layer(&layer_name)
         .await
-        .map_err(ApiError::InternalServerError)?;
+        .map_err(|e| match e {
+            tenant::storage_layer::layer::DownloadError::TimelineShutdown
+            | tenant::storage_layer::layer::DownloadError::DownloadCancelled => {
+                ApiError::ShuttingDown
+            }
+            other => ApiError::InternalServerError(other.into()),
+        })?;
 
     match downloaded {
         Some(true) => json_response(StatusCode::OK, ()),
@@ -3169,12 +3175,16 @@ async fn put_tenant_timeline_import_basebackup(
 
     let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Warn);
 
-    let span = info_span!("import_basebackup", tenant_id=%tenant_id, timeline_id=%timeline_id, base_lsn=%base_lsn, end_lsn=%end_lsn, pg_version=%pg_version);
+    let tenant_shard_id = TenantShardId::unsharded(tenant_id);
+
+    let span = info_span!("import_basebackup",
+        tenant_id=%tenant_id, timeline_id=%timeline_id, shard_id=%tenant_shard_id.shard_slug(),
+        base_lsn=%base_lsn, end_lsn=%end_lsn, pg_version=%pg_version);
     async move {
         let state = get_state(&request);
         let tenant = state
             .tenant_manager
-            .get_attached_tenant_shard(TenantShardId::unsharded(tenant_id))?;
+            .get_attached_tenant_shard(tenant_shard_id)?;
 
         let broker_client = state.broker_client.clone();
 
@@ -3383,7 +3393,17 @@ where
                             let status = response.status();
                             info!(%status, "Cancelled request finished successfully")
                         }
-                        Err(e) => error!("Cancelled request finished with an error: {e:?}"),
+                        Err(e) => match e {
+                            ApiError::ShuttingDown | ApiError::ResourceUnavailable(_) => {
+                                // Don't log this at error severity: they are normal during lifecycle of tenants/process
+                                info!("Cancelled request aborted for shutdown")
+                            }
+                            _ => {
+                                // Log these in a highly visible way, because we have no client to send the response to, but
+                                // would like to know that something went wrong.
+                                error!("Cancelled request finished with an error: {e:?}")
+                            }
+                        },
                     }
                 }
                 // only logging for cancelled panicked request handlers is the tracing_panic_hook,

pageserver/src/lib.rs

@@ -263,14 +263,6 @@ pub(crate) const TENANT_HEATMAP_BASENAME: &str = "heatmap-v1.json";
 /// data directory at pageserver startup can be automatically removed.
 pub(crate) const TEMP_FILE_SUFFIX: &str = "___temp";
 
-/// A marker file to mark that a timeline directory was not fully initialized.
-/// If a timeline directory with this marker is encountered at pageserver startup,
-/// the timeline directory and the marker file are both removed.
-/// Full path: `tenants/<tenant_id>/timelines/<timeline_id>___uninit`.
-pub(crate) const TIMELINE_UNINIT_MARK_SUFFIX: &str = "___uninit";
-
-pub(crate) const TIMELINE_DELETE_MARK_SUFFIX: &str = "___delete";
-
 pub fn is_temporary(path: &Utf8Path) -> bool {
     match path.file_name() {
         Some(name) => name.ends_with(TEMP_FILE_SUFFIX),
@@ -278,25 +270,6 @@ pub fn is_temporary(path: &Utf8Path) -> bool {
     }
 }
 
-fn ends_with_suffix(path: &Utf8Path, suffix: &str) -> bool {
-    match path.file_name() {
-        Some(name) => name.ends_with(suffix),
-        None => false,
-    }
-}
-
-// FIXME: DO NOT ADD new query methods like this, which will have a next step of parsing timelineid
-// from the directory name. Instead create type "UninitMark(TimelineId)" and only parse it once
-// from the name.
-
-pub(crate) fn is_uninit_mark(path: &Utf8Path) -> bool {
-    ends_with_suffix(path, TIMELINE_UNINIT_MARK_SUFFIX)
-}
-
-pub(crate) fn is_delete_mark(path: &Utf8Path) -> bool {
-    ends_with_suffix(path, TIMELINE_DELETE_MARK_SUFFIX)
-}
-
 /// During pageserver startup, we need to order operations not to exhaust tokio worker threads by
 /// blocking.
 ///

pageserver/src/metrics.rs

@@ -1,4 +1,13 @@
+use std::collections::HashMap;
+use std::num::NonZeroUsize;
+use std::pin::Pin;
+use std::sync::atomic::AtomicU64;
+use std::sync::{Arc, Mutex};
+use std::task::{Context, Poll};
+use std::time::{Duration, Instant};
+
 use enum_map::EnumMap;
+use futures::Future;
 use metrics::{
     register_counter_vec, register_gauge_vec, register_histogram, register_histogram_vec,
     register_int_counter, register_int_counter_pair_vec, register_int_counter_vec,
@@ -11,13 +20,27 @@ use pageserver_api::config::{
     PageServicePipeliningConfig, PageServicePipeliningConfigPipelined,
     PageServiceProtocolPipelinedExecutionStrategy,
 };
+use pageserver_api::models::InMemoryLayerInfo;
 use pageserver_api::shard::TenantShardId;
+use pin_project_lite::pin_project;
 use postgres_backend::{is_expected_io_error, QueryError};
 use pq_proto::framed::ConnectionError;
-use strum::{EnumCount, VariantNames};
+
+use strum::{EnumCount, IntoEnumIterator as _, VariantNames};
 use strum_macros::{IntoStaticStr, VariantNames};
 use utils::id::TimelineId;
 
+use crate::config::PageServerConf;
+use crate::context::{PageContentKind, RequestContext};
+use crate::pgdatadir_mapping::DatadirModificationStats;
+use crate::task_mgr::TaskKind;
+use crate::tenant::layer_map::LayerMap;
+use crate::tenant::mgr::TenantSlot;
+use crate::tenant::storage_layer::{InMemoryLayer, PersistentLayerDesc};
+use crate::tenant::tasks::BackgroundLoopKind;
+use crate::tenant::throttle::ThrottleResult;
+use crate::tenant::Timeline;
+
 /// Prometheus histogram buckets (in seconds) for operations in the critical
 /// path. In other words, operations that directly affect that latency of user
 /// queries.
@@ -94,11 +117,38 @@ pub(crate) static STORAGE_TIME_GLOBAL: Lazy<HistogramVec> = Lazy::new(|| {
     .expect("failed to define a metric")
 });
 
