wip

.devcontainer/Dockerfile.devcontainer

@@ -0,0 +1 @@
+FROM neondatabase/build-tools:pinned

.devcontainer/devcontainer.json

@@ -0,0 +1,23 @@
+// https://containers.dev/implementors/json_reference/
+{
+  "name": "Neon",
+  "build": {
+    "context": "..",
+    "dockerfile": "Dockerfile.devcontainer"
+  },
+
+  "postCreateCommand": {
+    "build neon": "BUILD_TYPE=debug CARGO_BUILD_FLAGS='--features=testing' mold -run make -s -j`nproc`",
+    "install python deps": "./scripts/pysync"
+  },
+
+  "customizations": {
+    "vscode": {
+      "extensions": [
+        "charliermarsh.ruff",
+        "github.vscode-github-actions",
+        "rust-lang.rust-analyzer"
+      ]
+    }
+  }
+}

.dockerignore

@@ -13,7 +13,6 @@
 # Directories
 !.cargo/
 !.config/
-!compute/
 !compute_tools/
 !control_plane/
 !libs/

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

@@ -62,7 +62,7 @@ jobs:
           #
           git config --global --add safe.directory ${{ github.workspace }}
           git config --global --add safe.directory ${GITHUB_WORKSPACE}
-          for r in 14 15 16 17; do
+          for r in 14 15 16; do
             git config --global --add safe.directory "${{ github.workspace }}/vendor/postgres-v$r"
             git config --global --add safe.directory "${GITHUB_WORKSPACE}/vendor/postgres-v$r"
           done
@@ -83,10 +83,6 @@ jobs:
         id: pg_v16_rev
         run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v16) >> $GITHUB_OUTPUT
 
-      - name: Set pg 17 revision for caching
-        id: pg_v17_rev
-        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v17) >> $GITHUB_OUTPUT
-
       # Set some environment variables used by all the steps.
       #
       # CARGO_FLAGS is extra options to pass to "cargo build", "cargo test" etc.
@@ -140,13 +136,6 @@ jobs:
           path: pg_install/v16
           key: v1-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-pg-${{ steps.pg_v16_rev.outputs.pg_rev }}-${{ hashFiles('Makefile', 'Dockerfile.build-tools') }}
 
-      - name: Cache postgres v17 build
-        id: cache_pg_17
-        uses: actions/cache@v4
-        with:
-          path: pg_install/v17
-          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-pg-${{ steps.pg_v17_rev.outputs.pg_rev }}-${{ hashFiles('Makefile', 'Dockerfile.build-tools') }}
-
       - name: Build postgres v14
         if: steps.cache_pg_14.outputs.cache-hit != 'true'
         run: mold -run make postgres-v14 -j$(nproc)
@@ -159,10 +148,6 @@ jobs:
         if: steps.cache_pg_16.outputs.cache-hit != 'true'
         run: mold -run make postgres-v16 -j$(nproc)
 
-      - name: Build postgres v17
-        if: steps.cache_pg_17.outputs.cache-hit != 'true'
-        run: mold -run make postgres-v17 -j$(nproc)
-
       - name: Build neon extensions
         run: mold -run make neon-pg-ext -j$(nproc)
 
@@ -225,7 +210,7 @@ jobs:
         run: |
           PQ_LIB_DIR=$(pwd)/pg_install/v16/lib
           export PQ_LIB_DIR
-          LD_LIBRARY_PATH=$(pwd)/pg_install/v17/lib
+          LD_LIBRARY_PATH=$(pwd)/pg_install/v16/lib
           export LD_LIBRARY_PATH
 
           #nextest does not yet support running doctests

.github/workflows/_push-to-acr.yml

@@ -52,5 +52,5 @@ jobs:
           for image in ${images}; do
             docker buildx imagetools create \
               -t ${{ inputs.registry_name }}.azurecr.io/neondatabase/${image}:${{ inputs.image_tag }} \
-                                                        neondatabase/${image}:${{ inputs.image_tag }}
+                                        neondatabase/${image}:${{ inputs.image_tag }}
           done

.github/workflows/build_and_test.yml

@@ -54,8 +54,8 @@ jobs:
       build-tag: ${{steps.build-tag.outputs.tag}}
 
     steps:
-      # Need `fetch-depth: 0` to count the number of commits in the branch
+      - name: Checkout
-      - uses: actions/checkout@v4
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -120,59 +120,6 @@ jobs:
       - name: Run mypy to check types
         run: poetry run mypy .
 
-  # Check that the vendor/postgres-* submodules point to the
-  # corresponding REL_*_STABLE_neon branches.
-  check-submodules:
-    runs-on: ubuntu-22.04
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - uses: dorny/paths-filter@v3
-        id: check-if-submodules-changed
-        with:
-          filters: |
-            vendor:
-              - 'vendor/**'
-
-      - name: Check vendor/postgres-v14 submodule reference
-        if: steps.check-if-submodules-changed.outputs.vendor == 'true'
-        uses: jtmullen/submodule-branch-check-action@v1
-        with:
-          path: "vendor/postgres-v14"
-          fetch_depth: "50"
-          sub_fetch_depth: "50"
-          pass_if_unchanged: true
-
-      - name: Check vendor/postgres-v15 submodule reference
-        if: steps.check-if-submodules-changed.outputs.vendor == 'true'
-        uses: jtmullen/submodule-branch-check-action@v1
-        with:
-          path: "vendor/postgres-v15"
-          fetch_depth: "50"
-          sub_fetch_depth: "50"
-          pass_if_unchanged: true
-
-      - name: Check vendor/postgres-v16 submodule reference
-        if: steps.check-if-submodules-changed.outputs.vendor == 'true'
-        uses: jtmullen/submodule-branch-check-action@v1
-        with:
-          path: "vendor/postgres-v16"
-          fetch_depth: "50"
-          sub_fetch_depth: "50"
-          pass_if_unchanged: true
-
-      - name: Check vendor/postgres-v17 submodule reference
-        if: steps.check-if-submodules-changed.outputs.vendor == 'true'
-        uses: jtmullen/submodule-branch-check-action@v1
-        with:
-          path: "vendor/postgres-v17"
-          fetch_depth: "50"
-          sub_fetch_depth: "50"
-          pass_if_unchanged: true
-
   check-codestyle-rust:
     needs: [ check-permissions, build-build-tools-image ]
     strategy:
@@ -212,10 +159,6 @@ jobs:
       # This will catch compiler & clippy warnings in all feature combinations.
       # TODO: use cargo hack for build and test as well, but, that's quite expensive.
       # NB: keep clippy args in sync with ./run_clippy.sh
-      #
-      # The only difference between "clippy --debug" and "clippy --release" is that in --release mode,
-      # #[cfg(debug_assertions)] blocks are not built. It's not worth building everything for second
-      # time just for that, so skip "clippy --release".
       - run: |
           CLIPPY_COMMON_ARGS="$( source .neon_clippy_args; echo "$CLIPPY_COMMON_ARGS")"
           if [ "$CLIPPY_COMMON_ARGS" = "" ]; then
@@ -225,6 +168,8 @@ jobs:
           echo "CLIPPY_COMMON_ARGS=${CLIPPY_COMMON_ARGS}" >> $GITHUB_ENV
       - name: Run cargo clippy (debug)
         run: cargo hack --feature-powerset clippy $CLIPPY_COMMON_ARGS
+      - name: Run cargo clippy (release)
+        run: cargo hack --feature-powerset clippy --release $CLIPPY_COMMON_ARGS
 
       - name: Check documentation generation
         run: cargo doc --workspace --no-deps --document-private-items
@@ -266,7 +211,7 @@ jobs:
       build-tag: ${{ needs.tag.outputs.build-tag }}
       build-type: ${{ matrix.build-type }}
       # Run tests on all Postgres versions in release builds and only on the latest version in debug builds
-      pg-versions: ${{ matrix.build-type == 'release' && '["v14", "v15", "v16", "v17"]' || '["v17"]' }}
+      pg-versions: ${{ matrix.build-type == 'release' && '["v14", "v15", "v16"]' || '["v16"]' }}
     secrets: inherit
 
   # Keep `benchmarks` job outside of `build-and-test-locally` workflow to make job failures non-blocking
@@ -412,7 +357,6 @@ jobs:
             })
 
   coverage-report:
-    if: ${{ !startsWith(github.ref_name, 'release') }}
     needs: [ check-permissions, build-build-tools-image, build-and-test-locally ]
     runs-on: [ self-hosted, small ]
     container:
@@ -429,8 +373,8 @@ jobs:
         coverage-html: ${{ steps.upload-coverage-report-new.outputs.report-url }}
         coverage-json: ${{ steps.upload-coverage-report-new.outputs.summary-json }}
     steps:
-      # Need `fetch-depth: 0` for differential coverage (to get diff between two commits)
+      - name: Checkout
-      - uses: actions/checkout@v4
+        uses: actions/checkout@v4
         with:
           submodules: true
           fetch-depth: 0
@@ -531,9 +475,11 @@ jobs:
     runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
 
     steps:
-      - uses: actions/checkout@v4
+      - name: Checkout
+        uses: actions/checkout@v4
         with:
           submodules: true
+          fetch-depth: 0
 
       - uses: ./.github/actions/set-docker-config-dir
       - uses: docker/setup-buildx-action@v3
@@ -602,28 +548,17 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        version:
+        version: [ v14, v15, v16 ]
-          # Much data was already generated on old PG versions with bullseye's
-          # libraries, the locales of which can cause data incompatibilities.
-          # However, new PG versions should check if they can be built on newer
-          # images, as that reduces the support burden of old and ancient
-          # distros.
-          - pg: v14
-            debian: bullseye-slim
-          - pg: v15
-            debian: bullseye-slim
-          - pg: v16
-            debian: bullseye-slim
-          - pg: v17
-            debian: bookworm-slim
         arch: [ x64, arm64 ]
 
     runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
 
     steps:
-      - uses: actions/checkout@v4
+      - name: Checkout
+        uses: actions/checkout@v4
         with:
           submodules: true
+          fetch-depth: 0
 
       - uses: ./.github/actions/set-docker-config-dir
       - uses: docker/setup-buildx-action@v3
@@ -658,46 +593,41 @@ jobs:
           context: .
           build-args: |
             GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
-            PG_VERSION=${{ matrix.version.pg }}
+            PG_VERSION=${{ matrix.version }}
             BUILD_TAG=${{ needs.tag.outputs.build-tag }}
             TAG=${{ needs.build-build-tools-image.outputs.image-tag }}
-            DEBIAN_FLAVOR=${{ matrix.version.debian }}
           provenance: false
           push: true
           pull: true
-          file: compute/Dockerfile.compute-node
+          file: Dockerfile.compute-node
-          cache-from: type=registry,ref=cache.neon.build/compute-node-${{ matrix.version.pg }}:cache-${{ matrix.arch }}
+          cache-from: type=registry,ref=cache.neon.build/compute-node-${{ matrix.version }}:cache-${{ matrix.arch }}
-          cache-to: ${{ github.ref_name == 'main' && format('type=registry,ref=cache.neon.build/compute-node-{0}:cache-{1},mode=max', matrix.version.pg, matrix.arch) || '' }}
+          cache-to: ${{ github.ref_name == 'main' && format('type=registry,ref=cache.neon.build/compute-node-{0}:cache-{1},mode=max', matrix.version, matrix.arch) || '' }}
           tags: |
-            neondatabase/compute-node-${{ matrix.version.pg }}:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}
+            neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}
 
       - name: Build neon extensions test image
-        if: matrix.version.pg == 'v16'
+        if: matrix.version == 'v16'
         uses: docker/build-push-action@v6
         with:
           context: .
           build-args: |
             GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
-            PG_VERSION=${{ matrix.version.pg }}
+            PG_VERSION=${{ matrix.version }}
             BUILD_TAG=${{ needs.tag.outputs.build-tag }}
             TAG=${{ needs.build-build-tools-image.outputs.image-tag }}
-            DEBIAN_FLAVOR=${{ matrix.version.debian }}
           provenance: false
           push: true
           pull: true
-          file: compute/Dockerfile.compute-node
+          file: Dockerfile.compute-node
           target: neon-pg-ext-test
-          cache-from: type=registry,ref=cache.neon.build/neon-test-extensions-${{ matrix.version.pg }}:cache-${{ matrix.arch }}
+          cache-from: type=registry,ref=cache.neon.build/neon-test-extensions-${{ matrix.version }}:cache-${{ matrix.arch }}
-          cache-to: ${{ github.ref_name == 'main' && format('type=registry,ref=cache.neon.build/neon-test-extensions-{0}:cache-{1},mode=max', matrix.version.pg, matrix.arch) || '' }}
+          cache-to: ${{ github.ref_name == 'main' && format('type=registry,ref=cache.neon.build/neon-test-extensions-{0}:cache-{1},mode=max', matrix.version, matrix.arch) || '' }}
           tags: |
-            neondatabase/neon-test-extensions-${{ matrix.version.pg }}:${{needs.tag.outputs.build-tag}}-${{ matrix.arch }}
+            neondatabase/neon-test-extensions-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}-${{ matrix.arch }}
 
       - name: Build compute-tools image
         # compute-tools are Postgres independent, so build it only once
-        # We pick 16, because that builds on debian 11 with older glibc (and is
+        if: matrix.version == 'v16'
-        # thus compatible with newer glibc), rather than 17 on Debian 12, as
-        # that isn't guaranteed to be compatible with Debian 11
-        if: matrix.version.pg == 'v16'
         uses: docker/build-push-action@v6
         with:
           target: compute-tools-image
@@ -706,11 +636,10 @@ jobs:
             GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
             BUILD_TAG=${{ needs.tag.outputs.build-tag }}
             TAG=${{ needs.build-build-tools-image.outputs.image-tag }}
-            DEBIAN_FLAVOR=${{ matrix.version.debian }}
           provenance: false
           push: true
           pull: true
-          file: compute/Dockerfile.compute-node
+          file: Dockerfile.compute-node
           tags: |
             neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}
 
@@ -720,7 +649,7 @@ jobs:
 
     strategy:
       matrix:
-        version: [ v14, v15, v16, v17 ]
+        version: [ v14, v15, v16 ]
 
     steps:
       - uses: docker/login-action@v3
@@ -742,7 +671,7 @@ jobs:
                                              neondatabase/neon-test-extensions-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-arm64
 
       - name: Create multi-arch compute-tools image
-        if: matrix.version == 'v17'
+        if: matrix.version == 'v16'
         run: |
           docker buildx imagetools create -t neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }} \
                                              neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}-x64 \
@@ -760,7 +689,7 @@ jobs:
                                                                                 neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}
 
       - name: Push multi-arch compute-tools image to ECR
-        if: matrix.version == 'v17'
+        if: matrix.version == 'v16'
         run: |
           docker buildx imagetools create -t 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{ needs.tag.outputs.build-tag }} \
                                                                                 neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}
@@ -771,12 +700,15 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        version: [ v14, v15, v16, v17 ]
+        version: [ v14, v15, v16 ]
     env:
       VM_BUILDER_VERSION: v0.29.3
 
     steps:
-      - uses: actions/checkout@v4
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
 
       - name: Downloading vm-builder
         run: |
@@ -798,7 +730,7 @@ jobs:
       - name: Build vm image
         run: |
           ./vm-builder \
-            -spec=compute/vm-image-spec.yaml \
+            -spec=vm-image-spec.yaml \
             -src=neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }} \
             -dst=neondatabase/vm-compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}
 
@@ -816,7 +748,10 @@ jobs:
     runs-on: ${{ fromJson(format('["self-hosted", "{0}"]', matrix.arch == 'arm64' && 'small-arm64' || 'small')) }}
 
     steps:
-      - uses: actions/checkout@v4
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
 
       - uses: ./.github/actions/set-docker-config-dir
       - uses: docker/login-action@v3
@@ -862,11 +797,8 @@ jobs:
     needs: [ check-permissions, tag, test-images, vm-compute-node-image ]
     runs-on: ubuntu-22.04
 
-    permissions:
-      id-token: write # for `aws-actions/configure-aws-credentials`
-
     env:
-      VERSIONS: v14 v15 v16 v17
+      VERSIONS: v14 v15 v16
 
     steps:
       - uses: docker/login-action@v3
@@ -907,26 +839,20 @@ jobs:
             done
           done
           docker buildx imagetools create -t neondatabase/neon-test-extensions-v16:latest \
-                                              neondatabase/neon-test-extensions-v16:${{ needs.tag.outputs.build-tag }}
+                                             neondatabase/neon-test-extensions-v16:${{ needs.tag.outputs.build-tag }}
-
-      - name: Configure AWS-prod credentials
-        if: github.ref_name == 'release'|| github.ref_name == 'release-proxy'
-        uses: aws-actions/configure-aws-credentials@v4
-        with:
-          aws-region: eu-central-1
-          mask-aws-account-id: true
-          role-to-assume: ${{ secrets.PROD_GHA_OIDC_ROLE }}
 
       - name: Login to prod ECR
         uses: docker/login-action@v3
         if: github.ref_name == 'release'|| github.ref_name == 'release-proxy'
         with:
           registry: 093970136003.dkr.ecr.eu-central-1.amazonaws.com
+          username: ${{ secrets.PROD_GHA_RUNNER_LIMITED_AWS_ACCESS_KEY_ID }}
+          password: ${{ secrets.PROD_GHA_RUNNER_LIMITED_AWS_SECRET_ACCESS_KEY }}
 
       - name: Copy all images to prod ECR
         if: github.ref_name == 'release'|| github.ref_name == 'release-proxy'
         run: |
-          for image in neon compute-tools {vm-,}compute-node-{v14,v15,v16,v17}; do
+          for image in neon compute-tools {vm-,}compute-node-{v14,v15,v16}; do
             docker buildx imagetools create -t 093970136003.dkr.ecr.eu-central-1.amazonaws.com/${image}:${{ needs.tag.outputs.build-tag }} \
                                                369495373322.dkr.ecr.eu-central-1.amazonaws.com/${image}:${{ needs.tag.outputs.build-tag }}
           done
@@ -938,7 +864,7 @@ jobs:
     with:
       client_id: ${{ vars.AZURE_DEV_CLIENT_ID }}
       image_tag: ${{ needs.tag.outputs.build-tag }}
-      images: neon compute-tools vm-compute-node-v14 vm-compute-node-v15 vm-compute-node-v16 vm-compute-node-v17 compute-node-v14 compute-node-v15 compute-node-v16 compute-node-v17
+      images: neon compute-tools vm-compute-node-v14 vm-compute-node-v15 vm-compute-node-v16 compute-node-v14 compute-node-v15 compute-node-v16
       registry_name: ${{ vars.AZURE_DEV_REGISTRY_NAME }}
       subscription_id: ${{ vars.AZURE_DEV_SUBSCRIPTION_ID }}
       tenant_id: ${{ vars.AZURE_TENANT_ID }}
@@ -950,7 +876,7 @@ jobs:
     with:
       client_id: ${{ vars.AZURE_PROD_CLIENT_ID }}
       image_tag: ${{ needs.tag.outputs.build-tag }}
-      images: neon compute-tools vm-compute-node-v14 vm-compute-node-v15 vm-compute-node-v16 vm-compute-node-v17 compute-node-v14 compute-node-v15 compute-node-v16 compute-node-v17
+      images: neon compute-tools vm-compute-node-v14 vm-compute-node-v15 vm-compute-node-v16 compute-node-v14 compute-node-v15 compute-node-v16
       registry_name: ${{ vars.AZURE_PROD_REGISTRY_NAME }}
       subscription_id: ${{ vars.AZURE_PROD_SUBSCRIPTION_ID }}
       tenant_id: ${{ vars.AZURE_TENANT_ID }}
@@ -1031,7 +957,6 @@ jobs:
 
   deploy:
     needs: [ check-permissions, promote-images, tag, build-and-test-locally, trigger-custom-extensions-build-and-wait, push-to-acr-dev, push-to-acr-prod ]
-    # `!failure() && !cancelled()` is required because the workflow depends on the job that can be skipped: `push-to-acr-dev` and `push-to-acr-prod`
     if: (github.ref_name == 'main' || github.ref_name == 'release' || github.ref_name == 'release-proxy') && !failure() && !cancelled()
 
     runs-on: [ self-hosted, small ]
@@ -1046,12 +971,15 @@ jobs:
           #
           git config --global --add safe.directory ${{ github.workspace }}
           git config --global --add safe.directory ${GITHUB_WORKSPACE}
-          for r in 14 15 16 17; do
+          for r in 14 15 16; do
             git config --global --add safe.directory "${{ github.workspace }}/vendor/postgres-v$r"
             git config --global --add safe.directory "${GITHUB_WORKSPACE}/vendor/postgres-v$r"
           done
 
-      - uses: actions/checkout@v4
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
 
       - name: Trigger deploy workflow
         env:
@@ -1130,8 +1058,7 @@ jobs:
   # The job runs on `release` branch and copies compatibility data and Neon artifact from the last *release PR* to the latest directory
   promote-compatibility-data:
     needs: [ deploy ]
-    # `!failure() && !cancelled()` is required because the workflow transitively depends on the job that can be skipped: `push-to-acr-dev` and `push-to-acr-prod`
+    if: github.ref_name == 'release'
-    if: github.ref_name == 'release' && !failure() && !cancelled()
 
     runs-on: ubuntu-22.04
     steps:
@@ -1190,7 +1117,6 @@ jobs:
 
               files_to_promote+=("s3://${BUCKET}/${s3_key}")
 
-              # TODO Add v17
               for pg_version in v14 v15 v16; do
                 # We run less tests for debug builds, so we don't need to promote them
                 if [ "${build_type}" == "debug" ] && { [ "${arch}" == "ARM64" ] || [ "${pg_version}" != "v16" ] ; }; then

.github/workflows/cloud-regress.yml

@@ -1,102 +0,0 @@
-name: Cloud Regression Test
-on:
-  schedule:
-    # * is a special character in YAML so you have to quote this string
-    #          ┌───────────── minute (0 - 59)
-    #          │ ┌───────────── hour (0 - 23)
-    #          │ │ ┌───────────── day of the month (1 - 31)
-    #          │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
-    #          │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
-    - cron:  '45 1 * * *' # run once a day, timezone is utc
-  workflow_dispatch: # adds ability to run this manually
-
-defaults:
-  run:
-    shell: bash -euxo pipefail {0}
-
-concurrency:
-  # Allow only one workflow
-  group: ${{ github.workflow }}
-  cancel-in-progress: true
-
-jobs:
-  regress:
-    env:
-      POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
-      DEFAULT_PG_VERSION: 16
-      TEST_OUTPUT: /tmp/test_output
-      BUILD_TYPE: remote
-      AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_DEV }}
-      AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_KEY_DEV }}
-
-    runs-on: us-east-2
-    container:
-      image: neondatabase/build-tools:pinned
-      options: --init
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Patch the test
-        run: |
-          cd "vendor/postgres-v${DEFAULT_PG_VERSION}"
-          patch -p1 < "../../compute/patches/cloud_regress_pg${DEFAULT_PG_VERSION}.patch"
-
-      - name: Generate a random password
-        id: pwgen
-        run: |
-          set +x
-          DBPASS=$(dd if=/dev/random bs=48 count=1 2>/dev/null | base64)
-          echo "::add-mask::${DBPASS//\//}"
-          echo DBPASS="${DBPASS//\//}" >> "${GITHUB_OUTPUT}"
-
-      - name: Change tests according to the generated password
-        env:
-          DBPASS: ${{ steps.pwgen.outputs.DBPASS }}
-        run: |
-          cd vendor/postgres-v"${DEFAULT_PG_VERSION}"/src/test/regress
-          for fname in sql/*.sql expected/*.out; do
-            sed -i.bak s/NEON_PASSWORD_PLACEHOLDER/"'${DBPASS}'"/ "${fname}"
-          done
-          for ph in $(grep NEON_MD5_PLACEHOLDER expected/password.out | awk '{print $3;}' | sort | uniq); do
-            USER=$(echo "${ph}" | cut -c 22-)
-            MD5=md5$(echo -n "${DBPASS}${USER}" | md5sum | awk '{print $1;}')
-            sed -i.bak "s/${ph}/${MD5}/" expected/password.out
-          done
-
-      - name: Download Neon artifact
-        uses: ./.github/actions/download
-        with:
-          name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
-          path: /tmp/neon/
-          prefix: latest
-
-      - name: Run the regression tests
-        uses: ./.github/actions/run-python-test-set
-        with:
-          build_type: ${{ env.BUILD_TYPE }}
-          test_selection: cloud_regress
-          pg_version: ${{ env.DEFAULT_PG_VERSION }}
-          extra_params: -m remote_cluster
-        env:
-          BENCHMARK_CONNSTR: ${{ secrets.PG_REGRESS_CONNSTR }}
-
-      - name: Create Allure report
-        id: create-allure-report
-        if: ${{ !cancelled() }}
-        uses: ./.github/actions/allure-report-generate
-
-      - name: Post to a Slack channel
-        if: ${{ github.event.schedule && failure() }}
-        uses: slackapi/slack-github-action@v1
-        with:
-          channel-id: "C033QLM5P7D" # on-call-staging-stream
-          slack-message: |
-            Periodic pg_regress on staging: ${{ job.status }}
-            <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
-            <${{ steps.create-allure-report.outputs.report-url }}|Allure report>
-        env:
-          SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
-

.github/workflows/neon_extra_builds.yml

@@ -72,10 +72,6 @@ jobs:
         id: pg_v16_rev
         run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v16) >> $GITHUB_OUTPUT
 
-      - name: Set pg 17 revision for caching
-        id: pg_v17_rev
-        run: echo pg_rev=$(git rev-parse HEAD:vendor/postgres-v17) >> $GITHUB_OUTPUT
-
       - name: Cache postgres v14 build
         id: cache_pg_14
         uses: actions/cache@v4
@@ -97,13 +93,6 @@ jobs:
           path: pg_install/v16
           key: v1-${{ runner.os }}-${{ runner.arch }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v16_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
 
-      - name: Cache postgres v17 build
-        id: cache_pg_17
-        uses: actions/cache@v4
-        with:
-          path: pg_install/v17
-          key: v1-${{ runner.os }}-${{ runner.arch }}-${{ env.BUILD_TYPE }}-pg-${{ steps.pg_v17_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
-
       - name: Set extra env for macOS
         run: |
           echo 'LDFLAGS=-L/usr/local/opt/openssl@3/lib' >> $GITHUB_ENV
@@ -131,10 +120,6 @@ jobs:
         if: steps.cache_pg_16.outputs.cache-hit != 'true'
         run: make postgres-v16 -j$(sysctl -n hw.ncpu)
 
-      - name: Build postgres v17
-        if: steps.cache_pg_17.outputs.cache-hit != 'true'
-        run: make postgres-v17 -j$(sysctl -n hw.ncpu)
-
       - name: Build neon extensions
         run: make neon-pg-ext -j$(sysctl -n hw.ncpu)
 
@@ -181,7 +166,7 @@ jobs:
         run: make walproposer-lib -j$(nproc)
 
       - name: Produce the build stats
-        run: PQ_LIB_DIR=$(pwd)/pg_install/v17/lib cargo build --all --release --timings -j$(nproc)
+        run: PQ_LIB_DIR=$(pwd)/pg_install/v16/lib cargo build --all --release --timings -j$(nproc)
 
       - name: Upload the build stats
         id: upload-stats

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

@@ -34,8 +34,8 @@ jobs:
       build-tag: ${{ steps.build-tag.outputs.tag }}
 
     steps:
-      # Need `fetch-depth: 0` to count the number of commits in the branch
+      - name: Checkout
-      - uses: actions/checkout@v4
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -102,12 +102,12 @@ jobs:
           # Default set of platforms to run e2e tests on
           platforms='["docker", "k8s"]'
 
-          # If the PR changes vendor/, pgxn/ or libs/vm_monitor/ directories, or compute/Dockerfile.compute-node, add k8s-neonvm to the list of platforms.
+          # If the PR changes vendor/, pgxn/ or libs/vm_monitor/ directories, or Dockerfile.compute-node, add k8s-neonvm to the list of platforms.
           # If the workflow run is not a pull request, add k8s-neonvm to the list.
           if [ "$GITHUB_EVENT_NAME" == "pull_request" ]; then
             for f in $(gh api "/repos/${GITHUB_REPOSITORY}/pulls/${PR_NUMBER}/files" --paginate --jq '.[].filename'); do
               case "$f" in
-                vendor/*|pgxn/*|libs/vm_monitor/*|compute/Dockerfile.compute-node)
+                vendor/*|pgxn/*|libs/vm_monitor/*|Dockerfile.compute-node)
                   platforms=$(echo "${platforms}" | jq --compact-output '. += ["k8s-neonvm"] | unique')
                   ;;
                 *)

.gitmodules

@@ -10,7 +10,3 @@
 	path = vendor/postgres-v16
 	url = https://github.com/neondatabase/postgres.git
 	branch = REL_16_STABLE_neon
-[submodule "vendor/postgres-v17"]
-	path = vendor/postgres-v17
-	url = https://github.com/neondatabase/postgres.git
-	branch = REL_17_STABLE_neon

Cargo.toml

@@ -76,6 +76,8 @@ clap = { version = "4.0", features = ["derive"] }
 comfy-table = "7.1"
 const_format = "0.2"
 crc32c = "0.6"
+crossbeam-deque = "0.8.5"
+crossbeam-utils = "0.8.5"
 dashmap = { version = "5.5.0", features = ["raw-api"] }
 either = "1.8"
 enum-map = "2.4.2"
@@ -93,7 +95,7 @@ hdrhistogram = "7.5.2"
 hex = "0.4"
 hex-literal = "0.4"
 hmac = "0.12.1"
-hostname = "0.4"
+hostname = "0.3.1"
 http = {version = "1.1.0", features = ["std"]}
 http-types = { version = "2", default-features = false }
 humantime = "2.1"
@@ -102,6 +104,7 @@ hyper = "0.14"
 tokio-tungstenite = "0.20.0"
 indexmap = "2"
 indoc = "2"
+inotify = "0.10.2"
 ipnet = "2.9.0"
 itertools = "0.10"
 jsonwebtoken = "9"
@@ -110,7 +113,7 @@ libc = "0.2"
 md5 = "0.7.0"
 measured = { version = "0.0.22", features=["lasso"] }
 measured-process = { version = "0.0.22" }
-memoffset = "0.9"
+memoffset = "0.8"
 nix = { version = "0.27", features = ["dir", "fs", "process", "socket", "signal", "poll"] }
 notify = "6.0.0"
 num_cpus = "1.15"
@@ -139,6 +142,7 @@ rpds = "0.13"
 rustc-hash = "1.1.0"
 rustls = "0.22"
 rustls-pemfile = "2"
+rustls-split = "0.3"
 scopeguard = "1.1"
 sysinfo = "0.29.2"
 sd-notify = "0.4.1"
@@ -160,6 +164,7 @@ strum_macros = "0.26"
 svg_fmt = "0.4.3"
 sync_wrapper = "0.1.2"
 tar = "0.4"
+task-local-extensions = "0.1.4"
 test-context = "0.3"
 thiserror = "1.0"
 tikv-jemallocator = "0.5"

Dockerfile

@@ -5,8 +5,6 @@
 ARG REPOSITORY=neondatabase
 ARG IMAGE=build-tools
 ARG TAG=pinned
-ARG DEFAULT_PG_VERSION=17
-ARG STABLE_PG_VERSION=16
 
 # Build Postgres
 FROM $REPOSITORY/$IMAGE:$TAG AS pg-build
@@ -15,7 +13,6 @@ WORKDIR /home/nonroot
 COPY --chown=nonroot vendor/postgres-v14 vendor/postgres-v14
 COPY --chown=nonroot vendor/postgres-v15 vendor/postgres-v15
 COPY --chown=nonroot vendor/postgres-v16 vendor/postgres-v16
-COPY --chown=nonroot vendor/postgres-v17 vendor/postgres-v17
 COPY --chown=nonroot pgxn pgxn
 COPY --chown=nonroot Makefile Makefile
 COPY --chown=nonroot scripts/ninstall.sh scripts/ninstall.sh
@@ -31,19 +28,16 @@ FROM $REPOSITORY/$IMAGE:$TAG AS build
 WORKDIR /home/nonroot
 ARG GIT_VERSION=local
 ARG BUILD_TAG
-ARG STABLE_PG_VERSION
 
 COPY --from=pg-build /home/nonroot/pg_install/v14/include/postgresql/server pg_install/v14/include/postgresql/server
 COPY --from=pg-build /home/nonroot/pg_install/v15/include/postgresql/server pg_install/v15/include/postgresql/server
 COPY --from=pg-build /home/nonroot/pg_install/v16/include/postgresql/server pg_install/v16/include/postgresql/server
-COPY --from=pg-build /home/nonroot/pg_install/v17/include/postgresql/server pg_install/v17/include/postgresql/server
 COPY --from=pg-build /home/nonroot/pg_install/v16/lib                       pg_install/v16/lib
-COPY --from=pg-build /home/nonroot/pg_install/v17/lib                       pg_install/v17/lib
 COPY --chown=nonroot . .
 