-pub(crate) static VEC_READ_NUM_LAYERS_VISITED: Lazy<Histogram> = Lazy::new(|| {
+/// Measures layers visited per read (i.e. read amplification).
+///
+/// NB: for a batch, we count all visited layers towards each read. While the cost of layer visits
+/// are amortized across the batch, and some layers may not intersect with a given key, each visited
+/// layer contributes directly to the observed latency for every read in the batch, which is what we
+/// care about.
+pub(crate) static LAYERS_PER_READ: Lazy<HistogramVec> = Lazy::new(|| {
+    register_histogram_vec!(
+        "pageserver_layers_per_read",
+        "Layers visited to serve a single read (read amplification). In a batch, all visited layers count towards every read.",
+        &["tenant_id", "shard_id", "timeline_id"],
+        // Low resolution to reduce cardinality.
+        vec![1.0, 5.0, 10.0, 25.0, 50.0, 100.0],
+    )
+    .expect("failed to define a metric")
+});
+
+pub(crate) static LAYERS_PER_READ_GLOBAL: Lazy<Histogram> = Lazy::new(|| {
     register_histogram!(
-        "pageserver_layers_visited_per_vectored_read_global",
-        "Average number of layers visited to reconstruct one key",
-        vec![1.0, 4.0, 8.0, 16.0, 32.0, 64.0, 128.0, 256.0, 512.0, 1024.0],
+        "pageserver_layers_per_read_global",
+        "Layers visited to serve a single read (read amplification). In a batch, all visited layers count towards every read.",
+        vec![1.0, 2.0, 4.0, 8.0, 16.0, 32.0, 64.0, 128.0, 256.0, 512.0, 1024.0],
+    )
+    .expect("failed to define a metric")
+});
+
+pub(crate) static DELTAS_PER_READ_GLOBAL: Lazy<Histogram> = Lazy::new(|| {
+    // We expect this to be low because of Postgres checkpoints. Let's see if that holds.
+    register_histogram!(
+        "pageserver_deltas_per_read_global",
+        "Number of delta pages applied to image page per read",
+        vec![0.0, 1.0, 2.0, 4.0, 8.0, 16.0, 32.0, 64.0, 128.0, 256.0],
     )
     .expect("failed to define a metric")
 });
@@ -443,18 +493,38 @@ static PITR_HISTORY_SIZE: Lazy<UIntGaugeVec> = Lazy::new(|| {
     .expect("failed to define a metric")
 });
 
-#[derive(strum_macros::EnumString, strum_macros::Display, strum_macros::IntoStaticStr)]
+#[derive(
+    strum_macros::EnumIter,
+    strum_macros::EnumString,
+    strum_macros::Display,
+    strum_macros::IntoStaticStr,
+)]
 #[strum(serialize_all = "kebab_case")]
-pub(crate) enum MetricLayerKind {
+pub(crate) enum LayerKind {
     Delta,
     Image,
 }
 
+#[derive(
+    strum_macros::EnumIter,
+    strum_macros::EnumString,
+    strum_macros::Display,
+    strum_macros::IntoStaticStr,
+)]
+#[strum(serialize_all = "kebab_case")]
+pub(crate) enum LayerLevel {
+    // We don't track the currently open ephemeral layer, since there's always exactly 1 and its
+    // size changes. See `TIMELINE_EPHEMERAL_BYTES`.
+    Frozen,
+    L0,
+    L1,
+}
+
 static TIMELINE_LAYER_SIZE: Lazy<UIntGaugeVec> = Lazy::new(|| {
     register_uint_gauge_vec!(
         "pageserver_layer_bytes",
-        "Sum of layer physical sizes in bytes",
-        &["tenant_id", "shard_id", "timeline_id", "kind"]
+        "Sum of frozen, L0, and L1 layer physical sizes in bytes (excluding the open ephemeral layer)",
+        &["tenant_id", "shard_id", "timeline_id", "level", "kind"]
     )
     .expect("failed to define a metric")
 });
@@ -462,8 +532,8 @@ static TIMELINE_LAYER_SIZE: Lazy<UIntGaugeVec> = Lazy::new(|| {
 static TIMELINE_LAYER_COUNT: Lazy<UIntGaugeVec> = Lazy::new(|| {
     register_uint_gauge_vec!(
         "pageserver_layer_count",
-        "Number of layers that exist",
-        &["tenant_id", "shard_id", "timeline_id", "kind"]
+        "Number of frozen, L0, and L1 layers (excluding the open ephemeral layer)",
+        &["tenant_id", "shard_id", "timeline_id", "level", "kind"]
     )
     .expect("failed to define a metric")
 });
@@ -2309,11 +2379,40 @@ pub(crate) struct WalIngestMetrics {
     pub(crate) records_observed: IntCounter,
     pub(crate) records_committed: IntCounter,
     pub(crate) records_filtered: IntCounter,
+    pub(crate) values_committed_metadata_images: IntCounter,
+    pub(crate) values_committed_metadata_deltas: IntCounter,
+    pub(crate) values_committed_data_images: IntCounter,
+    pub(crate) values_committed_data_deltas: IntCounter,
     pub(crate) gap_blocks_zeroed_on_rel_extend: IntCounter,
-    pub(crate) clear_vm_bits_unknown: IntCounterVec,
+}
+
+impl WalIngestMetrics {
+    pub(crate) fn inc_values_committed(&self, stats: &DatadirModificationStats) {
+        if stats.metadata_images > 0 {
+            self.values_committed_metadata_images
+                .inc_by(stats.metadata_images);
+        }
+        if stats.metadata_deltas > 0 {
+            self.values_committed_metadata_deltas
+                .inc_by(stats.metadata_deltas);
+        }
+        if stats.data_images > 0 {
+            self.values_committed_data_images.inc_by(stats.data_images);
+        }
+        if stats.data_deltas > 0 {
+            self.values_committed_data_deltas.inc_by(stats.data_deltas);
+        }
+    }
 }
 
 pub(crate) static WAL_INGEST: Lazy<WalIngestMetrics> = Lazy::new(|| {
+    let values_committed = register_int_counter_vec!(
+        "pageserver_wal_ingest_values_committed",
+        "Number of values committed to pageserver storage from WAL records",
+        &["class", "kind"],
+    )
+    .expect("failed to define a metric");
+
     WalIngestMetrics {
     bytes_received: register_int_counter!(
         "pageserver_wal_ingest_bytes_received",
@@ -2340,17 +2439,15 @@ pub(crate) static WAL_INGEST: Lazy<WalIngestMetrics> = Lazy::new(|| {
         "Number of WAL records filtered out due to sharding"
     )
     .expect("failed to define a metric"),
+    values_committed_metadata_images: values_committed.with_label_values(&["metadata", "image"]),
+    values_committed_metadata_deltas: values_committed.with_label_values(&["metadata", "delta"]),
+    values_committed_data_images: values_committed.with_label_values(&["data", "image"]),
+    values_committed_data_deltas: values_committed.with_label_values(&["data", "delta"]),
     gap_blocks_zeroed_on_rel_extend: register_int_counter!(
         "pageserver_gap_blocks_zeroed_on_rel_extend",
         "Total number of zero gap blocks written on relation extends"
     )
     .expect("failed to define a metric"),
-    clear_vm_bits_unknown: register_int_counter_vec!(
-        "pageserver_wal_ingest_clear_vm_bits_unknown",
-        "Number of ignored ClearVmBits operations due to unknown pages/relations",
-        &["entity"],
-    )
-    .expect("failed to define a metric"),
 }
 });
 