 ARG ADDITIONAL_RUSTFLAGS
 RUN set -e \
-    && PQ_LIB_DIR=$(pwd)/pg_install/v${STABLE_PG_VERSION}/lib RUSTFLAGS="-Clinker=clang -Clink-arg=-fuse-ld=mold -Clink-arg=-Wl,--no-rosegment ${ADDITIONAL_RUSTFLAGS}" cargo build \
+    && PQ_LIB_DIR=$(pwd)/pg_install/v16/lib RUSTFLAGS="-Clinker=clang -Clink-arg=-fuse-ld=mold -Clink-arg=-Wl,--no-rosegment ${ADDITIONAL_RUSTFLAGS}" cargo build \
       --bin pg_sni_router  \
       --bin pageserver  \
       --bin pagectl  \
@@ -58,7 +52,6 @@ RUN set -e \
 # Build final image
 #
 FROM debian:bullseye-slim
-ARG DEFAULT_PG_VERSION
 WORKDIR /data
 
 RUN set -e \
@@ -84,7 +77,6 @@ COPY --from=build --chown=neon:neon /home/nonroot/target/release/storage_scrubbe
 COPY --from=pg-build /home/nonroot/pg_install/v14 /usr/local/v14/
 COPY --from=pg-build /home/nonroot/pg_install/v15 /usr/local/v15/
 COPY --from=pg-build /home/nonroot/pg_install/v16 /usr/local/v16/
-COPY --from=pg-build /home/nonroot/pg_install/v17 /usr/local/v17/
 COPY --from=pg-build /home/nonroot/postgres_install.tar.gz /data/
 
 # By default, pageserver uses `.neon/` working directory in WORKDIR, so create one and fill it with the dummy config.
@@ -101,7 +93,7 @@ RUN mkdir -p /data/.neon/ && \
 
 # When running a binary that links with libpq, default to using our most recent postgres version.  Binaries
 # that want a particular postgres version will select it explicitly: this is just a default.
-ENV LD_LIBRARY_PATH=/usr/local/v${DEFAULT_PG_VERSION}/lib
+ENV LD_LIBRARY_PATH=/usr/local/v16/lib
 
 
 VOLUME ["/data"]

Dockerfile.compute-node

@@ -3,15 +3,13 @@ ARG REPOSITORY=neondatabase
 ARG IMAGE=build-tools
 ARG TAG=pinned
 ARG BUILD_TAG
-ARG DEBIAN_FLAVOR=bullseye-slim
 
 #########################################################################################
 #
 # Layer "build-deps"
 #
 #########################################################################################
-FROM debian:$DEBIAN_FLAVOR AS build-deps
+FROM debian:bullseye-slim AS build-deps
-ARG DEBIAN_FLAVOR
 RUN apt update &&  \
     apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev \
     zlib1g-dev libxml2-dev libcurl4-openssl-dev libossp-uuid-dev wget pkg-config libssl-dev \
@@ -57,27 +55,22 @@ RUN cd postgres && \
     # We could add the additional grant statements to the postgres repository but it would be hard to maintain,
     # whenever we need to pick up a new postgres version and we want to limit the changes in our postgres fork,
     # so we do it here.
+    old_list="pg_stat_statements--1.0--1.1.sql pg_stat_statements--1.1--1.2.sql pg_stat_statements--1.2--1.3.sql pg_stat_statements--1.3--1.4.sql pg_stat_statements--1.4--1.5.sql pg_stat_statements--1.4.sql pg_stat_statements--1.5--1.6.sql"; \
+    # the first loop is for pg_stat_statement extension version <= 1.6
     for file in /usr/local/pgsql/share/extension/pg_stat_statements--*.sql; do \
         filename=$(basename "$file"); \
-        # Note that there are no downgrade scripts for pg_stat_statements, so we \
+        if echo "$old_list" | grep -q -F "$filename"; then \
-        # don't have to modify any downgrade paths or (much) older versions: we only \
-        # have to make sure every creation of the pg_stat_statements_reset function \
-        # also adds execute permissions to the neon_superuser.
-        case $filename in \
-          pg_stat_statements--1.4.sql) \
-            # pg_stat_statements_reset is first created with 1.4
             echo 'GRANT EXECUTE ON FUNCTION pg_stat_statements_reset() TO neon_superuser;' >> $file; \
-            ;; \
+        fi; \
-          pg_stat_statements--1.6--1.7.sql) \
+    done; \
-            # Then with the 1.6-1.7 migration it is re-created with a new signature, thus add the permissions back
+    # the second loop is for pg_stat_statement extension versions >= 1.7,
+    # where pg_stat_statement_reset() got 3 additional arguments
+    for file in /usr/local/pgsql/share/extension/pg_stat_statements--*.sql; do \
+        filename=$(basename "$file"); \
+        if ! echo "$old_list" | grep -q -F "$filename"; then \
             echo 'GRANT EXECUTE ON FUNCTION pg_stat_statements_reset(Oid, Oid, bigint) TO neon_superuser;' >> $file; \
-            ;; \
+        fi; \
-          pg_stat_statements--1.10--1.11.sql) \
+    done
-            # Then with the 1.10-1.11 migration it is re-created with a new signature again, thus add the permissions back
-            echo 'GRANT EXECUTE ON FUNCTION pg_stat_statements_reset(Oid, Oid, bigint, boolean) TO neon_superuser;' >> $file; \
-            ;; \
-        esac; \
-    done;
 
 #########################################################################################
 #
@@ -86,7 +79,6 @@ RUN cd postgres && \
 #
 #########################################################################################
 FROM build-deps AS postgis-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 RUN apt update && \
     apt install -y cmake gdal-bin libboost-dev libboost-thread-dev libboost-filesystem-dev \
@@ -95,11 +87,7 @@ RUN apt update && \
     protobuf-c-compiler xsltproc
 
 # SFCGAL > 1.3 requires CGAL > 5.2, Bullseye's libcgal-dev is 5.2
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://gitlab.com/Oslandia/SFCGAL/-/archive/v1.3.10/SFCGAL-v1.3.10.tar.gz -O SFCGAL.tar.gz && \
-    mkdir -p /sfcgal && \
-    echo "Postgis doensn't yet support PG17 (needs 3.4.3, if not higher)" && exit 0;; \
-    esac && \
-    wget https://gitlab.com/Oslandia/SFCGAL/-/archive/v1.3.10/SFCGAL-v1.3.10.tar.gz -O SFCGAL.tar.gz && \
     echo "4e39b3b2adada6254a7bdba6d297bb28e1a9835a9f879b74f37e2dab70203232 SFCGAL.tar.gz" | sha256sum --check && \
     mkdir sfcgal-src && cd sfcgal-src && tar xzf ../SFCGAL.tar.gz --strip-components=1 -C . && \
     cmake -DCMAKE_BUILD_TYPE=Release . && make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -108,10 +96,7 @@ RUN case "${PG_VERSION}" in "v17") \
 
 ENV PATH="/usr/local/pgsql/bin:$PATH"
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.3.tar.gz -O postgis.tar.gz && \
-    echo "Postgis doensn't yet support PG17 (needs 3.4.3, if not higher)" && exit 0;; \
-    esac && \
-    wget https://download.osgeo.org/postgis/source/postgis-3.3.3.tar.gz -O postgis.tar.gz && \
     echo "74eb356e3f85f14233791013360881b6748f78081cc688ff9d6f0f673a762d13 postgis.tar.gz" | sha256sum --check && \
     mkdir postgis-src && cd postgis-src && tar xzf ../postgis.tar.gz --strip-components=1 -C . && \
     find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
@@ -137,10 +122,7 @@ RUN case "${PG_VERSION}" in "v17") \
     cp /usr/local/pgsql/share/extension/address_standardizer.control /extensions/postgis && \
     cp /usr/local/pgsql/share/extension/address_standardizer_data_us.control /extensions/postgis
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/pgRouting/pgrouting/archive/v3.4.2.tar.gz -O pgrouting.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/pgRouting/pgrouting/archive/v3.4.2.tar.gz -O pgrouting.tar.gz && \
     echo "cac297c07d34460887c4f3b522b35c470138760fe358e351ad1db4edb6ee306e pgrouting.tar.gz" | sha256sum --check && \
     mkdir pgrouting-src && cd pgrouting-src && tar xzf ../pgrouting.tar.gz --strip-components=1 -C . && \
     mkdir build && cd build && \
@@ -160,19 +142,12 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS plv8-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN apt update && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    apt update && \
     apt install -y ninja-build python3-dev libncurses5 binutils clang
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.10.tar.gz -O plv8.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.10.tar.gz -O plv8.tar.gz && \
     echo "7096c3290928561f0d4901b7a52794295dc47f6303102fae3f8e42dd575ad97d plv8.tar.gz" | sha256sum --check && \
     mkdir plv8-src && cd plv8-src && tar xzf ../plv8.tar.gz --strip-components=1 -C . && \
     # generate and copy upgrade scripts
@@ -197,13 +172,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS h3-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN case "$(uname -m)" in \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    case "$(uname -m)" in \
       "x86_64") \
         export CMAKE_CHECKSUM=739d372726cb23129d57a539ce1432453448816e345e1545f6127296926b6754 \
         ;; \
@@ -221,11 +192,7 @@ RUN case "${PG_VERSION}" in "v17") \
       && /tmp/cmake-install.sh --skip-license --prefix=/usr/local/ \
       && rm /tmp/cmake-install.sh
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz && \
-        mkdir -p /h3/usr/ && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz && \
     echo "ec99f1f5974846bde64f4513cf8d2ea1b8d172d2218ab41803bf6a63532272bc h3.tar.gz" | sha256sum --check && \
     mkdir h3-src && cd h3-src && tar xzf ../h3.tar.gz --strip-components=1 -C . && \
     mkdir build && cd build && \
@@ -235,10 +202,7 @@ RUN case "${PG_VERSION}" in "v17") \
     cp -R /h3/usr / && \
     rm -rf build
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.1.3.tar.gz -O h3-pg.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.1.3.tar.gz -O h3-pg.tar.gz && \
     echo "5c17f09a820859ffe949f847bebf1be98511fb8f1bd86f94932512c00479e324 h3-pg.tar.gz" | sha256sum --check && \
     mkdir h3-pg-src && cd h3-pg-src && tar xzf ../h3-pg.tar.gz --strip-components=1 -C . && \
     export PATH="/usr/local/pgsql/bin:$PATH" && \
@@ -254,13 +218,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS unit-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -O postgresql-unit.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -O postgresql-unit.tar.gz && \
     echo "411d05beeb97e5a4abf17572bfcfbb5a68d98d1018918feff995f6ee3bb03e79 postgresql-unit.tar.gz" | sha256sum --check && \
     mkdir postgresql-unit-src && cd postgresql-unit-src && tar xzf ../postgresql-unit.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -279,18 +239,14 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS vector-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-COPY compute/patches/pgvector.patch /pgvector.patch
+COPY patches/pgvector.patch /pgvector.patch
 
 # By default, pgvector Makefile uses `-march=native`. We don't want that,
 # because we build the images on different machines than where we run them.
 # Pass OPTFLAGS="" to remove it.
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.7.2.tar.gz -O pgvector.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.7.2.tar.gz -O pgvector.tar.gz && \
     echo "617fba855c9bcb41a2a9bc78a78567fd2e147c72afd5bf9d37b31b9591632b30 pgvector.tar.gz" | sha256sum --check && \
     mkdir pgvector-src && cd pgvector-src && tar xzf ../pgvector.tar.gz --strip-components=1 -C . && \
     patch -p1 < /pgvector.patch && \
@@ -305,14 +261,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pgjwt-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 # 9742dab1b2f297ad3811120db7b21451bca2d3c9 made on 13/11/2021
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/michelp/pgjwt/archive/9742dab1b2f297ad3811120db7b21451bca2d3c9.tar.gz -O pgjwt.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/michelp/pgjwt/archive/9742dab1b2f297ad3811120db7b21451bca2d3c9.tar.gz -O pgjwt.tar.gz && \
     echo "cfdefb15007286f67d3d45510f04a6a7a495004be5b3aecb12cda667e774203f pgjwt.tar.gz" | sha256sum --check && \
     mkdir pgjwt-src && cd pgjwt-src && tar xzf ../pgjwt.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -325,13 +277,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS hypopg-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/HypoPG/hypopg/archive/refs/tags/1.4.0.tar.gz -O hypopg.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/HypoPG/hypopg/archive/refs/tags/1.4.0.tar.gz -O hypopg.tar.gz && \
     echo "0821011743083226fc9b813c1f2ef5897a91901b57b6bea85a78e466187c6819 hypopg.tar.gz" | sha256sum --check && \
     mkdir hypopg-src && cd hypopg-src && tar xzf ../hypopg.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -345,13 +293,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-hashids-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/iCyberon/pg_hashids/archive/refs/tags/v1.2.1.tar.gz -O pg_hashids.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/iCyberon/pg_hashids/archive/refs/tags/v1.2.1.tar.gz -O pg_hashids.tar.gz && \
     echo "74576b992d9277c92196dd8d816baa2cc2d8046fe102f3dcd7f3c3febed6822a pg_hashids.tar.gz" | sha256sum --check && \
     mkdir pg_hashids-src && cd pg_hashids-src && tar xzf ../pg_hashids.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
@@ -365,15 +309,11 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS rum-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-COPY compute/patches/rum.patch /rum.patch
+COPY patches/rum.patch /rum.patch
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/postgrespro/rum/archive/refs/tags/1.3.13.tar.gz -O rum.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/postgrespro/rum/archive/refs/tags/1.3.13.tar.gz -O rum.tar.gz && \
     echo "6ab370532c965568df6210bd844ac6ba649f53055e48243525b0b7e5c4d69a7d rum.tar.gz" | sha256sum --check && \
     mkdir rum-src && cd rum-src && tar xzf ../rum.tar.gz --strip-components=1 -C . && \
     patch -p1 < /rum.patch && \
@@ -388,13 +328,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pgtap-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/theory/pgtap/archive/refs/tags/v1.2.0.tar.gz -O pgtap.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/theory/pgtap/archive/refs/tags/v1.2.0.tar.gz -O pgtap.tar.gz && \
     echo "9c7c3de67ea41638e14f06da5da57bac6f5bd03fea05c165a0ec862205a5c052 pgtap.tar.gz" | sha256sum --check && \
     mkdir pgtap-src && cd pgtap-src && tar xzf ../pgtap.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -408,13 +344,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS ip4r-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/RhodiumToad/ip4r/archive/refs/tags/2.4.2.tar.gz -O ip4r.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/RhodiumToad/ip4r/archive/refs/tags/2.4.2.tar.gz -O ip4r.tar.gz && \
     echo "0f7b1f159974f49a47842a8ab6751aecca1ed1142b6d5e38d81b064b2ead1b4b ip4r.tar.gz" | sha256sum --check && \
     mkdir ip4r-src && cd ip4r-src && tar xzf ../ip4r.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -428,13 +360,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS prefix-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/dimitri/prefix/archive/refs/tags/v1.2.10.tar.gz -O prefix.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/dimitri/prefix/archive/refs/tags/v1.2.10.tar.gz -O prefix.tar.gz && \
     echo "4342f251432a5f6fb05b8597139d3ccde8dcf87e8ca1498e7ee931ca057a8575 prefix.tar.gz" | sha256sum --check && \
     mkdir prefix-src && cd prefix-src && tar xzf ../prefix.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -448,13 +376,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS hll-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/citusdata/postgresql-hll/archive/refs/tags/v2.18.tar.gz -O hll.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/citusdata/postgresql-hll/archive/refs/tags/v2.18.tar.gz -O hll.tar.gz && \
     echo "e2f55a6f4c4ab95ee4f1b4a2b73280258c5136b161fe9d059559556079694f0e hll.tar.gz" | sha256sum --check && \
     mkdir hll-src && cd hll-src && tar xzf ../hll.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
@@ -468,13 +392,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS plpgsql-check-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/okbob/plpgsql_check/archive/refs/tags/v2.5.3.tar.gz -O plpgsql_check.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/okbob/plpgsql_check/archive/refs/tags/v2.5.3.tar.gz -O plpgsql_check.tar.gz && \
     echo "6631ec3e7fb3769eaaf56e3dfedb829aa761abf163d13dba354b4c218508e1c0 plpgsql_check.tar.gz" | sha256sum --check && \
     mkdir plpgsql_check-src && cd plpgsql_check-src && tar xzf ../plpgsql_check.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
@@ -493,10 +413,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 ARG PG_VERSION
 ENV PATH="/usr/local/pgsql/bin:$PATH"
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN case "${PG_VERSION}" in \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    case "${PG_VERSION}" in \
       "v14" | "v15") \
         export TIMESCALEDB_VERSION=2.10.1 \
         export TIMESCALEDB_CHECKSUM=6fca72a6ed0f6d32d2b3523951ede73dc5f9b0077b38450a029a5f411fdb8c73 \
@@ -529,10 +446,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 ARG PG_VERSION
 ENV PATH="/usr/local/pgsql/bin:$PATH"
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN case "${PG_VERSION}" in \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    case "${PG_VERSION}" in \
       "v14") \
         export PG_HINT_PLAN_VERSION=14_1_4_1 \
         export PG_HINT_PLAN_CHECKSUM=c3501becf70ead27f70626bce80ea401ceac6a77e2083ee5f3ff1f1444ec1ad1 \
@@ -545,9 +459,6 @@ RUN case "${PG_VERSION}" in "v17") \
         export PG_HINT_PLAN_VERSION=16_1_6_0 \
         export PG_HINT_PLAN_CHECKSUM=fc85a9212e7d2819d4ae4ac75817481101833c3cfa9f0fe1f980984e12347d00 \
         ;; \
-      "v17") \
-        echo "TODO: PG17 pg_hint_plan support" && exit 0 \
-        ;; \
       *) \
         echo "Export the valid PG_HINT_PLAN_VERSION variable" && exit 1 \
         ;; \
@@ -567,14 +478,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-cron-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/citusdata/pg_cron/archive/refs/tags/v1.6.0.tar.gz -O pg_cron.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/citusdata/pg_cron/archive/refs/tags/v1.6.0.tar.gz -O pg_cron.tar.gz && \
     echo "383a627867d730222c272bfd25cd5e151c578d73f696d32910c7db8c665cc7db pg_cron.tar.gz" | sha256sum --check && \
     mkdir pg_cron-src && cd pg_cron-src && tar xzf ../pg_cron.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -588,13 +495,9 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS rdkit-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN apt-get update && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    apt-get update && \
     apt-get install -y \
         cmake \
         libboost-iostreams1.74-dev \
@@ -604,10 +507,7 @@ RUN case "${PG_VERSION}" in "v17") \
         libeigen3-dev
 