@@ -2590,10 +2687,7 @@ pub(crate) struct TimelineMetrics {
     pub disk_consistent_lsn_gauge: IntGauge,
     pub pitr_history_size: UIntGauge,
     pub archival_size: UIntGauge,
-    pub(crate) layer_size_image: UIntGauge,
-    pub(crate) layer_count_image: UIntGauge,
-    pub(crate) layer_size_delta: UIntGauge,
-    pub(crate) layer_count_delta: UIntGauge,
+    pub layers_per_read: Histogram,
     pub standby_horizon_gauge: IntGauge,
     pub resident_physical_size_gauge: UIntGauge,
     pub visible_physical_size_gauge: UIntGauge,
@@ -2691,40 +2785,8 @@ impl TimelineMetrics {
             .get_metric_with_label_values(&[&tenant_id, &shard_id, &timeline_id])
             .unwrap();
 
-        let layer_size_image = TIMELINE_LAYER_SIZE
-            .get_metric_with_label_values(&[
-                &tenant_id,
-                &shard_id,
-                &timeline_id,
-                MetricLayerKind::Image.into(),
-            ])
-            .unwrap();
-
-        let layer_count_image = TIMELINE_LAYER_COUNT
-            .get_metric_with_label_values(&[
-                &tenant_id,
-                &shard_id,
-                &timeline_id,
-                MetricLayerKind::Image.into(),
-            ])
-            .unwrap();
-
-        let layer_size_delta = TIMELINE_LAYER_SIZE
-            .get_metric_with_label_values(&[
-                &tenant_id,
-                &shard_id,
-                &timeline_id,
-                MetricLayerKind::Delta.into(),
-            ])
-            .unwrap();
-
-        let layer_count_delta = TIMELINE_LAYER_COUNT
-            .get_metric_with_label_values(&[
-                &tenant_id,
-                &shard_id,
-                &timeline_id,
-                MetricLayerKind::Delta.into(),
-            ])
+        let layers_per_read = LAYERS_PER_READ
+            .get_metric_with_label_values(&[&tenant_id, &shard_id, &timeline_id])
             .unwrap();
 
         let standby_horizon_gauge = STANDBY_HORIZON
@@ -2791,10 +2853,7 @@ impl TimelineMetrics {
             disk_consistent_lsn_gauge,
             pitr_history_size,
             archival_size,
-            layer_size_image,
-            layer_count_image,
-            layer_size_delta,
-            layer_count_delta,
+            layers_per_read,
             standby_horizon_gauge,
             resident_physical_size_gauge,
             visible_physical_size_gauge,
@@ -2837,6 +2896,92 @@ impl TimelineMetrics {
             .add(duration);
     }
 
+    /// Generates TIMELINE_LAYER labels for a persistent layer.
+    fn make_layer_labels(&self, layer_desc: &PersistentLayerDesc) -> [&str; 5] {
+        let level = match LayerMap::is_l0(&layer_desc.key_range, layer_desc.is_delta()) {
+            true => LayerLevel::L0,
+            false => LayerLevel::L1,
+        };
+        let kind = match layer_desc.is_delta() {
+            true => LayerKind::Delta,
+            false => LayerKind::Image,
+        };
+        [
+            &self.tenant_id,
+            &self.shard_id,
+            &self.timeline_id,
+            level.into(),
+            kind.into(),
+        ]
+    }
+
+    /// Generates TIMELINE_LAYER labels for a frozen ephemeral layer.
+    fn make_frozen_layer_labels(&self, _layer: &InMemoryLayer) -> [&str; 5] {
+        [
+            &self.tenant_id,
+            &self.shard_id,
+            &self.timeline_id,
+            LayerLevel::Frozen.into(),
+            LayerKind::Delta.into(), // by definition
+        ]
+    }
+
+    /// Removes a frozen ephemeral layer to TIMELINE_LAYER metrics.
+    pub fn dec_frozen_layer(&self, layer: &InMemoryLayer) {
+        assert!(matches!(layer.info(), InMemoryLayerInfo::Frozen { .. }));
+        let labels = self.make_frozen_layer_labels(layer);
+        let size = layer.try_len().expect("frozen layer should have no writer");
+        TIMELINE_LAYER_COUNT
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .dec();
+        TIMELINE_LAYER_SIZE
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .sub(size);
+    }
+
+    /// Adds a frozen ephemeral layer to TIMELINE_LAYER metrics.
+    pub fn inc_frozen_layer(&self, layer: &InMemoryLayer) {
+        assert!(matches!(layer.info(), InMemoryLayerInfo::Frozen { .. }));
+        let labels = self.make_frozen_layer_labels(layer);
+        let size = layer.try_len().expect("frozen layer should have no writer");
+        TIMELINE_LAYER_COUNT
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .inc();
+        TIMELINE_LAYER_SIZE
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .add(size);
+    }
+
+    /// Removes a persistent layer from TIMELINE_LAYER metrics.
+    pub fn dec_layer(&self, layer_desc: &PersistentLayerDesc) {
+        let labels = self.make_layer_labels(layer_desc);
+        TIMELINE_LAYER_COUNT
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .dec();
+        TIMELINE_LAYER_SIZE
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .sub(layer_desc.file_size);
+    }
+
+    /// Adds a persistent layer to TIMELINE_LAYER metrics.
+    pub fn inc_layer(&self, layer_desc: &PersistentLayerDesc) {
+        let labels = self.make_layer_labels(layer_desc);
+        TIMELINE_LAYER_COUNT
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .inc();
+        TIMELINE_LAYER_SIZE
+            .get_metric_with_label_values(&labels)
+            .unwrap()
+            .add(layer_desc.file_size);
+    }
+
     pub(crate) fn shutdown(&self) {
         let was_shutdown = self
             .shutdown
@@ -2869,30 +3014,16 @@ impl TimelineMetrics {
         let _ = TIMELINE_ARCHIVE_SIZE.remove_label_values(&[tenant_id, shard_id, timeline_id]);
         let _ = PITR_HISTORY_SIZE.remove_label_values(&[tenant_id, shard_id, timeline_id]);
 
-        let _ = TIMELINE_LAYER_SIZE.remove_label_values(&[
-            tenant_id,
-            shard_id,
-            timeline_id,
-            MetricLayerKind::Image.into(),
-        ]);
-        let _ = TIMELINE_LAYER_COUNT.remove_label_values(&[
-            tenant_id,
-            shard_id,
-            timeline_id,
-            MetricLayerKind::Image.into(),
-        ]);
-        let _ = TIMELINE_LAYER_SIZE.remove_label_values(&[
-            tenant_id,
-            shard_id,
-            timeline_id,
-            MetricLayerKind::Delta.into(),
-        ]);
-        let _ = TIMELINE_LAYER_COUNT.remove_label_values(&[
-            tenant_id,
-            shard_id,
-            timeline_id,
-            MetricLayerKind::Delta.into(),
-        ]);
+        for ref level in LayerLevel::iter() {
+            for ref kind in LayerKind::iter() {
+                let labels: [&str; 5] =
+                    [tenant_id, shard_id, timeline_id, level.into(), kind.into()];
+                let _ = TIMELINE_LAYER_SIZE.remove_label_values(&labels);
+                let _ = TIMELINE_LAYER_COUNT.remove_label_values(&labels);
+            }
+        }
+
+        let _ = LAYERS_PER_READ.remove_label_values(&[tenant_id, shard_id, timeline_id]);
 
         let _ = EVICTIONS.remove_label_values(&[tenant_id, shard_id, timeline_id]);
         let _ = AUX_FILE_SIZE.remove_label_values(&[tenant_id, shard_id, timeline_id]);
@@ -2974,24 +3105,6 @@ pub(crate) fn remove_tenant_metrics(tenant_shard_id: &TenantShardId) {
     // we leave the BROKEN_TENANTS_SET entry if any
 }
 