 ENV PATH="/usr/local/pgsql/bin/:/usr/local/pgsql/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/rdkit/rdkit/archive/refs/tags/Release_2023_03_3.tar.gz -O rdkit.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/rdkit/rdkit/archive/refs/tags/Release_2023_03_3.tar.gz -O rdkit.tar.gz && \
     echo "bdbf9a2e6988526bfeb8c56ce3cdfe2998d60ac289078e2215374288185e8c8d rdkit.tar.gz" | sha256sum --check && \
     mkdir rdkit-src && cd rdkit-src && tar xzf ../rdkit.tar.gz --strip-components=1 -C . && \
     cmake \
@@ -644,14 +544,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-uuidv7-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/fboulnois/pg_uuidv7/archive/refs/tags/v1.0.1.tar.gz -O pg_uuidv7.tar.gz && \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/fboulnois/pg_uuidv7/archive/refs/tags/v1.0.1.tar.gz -O pg_uuidv7.tar.gz && \
     echo "0d0759ab01b7fb23851ecffb0bce27822e1868a4a5819bfd276101c716637a7a pg_uuidv7.tar.gz" | sha256sum --check && \
     mkdir pg_uuidv7-src && cd pg_uuidv7-src && tar xzf ../pg_uuidv7.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -665,14 +561,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-roaringbitmap-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/ChenHuajun/pg_roaringbitmap/archive/refs/tags/v0.5.4.tar.gz -O pg_roaringbitmap.tar.gz && \
-    echo "v17 extensions is not supported yet by pg_roaringbitmap. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/ChenHuajun/pg_roaringbitmap/archive/refs/tags/v0.5.4.tar.gz -O pg_roaringbitmap.tar.gz && \
     echo "b75201efcb1c2d1b014ec4ae6a22769cc7a224e6e406a587f5784a37b6b5a2aa pg_roaringbitmap.tar.gz" | sha256sum --check && \
     mkdir pg_roaringbitmap-src && cd pg_roaringbitmap-src && tar xzf ../pg_roaringbitmap.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -686,14 +578,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-semver-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/theory/pg-semver/archive/refs/tags/v0.32.1.tar.gz -O pg_semver.tar.gz && \
-    echo "v17 is not supported yet by pg_semver. Quit" && exit 0;; \
-    esac && \
-    wget https://github.com/theory/pg-semver/archive/refs/tags/v0.32.1.tar.gz -O pg_semver.tar.gz && \
     echo "fbdaf7512026d62eec03fad8687c15ed509b6ba395bff140acd63d2e4fbe25d7 pg_semver.tar.gz" | sha256sum --check && \
     mkdir pg_semver-src && cd pg_semver-src && tar xzf ../pg_semver.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -732,14 +620,10 @@ RUN case "${PG_VERSION}" in \
 #
 #########################################################################################
 FROM build-deps AS pg-anon-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget  https://github.com/neondatabase/postgresql_anonymizer/archive/refs/tags/neon_1.1.1.tar.gz -O pg_anon.tar.gz && \
-    echo "postgresql_anonymizer does not yet support PG17" && exit 0;; \
-    esac && \
-    wget  https://github.com/neondatabase/postgresql_anonymizer/archive/refs/tags/neon_1.1.1.tar.gz -O pg_anon.tar.gz && \
     echo "321ea8d5c1648880aafde850a2c576e4a9e7b9933a34ce272efc839328999fa9  pg_anon.tar.gz" | sha256sum --check && \
     mkdir pg_anon-src && cd pg_anon-src && tar xzf ../pg_anon.tar.gz --strip-components=1 -C . && \
     find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
@@ -757,7 +641,6 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS rust-extensions-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 RUN apt-get update && \
@@ -768,11 +651,9 @@ ENV HOME=/home/nonroot
 ENV PATH="/home/nonroot/.cargo/bin:/usr/local/pgsql/bin/:$PATH"
 USER nonroot
 WORKDIR /home/nonroot
+ARG PG_VERSION
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux-gnu/rustup-init && \
-    echo "v17 is not supported yet by pgrx. Quit" && exit 0;; \
-    esac && \
-    curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux-gnu/rustup-init && \
     chmod +x rustup-init && \
     ./rustup-init -y --no-modify-path --profile minimal --default-toolchain stable && \
     rm rustup-init && \
@@ -791,10 +672,7 @@ USER root
 FROM rust-extensions-build AS pg-jsonschema-pg-build
 ARG PG_VERSION
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/supabase/pg_jsonschema/archive/refs/tags/v0.3.1.tar.gz -O pg_jsonschema.tar.gz && \
-    echo "pg_jsonschema does not yet have a release that supports pg17" && exit 0;; \
-    esac && \
-    wget https://github.com/supabase/pg_jsonschema/archive/refs/tags/v0.3.1.tar.gz -O pg_jsonschema.tar.gz && \
     echo "61df3db1ed83cf24f6aa39c826f8818bfa4f0bd33b587fd6b2b1747985642297 pg_jsonschema.tar.gz" | sha256sum --check && \
     mkdir pg_jsonschema-src && cd pg_jsonschema-src && tar xzf ../pg_jsonschema.tar.gz --strip-components=1 -C . && \
     # see commit 252b3685a27a0f4c31a0f91e983c6314838e89e8
@@ -816,10 +694,7 @@ RUN case "${PG_VERSION}" in "v17") \
 FROM rust-extensions-build AS pg-graphql-pg-build
 ARG PG_VERSION
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/supabase/pg_graphql/archive/refs/tags/v1.5.7.tar.gz -O pg_graphql.tar.gz && \
-    echo "pg_graphql does not yet have a release that supports pg17 as of now" && exit 0;; \
-    esac && \
-    wget https://github.com/supabase/pg_graphql/archive/refs/tags/v1.5.7.tar.gz -O pg_graphql.tar.gz && \
     echo "2b3e567a5b31019cb97ae0e33263c1bcc28580be5a444ac4c8ece5c4be2aea41 pg_graphql.tar.gz" | sha256sum --check && \
     mkdir pg_graphql-src && cd pg_graphql-src && tar xzf ../pg_graphql.tar.gz --strip-components=1 -C . && \
     sed -i 's/pgrx = "=0.11.3"/pgrx = { version = "0.11.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
@@ -839,10 +714,7 @@ FROM rust-extensions-build AS pg-tiktoken-pg-build
 ARG PG_VERSION
 
 # 26806147b17b60763039c6a6878884c41a262318 made on 26/09/2023
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/kelvich/pg_tiktoken/archive/26806147b17b60763039c6a6878884c41a262318.tar.gz -O pg_tiktoken.tar.gz && \
-    echo "pg_tiktoken does not have versions, nor support for pg17" && exit 0;; \
-    esac && \
-    wget https://github.com/kelvich/pg_tiktoken/archive/26806147b17b60763039c6a6878884c41a262318.tar.gz -O pg_tiktoken.tar.gz && \
     echo "e64e55aaa38c259512d3e27c572da22c4637418cf124caba904cd50944e5004e pg_tiktoken.tar.gz" | sha256sum --check && \
     mkdir pg_tiktoken-src && cd pg_tiktoken-src && tar xzf ../pg_tiktoken.tar.gz --strip-components=1 -C . && \
     # TODO update pgrx version in the pg_tiktoken repo and remove this line
@@ -861,10 +733,7 @@ RUN case "${PG_VERSION}" in "v17") \
 FROM rust-extensions-build AS pg-pgx-ulid-build
 ARG PG_VERSION
 
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/pksunkara/pgx_ulid/archive/refs/tags/v0.1.5.tar.gz -O pgx_ulid.tar.gz && \
-    echo "pgx_ulid does not support pg17 as of the latest version (0.1.5)" && exit 0;; \
-    esac && \
-    wget https://github.com/pksunkara/pgx_ulid/archive/refs/tags/v0.1.5.tar.gz -O pgx_ulid.tar.gz && \
     echo "9d1659a2da65af0133d5451c454de31b37364e3502087dadf579f790bc8bef17 pgx_ulid.tar.gz" | sha256sum --check && \
     mkdir pgx_ulid-src && cd pgx_ulid-src && tar xzf ../pgx_ulid.tar.gz --strip-components=1 -C . && \
     sed -i 's/pgrx       = "^0.11.2"/pgrx = { version = "=0.11.3", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
@@ -879,14 +748,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #########################################################################################
 
 FROM build-deps AS wal2json-pg-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/eulerto/wal2json/archive/refs/tags/wal2json_2_5.tar.gz && \
-    echo "We'll need to update wal2json to 2.6+ for pg17 support" && exit 0;; \
-    esac && \
-    wget https://github.com/eulerto/wal2json/archive/refs/tags/wal2json_2_5.tar.gz && \
     echo "b516653575541cf221b99cf3f8be9b6821f6dbcfc125675c85f35090f824f00e wal2json_2_5.tar.gz" | sha256sum --check && \
     mkdir wal2json-src && cd wal2json-src && tar xzf ../wal2json_2_5.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -899,14 +764,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-ivm-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/sraoss/pg_ivm/archive/refs/tags/v1.7.tar.gz -O pg_ivm.tar.gz && \
-    echo "We'll need to update pg_ivm to 1.9+ for pg17 support" && exit 0;; \
-    esac && \
-    wget https://github.com/sraoss/pg_ivm/archive/refs/tags/v1.7.tar.gz -O pg_ivm.tar.gz && \
     echo "ebfde04f99203c7be4b0e873f91104090e2e83e5429c32ac242d00f334224d5e pg_ivm.tar.gz" | sha256sum --check && \
     mkdir pg_ivm-src && cd pg_ivm-src && tar xzf ../pg_ivm.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -920,14 +781,10 @@ RUN case "${PG_VERSION}" in "v17") \
 #
 #########################################################################################
 FROM build-deps AS pg-partman-build
-ARG PG_VERSION
 COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
 
 ENV PATH="/usr/local/pgsql/bin/:$PATH"
-RUN case "${PG_VERSION}" in "v17") \
+RUN wget https://github.com/pgpartman/pg_partman/archive/refs/tags/v5.0.1.tar.gz -O pg_partman.tar.gz && \
-    echo "pg_partman doesn't support PG17 yet" && exit 0;; \
-    esac && \
-    wget https://github.com/pgpartman/pg_partman/archive/refs/tags/v5.0.1.tar.gz -O pg_partman.tar.gz && \
     echo "75b541733a9659a6c90dbd40fccb904a630a32880a6e3044d0c4c5f4c8a65525 pg_partman.tar.gz" | sha256sum --check && \
     mkdir pg_partman-src && cd pg_partman-src && tar xzf ../pg_partman.tar.gz --strip-components=1 -C . && \
     make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -997,8 +854,8 @@ RUN make -j $(getconf _NPROCESSORS_ONLN) \
     case "${PG_VERSION}" in \
         "v14" | "v15") \
         ;; \
-        "v16" | "v17") \
+        "v16") \
-            echo "Skipping HNSW for PostgreSQL ${PG_VERSION}" && exit 0 \
+            echo "Skipping HNSW for PostgreSQL 16" && exit 0 \
         ;; \
         *) \
             echo "unexpected PostgreSQL version" && exit 1 \
@@ -1029,47 +886,10 @@ RUN cd compute_tools && mold -run cargo build --locked --profile release-line-de
 #
 #########################################################################################
 
-FROM debian:$DEBIAN_FLAVOR AS compute-tools-image
+FROM debian:bullseye-slim AS compute-tools-image
-ARG DEBIAN_FLAVOR
 
 COPY --from=compute-tools /home/nonroot/target/release-line-debug-size-lto/compute_ctl /usr/local/bin/compute_ctl
 
-#########################################################################################
-#
-# Layer "pgbouncer"
-#
-#########################################################################################
-
-FROM debian:$DEBIAN_FLAVOR AS pgbouncer
-ARG DEBIAN_FLAVOR
-RUN set -e \
-    && apt-get update \
-    && apt-get install -y \
-        build-essential \
-        git \
-        libevent-dev \
-        libtool \
-        pkg-config
-
-# Use `dist_man_MANS=` to skip manpage generation (which requires python3/pandoc)
-ENV PGBOUNCER_TAG=pgbouncer_1_22_1
-RUN set -e \
-    && git clone --recurse-submodules --depth 1 --branch ${PGBOUNCER_TAG} https://github.com/pgbouncer/pgbouncer.git pgbouncer \
-    && cd pgbouncer \
-    && ./autogen.sh \
-    && LDFLAGS=-static ./configure --prefix=/usr/local/pgbouncer --without-openssl \
-    && make -j $(nproc) dist_man_MANS= \
-    && make install dist_man_MANS=
-
-#########################################################################################
-#
-# Layers "postgres-exporter" and "sql-exporter"
-#
-#########################################################################################
-
-FROM quay.io/prometheuscommunity/postgres-exporter:v0.12.1 AS postgres-exporter
-FROM burningalchemist/sql_exporter:0.13 AS sql-exporter
-
 #########################################################################################
 #
 # Clean up postgres folder before inclusion
@@ -1079,7 +899,7 @@ FROM neon-pg-ext-build AS postgres-cleanup-layer
 COPY --from=neon-pg-ext-build /usr/local/pgsql /usr/local/pgsql
 
 # Remove binaries from /bin/ that we won't use (or would manually copy & install otherwise)
-RUN cd /usr/local/pgsql/bin && rm -f ecpg raster2pgsql shp2pgsql pgtopo_export pgtopo_import pgsql2shp
+RUN cd /usr/local/pgsql/bin && rm ecpg raster2pgsql shp2pgsql pgtopo_export pgtopo_import pgsql2shp
 
 # Remove headers that we won't need anymore - we've completed installation of all extensions
 RUN rm -r /usr/local/pgsql/include
@@ -1098,10 +918,7 @@ RUN rm /usr/local/pgsql/lib/lib*.a
 
 FROM neon-pg-ext-build AS neon-pg-ext-test
 ARG PG_VERSION
-RUN case "${PG_VERSION}" in "v17") \
+RUN mkdir /ext-src
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    mkdir /ext-src
 
 #COPY --from=postgis-build /postgis.tar.gz /ext-src/
 #COPY --from=postgis-build /sfcgal/* /usr
@@ -1117,7 +934,7 @@ COPY --from=pgjwt-pg-build /pgjwt.tar.gz /ext-src
 COPY --from=hypopg-pg-build /hypopg.tar.gz /ext-src
 COPY --from=pg-hashids-pg-build /pg_hashids.tar.gz /ext-src
 COPY --from=rum-pg-build /rum.tar.gz /ext-src
-COPY compute/patches/rum.patch /ext-src
+COPY patches/rum.patch /ext-src
 #COPY --from=pgtap-pg-build /pgtap.tar.gz /ext-src
 COPY --from=ip4r-pg-build /ip4r.tar.gz /ext-src
 COPY --from=prefix-pg-build /prefix.tar.gz /ext-src
@@ -1125,9 +942,9 @@ COPY --from=hll-pg-build /hll.tar.gz /ext-src
 COPY --from=plpgsql-check-pg-build /plpgsql_check.tar.gz /ext-src
 #COPY --from=timescaledb-pg-build /timescaledb.tar.gz /ext-src
 COPY --from=pg-hint-plan-pg-build /pg_hint_plan.tar.gz /ext-src
-COPY compute/patches/pg_hint_plan.patch /ext-src
+COPY patches/pg_hint_plan.patch /ext-src
 COPY --from=pg-cron-pg-build /pg_cron.tar.gz /ext-src
-COPY compute/patches/pg_cron.patch /ext-src
+COPY patches/pg_cron.patch /ext-src
 #COPY --from=pg-pgx-ulid-build /home/nonroot/pgx_ulid.tar.gz /ext-src
 #COPY --from=rdkit-pg-build /rdkit.tar.gz /ext-src
 COPY --from=pg-uuidv7-pg-build /pg_uuidv7.tar.gz /ext-src
@@ -1136,42 +953,21 @@ COPY --from=pg-semver-pg-build /pg_semver.tar.gz /ext-src
 #COPY --from=pg-embedding-pg-build /home/nonroot/pg_embedding-src/ /ext-src
 #COPY --from=wal2json-pg-build /wal2json_2_5.tar.gz /ext-src
 COPY --from=pg-anon-pg-build /pg_anon.tar.gz /ext-src
-COPY compute/patches/pg_anon.patch /ext-src
+COPY patches/pg_anon.patch /ext-src
 COPY --from=pg-ivm-build /pg_ivm.tar.gz /ext-src
 COPY --from=pg-partman-build /pg_partman.tar.gz /ext-src
-RUN case "${PG_VERSION}" in "v17") \
+RUN cd /ext-src/ && for f in *.tar.gz; \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    cd /ext-src/ && for f in *.tar.gz; \
     do echo $f; dname=$(echo $f | sed 's/\.tar.*//')-src; \
     rm -rf $dname; mkdir $dname; tar xzf $f --strip-components=1 -C $dname \
     || exit 1; rm -f $f; done
-RUN case "${PG_VERSION}" in "v17") \
+RUN cd /ext-src/pgvector-src && patch -p1 <../pgvector.patch
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
+RUN cd /ext-src/rum-src && patch -p1 <../rum.patch
-    esac && \
-    cd /ext-src/rum-src && patch -p1 <../rum.patch
-RUN case "${PG_VERSION}" in "v17") \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    cd /ext-src/pgvector-src && patch -p1 <../pgvector.patch
 # cmake is required for the h3 test
-RUN case "${PG_VERSION}" in "v17") \
+RUN apt-get update && apt-get install -y cmake
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
+RUN cd /ext-src/pg_hint_plan-src && patch -p1 < /ext-src/pg_hint_plan.patch
-    esac && \
-    apt-get update && apt-get install -y cmake
-RUN case "${PG_VERSION}" in "v17") \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    cd /ext-src/pg_hint_plan-src && patch -p1 < /ext-src/pg_hint_plan.patch
 COPY --chmod=755 docker-compose/run-tests.sh /run-tests.sh
-RUN case "${PG_VERSION}" in "v17") \
+RUN patch -p1 </ext-src/pg_anon.patch
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
+RUN patch -p1 </ext-src/pg_cron.patch
-    esac && \
-    patch -p1 </ext-src/pg_anon.patch
-RUN case "${PG_VERSION}" in "v17") \
-    echo "v17 extensions are not supported yet. Quit" && exit 0;; \
-    esac && \
-    patch -p1 </ext-src/pg_cron.patch
 ENV PATH=/usr/local/pgsql/bin:$PATH
 ENV PGHOST=compute
 ENV PGPORT=55433
@@ -1183,9 +979,7 @@ ENV PGDATABASE=postgres
 # Put it all together into the final image
 #
 #########################################################################################
-FROM debian:$DEBIAN_FLAVOR
+FROM debian:bullseye-slim
-ARG DEBIAN_FLAVOR
-ENV DEBIAN_FLAVOR=$DEBIAN_FLAVOR
 # Add user postgres
 RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
     echo "postgres:test_console_pass" | chpasswd && \
@@ -1201,50 +995,23 @@ RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
 COPY --from=postgres-cleanup-layer --chown=postgres /usr/local/pgsql /usr/local
 COPY --from=compute-tools --chown=postgres /home/nonroot/target/release-line-debug-size-lto/compute_ctl /usr/local/bin/compute_ctl
 