-use futures::Future;
-use pin_project_lite::pin_project;
-use std::collections::HashMap;
-use std::num::NonZeroUsize;
-use std::pin::Pin;
-use std::sync::atomic::AtomicU64;
-use std::sync::{Arc, Mutex};
-use std::task::{Context, Poll};
-use std::time::{Duration, Instant};
-
-use crate::config::PageServerConf;
-use crate::context::{PageContentKind, RequestContext};
-use crate::task_mgr::TaskKind;
-use crate::tenant::mgr::TenantSlot;
-use crate::tenant::tasks::BackgroundLoopKind;
-use crate::tenant::throttle::ThrottleResult;
-use crate::tenant::Timeline;
-
 /// Maintain a per timeline gauge in addition to the global gauge.
 pub(crate) struct PerTimelineRemotePhysicalSizeGauge {
     last_set: AtomicU64,
@@ -3862,7 +3975,8 @@ pub fn preinitialize_metrics(conf: &'static PageServerConf) {
 
     // histograms
     [
-        &VEC_READ_NUM_LAYERS_VISITED,
+        &LAYERS_PER_READ_GLOBAL,
+        &DELTAS_PER_READ_GLOBAL,
         &WAIT_LSN_TIME,
         &WAL_REDO_TIME,
         &WAL_REDO_RECORDS_HISTOGRAM,

pageserver/src/page_service.rs

@@ -1280,8 +1280,6 @@ impl PageServerHandler {
                 }
                 Ok(())
             }
-            // and log the info! line inside the request span
-            .instrument(span.clone())
             .await?;
         }
         Ok(())
@@ -2037,6 +2035,12 @@ impl PageServerHandler {
             .get(tenant_id, timeline_id, ShardSelector::Zero)
             .await?;
 
+        if timeline.is_archived() == Some(true) {
+            // TODO after a grace period, turn this log line into a hard error
+            tracing::warn!("timeline {tenant_id}/{timeline_id} is archived, but got basebackup request for it.");
+            //return Err(QueryError::NotFound("timeline is archived".into()))
+        }
+
         let latest_gc_cutoff_lsn = timeline.get_latest_gc_cutoff_lsn();
         if let Some(lsn) = lsn {
             // Backup was requested at a particular LSN. Wait for it to arrive.

pageserver/src/pgdatadir_mapping.rs

@@ -48,7 +48,7 @@ use tracing::{debug, trace, warn};
 use utils::bin_ser::DeserializeError;
 use utils::pausable_failpoint;
 use utils::{bin_ser::BeSer, lsn::Lsn};
-use wal_decoder::serialized_batch::SerializedValueBatch;
+use wal_decoder::serialized_batch::{SerializedValueBatch, ValueMeta};
 
 /// Max delta records appended to the AUX_FILES_KEY (for aux v1). The write path will write a full image once this threshold is reached.
 pub const MAX_AUX_FILE_DELTAS: usize = 1024;
@@ -1297,6 +1297,26 @@ impl DatadirModification<'_> {
             .is_some_and(|b| b.has_data())
     }
 
+    /// Returns statistics about the currently pending modifications.
+    pub(crate) fn stats(&self) -> DatadirModificationStats {
+        let mut stats = DatadirModificationStats::default();
+        for (_, _, value) in self.pending_metadata_pages.values().flatten() {
+            match value {
+                Value::Image(_) => stats.metadata_images += 1,
+                Value::WalRecord(r) if r.will_init() => stats.metadata_images += 1,
+                Value::WalRecord(_) => stats.metadata_deltas += 1,
+            }
+        }
+        for valuemeta in self.pending_data_batch.iter().flat_map(|b| &b.metadata) {
+            match valuemeta {
+                ValueMeta::Serialized(s) if s.will_init => stats.data_images += 1,
+                ValueMeta::Serialized(_) => stats.data_deltas += 1,
+                ValueMeta::Observed(_) => {}
+            }
+        }
+        stats
+    }
+
     /// Set the current lsn
     pub(crate) fn set_lsn(&mut self, lsn: Lsn) -> anyhow::Result<()> {
         ensure!(
@@ -2317,6 +2337,15 @@ impl DatadirModification<'_> {
     }
 }
 
+/// Statistics for a DatadirModification.
+#[derive(Default)]
+pub struct DatadirModificationStats {
+    pub metadata_images: u64,
+    pub metadata_deltas: u64,
+    pub data_images: u64,
+    pub data_deltas: u64,
+}
+
 /// This struct facilitates accessing either a committed key from the timeline at a
 /// specific LSN, or the latest uncommitted key from a pending modification.
 ///

pageserver/src/tenant.rs

@@ -46,6 +46,7 @@ use std::sync::atomic::AtomicBool;
 use std::sync::Weak;
 use std::time::SystemTime;
 use storage_broker::BrokerClientChannel;
+use timeline::compaction::CompactionOutcome;
 use timeline::compaction::GcCompactionQueue;
 use timeline::import_pgdata;
 use timeline::offload::offload_timeline;
@@ -95,7 +96,6 @@ use crate::context::{DownloadBehavior, RequestContext};
 use crate::deletion_queue::DeletionQueueClient;
 use crate::deletion_queue::DeletionQueueError;
 use crate::import_datadir;
-use crate::is_uninit_mark;
 use crate::l0_flush::L0FlushGlobalState;
 use crate::metrics::CONCURRENT_INITDBS;
 use crate::metrics::INITDB_RUN_TIME;
@@ -1793,11 +1793,7 @@ impl Tenant {
             let entry = entry.context("read timeline dir entry")?;
             let entry_path = entry.path();
 
-            let purge = if crate::is_temporary(entry_path)
-                // TODO: remove uninit mark code (https://github.com/neondatabase/neon/issues/5718)
-                || is_uninit_mark(entry_path)
-                || crate::is_delete_mark(entry_path)
-            {
+            let purge = if crate::is_temporary(entry_path) {
                 true
             } else {
                 match TimelineId::try_from(entry_path.file_name()) {
@@ -2426,7 +2422,7 @@ impl Tenant {
         // Make sure the freeze_and_flush reaches remote storage.
         tline.remote_client.wait_completion().await.unwrap();
 
-        let tl = uninit_tl.finish_creation()?;
+        let tl = uninit_tl.finish_creation().await?;
         // The non-test code would call tl.activate() here.
         tl.set_state(TimelineState::Active);
         Ok(tl)
@@ -2912,10 +2908,10 @@ impl Tenant {
         self: &Arc<Self>,
         cancel: &CancellationToken,
         ctx: &RequestContext,
-    ) -> Result<bool, timeline::CompactionError> {
+    ) -> Result<CompactionOutcome, timeline::CompactionError> {
         // Don't start doing work during shutdown, or when broken, we do not need those in the logs
         if !self.is_active() {
-            return Ok(false);
+            return Ok(CompactionOutcome::Done);
         }
 