-# pgbouncer and its config
-COPY --from=pgbouncer         /usr/local/pgbouncer/bin/pgbouncer /usr/local/bin/pgbouncer
-COPY --chmod=0666 --chown=postgres compute/etc/pgbouncer.ini /etc/pgbouncer.ini
-
-# Metrics exporter binaries and  configuration files
-COPY --from=postgres-exporter /bin/postgres_exporter /bin/postgres_exporter
-COPY --from=sql-exporter      /bin/sql_exporter      /bin/sql_exporter
-
-COPY --chmod=0644 compute/etc/sql_exporter.yml               /etc/sql_exporter.yml
-COPY --chmod=0644 compute/etc/neon_collector.yml             /etc/neon_collector.yml
-COPY --chmod=0644 compute/etc/sql_exporter_autoscaling.yml   /etc/sql_exporter_autoscaling.yml
-COPY --chmod=0644 compute/etc/neon_collector_autoscaling.yml /etc/neon_collector_autoscaling.yml
-
 # Create remote extension download directory
 RUN mkdir /usr/local/download_extensions && chown -R postgres:postgres /usr/local/download_extensions
 
 # Install:
 # libreadline8 for psql
+# libicu67, locales for collations (including ICU and plpgsql_check)
 # liblz4-1 for lz4
 # libossp-uuid16 for extension ossp-uuid
-# libgeos, libsfcgal1, and libprotobuf-c1 for PostGIS
+# libgeos, libgdal, libsfcgal1, libproj and libprotobuf-c1 for PostGIS
 # libxml2, libxslt1.1 for xml2
 # libzstd1 for zstd
 # libboost* for rdkit
 # ca-certificates for communicating with s3 by compute_ctl
-
+RUN apt update &&  \
-
-RUN apt update && \
-    case $DEBIAN_FLAVOR in \
-      # Version-specific installs for Bullseye (PG14-PG16):
-      # libicu67, locales for collations (including ICU and plpgsql_check)
-      # libgdal28, libproj19 for PostGIS
-      bullseye*) \
-        VERSION_INSTALLS="libicu67 libgdal28 libproj19"; \
-      ;; \
-      # Version-specific installs for Bookworm (PG17):
-      # libicu72, locales for collations (including ICU and plpgsql_check)
-      # libgdal32, libproj25 for PostGIS
-      bookworm*) \
-        VERSION_INSTALLS="libicu72 libgdal32 libproj25"; \
-      ;; \
-    esac && \
     apt install --no-install-recommends -y \
         gdb \
+        libicu67 \
         liblz4-1 \
         libreadline8 \
         libboost-iostreams1.74.0 \
@@ -1253,6 +1020,8 @@ RUN apt update && \
         libboost-system1.74.0 \
         libossp-uuid16 \
         libgeos-c1v5 \
+        libgdal28 \
+        libproj19 \
         libprotobuf-c1 \
         libsfcgal1 \
         libxml2 \
@@ -1261,8 +1030,7 @@ RUN apt update && \
         libcurl4-openssl-dev \
         locales \
         procps \
-        ca-certificates \
+        ca-certificates && \
-        $VERSION_INSTALLS && \
     rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
     localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8
 

Makefile

@@ -119,8 +119,6 @@ $(POSTGRES_INSTALL_DIR)/build/%/config.status:
 # I'm not sure why it wouldn't work, but this is the only place (apart from
 # the "build-all-versions" entry points) where direct mention of PostgreSQL
 # versions is used.
-.PHONY: postgres-configure-v17
-postgres-configure-v17: $(POSTGRES_INSTALL_DIR)/build/v17/config.status
 .PHONY: postgres-configure-v16
 postgres-configure-v16: $(POSTGRES_INSTALL_DIR)/build/v16/config.status
 .PHONY: postgres-configure-v15
@@ -217,31 +215,29 @@ neon-pg-clean-ext-%:
 # they depend on openssl and other libraries that are not included in our
 # Rust build.
 .PHONY: walproposer-lib
-walproposer-lib: neon-pg-ext-v17
+walproposer-lib: neon-pg-ext-v16
 	+@echo "Compiling walproposer-lib"
 	mkdir -p $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
-	$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v17/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
+	$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v16/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
 		-C $(POSTGRES_INSTALL_DIR)/build/walproposer-lib \
 		-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile walproposer-lib
-	cp $(POSTGRES_INSTALL_DIR)/v17/lib/libpgport.a $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
+	cp $(POSTGRES_INSTALL_DIR)/v16/lib/libpgport.a $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
-	cp $(POSTGRES_INSTALL_DIR)/v17/lib/libpgcommon.a $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
+	cp $(POSTGRES_INSTALL_DIR)/v16/lib/libpgcommon.a $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
+ifeq ($(UNAME_S),Linux)
 	$(AR) d $(POSTGRES_INSTALL_DIR)/build/walproposer-lib/libpgport.a \
 		pg_strong_random.o
 	$(AR) d $(POSTGRES_INSTALL_DIR)/build/walproposer-lib/libpgcommon.a \
-		checksum_helper.o \
+		pg_crc32c.o \
-		cryptohash_openssl.o \
 		hmac_openssl.o \
+		cryptohash_openssl.o \
+		scram-common.o \
 		md5_common.o \
-		parse_manifest.o \
+		checksum_helper.o
-		scram-common.o
-ifeq ($(UNAME_S),Linux)
-	$(AR) d $(POSTGRES_INSTALL_DIR)/build/walproposer-lib/libpgcommon.a \
-		pg_crc32c.o
 endif
 
 .PHONY: walproposer-lib-clean
 walproposer-lib-clean:
-	$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v17/bin/pg_config \
+	$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v16/bin/pg_config \
 		-C $(POSTGRES_INSTALL_DIR)/build/walproposer-lib \
 		-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile clean
 
@@ -249,44 +245,38 @@ walproposer-lib-clean:
 neon-pg-ext: \
 	neon-pg-ext-v14 \
 	neon-pg-ext-v15 \
-	neon-pg-ext-v16 \
+	neon-pg-ext-v16
-	neon-pg-ext-v17
 
 .PHONY: neon-pg-clean-ext
 neon-pg-clean-ext: \
 	neon-pg-clean-ext-v14 \
 	neon-pg-clean-ext-v15 \
-	neon-pg-clean-ext-v16 \
+	neon-pg-clean-ext-v16
-	neon-pg-clean-ext-v17
 
 # shorthand to build all Postgres versions
 .PHONY: postgres
 postgres: \
 	postgres-v14 \
 	postgres-v15 \
-	postgres-v16 \
+	postgres-v16
-	postgres-v17
 
 .PHONY: postgres-headers
 postgres-headers: \
 	postgres-headers-v14 \
 	postgres-headers-v15 \
-	postgres-headers-v16 \
+	postgres-headers-v16
-	postgres-headers-v17
 
 .PHONY: postgres-clean
 postgres-clean: \
 	postgres-clean-v14 \
 	postgres-clean-v15 \
-	postgres-clean-v16 \
+	postgres-clean-v16
-	postgres-clean-v17
 
 .PHONY: postgres-check
 postgres-check: \
 	postgres-check-v14 \
 	postgres-check-v15 \
-	postgres-check-v16 \
+	postgres-check-v16
-	postgres-check-v17
 
 # This doesn't remove the effects of 'configure'.
 .PHONY: clean
@@ -331,13 +321,13 @@ postgres-%-pgindent: postgres-%-pg-bsd-indent postgres-%-typedefs.list
 	rm -f pg*.BAK
 
 # Indent pxgn/neon.
-.PHONY: neon-pgindent
+.PHONY: pgindent
-neon-pgindent: postgres-v17-pg-bsd-indent neon-pg-ext-v17
+neon-pgindent: postgres-v16-pg-bsd-indent neon-pg-ext-v16
-	$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v17/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
+	$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v16/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
-		FIND_TYPEDEF=$(ROOT_PROJECT_DIR)/vendor/postgres-v17/src/tools/find_typedef \
+		FIND_TYPEDEF=$(ROOT_PROJECT_DIR)/vendor/postgres-v16/src/tools/find_typedef \
-		INDENT=$(POSTGRES_INSTALL_DIR)/build/v17/src/tools/pg_bsd_indent/pg_bsd_indent \
+		INDENT=$(POSTGRES_INSTALL_DIR)/build/v16/src/tools/pg_bsd_indent/pg_bsd_indent \
-		PGINDENT_SCRIPT=$(ROOT_PROJECT_DIR)/vendor/postgres-v17/src/tools/pgindent/pgindent \
+		PGINDENT_SCRIPT=$(ROOT_PROJECT_DIR)/vendor/postgres-v16/src/tools/pgindent/pgindent \
-		-C $(POSTGRES_INSTALL_DIR)/build/neon-v17 \
+		-C $(POSTGRES_INSTALL_DIR)/build/neon-v16 \
 		-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile pgindent
 
 

compute/README.md

@@ -1,21 +0,0 @@
-This directory contains files that are needed to build the compute
-images, or included in the compute images.
-
-Dockerfile.compute-node
-	To build the compute image
-
-vm-image-spec.yaml
-	Instructions for vm-builder, to turn the compute-node image into
-	corresponding vm-compute-node image.
-
-etc/
-	Configuration files included in /etc in the compute image
-
-patches/
-	Some extensions need to be patched to work with Neon. This
-	directory contains such patches. They are applied to the extension
-	sources in Dockerfile.compute-node
-
-In addition to these, postgres itself, the neon postgres extension,
-and compute_ctl are built and copied into the compute image by
-Dockerfile.compute-node.

compute/etc/neon_collector.yml

@@ -1,247 +0,0 @@
-collector_name: neon_collector
-metrics:
-- metric_name: lfc_misses
-  type: gauge
-  help: 'lfc_misses'
-  key_labels:
-  values: [lfc_misses]
-  query: |
-    select lfc_value as lfc_misses from neon.neon_lfc_stats where lfc_key='file_cache_misses';
-
-- metric_name: lfc_used
-  type: gauge
-  help: 'LFC chunks used (chunk = 1MB)'
-  key_labels:
-  values: [lfc_used]
-  query: |
-    select lfc_value as lfc_used from neon.neon_lfc_stats where lfc_key='file_cache_used';
-
-- metric_name: lfc_hits
-  type: gauge
-  help: 'lfc_hits'
-  key_labels:
-  values: [lfc_hits]
-  query: |
-    select lfc_value as lfc_hits from neon.neon_lfc_stats where lfc_key='file_cache_hits';
-
-- metric_name: lfc_writes
-  type: gauge
-  help: 'lfc_writes'
-  key_labels:
-  values: [lfc_writes]
-  query: |
-    select lfc_value as lfc_writes from neon.neon_lfc_stats where lfc_key='file_cache_writes';
-
-- metric_name: lfc_cache_size_limit
-  type: gauge
-  help: 'LFC cache size limit in bytes'
-  key_labels:
-  values: [lfc_cache_size_limit]
-  query: |
-    select pg_size_bytes(current_setting('neon.file_cache_size_limit')) as lfc_cache_size_limit;
-
-- metric_name: connection_counts
-  type: gauge
-  help: 'Connection counts'
-  key_labels:
-    - datname
-    - state
-  values: [count]
-  query: |
-    select datname, state, count(*) as count from pg_stat_activity where state <> '' group by datname, state;
-
-- metric_name: pg_stats_userdb
-  type: gauge
-  help: 'Stats for several oldest non-system dbs'
-  key_labels:
-    - datname
-  value_label: kind
-  values:
-    - db_size
-    - deadlocks
-    # Rows
-    - inserted
-    - updated
-    - deleted
-  # We export stats for 10 non-system database. Without this limit
-  # it is too easy to abuse the system by creating lots of databases.
-  query: |
-    select pg_database_size(datname) as db_size, deadlocks,
-       tup_inserted as inserted, tup_updated as updated, tup_deleted as deleted,
-       datname
-     from pg_stat_database
-     where datname IN (
-       select datname
-       from pg_database
-       where datname <> 'postgres' and not datistemplate
-       order by oid
-       limit 10
-     );
-
-- metric_name: max_cluster_size
-  type: gauge
-  help: 'neon.max_cluster_size setting'
-  key_labels:
-  values: [max_cluster_size]
-  query: |
-    select setting::int as max_cluster_size from pg_settings where name = 'neon.max_cluster_size';
-
-- metric_name: db_total_size
-  type: gauge
-  help: 'Size of all databases'
-  key_labels:
-  values: [total]
-  query: |
-    select sum(pg_database_size(datname)) as total from pg_database;
-
-# DEPRECATED
-- metric_name: lfc_approximate_working_set_size
-  type: gauge
-  help: 'Approximate working set size in pages of 8192 bytes'
-  key_labels:
-  values: [approximate_working_set_size]
-  query: |
-    select neon.approximate_working_set_size(false) as approximate_working_set_size;
-
-- metric_name: lfc_approximate_working_set_size_windows
-  type: gauge
-  help: 'Approximate working set size in pages of 8192 bytes'
-  key_labels: [duration]
-  values: [size]
-  # NOTE: This is the "public" / "human-readable" version. Here, we supply a small selection
-  # of durations in a pretty-printed form.
-  query: |
-    select
-      x as duration,
-      neon.approximate_working_set_size_seconds(extract('epoch' from x::interval)::int) as size
-    from
-      (values ('5m'),('15m'),('1h')) as t (x);
-
-- metric_name: compute_current_lsn
-  type: gauge
-  help: 'Current LSN of the database'
-  key_labels:
-  values: [lsn]
-  query: |
-    select
-      case
-        when pg_catalog.pg_is_in_recovery()
-        then (pg_last_wal_replay_lsn() - '0/0')::FLOAT8
-        else (pg_current_wal_lsn() - '0/0')::FLOAT8
-      end as lsn;
-
-- metric_name: compute_receive_lsn
-  type: gauge
-  help: 'Returns the last write-ahead log location that has been received and synced to disk by streaming replication'
-  key_labels:
-  values: [lsn]
-  query: |
-    SELECT
-      CASE
-        WHEN pg_catalog.pg_is_in_recovery()
-        THEN (pg_last_wal_receive_lsn() - '0/0')::FLOAT8
-        ELSE 0
-      END AS lsn;
-
-- metric_name: replication_delay_bytes
-  type: gauge
-  help: 'Bytes between received and replayed LSN'
-  key_labels:
-  values: [replication_delay_bytes]
-  # We use a GREATEST call here because this calculation can be negative.
-  # The calculation is not atomic, meaning after we've gotten the receive
-  # LSN, the replay LSN may have advanced past the receive LSN we
-  # are using for the calculation.
-  query: |
-    SELECT GREATEST(0, pg_wal_lsn_diff(pg_last_wal_receive_lsn(), pg_last_wal_replay_lsn())) AS replication_delay_bytes;
-
-- metric_name: replication_delay_seconds
-  type: gauge
-  help: 'Time since last LSN was replayed'
-  key_labels:
-  values: [replication_delay_seconds]
-  query: |
-    SELECT
-      CASE
-        WHEN pg_last_wal_receive_lsn() = pg_last_wal_replay_lsn() THEN 0
-        ELSE GREATEST (0, EXTRACT (EPOCH FROM now() - pg_last_xact_replay_timestamp()))
-     END AS replication_delay_seconds;
-
-- metric_name: checkpoints_req
-  type: gauge
-  help: 'Number of requested checkpoints'
-  key_labels:
-  values: [checkpoints_req]
-  query: |
-    SELECT checkpoints_req FROM pg_stat_bgwriter;
-
-- metric_name: checkpoints_timed
-  type: gauge
-  help: 'Number of scheduled checkpoints'
-  key_labels:
-  values: [checkpoints_timed]
-  query: |
-    SELECT checkpoints_timed FROM pg_stat_bgwriter;
-
-- metric_name: compute_logical_snapshot_files
-  type: gauge
-  help: 'Number of snapshot files in pg_logical/snapshot'
-  key_labels:
-    - timeline_id
-  values: [num_logical_snapshot_files]
-  query: |
-    SELECT
-      (SELECT setting FROM pg_settings WHERE name = 'neon.timeline_id') AS timeline_id,
-      -- Postgres creates temporary snapshot files of the form %X-%X.snap.%d.tmp. These
-      -- temporary snapshot files are renamed to the actual snapshot files after they are
-      -- completely built. We only WAL-log the completely built snapshot files.
-      (SELECT COUNT(*) FROM pg_ls_logicalsnapdir() WHERE name LIKE '%.snap') AS num_logical_snapshot_files;
-
-# In all the below metrics, we cast LSNs to floats because Prometheus only supports floats.
-# It's probably fine because float64 can store integers from -2^53 to +2^53 exactly.
-
-# Number of slots is limited by max_replication_slots, so collecting position for all of them shouldn't be bad.
-- metric_name: logical_slot_restart_lsn
-  type: gauge
-  help: 'restart_lsn of logical slots'
-  key_labels:
-    - slot_name
-  values: [restart_lsn]
-  query: |
-    select slot_name, (restart_lsn - '0/0')::FLOAT8 as restart_lsn
-    from pg_replication_slots
-    where slot_type = 'logical';
-
-- metric_name: compute_subscriptions_count
-  type: gauge
-  help: 'Number of logical replication subscriptions grouped by enabled/disabled'
-  key_labels:
-    - enabled
-  values: [subscriptions_count]
-  query: |
-    select subenabled::text as enabled, count(*) as subscriptions_count
-    from pg_subscription
-    group by subenabled;
-
-- metric_name: retained_wal
-  type: gauge
-  help: 'Retained WAL in inactive replication slots'
-  key_labels:
-    - slot_name
-  values: [retained_wal]
-  query: |
-    SELECT slot_name, pg_wal_lsn_diff(pg_current_wal_lsn(), restart_lsn)::FLOAT8 AS retained_wal
-    FROM pg_replication_slots
-    WHERE active = false;
-
-- metric_name: wal_is_lost
-  type: gauge
-  help: 'Whether or not the replication slot wal_status is lost'
-  key_labels:
-    - slot_name
-  values: [wal_is_lost]
-  query: |
-    SELECT slot_name,
-           CASE WHEN wal_status = 'lost' THEN 1 ELSE 0 END AS wal_is_lost
-    FROM pg_replication_slots;
-