         {
@@ -2929,7 +2925,7 @@ impl Tenant {
             // to AttachedSingle state.
             if !conf.location.may_upload_layers_hint() {
                 info!("Skipping compaction in location state {:?}", conf.location);
-                return Ok(false);
+                return Ok(CompactionOutcome::Done);
             }
         }
 
@@ -2972,7 +2968,7 @@ impl Tenant {
         // Before doing any I/O work, check our circuit breaker
         if self.compaction_circuit_breaker.lock().unwrap().is_broken() {
             info!("Skipping compaction due to previous failures");
-            return Ok(false);
+            return Ok(CompactionOutcome::Done);
         }
 
         let mut has_pending_task = false;
@@ -2980,10 +2976,10 @@ impl Tenant {
         for (timeline_id, timeline, (can_compact, can_offload)) in &timelines_to_compact_or_offload
         {
             // pending_task_left == None: cannot compact, maybe still pending tasks
-            // pending_task_left == Some(true): compaction task left
-            // pending_task_left == Some(false): no compaction task left
+            // pending_task_left == Some(Pending): compaction task left
+            // pending_task_left == Some(Done): no compaction task left
             let pending_task_left = if *can_compact {
-                let has_pending_l0_compaction_task = timeline
+                let compaction_outcome = timeline
                     .compact(cancel, EnumSet::empty(), ctx)
                     .instrument(info_span!("compact_timeline", %timeline_id))
                     .await
@@ -3001,27 +2997,27 @@ impl Tenant {
                                 .fail(&CIRCUIT_BREAKERS_BROKEN, e);
                         }
                     })?;
-                if has_pending_l0_compaction_task {
-                    Some(true)
+                if let CompactionOutcome::Pending = compaction_outcome {
+                    Some(CompactionOutcome::Pending)
                 } else {
                     let queue = {
                         let guard = self.scheduled_compaction_tasks.lock().unwrap();
                         guard.get(timeline_id).cloned()
                     };
                     if let Some(queue) = queue {
-                        let has_pending_tasks = queue
+                        let outcome = queue
                             .iteration(cancel, ctx, &self.gc_block, timeline)
                             .await?;
-                        Some(has_pending_tasks)
+                        Some(outcome)
                     } else {
-                        Some(false)
+                        Some(CompactionOutcome::Done)
                     }
                 }
             } else {
                 None
             };
-            has_pending_task |= pending_task_left.unwrap_or(false);
-            if pending_task_left == Some(false) && *can_offload {
+            has_pending_task |= pending_task_left == Some(CompactionOutcome::Pending);
+            if pending_task_left == Some(CompactionOutcome::Done) && *can_offload {
                 pausable_failpoint!("before-timeline-auto-offload");
                 match offload_timeline(self, timeline)
                     .instrument(info_span!("offload_timeline", %timeline_id))
@@ -3041,7 +3037,11 @@ impl Tenant {
             .unwrap()
             .success(&CIRCUIT_BREAKERS_UNBROKEN);
 
-        Ok(has_pending_task)
+        Ok(if has_pending_task {
+            CompactionOutcome::Pending
+        } else {
+            CompactionOutcome::Done
+        })
     }
 
     /// Cancel scheduled compaction tasks
@@ -3816,6 +3816,13 @@ impl Tenant {
             .unwrap_or(self.conf.default_tenant_conf.compaction_threshold)
     }
 
+    pub fn get_compaction_upper_limit(&self) -> usize {
+        let tenant_conf = self.tenant_conf.load().tenant_conf.clone();
+        tenant_conf
+            .compaction_upper_limit
+            .unwrap_or(self.conf.default_tenant_conf.compaction_upper_limit)
+    }
+
     pub fn get_gc_horizon(&self) -> u64 {
         let tenant_conf = self.tenant_conf.load().tenant_conf.clone();
         tenant_conf
@@ -4695,7 +4702,7 @@ impl Tenant {
             )
             .await?;
 
-        let new_timeline = uninitialized_timeline.finish_creation()?;
+        let new_timeline = uninitialized_timeline.finish_creation().await?;
 
         // Root timeline gets its layers during creation and uploads them along with the metadata.
         // A branch timeline though, when created, can get no writes for some time, hence won't get any layers created.
@@ -4885,10 +4892,11 @@ impl Tenant {
         }
 
         // this new directory is very temporary, set to remove it immediately after bootstrap, we don't need it
+        let pgdata_path_deferred = pgdata_path.clone();
         scopeguard::defer! {
-            if let Err(e) = fs::remove_dir_all(&pgdata_path) {
+            if let Err(e) = fs::remove_dir_all(&pgdata_path_deferred) {
                 // this is unlikely, but we will remove the directory on pageserver restart or another bootstrap call
-                error!("Failed to remove temporary initdb directory '{pgdata_path}': {e}");
+                error!("Failed to remove temporary initdb directory '{pgdata_path_deferred}': {e}");
             }
         }
         if let Some(existing_initdb_timeline_id) = load_existing_initdb {
@@ -4955,7 +4963,7 @@ impl Tenant {
             pgdata_lsn,
             pg_version,
         );
-        let raw_timeline = self
+        let mut raw_timeline = self
             .prepare_new_timeline(
                 timeline_id,
                 &new_metadata,
@@ -4966,42 +4974,33 @@ impl Tenant {
             .await?;
 
         let tenant_shard_id = raw_timeline.owning_tenant.tenant_shard_id;
-        let unfinished_timeline = raw_timeline.raw_timeline()?;
-
-        // Flush the new layer files to disk, before we make the timeline as available to
-        // the outside world.
-        //
-        // Flush loop needs to be spawned in order to be able to flush.
-        unfinished_timeline.maybe_spawn_flush_loop();
-
-        import_datadir::import_timeline_from_postgres_datadir(
-            unfinished_timeline,
-            &pgdata_path,
-            pgdata_lsn,
-            ctx,
-        )
-        .await
-        .with_context(|| {
-            format!("Failed to import pgdatadir for timeline {tenant_shard_id}/{timeline_id}")
-        })?;
-
-        fail::fail_point!("before-checkpoint-new-timeline", |_| {
-            Err(CreateTimelineError::Other(anyhow::anyhow!(
-                "failpoint before-checkpoint-new-timeline"
-            )))
-        });
-
-        unfinished_timeline
-            .freeze_and_flush()
-            .await
-            .with_context(|| {
-                format!(
-                    "Failed to flush after pgdatadir import for timeline {tenant_shard_id}/{timeline_id}"
+        raw_timeline
+            .write(|unfinished_timeline| async move {
+                import_datadir::import_timeline_from_postgres_datadir(
+                    &unfinished_timeline,
+                    &pgdata_path,
+                    pgdata_lsn,
+                    ctx,
                 )
-            })?;
+                .await
+                .with_context(|| {
+                    format!(
+                        "Failed to import pgdatadir for timeline {tenant_shard_id}/{timeline_id}"
+                    )
+                })?;
+
+                fail::fail_point!("before-checkpoint-new-timeline", |_| {
+                    Err(CreateTimelineError::Other(anyhow::anyhow!(
+                        "failpoint before-checkpoint-new-timeline"
+                    )))
+                });
+
+                Ok(())
+            })
+            .await?;
 
         // All done!
-        let timeline = raw_timeline.finish_creation()?;
+        let timeline = raw_timeline.finish_creation().await?;
 
         // Callers are responsible to wait for uploads to complete and for activating the timeline.
 