compute/etc/neon_collector_autoscaling.yml

@@ -1,55 +0,0 @@
-collector_name: neon_collector_autoscaling
-metrics:
-- metric_name: lfc_misses
-  type: gauge
-  help: 'lfc_misses'
-  key_labels:
-  values: [lfc_misses]
-  query: |
-    select lfc_value as lfc_misses from neon.neon_lfc_stats where lfc_key='file_cache_misses';
-
-- metric_name: lfc_used
-  type: gauge
-  help: 'LFC chunks used (chunk = 1MB)'
-  key_labels:
-  values: [lfc_used]
-  query: |
-    select lfc_value as lfc_used from neon.neon_lfc_stats where lfc_key='file_cache_used';
-
-- metric_name: lfc_hits
-  type: gauge
-  help: 'lfc_hits'
-  key_labels:
-  values: [lfc_hits]
-  query: |
-    select lfc_value as lfc_hits from neon.neon_lfc_stats where lfc_key='file_cache_hits';
-
-- metric_name: lfc_writes
-  type: gauge
-  help: 'lfc_writes'
-  key_labels:
-  values: [lfc_writes]
-  query: |
-    select lfc_value as lfc_writes from neon.neon_lfc_stats where lfc_key='file_cache_writes';
-
-- metric_name: lfc_cache_size_limit
-  type: gauge
-  help: 'LFC cache size limit in bytes'
-  key_labels:
-  values: [lfc_cache_size_limit]
-  query: |
-    select pg_size_bytes(current_setting('neon.file_cache_size_limit')) as lfc_cache_size_limit;
-
-- metric_name: lfc_approximate_working_set_size_windows
-  type: gauge
-  help: 'Approximate working set size in pages of 8192 bytes'
-  key_labels: [duration_seconds]
-  values: [size]
-  # NOTE: This is the "internal" / "machine-readable" version. This outputs the working set
-  # size looking back 1..60 minutes, labeled with the number of minutes.
-  query: |
-    select
-      x::text as duration_seconds,
-      neon.approximate_working_set_size_seconds(x) as size
-    from
-      (select generate_series * 60 as x from generate_series(1, 60)) as t (x);

compute/etc/pgbouncer.ini

@@ -1,17 +0,0 @@
-[databases]
-*=host=localhost port=5432 auth_user=cloud_admin
-[pgbouncer]
-listen_port=6432
-listen_addr=0.0.0.0
-auth_type=scram-sha-256
-auth_user=cloud_admin
-auth_dbname=postgres
-client_tls_sslmode=disable
-server_tls_sslmode=disable
-pool_mode=transaction
-max_client_conn=10000
-default_pool_size=64
-max_prepared_statements=0
-admin_users=postgres
-unix_socket_dir=/tmp/
-unix_socket_mode=0777

compute/etc/sql_exporter.yml

@@ -1,33 +0,0 @@
-# Configuration for sql_exporter
-# Global defaults.
-global:
-  # If scrape_timeout <= 0, no timeout is set unless Prometheus provides one. The default is 10s.
-  scrape_timeout: 10s
-  # Subtracted from Prometheus' scrape_timeout to give us some headroom and prevent Prometheus from timing out first.
-  scrape_timeout_offset: 500ms
-  # Minimum interval between collector runs: by default (0s) collectors are executed on every scrape.
-  min_interval: 0s
-  # Maximum number of open connections to any one target. Metric queries will run concurrently on multiple connections,
-  # as will concurrent scrapes.
-  max_connections: 1
-  # Maximum number of idle connections to any one target. Unless you use very long collection intervals, this should
-  # always be the same as max_connections.
-  max_idle_connections: 1
-  # Maximum number of maximum amount of time a connection may be reused. Expired connections may be closed lazily before reuse.
-  # If 0, connections are not closed due to a connection's age.
-  max_connection_lifetime: 5m
-
-# The target to monitor and the collectors to execute on it.
-target:
-  # Data source name always has a URI schema that matches the driver name. In some cases (e.g. MySQL)
-  # the schema gets dropped or replaced to match the driver expected DSN format.
-  data_source_name: 'postgresql://cloud_admin@127.0.0.1:5432/postgres?sslmode=disable&application_name=sql_exporter'
-
-  # Collectors (referenced by name) to execute on the target.
-  # Glob patterns are supported (see <https://pkg.go.dev/path/filepath#Match> for syntax).
-  collectors: [neon_collector]
-
-# Collector files specifies a list of globs. One collector definition is read from each matching file.
-# Glob patterns are supported (see <https://pkg.go.dev/path/filepath#Match> for syntax).
-collector_files:
-  - "neon_collector.yml"

compute/etc/sql_exporter_autoscaling.yml

@@ -1,33 +0,0 @@
-# Configuration for sql_exporter for autoscaling-agent
-# Global defaults.
-global:
-  # If scrape_timeout <= 0, no timeout is set unless Prometheus provides one. The default is 10s.
-  scrape_timeout: 10s
-  # Subtracted from Prometheus' scrape_timeout to give us some headroom and prevent Prometheus from timing out first.
-  scrape_timeout_offset: 500ms
-  # Minimum interval between collector runs: by default (0s) collectors are executed on every scrape.
-  min_interval: 0s
-  # Maximum number of open connections to any one target. Metric queries will run concurrently on multiple connections,
-  # as will concurrent scrapes.
-  max_connections: 1
-  # Maximum number of idle connections to any one target. Unless you use very long collection intervals, this should
-  # always be the same as max_connections.
-  max_idle_connections: 1
-  # Maximum number of maximum amount of time a connection may be reused. Expired connections may be closed lazily before reuse.
-  # If 0, connections are not closed due to a connection's age.
-  max_connection_lifetime: 5m
-
-# The target to monitor and the collectors to execute on it.
-target:
-  # Data source name always has a URI schema that matches the driver name. In some cases (e.g. MySQL)
-  # the schema gets dropped or replaced to match the driver expected DSN format.
-  data_source_name: 'postgresql://cloud_admin@127.0.0.1:5432/postgres?sslmode=disable&application_name=sql_exporter_autoscaling'
-
-  # Collectors (referenced by name) to execute on the target.
-  # Glob patterns are supported (see <https://pkg.go.dev/path/filepath#Match> for syntax).
-  collectors: [neon_collector_autoscaling]
-
-# Collector files specifies a list of globs. One collector definition is read from each matching file.
-# Glob patterns are supported (see <https://pkg.go.dev/path/filepath#Match> for syntax).
-collector_files:
-  - "neon_collector_autoscaling.yml"

compute/vm-image-spec.yaml

@@ -1,112 +0,0 @@
-# Supplemental file for neondatabase/autoscaling's vm-builder, for producing the VM compute image.
----
-commands:
-  - name: cgconfigparser
-    user: root
-    sysvInitAction: sysinit
-    shell: 'cgconfigparser -l /etc/cgconfig.conf -s 1664'
-  # restrict permissions on /neonvm/bin/resize-swap, because we grant access to compute_ctl for
-  # running it as root.
-  - name: chmod-resize-swap
-    user: root
-    sysvInitAction: sysinit
-    shell: 'chmod 711 /neonvm/bin/resize-swap'
-  - name: pgbouncer
-    user: postgres
-    sysvInitAction: respawn
-    shell: '/usr/local/bin/pgbouncer /etc/pgbouncer.ini'
-  - name: postgres-exporter
-    user: nobody
-    sysvInitAction: respawn
-    shell: 'DATA_SOURCE_NAME="user=cloud_admin sslmode=disable dbname=postgres application_name=postgres-exporter" /bin/postgres_exporter'
-  - name: sql-exporter
-    user: nobody
-    sysvInitAction: respawn
-    shell: '/bin/sql_exporter -config.file=/etc/sql_exporter.yml -web.listen-address=:9399'
-  - name: sql-exporter-autoscaling
-    user: nobody
-    sysvInitAction: respawn
-    shell: '/bin/sql_exporter -config.file=/etc/sql_exporter_autoscaling.yml -web.listen-address=:9499'
-shutdownHook: |
-  su -p postgres --session-command '/usr/local/bin/pg_ctl stop -D /var/db/postgres/compute/pgdata -m fast --wait -t 10'
-files:
-  - filename: compute_ctl-resize-swap
-    content: |
-      # Allow postgres user (which is what compute_ctl runs as) to run /neonvm/bin/resize-swap
-      # as root without requiring entering a password (NOPASSWD), regardless of hostname (ALL)
-      postgres ALL=(root) NOPASSWD: /neonvm/bin/resize-swap
-  - filename: cgconfig.conf
-    content: |
-      # Configuration for cgroups in VM compute nodes
-      group neon-postgres {
-          perm {
-              admin {
-                  uid = postgres;
-              }
-              task {
-                  gid = users;
-              }
-          }
-          memory {}
-      }
-build: |
-  # Build cgroup-tools
-  #
-  # At time of writing (2023-03-14), debian bullseye has a version of cgroup-tools (technically
-  # libcgroup) that doesn't support cgroup v2 (version 0.41-11). Unfortunately, the vm-monitor
-  # requires cgroup v2, so we'll build cgroup-tools ourselves.
-  FROM debian:bullseye-slim as libcgroup-builder
-  ENV LIBCGROUP_VERSION=v2.0.3
-
-  RUN set -exu \
-      && apt update \
-      && apt install --no-install-recommends -y \
-          git \
-          ca-certificates \
-          automake \
-          cmake \
-          make \
-          gcc \
-          byacc \
-          flex \
-          libtool \
-          libpam0g-dev \
-      && git clone --depth 1 -b $LIBCGROUP_VERSION https://github.com/libcgroup/libcgroup \
-      && INSTALL_DIR="/libcgroup-install" \
-      && mkdir -p "$INSTALL_DIR/bin" "$INSTALL_DIR/include" \
-      && cd libcgroup \
-      # extracted from bootstrap.sh, with modified flags:
-      && (test -d m4 || mkdir m4) \
-      && autoreconf -fi \
-      && rm -rf autom4te.cache \
-      && CFLAGS="-O3" ./configure --prefix="$INSTALL_DIR" --sysconfdir=/etc --localstatedir=/var --enable-opaque-hierarchy="name=systemd" \
-      # actually build the thing...
-      && make install
-merge: |
-  # tweak nofile limits
-  RUN set -e \
-      && echo 'fs.file-max = 1048576' >>/etc/sysctl.conf \
-      && test ! -e /etc/security || ( \
-         echo '*    - nofile 1048576' >>/etc/security/limits.conf \
-      && echo 'root - nofile 1048576' >>/etc/security/limits.conf \
-         )
-
-  # Allow postgres user (compute_ctl) to run swap resizer.
-  # Need to install sudo in order to allow this.
-  #
-  # Also, remove the 'read' permission from group/other on /neonvm/bin/resize-swap, just to be safe.
-  RUN set -e \
-      && apt update \
-      && apt install --no-install-recommends -y \
-             sudo \
-      && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
-  COPY compute_ctl-resize-swap /etc/sudoers.d/compute_ctl-resize-swap
-
-  COPY cgconfig.conf /etc/cgconfig.conf
-
-  RUN set -e \
-      && chmod 0644 /etc/cgconfig.conf
-
-  COPY --from=libcgroup-builder /libcgroup-install/bin/*  /usr/bin/
-  COPY --from=libcgroup-builder /libcgroup-install/lib/*  /usr/lib/
-  COPY --from=libcgroup-builder /libcgroup-install/sbin/* /usr/sbin/

compute_tools/Cargo.toml

@@ -11,6 +11,7 @@ testing = []
 
 [dependencies]
 anyhow.workspace = true
+async-compression.workspace = true
 chrono.workspace = true
 cfg-if.workspace = true
 clap.workspace = true
@@ -23,6 +24,7 @@ num_cpus.workspace = true
 opentelemetry.workspace = true
 postgres.workspace = true
 regex.workspace = true
+serde.workspace = true
 serde_json.workspace = true
 signal-hook.workspace = true
 tar.workspace = true
@@ -41,6 +43,7 @@ url.workspace = true
 compute_api.workspace = true
 utils.workspace = true
 workspace_hack.workspace = true
+toml_edit.workspace = true
 remote_storage = { version = "0.1", path = "../libs/remote_storage/" }
 vm_monitor = { version = "0.1", path = "../libs/vm_monitor/" }
 zstd = "0.13"

compute_tools/src/compute.rs

@@ -1052,19 +1052,26 @@ impl ComputeNode {
         let pg_process = self.start_postgres(pspec.storage_auth_token.clone())?;
 
         let config_time = Utc::now();
-        if pspec.spec.mode == ComputeMode::Primary && !pspec.spec.skip_pg_catalog_updates {
+        if pspec.spec.mode == ComputeMode::Primary {
-            let pgdata_path = Path::new(&self.pgdata);
+            if !pspec.spec.skip_pg_catalog_updates {
-            // temporarily reset max_cluster_size in config
+                let pgdata_path = Path::new(&self.pgdata);
-            // to avoid the possibility of hitting the limit, while we are applying config:
+                // temporarily reset max_cluster_size in config
-            // creating new extensions, roles, etc...
+                // to avoid the possibility of hitting the limit, while we are applying config:
-            config::with_compute_ctl_tmp_override(pgdata_path, "neon.max_cluster_size=-1", || {
+                // creating new extensions, roles, etc...
+                config::with_compute_ctl_tmp_override(
+                    pgdata_path,
+                    "neon.max_cluster_size=-1",
+                    || {
+                        self.pg_reload_conf()?;
+
+                        self.apply_config(&compute_state)?;
+
+                        Ok(())
+                    },
+                )?;
                 self.pg_reload_conf()?;
-
+            }
-                self.apply_config(&compute_state)?;
+            self.post_apply_config()?;
-
-                Ok(())
-            })?;
-            self.pg_reload_conf()?;
         }
 
         let startup_end_time = Utc::now();

compute_tools/src/extension_server.rs

@@ -124,7 +124,6 @@ fn parse_pg_version(human_version: &str) -> &str {
             "14" => return "v14",
             "15" => return "v15",
             "16" => return "v16",
-            "17" => return "v17",
             _ => {}
         },
         _ => {}

compute_tools/src/migrations/0011-grant_pg_show_replication_origin_status_to_neon_superuser.sql

@@ -1 +0,0 @@
-GRANT EXECUTE ON FUNCTION pg_show_replication_origin_status TO neon_superuser;

compute_tools/src/spec.rs

@@ -793,9 +793,6 @@ pub fn handle_migrations(client: &mut Client) -> Result<()> {
         include_str!(
             "./migrations/0010-grant_snapshot_synchronization_funcs_to_neon_superuser.sql"
         ),
-        include_str!(
-            "./migrations/0011-grant_pg_show_replication_origin_status_to_neon_superuser.sql"
-        ),
     ];
 
     MigrationRunner::new(client, &migrations).run_migrations()?;

control_plane/Cargo.toml

@@ -9,9 +9,13 @@ anyhow.workspace = true
 camino.workspace = true
 clap.workspace = true
 comfy-table.workspace = true
+futures.workspace = true
+git-version.workspace = true
 humantime.workspace = true
 nix.workspace = true
 once_cell.workspace = true
+postgres.workspace = true
+hex.workspace = true
 humantime-serde.workspace = true
 hyper.workspace = true
 regex.workspace = true
@@ -19,6 +23,8 @@ reqwest = { workspace = true, features = ["blocking", "json"] }
 scopeguard.workspace = true
 serde.workspace = true
 serde_json.workspace = true
+serde_with.workspace = true
+tar.workspace = true
 thiserror.workspace = true
 toml.workspace = true
 toml_edit.workspace = true

control_plane/src/background_process.rs

@@ -151,7 +151,7 @@ where
                     print!(".");
                     io::stdout().flush().unwrap();
                 }
-                tokio::time::sleep(RETRY_INTERVAL).await;
+                thread::sleep(RETRY_INTERVAL);
             }
             Err(e) => {
                 println!("error starting process {process_name:?}: {e:#}");

control_plane/src/bin/neon_local.rs

@@ -34,14 +34,12 @@ use safekeeper_api::{
     DEFAULT_HTTP_LISTEN_PORT as DEFAULT_SAFEKEEPER_HTTP_PORT,
     DEFAULT_PG_LISTEN_PORT as DEFAULT_SAFEKEEPER_PG_PORT,
 };
-use std::borrow::Cow;
 use std::collections::{BTreeSet, HashMap};
 use std::path::PathBuf;
 use std::process::exit;
 use std::str::FromStr;
 use std::time::Duration;
 use storage_broker::DEFAULT_LISTEN_ADDR as DEFAULT_BROKER_ADDR;
-use tokio::task::JoinSet;
 use url::Host;
 use utils::{
     auth::{Claims, Scope},
@@ -89,35 +87,34 @@ fn main() -> Result<()> {
 
     // Check for 'neon init' command first.
     let subcommand_result = if sub_name == "init" {
-        handle_init(sub_args).map(|env| Some(Cow::Owned(env)))
+        handle_init(sub_args).map(Some)
     } else {
         // all other commands need an existing config
-
+        let mut env =
-        let env = LocalEnv::load_config(&local_env::base_path()).context("Error loading config")?;
+            LocalEnv::load_config(&local_env::base_path()).context("Error loading config")?;
         let original_env = env.clone();
-        let env = Box::leak(Box::new(env));
+
         let rt = tokio::runtime::Builder::new_current_thread()
             .enable_all()
             .build()
             .unwrap();
 
         let subcommand_result = match sub_name {
-            "tenant" => rt.block_on(handle_tenant(sub_args, env)),
+            "tenant" => rt.block_on(handle_tenant(sub_args, &mut env)),
-            "timeline" => rt.block_on(handle_timeline(sub_args, env)),
+            "timeline" => rt.block_on(handle_timeline(sub_args, &mut env)),
-            "start" => rt.block_on(handle_start_all(env, get_start_timeout(sub_args))),
+            "start" => rt.block_on(handle_start_all(&env, get_start_timeout(sub_args))),
-            "stop" => rt.block_on(handle_stop_all(sub_args, env)),
+            "stop" => rt.block_on(handle_stop_all(sub_args, &env)),
-            "pageserver" => rt.block_on(handle_pageserver(sub_args, env)),
+            "pageserver" => rt.block_on(handle_pageserver(sub_args, &env)),
-            "storage_controller" => rt.block_on(handle_storage_controller(sub_args, env)),
+            "storage_controller" => rt.block_on(handle_storage_controller(sub_args, &env)),
-            "storage_broker" => rt.block_on(handle_storage_broker(sub_args, env)),
+            "safekeeper" => rt.block_on(handle_safekeeper(sub_args, &env)),
-            "safekeeper" => rt.block_on(handle_safekeeper(sub_args, env)),
+            "endpoint" => rt.block_on(handle_endpoint(sub_args, &env)),
-            "endpoint" => rt.block_on(handle_endpoint(sub_args, env)),
+            "mappings" => handle_mappings(sub_args, &mut env),
-            "mappings" => handle_mappings(sub_args, env),
             "pg" => bail!("'pg' subcommand has been renamed to 'endpoint'"),
             _ => bail!("unexpected subcommand {sub_name}"),
         };
 
-        if &original_env != env {
+        if original_env != env {
-            subcommand_result.map(|()| Some(Cow::Borrowed(env)))
+            subcommand_result.map(|()| Some(env))
         } else {
             subcommand_result.map(|()| None)
         }
@@ -1248,122 +1245,49 @@ async fn handle_safekeeper(sub_match: &ArgMatches, env: &local_env::LocalEnv) ->
     Ok(())
 }
 
-async fn handle_storage_broker(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
-    let (sub_name, sub_args) = match sub_match.subcommand() {
-        Some(broker_command_data) => broker_command_data,
-        None => bail!("no broker subcommand provided"),
-    };
-
-    match sub_name {
-        "start" => {
-            if let Err(e) = broker::start_broker_process(env, get_start_timeout(sub_args)).await {
-                eprintln!("broker start failed: {e}");
-                exit(1);
-            }
-        }
-
-        "stop" => {
-            if let Err(e) = broker::stop_broker_process(env) {
-                eprintln!("broker stop failed: {e}");
-                exit(1);
-            }
-        }
-
-        _ => bail!("Unexpected broker subcommand '{}'", sub_name),
-    }
-    Ok(())
-}
-
 async fn handle_start_all(
-    env: &'static local_env::LocalEnv,
+    env: &local_env::LocalEnv,
     retry_timeout: &Duration,
 ) -> anyhow::Result<()> {
-    let Err(errors) = handle_start_all_impl(env, *retry_timeout).await else {
-        neon_start_status_check(env, retry_timeout)
-            .await
-            .context("status check after successful startup of all services")?;
-        return Ok(());
-    };
-
-    eprintln!("startup failed because one or more services could not be started");
-
-    for e in errors {
-        eprintln!("{e}");
-        let debug_repr = format!("{e:?}");
-        for line in debug_repr.lines() {
-            eprintln!("  {line}");
-        }
-    }
-
-    try_stop_all(env, true).await;
-
-    exit(2);
-}
-
-/// Returns Ok() if and only if all services could be started successfully.
-/// Otherwise, returns the list of errors that occurred during startup.
-async fn handle_start_all_impl(
-    env: &'static local_env::LocalEnv,
-    retry_timeout: Duration,
-) -> Result<(), Vec<anyhow::Error>> {
     // Endpoints are not started automatically
 
-    let mut js = JoinSet::new();
+    broker::start_broker_process(env, retry_timeout).await?;
 
-    // force infalliblity through closure
+    // Only start the storage controller if the pageserver is configured to need it
-    #[allow(clippy::redundant_closure_call)]
+    if env.control_plane_api.is_some() {
-    (|| {
+        let storage_controller = StorageController::from_env(env);
-        js.spawn(async move {
+        if let Err(e) = storage_controller
-            let retry_timeout = retry_timeout;
+            .start(NeonStorageControllerStartArgs::with_default_instance_id(
-            broker::start_broker_process(env, &retry_timeout).await
+                (*retry_timeout).into(),
-        });
+            ))
-
+            .await
-        // Only start the storage controller if the pageserver is configured to need it
+        {
-        if env.control_plane_api.is_some() {
+            eprintln!("storage_controller start failed: {:#}", e);
-            js.spawn(async move {
+            try_stop_all(env, true).await;
-                let storage_controller = StorageController::from_env(env);
+            exit(1);
-                storage_controller
-                    .start(NeonStorageControllerStartArgs::with_default_instance_id(
-                        retry_timeout.into(),
-                    ))
-                    .await
-                    .map_err(|e| e.context("start storage_controller"))
-            });
-        }
-
-        for ps_conf in &env.pageservers {
-            js.spawn(async move {
-                let pageserver = PageServerNode::from_env(env, ps_conf);
-                pageserver
-                    .start(&retry_timeout)
-                    .await
-                    .map_err(|e| e.context(format!("start pageserver {}", ps_conf.id)))
-            });
-        }
-
-        for node in env.safekeepers.iter() {
-            js.spawn(async move {
-                let safekeeper = SafekeeperNode::from_env(env, node);
-                safekeeper
-                    .start(vec![], &retry_timeout)
-                    .await
-                    .map_err(|e| e.context(format!("start safekeeper {}", safekeeper.id)))
-            });
-        }
-    })();
-
-    let mut errors = Vec::new();
-    while let Some(result) = js.join_next().await {
-        let result = result.expect("we don't panic or cancel the tasks");
-        if let Err(e) = result {
-            errors.push(e);
         }
     }
 
-    if !errors.is_empty() {
+    for ps_conf in &env.pageservers {
-        return Err(errors);
+        let pageserver = PageServerNode::from_env(env, ps_conf);
+        if let Err(e) = pageserver.start(retry_timeout).await {
+            eprintln!("pageserver {} start failed: {:#}", ps_conf.id, e);
+            try_stop_all(env, true).await;
+            exit(1);
+        }
     }
 
+    for node in env.safekeepers.iter() {
+        let safekeeper = SafekeeperNode::from_env(env, node);
+        if let Err(e) = safekeeper.start(vec![], retry_timeout).await {
+            eprintln!("safekeeper {} start failed: {:#}", safekeeper.id, e);
+            try_stop_all(env, false).await;
+            exit(1);
+        }
+    }
+
+    neon_start_status_check(env, retry_timeout).await?;
+
     Ok(())
 }
 
@@ -1748,19 +1672,6 @@ fn cli() -> Command {
                             .arg(stop_mode_arg.clone())
                             .arg(instance_id))
         )
-        .subcommand(
-            Command::new("storage_broker")
-                .arg_required_else_help(true)
-                .about("Manage broker")
-                .subcommand(Command::new("start")
-                            .about("Start broker")
-                            .arg(timeout_arg.clone())
-                )
-                .subcommand(Command::new("stop")
-                            .about("Stop broker")
-                            .arg(stop_mode_arg.clone())
-                )
-        )
         .subcommand(
             Command::new("safekeeper")
                 .arg_required_else_help(true)

control_plane/src/endpoint.rs

@@ -702,7 +702,7 @@ impl Endpoint {
                     }
                 }
             }
-            tokio::time::sleep(ATTEMPT_INTERVAL).await;
+            std::thread::sleep(ATTEMPT_INTERVAL);
         }
 
         // disarm the scopeguard, let the child outlive this function (and neon_local invoction)

control_plane/src/local_env.rs

@@ -342,7 +342,7 @@ impl LocalEnv {
 
         #[allow(clippy::manual_range_patterns)]
         match pg_version {
-            14 | 15 | 16 | 17 => Ok(path.join(format!("v{pg_version}"))),
+            14 | 15 | 16 => Ok(path.join(format!("v{pg_version}"))),
             _ => bail!("Unsupported postgres version: {}", pg_version),
         }
     }

control_plane/src/pageserver.rs

@@ -17,7 +17,9 @@ use std::time::Duration;
 
 use anyhow::{bail, Context};
 use camino::Utf8PathBuf;
-use pageserver_api::models::{self, AuxFilePolicy, TenantInfo, TimelineInfo};
+use pageserver_api::models::{
+    self, AuxFilePolicy, LocationConfig, TenantHistorySize, TenantInfo, TimelineInfo,
+};
 use pageserver_api::shard::TenantShardId;
 use pageserver_client::mgmt_api;
 use postgres_backend::AuthType;
@@ -322,6 +324,22 @@ impl PageServerNode {
         background_process::stop_process(immediate, "pageserver", &self.pid_file())
     }
 
+    pub async fn page_server_psql_client(
+        &self,
+    ) -> anyhow::Result<(
+        tokio_postgres::Client,
+        tokio_postgres::Connection<tokio_postgres::Socket, tokio_postgres::tls::NoTlsStream>,
+    )> {
+        let mut config = self.pg_connection_config.clone();
+        if self.conf.pg_auth_type == AuthType::NeonJWT {
+            let token = self
+                .env
+                .generate_auth_token(&Claims::new(None, Scope::PageServerApi))?;
+            config = config.set_password(Some(token));
+        }
+        Ok(config.connect_no_tls().await?)
+    }
+
     pub async fn check_status(&self) -> mgmt_api::Result<()> {
         self.http_client.status().await
     }
@@ -522,6 +540,19 @@ impl PageServerNode {
         Ok(())
     }
 
+    pub async fn location_config(
+        &self,
+        tenant_shard_id: TenantShardId,
+        config: LocationConfig,
+        flush_ms: Option<Duration>,
+        lazy: bool,
+    ) -> anyhow::Result<()> {
+        Ok(self
+            .http_client
+            .location_config(tenant_shard_id, config, flush_ms, lazy)
+            .await?)
+    }
+
     pub async fn timeline_list(
         &self,
         tenant_shard_id: &TenantShardId,
@@ -605,4 +636,14 @@ impl PageServerNode {
 
         Ok(())
     }
+
+    pub async fn tenant_synthetic_size(
+        &self,
+        tenant_shard_id: TenantShardId,
+    ) -> anyhow::Result<TenantHistorySize> {
+        Ok(self
+            .http_client
+            .tenant_synthetic_size(tenant_shard_id)
+            .await?)
+    }
 }

control_plane/src/postgresql_conf.rs

@@ -4,10 +4,13 @@
 /// NOTE: This doesn't implement the full, correct postgresql.conf syntax. Just
 /// enough to extract a few settings we need in Neon, assuming you don't do
 /// funny stuff like include-directives or funny escaping.
+use anyhow::{bail, Context, Result};
 use once_cell::sync::Lazy;
 use regex::Regex;
 use std::collections::HashMap;
 use std::fmt;
+use std::io::BufRead;
+use std::str::FromStr;
 
 /// In-memory representation of a postgresql.conf file
 #[derive(Default, Debug)]
@@ -16,16 +19,84 @@ pub struct PostgresConf {
     hash: HashMap<String, String>,
 }
 
+static CONF_LINE_RE: Lazy<Regex> = Lazy::new(|| Regex::new(r"^((?:\w|\.)+)\s*=\s*(\S+)$").unwrap());
+
 impl PostgresConf {
     pub fn new() -> PostgresConf {
         PostgresConf::default()
     }
 
+    /// Read file into memory
+    pub fn read(read: impl std::io::Read) -> Result<PostgresConf> {
+        let mut result = Self::new();
+
+        for line in std::io::BufReader::new(read).lines() {
+            let line = line?;
+
+            // Store each line in a vector, in original format
+            result.lines.push(line.clone());
+
+            // Also parse each line and insert key=value lines into a hash map.
+            //
+            // FIXME: This doesn't match exactly the flex/bison grammar in PostgreSQL.
+            // But it's close enough for our usage.
+            let line = line.trim();
+            if line.starts_with('#') {
+                // comment, ignore
+                continue;
+            } else if let Some(caps) = CONF_LINE_RE.captures(line) {
+                let name = caps.get(1).unwrap().as_str();
+                let raw_val = caps.get(2).unwrap().as_str();
+
+                if let Ok(val) = deescape_str(raw_val) {
+                    // Note: if there's already an entry in the hash map for
+                    // this key, this will replace it. That's the behavior what
+                    // we want; when PostgreSQL reads the file, each line
+                    // overrides any previous value for the same setting.
+                    result.hash.insert(name.to_string(), val.to_string());
+                }
+            }
+        }
+        Ok(result)
+    }
+
     /// Return the current value of 'option'
     pub fn get(&self, option: &str) -> Option<&str> {
         self.hash.get(option).map(|x| x.as_ref())
     }
 
+    /// Return the current value of a field, parsed to the right datatype.
+    ///
+    /// This calls the FromStr::parse() function on the value of the field. If
+    /// the field does not exist, or parsing fails, returns an error.
+    ///
+    pub fn parse_field<T>(&self, field_name: &str, context: &str) -> Result<T>
+    where
+        T: FromStr,
+        <T as FromStr>::Err: std::error::Error + Send + Sync + 'static,
+    {
+        self.get(field_name)
+            .with_context(|| format!("could not find '{}' option {}", field_name, context))?
+            .parse::<T>()
+            .with_context(|| format!("could not parse '{}' option {}", field_name, context))
+    }
+
+    pub fn parse_field_optional<T>(&self, field_name: &str, context: &str) -> Result<Option<T>>
+    where
+        T: FromStr,
+        <T as FromStr>::Err: std::error::Error + Send + Sync + 'static,
+    {
+        if let Some(val) = self.get(field_name) {
+            let result = val
+                .parse::<T>()
+                .with_context(|| format!("could not parse '{}' option {}", field_name, context))?;
+
+            Ok(Some(result))
+        } else {
+            Ok(None)
+        }
+    }
+
     ///
     /// Note: if you call this multiple times for the same option, the config
     /// file will a line for each call. It would be nice to have a function
@@ -83,8 +154,48 @@ fn escape_str(s: &str) -> String {
     }
 }
 