@@ -5469,6 +5468,7 @@ pub(crate) mod harness {
                 compaction_target_size: Some(tenant_conf.compaction_target_size),
                 compaction_period: Some(tenant_conf.compaction_period),
                 compaction_threshold: Some(tenant_conf.compaction_threshold),
+                compaction_upper_limit: Some(tenant_conf.compaction_upper_limit),
                 compaction_algorithm: Some(tenant_conf.compaction_algorithm),
                 l0_flush_delay_threshold: tenant_conf.l0_flush_delay_threshold,
                 l0_flush_stall_threshold: tenant_conf.l0_flush_stall_threshold,
@@ -5491,6 +5491,9 @@ pub(crate) mod harness {
                 image_layer_creation_check_threshold: Some(
                     tenant_conf.image_layer_creation_check_threshold,
                 ),
+                image_creation_preempt_threshold: Some(
+                    tenant_conf.image_creation_preempt_threshold,
+                ),
                 lsn_lease_length: Some(tenant_conf.lsn_lease_length),
                 lsn_lease_length_for_ts: Some(tenant_conf.lsn_lease_length_for_ts),
                 timeline_offloading: Some(tenant_conf.timeline_offloading),

pageserver/src/tenant/config.rs

@@ -277,6 +277,10 @@ pub struct TenantConfOpt {
     #[serde(default)]
     pub compaction_threshold: Option<usize>,
 
+    #[serde(skip_serializing_if = "Option::is_none")]
+    #[serde(default)]
+    pub compaction_upper_limit: Option<usize>,
+
     #[serde(skip_serializing_if = "Option::is_none")]
     #[serde(default)]
     pub compaction_algorithm: Option<CompactionAlgorithmSettings>,
@@ -353,6 +357,9 @@ pub struct TenantConfOpt {
     #[serde(skip_serializing_if = "Option::is_none")]
     pub image_layer_creation_check_threshold: Option<u8>,
 
+    #[serde(skip_serializing_if = "Option::is_none")]
+    pub image_creation_preempt_threshold: Option<usize>,
+
     #[serde(skip_serializing_if = "Option::is_none")]
     #[serde(with = "humantime_serde")]
     #[serde(default)]
@@ -401,6 +408,9 @@ impl TenantConfOpt {
             compaction_threshold: self
                 .compaction_threshold
                 .unwrap_or(global_conf.compaction_threshold),
+            compaction_upper_limit: self
+                .compaction_upper_limit
+                .unwrap_or(global_conf.compaction_upper_limit),
             compaction_algorithm: self
                 .compaction_algorithm
                 .as_ref()
@@ -446,6 +456,9 @@ impl TenantConfOpt {
             image_layer_creation_check_threshold: self
                 .image_layer_creation_check_threshold
                 .unwrap_or(global_conf.image_layer_creation_check_threshold),
+            image_creation_preempt_threshold: self
+                .image_creation_preempt_threshold
+                .unwrap_or(global_conf.image_creation_preempt_threshold),
             lsn_lease_length: self
                 .lsn_lease_length
                 .unwrap_or(global_conf.lsn_lease_length),
@@ -478,6 +491,7 @@ impl TenantConfOpt {
             mut compaction_target_size,
             mut compaction_period,
             mut compaction_threshold,
+            mut compaction_upper_limit,
             mut compaction_algorithm,
             mut l0_flush_delay_threshold,
             mut l0_flush_stall_threshold,
@@ -496,6 +510,7 @@ impl TenantConfOpt {
             mut lazy_slru_download,
             mut timeline_get_throttle,
             mut image_layer_creation_check_threshold,
+            mut image_creation_preempt_threshold,
             mut lsn_lease_length,
             mut lsn_lease_length_for_ts,
             mut timeline_offloading,
@@ -519,6 +534,9 @@ impl TenantConfOpt {
             .map(|v| humantime::parse_duration(&v))?
             .apply(&mut compaction_period);
         patch.compaction_threshold.apply(&mut compaction_threshold);
+        patch
+            .compaction_upper_limit
+            .apply(&mut compaction_upper_limit);
         patch.compaction_algorithm.apply(&mut compaction_algorithm);
         patch
             .l0_flush_delay_threshold
@@ -567,6 +585,9 @@ impl TenantConfOpt {
         patch
             .image_layer_creation_check_threshold
             .apply(&mut image_layer_creation_check_threshold);
+        patch
+            .image_creation_preempt_threshold
+            .apply(&mut image_creation_preempt_threshold);
         patch
             .lsn_lease_length
             .map(|v| humantime::parse_duration(&v))?
@@ -596,6 +617,7 @@ impl TenantConfOpt {
             compaction_target_size,
             compaction_period,
             compaction_threshold,
+            compaction_upper_limit,
             compaction_algorithm,
             l0_flush_delay_threshold,
             l0_flush_stall_threshold,
@@ -614,6 +636,7 @@ impl TenantConfOpt {
             lazy_slru_download,
             timeline_get_throttle,
             image_layer_creation_check_threshold,
+            image_creation_preempt_threshold,
             lsn_lease_length,
             lsn_lease_length_for_ts,
             timeline_offloading,
@@ -657,6 +680,7 @@ impl From<TenantConfOpt> for models::TenantConfig {
             compaction_target_size: value.compaction_target_size,
             compaction_period: value.compaction_period.map(humantime),
             compaction_threshold: value.compaction_threshold,
+            compaction_upper_limit: value.compaction_upper_limit,
             l0_flush_delay_threshold: value.l0_flush_delay_threshold,
             l0_flush_stall_threshold: value.l0_flush_stall_threshold,
             l0_flush_wait_upload: value.l0_flush_wait_upload,
@@ -676,6 +700,7 @@ impl From<TenantConfOpt> for models::TenantConfig {
             lazy_slru_download: value.lazy_slru_download,
             timeline_get_throttle: value.timeline_get_throttle,
             image_layer_creation_check_threshold: value.image_layer_creation_check_threshold,
+            image_creation_preempt_threshold: value.image_creation_preempt_threshold,
             lsn_lease_length: value.lsn_lease_length.map(humantime),
             lsn_lease_length_for_ts: value.lsn_lease_length_for_ts.map(humantime),
             timeline_offloading: value.timeline_offloading,

pageserver/src/tenant/remote_timeline_client.rs

@@ -437,8 +437,7 @@ impl RemoteTimelineClient {
             .conf
             .remote_storage_config
             .as_ref()
-            .and_then(|r| r.concurrency_limit())
-            .unwrap_or(0);
+            .map_or(0, |r| r.concurrency_limit());
         let mut upload_queue = self.upload_queue.lock().unwrap();
         upload_queue.initialize_with_current_remote_index_part(index_part, inprogress_limit)?;
         self.update_remote_physical_size_gauge(Some(index_part));
@@ -461,8 +460,7 @@ impl RemoteTimelineClient {
             .conf
             .remote_storage_config
             .as_ref()
-            .and_then(|r| r.concurrency_limit())
-            .unwrap_or(0);
+            .map_or(0, |r| r.concurrency_limit());
         let mut upload_queue = self.upload_queue.lock().unwrap();
         upload_queue.initialize_empty_remote(local_metadata, inprogress_limit)?;
         self.update_remote_physical_size_gauge(None);
@@ -484,8 +482,7 @@ impl RemoteTimelineClient {
             .conf
             .remote_storage_config
             .as_ref()
-            .and_then(|r| r.concurrency_limit())
-            .unwrap_or(0);
+            .map_or(0, |r| r.concurrency_limit());
 
         let mut upload_queue = self.upload_queue.lock().unwrap();
         upload_queue.initialize_with_current_remote_index_part(index_part, inprogress_limit)?;

pageserver/src/tenant/secondary/downloader.rs

@@ -673,12 +673,30 @@ impl<'a> TenantDownloader<'a> {
             HeatMapDownload::Modified(m) => m,
         };
 
-        let heatmap = serde_json::from_slice::<HeatMapTenant>(&heatmap_bytes)?;
-
-        // Save the heatmap: this will be useful on restart, allowing us to reconstruct
-        // layer metadata without having to re-download it.
+        // Heatmap storage location
         let heatmap_path = self.conf.tenant_heatmap_path(tenant_shard_id);
 
+        let last_heatmap = if last_download.is_none() {
+            match load_heatmap(&heatmap_path, ctx).await {
+                Ok(htm) => htm,
+                Err(e) => {
+                    tracing::warn!("Couldn't load heatmap from {heatmap_path}: {e:?}");
+                    None
+                }
+            }
+        } else {
+            None
+        };
+
+        let last_heatmap_timelines = last_heatmap.as_ref().map(|htm| {
+            htm.timelines
+                .iter()
+                .map(|tl| (tl.timeline_id, tl))
+                .collect::<HashMap<_, _>>()
+        });
+
+        let heatmap = serde_json::from_slice::<HeatMapTenant>(&heatmap_bytes)?;
+
         let temp_path = path_with_suffix_extension(&heatmap_path, TEMP_FILE_SUFFIX);
         let context_msg = format!("write tenant {tenant_shard_id} heatmap to {heatmap_path}");
         let heatmap_path_bg = heatmap_path.clone();
@@ -707,10 +725,17 @@ impl<'a> TenantDownloader<'a> {
             let timeline_state = match timeline_state {
                 Some(t) => t,
                 None => {
+                    let last_heatmap =
+                        last_heatmap_timelines
+                            .as_ref()
+                            .and_then(|last_heatmap_timelines| {
+                                last_heatmap_timelines.get(&timeline.timeline_id).copied()
+                            });
                     // We have no existing state: need to scan local disk for layers first.
                     let timeline_state = init_timeline_state(
                         self.conf,
                         tenant_shard_id,
+                        last_heatmap,
                         timeline,
                         &self.secondary_state.resident_size_metric,
                     )
@@ -1079,12 +1104,12 @@ impl<'a> TenantDownloader<'a> {
                 }
             }
 
-            if on_disk.metadata.generation_file_size() != on_disk.metadata.generation_file_size() {
+            if on_disk.metadata.generation_file_size() != layer.metadata.generation_file_size() {
                 tracing::info!(
                     "Re-downloading layer {} with changed size or generation: {:?}->{:?}",
                     layer.name,
                     on_disk.metadata.generation_file_size(),
-                    on_disk.metadata.generation_file_size()
+                    layer.metadata.generation_file_size()
                 );
                 return LayerAction::Download;
             }
@@ -1277,6 +1302,7 @@ impl<'a> TenantDownloader<'a> {
 async fn init_timeline_state(
     conf: &'static PageServerConf,
     tenant_shard_id: &TenantShardId,
+    last_heatmap: Option<&HeatMapTimeline>,
     heatmap: &HeatMapTimeline,
     resident_metric: &UIntGauge,
 ) -> SecondaryDetailTimeline {
@@ -1306,6 +1332,13 @@ async fn init_timeline_state(
     let heatmap_metadata: HashMap<&LayerName, &HeatMapLayer> =
         heatmap.layers.iter().map(|l| (&l.name, l)).collect();
 
+    let last_heatmap_metadata: HashMap<&LayerName, &HeatMapLayer> =
+        if let Some(last_heatmap) = last_heatmap {
+            last_heatmap.layers.iter().map(|l| (&l.name, l)).collect()
+        } else {
+            HashMap::new()
+        };
+
     while let Some(dentry) = dir
         .next_entry()
         .await
@@ -1339,18 +1372,32 @@ async fn init_timeline_state(
         match LayerName::from_str(file_name) {
             Ok(name) => {
                 let remote_meta = heatmap_metadata.get(&name);
+                let last_meta = last_heatmap_metadata.get(&name);
+                let mut remove = false;
                 match remote_meta {
                     Some(remote_meta) => {
+                        let last_meta_generation_file_size = last_meta
+                            .map(|m| m.metadata.generation_file_size())
+                            .unwrap_or(remote_meta.metadata.generation_file_size());
                         // TODO: checksums for layers (https://github.com/neondatabase/neon/issues/2784)
-                        if local_meta.len() != remote_meta.metadata.file_size {
-                            // This should not happen, because we do crashsafe write-then-rename when downloading
-                            // layers, and layers in remote storage are immutable.  Remove the local file because
-                            // we cannot trust it.
-                            tracing::warn!(
+                        if remote_meta.metadata.generation_file_size()
+                            != last_meta_generation_file_size
+                        {
+                            tracing::info!(
+                                "Removing local layer {name} as on-disk json metadata has different generation or file size from remote: {:?} -> {:?}",
+                                last_meta_generation_file_size,
+                                remote_meta.metadata.generation_file_size()
+                            );
+                            remove = true;
+                        } else if local_meta.len() != remote_meta.metadata.file_size {
+                            // This can happen in the presence of race conditions: the remote and on-disk metadata have changed, but we haven't had
+                            // the chance yet to download the new layer to disk, before the process restarted.
+                            tracing::info!(
                                 "Removing local layer {name} with unexpected local size {} != {}",
                                 local_meta.len(),
                                 remote_meta.metadata.file_size
                             );
+                            remove = true;
                         } else {
                             // We expect the access time to be initialized immediately afterwards, when
                             // the latest heatmap is applied to the state.
@@ -1372,15 +1419,18 @@ async fn init_timeline_state(
                             "Removing secondary local layer {} because it's absent in heatmap",
                             name
                         );
-                        tokio::fs::remove_file(&dentry.path())
-                            .await
-                            .or_else(fs_ext::ignore_not_found)
-                            .fatal_err(&format!(
-                                "Removing layer {}",
-                                dentry.path().to_string_lossy()
-                            ));
+                        remove = true;
                     }
                 }
+                if remove {
+                    tokio::fs::remove_file(&dentry.path())
+                        .await
+                        .or_else(fs_ext::ignore_not_found)
+                        .fatal_err(&format!(
+                            "Removing layer {}",
+                            dentry.path().to_string_lossy()
+                        ));
+                }
             }
             Err(_) => {
                 // Ignore it.
@@ -1391,3 +1441,18 @@ async fn init_timeline_state(
 
     detail
 }
+
+/// Loads a json-encoded heatmap file from the provided on-disk path
+async fn load_heatmap(
+    path: &Utf8PathBuf,
+    ctx: &RequestContext,
+) -> Result<Option<HeatMapTenant>, anyhow::Error> {
+    let mut file = match VirtualFile::open(path, ctx).await {
+        Ok(file) => file,
+        Err(e) if e.kind() == std::io::ErrorKind::NotFound => return Ok(None),
+        Err(e) => Err(e)?,
+    };
+    let st = file.read_to_string(ctx).await?;
+    let htm = serde_json::from_str(&st)?;
+    Ok(Some(htm))
+}

pageserver/src/tenant/secondary/heatmap_uploader.rs

@@ -9,13 +9,14 @@ use crate::{
     metrics::SECONDARY_MODE,
     tenant::{
         config::AttachmentMode,
-        mgr::GetTenantError,
-        mgr::TenantManager,
+        mgr::{GetTenantError, TenantManager},
         remote_timeline_client::remote_heatmap_path,
         span::debug_assert_current_span_has_tenant_id,
         tasks::{warn_when_period_overrun, BackgroundLoopKind},
         Tenant,
     },
+    virtual_file::VirtualFile,
+    TEMP_FILE_SUFFIX,
 };
 
 use futures::Future;
@@ -32,7 +33,10 @@ use super::{
 };
 use tokio_util::sync::CancellationToken;
 use tracing::{info_span, instrument, Instrument};
-use utils::{backoff, completion::Barrier, yielding_loop::yielding_loop};
+use utils::{
+    backoff, completion::Barrier, crashsafe::path_with_suffix_extension,
+    yielding_loop::yielding_loop,
+};
 
 pub(super) async fn heatmap_uploader_task(
     tenant_manager: Arc<TenantManager>,
@@ -461,6 +465,18 @@ async fn upload_tenant_heatmap(
         }
     }
 
+    // After a successful upload persist the fresh heatmap to disk.
+    // When restarting, the tenant will read the heatmap from disk
+    // and additively generate a new heatmap (see [`Timeline::generate_heatmap`]).
+    // If the heatmap is stale, the additive generation can lead to keeping previously
+    // evicted timelines on the secondarie's disk.
+    let tenant_shard_id = tenant.get_tenant_shard_id();
+    let heatmap_path = tenant.conf.tenant_heatmap_path(tenant_shard_id);
+    let temp_path = path_with_suffix_extension(&heatmap_path, TEMP_FILE_SUFFIX);
+    if let Err(err) = VirtualFile::crashsafe_overwrite(heatmap_path, temp_path, bytes).await {
+        tracing::warn!("Non fatal IO error writing to disk after heatmap upload: {err}");
+    }
+
     tracing::info!("Successfully uploaded {size} byte heatmap to {path}");
 
     Ok(UploadHeatmapOutcome::Uploaded(LastUploadState {

pageserver/src/tenant/storage_layer.rs

@@ -33,6 +33,7 @@ use utils::sync::gate::GateGuard;
 
 use utils::lsn::Lsn;
 
+pub use batch_split_writer::{BatchLayerWriter, SplitDeltaLayerWriter, SplitImageLayerWriter};
 pub use delta_layer::{DeltaLayer, DeltaLayerWriter, ValueRef};
 pub use image_layer::{ImageLayer, ImageLayerWriter};
 pub use inmemory_layer::InMemoryLayer;
@@ -79,6 +80,16 @@ pub(crate) struct ValueReconstructState {
     pub(crate) img: Option<(Lsn, Bytes)>,
 }
 
+impl ValueReconstructState {
+    /// Returns the number of page deltas applied to the page image.
+    pub fn num_deltas(&self) -> usize {
+        match self.img {
+            Some(_) => self.records.len(),
+            None => self.records.len() - 1, // omit will_init record
+        }
+    }
+}
+
 #[derive(Clone, Copy, Debug, Default, Eq, PartialEq)]
 pub(crate) enum ValueReconstructSituation {
     Complete,

pageserver/src/tenant/storage_layer/batch_split_writer.rs

@@ -87,6 +87,23 @@ impl BatchLayerWriter {
         ));
     }
 
+    pub(crate) async fn finish(
+        self,
+        tline: &Arc<Timeline>,
+        ctx: &RequestContext,
+    ) -> anyhow::Result<Vec<ResidentLayer>> {
+        let res = self
+            .finish_with_discard_fn(tline, ctx, |_| async { false })
+            .await?;
+        let mut output = Vec::new();
+        for r in res {
+            if let BatchWriterResult::Produced(layer) = r {
+                output.push(layer);
+            }
+        }
+        Ok(output)
+    }
+
     pub(crate) async fn finish_with_discard_fn<D, F>(
         self,
         tline: &Arc<Timeline>,
@@ -149,6 +166,10 @@ impl BatchLayerWriter {
         // END: catch every error and do the recovery in the above section
         Ok(generated_layers)
     }
+
+    pub fn pending_layer_num(&self) -> usize {
+        self.generated_layer_writers.len()
+    }
 }
 
 /// An image writer that takes images and produces multiple image layers.

pageserver/src/tenant/storage_layer/layer.rs

@@ -340,7 +340,7 @@ impl Layer {
     /// Download the layer if evicted.
     ///
     /// Will not error when the layer is already downloaded.
-    pub(crate) async fn download(&self) -> anyhow::Result<()> {
+    pub(crate) async fn download(&self) -> Result<(), DownloadError> {
         self.0.get_or_maybe_download(true, None).await?;
         Ok(())
     }
@@ -701,13 +701,7 @@ impl Drop for LayerInner {
         if let Some(timeline) = timeline.as_ref() {
             // Only need to decrement metrics if the timeline still exists: otherwise
             // it will have already de-registered these metrics via TimelineMetrics::shutdown
-            if self.desc.is_delta() {
-                timeline.metrics.layer_count_delta.dec();
-                timeline.metrics.layer_size_delta.sub(self.desc.file_size);
-            } else {
-                timeline.metrics.layer_count_image.dec();
-                timeline.metrics.layer_size_image.sub(self.desc.file_size);
-            }
+            timeline.metrics.dec_layer(&self.desc);
 
             if matches!(self.access_stats.visibility(), LayerVisibilityHint::Visible) {
                 debug_assert!(
@@ -817,13 +811,7 @@ impl LayerInner {
         };
 
         // This object acts as a RAII guard on these metrics: increment on construction
-        if desc.is_delta() {
-            timeline.metrics.layer_count_delta.inc();
-            timeline.metrics.layer_size_delta.add(desc.file_size);
-        } else {
-            timeline.metrics.layer_count_image.inc();
-            timeline.metrics.layer_size_image.add(desc.file_size);
-        }
+        timeline.metrics.inc_layer(&desc);
 
         // New layers are visible by default. This metric is later updated on drop or in set_visibility
         timeline