+/// De-escape a possibly-quoted value.
+///
+/// See `DeescapeQuotedString` function in PostgreSQL sources for how PostgreSQL
+/// does this.
+fn deescape_str(s: &str) -> Result<String> {
+    // If the string has a quote at the beginning and end, strip them out.
+    if s.len() >= 2 && s.starts_with('\'') && s.ends_with('\'') {
+        let mut result = String::new();
+
+        let mut iter = s[1..(s.len() - 1)].chars().peekable();
+        while let Some(c) = iter.next() {
+            let newc = if c == '\\' {
+                match iter.next() {
+                    Some('b') => '\x08',
+                    Some('f') => '\x0c',
+                    Some('n') => '\n',
+                    Some('r') => '\r',
+                    Some('t') => '\t',
+                    Some('0'..='7') => {
+                        // TODO
+                        bail!("octal escapes not supported");
+                    }
+                    Some(n) => n,
+                    None => break,
+                }
+            } else if c == '\'' && iter.peek() == Some(&'\'') {
+                // doubled quote becomes just one quote
+                iter.next().unwrap()
+            } else {
+                c
+            };
+
+            result.push(newc);
+        }
+        Ok(result)
+    } else {
+        Ok(s.to_string())
+    }
+}
+
 #[test]
-fn test_postgresql_conf_escapes() -> anyhow::Result<()> {
+fn test_postgresql_conf_escapes() -> Result<()> {
     assert_eq!(escape_str("foo bar"), "'foo bar'");
     // these don't need to be quoted
     assert_eq!(escape_str("foo"), "foo");
@@ -103,5 +214,13 @@ fn test_postgresql_conf_escapes() -> anyhow::Result<()> {
     assert_eq!(escape_str("fo\\o"), "'fo\\\\o'");
     assert_eq!(escape_str("10 cats"), "'10 cats'");
 
+    // Test de-escaping
+    assert_eq!(deescape_str(&escape_str("foo"))?, "foo");
+    assert_eq!(deescape_str(&escape_str("fo'o\nba\\r"))?, "fo'o\nba\\r");
+    assert_eq!(deescape_str("'\\b\\f\\n\\r\\t'")?, "\x08\x0c\n\r\t");
+
+    // octal-escapes are currently not supported
+    assert!(deescape_str("'foo\\7\\07\\007'").is_err());
+
     Ok(())
 }

control_plane/src/storage_controller.rs

@@ -28,7 +28,6 @@ use utils::{
     auth::{encode_from_key_file, Claims, Scope},
     id::{NodeId, TenantId},
 };
-use whoami::username;
 
 pub struct StorageController {
     env: LocalEnv,
@@ -184,7 +183,7 @@ impl StorageController {
     /// to other versions if that one isn't found.  Some automated tests create circumstances
     /// where only one version is available in pg_distrib_dir, such as `test_remote_extensions`.
     async fn get_pg_dir(&self, dir_name: &str) -> anyhow::Result<Utf8PathBuf> {
-        let prefer_versions = [STORAGE_CONTROLLER_POSTGRES_VERSION, 16, 15, 14];
+        let prefer_versions = [STORAGE_CONTROLLER_POSTGRES_VERSION, 15, 14];
 
         for v in prefer_versions {
             let path = Utf8PathBuf::from_path_buf(self.env.pg_dir(v, dir_name)?).unwrap();
@@ -212,16 +211,7 @@ impl StorageController {
     /// Readiness check for our postgres process
     async fn pg_isready(&self, pg_bin_dir: &Utf8Path, postgres_port: u16) -> anyhow::Result<bool> {
         let bin_path = pg_bin_dir.join("pg_isready");
-        let args = [
+        let args = ["-h", "localhost", "-p", &format!("{}", postgres_port)];
-            "-h",
-            "localhost",
-            "-U",
-            &username(),
-            "-d",
-            DB_NAME,
-            "-p",
-            &format!("{}", postgres_port),
-        ];
         let exitcode = Command::new(bin_path).args(args).spawn()?.wait().await?;
 
         Ok(exitcode.success())
@@ -235,11 +225,7 @@ impl StorageController {
     ///
     /// Returns the database url
     pub async fn setup_database(&self, postgres_port: u16) -> anyhow::Result<String> {
-        let database_url = format!(
+        let database_url = format!("postgresql://localhost:{}/{DB_NAME}", postgres_port);
-            "postgresql://{}@localhost:{}/{DB_NAME}",
-            &username(),
-            postgres_port
-        );
 
         let pg_bin_dir = self.get_pg_bin_dir().await?;
         let createdb_path = pg_bin_dir.join("createdb");
@@ -249,10 +235,6 @@ impl StorageController {
                 "localhost",
                 "-p",
                 &format!("{}", postgres_port),
-                "-U",
-                &username(),
-                "-O",
-                &username(),
                 DB_NAME,
             ])
             .output()
@@ -289,7 +271,7 @@ impl StorageController {
             // But tokio-postgres fork doesn't have this upstream commit:
             // https://github.com/sfackler/rust-postgres/commit/cb609be758f3fb5af537f04b584a2ee0cebd5e79
             // => we should rebase our fork => TODO https://github.com/neondatabase/neon/issues/8399
-            .user(&username())
+            .user(&whoami::username())
             .dbname(DB_NAME)
             .connect(tokio_postgres::NoTls)
             .await
@@ -346,19 +328,6 @@ impl StorageController {
             let pg_log_path = pg_data_path.join("postgres.log");
 
             if !tokio::fs::try_exists(&pg_data_path).await? {
-                let initdb_args = [
-                    "-D",
-                    pg_data_path.as_ref(),
-                    "--username",
-                    &username(),
-                    "--no-sync",
-                    "--no-instructions",
-                ];
-                tracing::info!(
-                    "Initializing storage controller database with args: {:?}",
-                    initdb_args
-                );
-
                 // Initialize empty database
                 let initdb_path = pg_bin_dir.join("initdb");
                 let mut child = Command::new(&initdb_path)
@@ -366,7 +335,7 @@ impl StorageController {
                         ("LD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
                         ("DYLD_LIBRARY_PATH".to_owned(), pg_lib_dir.to_string()),
                     ])
-                    .args(initdb_args)
+                    .args(["-D", pg_data_path.as_ref()])
                     .spawn()
                     .expect("Failed to spawn initdb");
                 let status = child.wait().await?;
@@ -395,14 +364,8 @@ impl StorageController {
                 pg_data_path.as_ref(),
                 "-l",
                 pg_log_path.as_ref(),
-                "-U",
-                &username(),
                 "start",
             ];
-            tracing::info!(
-                "Starting storage controller database with args: {:?}",
-                db_start_args
-            );
 
             background_process::start_process(
                 "storage_controller_db",

control_plane/storcon_cli/Cargo.toml

@@ -11,11 +11,14 @@ clap.workspace = true
 comfy-table.workspace = true
 futures.workspace = true
 humantime.workspace = true
+hyper.workspace = true
 pageserver_api.workspace = true
 pageserver_client.workspace = true
 reqwest.workspace = true
+serde.workspace = true
 serde_json = { workspace = true, features = ["raw_value"] }
 storage_controller_client.workspace = true
+thiserror.workspace = true
 tokio.workspace = true
 tracing.workspace = true
 utils.workspace = true

control_plane/storcon_cli/src/main.rs

@@ -4,8 +4,8 @@ use std::{str::FromStr, time::Duration};
 use clap::{Parser, Subcommand};
 use pageserver_api::{
     controller_api::{
-        AvailabilityZone, NodeAvailabilityWrapper, NodeDescribeResponse, NodeShardResponse,
+        NodeAvailabilityWrapper, NodeDescribeResponse, NodeShardResponse, ShardSchedulingPolicy,
-        ShardSchedulingPolicy, TenantCreateRequest, TenantDescribeResponse, TenantPolicyRequest,
+        TenantCreateRequest, TenantDescribeResponse, TenantPolicyRequest,
     },
     models::{
         EvictionPolicy, EvictionPolicyLayerAccessThreshold, LocationConfigSecondary,
@@ -339,7 +339,7 @@ async fn main() -> anyhow::Result<()> {
                         listen_pg_port,
                         listen_http_addr,
                         listen_http_port,
-                        availability_zone_id: AvailabilityZone(availability_zone_id),
+                        availability_zone_id,
                     }),
                 )
                 .await?;

docker-compose/README.md

@@ -2,8 +2,8 @@
 # Example docker compose configuration
 
 The configuration in this directory is used for testing Neon docker images: it is
-not intended for deploying a usable system. To run a development environment where
+not intended for deploying a usable system.  To run a development environment where
-you can experiment with a miniature Neon system, use `cargo neon` rather than container images.
+you can experiment with a minature Neon system, use `cargo neon` rather than container images.
 
 This configuration does not start the storage controller, because the controller
 needs a way to reconfigure running computes, and no such thing exists in this setup.

docs/rfcs/038-independent-compute-release.md

@@ -1,343 +0,0 @@
-# Independent compute release
-
-Created at: 2024-08-30. Author: Alexey Kondratov (@ololobus)
-
-## Summary
-
-This document proposes an approach to fully independent compute release flow. It attempts to
-cover the following features:
-
-- Process is automated as much as possible to minimize human errors.
-- Compute<->storage protocol compatibility is ensured.
-- A transparent release history is available with an easy rollback strategy.
-- Although not in the scope of this document, there is a viable way to extend the proposed release
-  flow to achieve the canary and/or blue-green deployment strategies.
-
-## Motivation
-
-Previously, the compute release was tightly coupled to the storage release. This meant that once
-some storage nodes got restarted with a newer version, all new compute starts using these nodes
-automatically got a new version. Thus, two releases happen in parallel, which increases the blast
-radius and makes ownership fuzzy.
-
-Now, we practice a manual v0 independent compute release flow -- after getting a new compute release
-image and tag, we pin it region by region using Admin UI. It's better, but it still has its own flaws:
-
-1. It's a simple but fairly manual process, as you need to click through a few pages.
-2. It's prone to human errors, e.g., you could mistype or copy the wrong compute tag.
-3. We now require an additional approval in the Admin UI, which partially solves the 2.,
-   but also makes the whole process pretty annoying, as you constantly need to go back
-   and forth between two people.
-
-## Non-goals
-
-It's not the goal of this document to propose a design for some general-purpose release tool like Helm.
-The document considers how the current compute fleet is orchestrated at Neon. Even if we later
-decide to split the control plane further (e.g., introduce a separate compute controller), the proposed
-release process shouldn't change much, i.e., the releases table and API will reside in
-one of the parts.
-
-Achieving the canary and/or blue-green deploy strategies is out of the scope of this document. They
-were kept in mind, though, so it's expected that the proposed approach will lay down the foundation
-for implementing them in future iterations.
-
-## Impacted components
-
-Compute, control plane, CI, observability (some Grafana dashboards may require changes).
-
-## Prior art
-
-One of the very close examples is how Helm tracks [releases history](https://helm.sh/docs/helm/helm_history/).
-
-In the code:
-
-- [Release](https://github.com/helm/helm/blob/2b30cf4b61d587d3f7594102bb202b787b9918db/pkg/release/release.go#L20-L43)
-- [Release info](https://github.com/helm/helm/blob/2b30cf4b61d587d3f7594102bb202b787b9918db/pkg/release/info.go#L24-L40)
-- [Release status](https://github.com/helm/helm/blob/2b30cf4b61d587d3f7594102bb202b787b9918db/pkg/release/status.go#L18-L42)
-
-TL;DR it has several important attributes:
-
-- Revision -- unique release ID/primary key. It is not the same as the application version,
-  because the same version can be deployed several times, e.g., after a newer version rollback.
-- App version -- version of the application chart/code.
-- Config -- set of overrides to the default config of the application.
-- Status -- current status of the release in the history.
-- Timestamps -- tracks when a release was created and deployed.
-
-## Proposed implementation
-
-### Separate release branch
-
-We will use a separate release branch, `release-compute`, to have a clean history for releases and commits.
-In order to avoid confusion with storage releases, we will use a different prefix for compute [git release
-tags](https://github.com/neondatabase/neon/releases) -- `release-compute-XXXX`. We will use the same tag for
-Docker images as well. The `neondatabase/compute-node-v16:release-compute-XXXX` looks longer and a bit redundant,
-but it's better to have image and git tags in sync.
-
-Currently, control plane relies on the numeric compute and storage release versions to decide on compute->storage
-compatibility. Once we implement this proposal, we should drop this code as release numbers will be completely
-independent. The only constraint we want is that it must monotonically increase within the same release branch.
-
-### Compute config/settings manifest
-
-We will create a new sub-directory `compute` and file `compute/manifest.yaml` with a structure:
-
-```yaml
-pg_settings:
-  # Common settings for primaries and secondaries of all versions.
-  common:
-    wal_log_hints: "off"
-    max_wal_size: "1024"
-
-  per_version:
-    14:
-      # Common settings for both replica and primary of version PG 14
-      common:
-        shared_preload_libraries: "neon,pg_stat_statements,extension_x"
-    15:
-      common:
-        shared_preload_libraries: "neon,pg_stat_statements,extension_x"
-      # Settings that should be applied only to
-      replica:
-        # Available only starting Postgres 15th
-        recovery_prefetch: "off"
-    # ...
-    17:
-      common:
-        # For example, if third-party `extension_x` is not yet available for PG 17
-        shared_preload_libraries: "neon,pg_stat_statements"
-      replica:
-        recovery_prefetch: "off"
-```
-
-**N.B.** Setting value should be a string with `on|off` for booleans and a number (as a string)
-without units for all numeric settings. That's how the control plane currently operates.
-
-The priority of settings will be (a higher number is a higher priority):
-
-1. Any static and hard-coded settings in the control plane
-2. `pg_settings->common`
-3. Per-version `common`
-4. Per-version `replica`
-5. Any per-user/project/endpoint overrides in the control plane
-6. Any dynamic setting calculated based on the compute size
-
-**N.B.** For simplicity, we do not do any custom logic for `shared_preload_libraries`, so it's completely
-overridden if specified on some level. Make sure that you include all necessary extensions in it when you
-do any overrides.
-
-**N.B.** There is a tricky question about what to do with custom compute image pinning we sometimes
-do for particular projects and customers. That's usually some ad-hoc work and images are based on
-the latest compute image, so it's relatively safe to assume that we could use settings from the latest compute
-release. If for some reason that's not true, and further overrides are needed, it's also possible to do
-on the project level together with pinning the image, so it's on-call/engineer/support responsibility to
-ensure that compute starts with the specified custom image. The only real risk is that compute image will get
-stale and settings from new releases will drift away, so eventually it will get something incompatible,
-but i) this is some operational issue, as we do not want stale images anyway, and ii) base settings
-receive something really new so rarely that the chance of this happening is very low. If we want to solve it completely,
-then together with pinning the image we could also pin the matching release revision in the control plane.
-
-The compute team will own the content of `compute/manifest.yaml`.
-
-### Control plane: releases table
-
-In order to store information about releases, the control plane will use a table `compute_releases` with the following
-schema:
-
-```sql
-CREATE TABLE compute_releases (
-  -- Unique release ID
-  -- N.B. Revision won't by synchronized across all regions, because all control planes are technically independent
-  -- services. We have the same situation with Helm releases as well because they could be deployed and rolled back
-  -- independently in different clusters.
-  revision BIGSERIAL PRIMARY KEY,
-  -- Numeric version of the compute image, e.g. 9057
-  version BIGINT NOT NULL,
-  -- Compute image tag, e.g. `release-9057`
-  tag TEXT NOT NULL,
-  -- Current release status. Currently, it will be a simple enum
-  -- * `deployed` -- release is deployed and used for new compute starts.
-  --                 Exactly one release can have this status at a time.
-  -- * `superseded` -- release has been replaced by a newer one.
-  -- But we can always extend it in the future when we need more statuses
-  -- for more complex deployment strategies.
-  status TEXT NOT NULL,
-  -- Any additional metadata for compute in the corresponding release
-  manifest JSONB NOT NULL,
-  -- Timestamp when release record was created in the control plane database
-  created_at TIMESTAMP NOT NULL DEFAULT now(),
-  -- Timestamp when release deployment was finished
-  deployed_at TIMESTAMP
-);
-```
-
-We keep track of the old releases not only for the sake of audit, but also because we usually have ~30% of
-old computes started using the image from one of the previous releases. Yet, when users want to reconfigure
-them without restarting, the control plane needs to know what settings are applicable to them, so we also need
-information about the previous releases that are readily available. There could be some other auxiliary info
-needed as well: supported extensions, compute flags, etc.
-
-**N.B.** Here, we can end up in an ambiguous situation when the same compute image is deployed twice, e.g.,
-it was deployed once, then rolled back, and then deployed again, potentially with a different manifest. Yet,
-we could've started some computes with the first deployment and some with the second. Thus, when we need to
-look up the manifest for the compute by its image tag, we will see two records in the table with the same tag,
-but different revision numbers. We can assume that this could happen only in case of rollbacks, so we
-can just take the latest revision for the given tag.
-
-### Control plane: management API
-
-The control plane will implement new API methods to manage releases:
-
-1. `POST /management/api/v2/compute_releases` to create a new release. With payload
-
-   ```json
-    {
-      "version": 9057,
-      "tag": "release-9057",
-      "manifest": {}
-    }
-   ```
-
-   and response
-
-   ```json
-    {
-      "revision": 53,
-      "version": 9057,
-      "tag": "release-9057",
-      "status": "deployed",
-      "manifest": {},
-      "created_at": "2024-08-15T15:52:01.0000Z",
-      "deployed_at": "2024-08-15T15:52:01.0000Z",
-    }
-   ```
-
-   Here, we can actually mix-in custom (remote) extensions metadata into the `manifest`, so that the control plane
-   will get information about all available extensions not bundled into compute image. The corresponding
-   workflow in `neondatabase/build-custom-extensions` should produce it as an artifact and make
-   it accessible to the workflow in the `neondatabase/infra`. See the complete release flow below. Doing that,
-   we put a constraint that new custom extension requires new compute release, which is good for the safety,
-   but is not exactly what we want operational-wise (we want to be able to deploy new extensions without new
-   images). Yet, it can be solved incrementally: v0 -- do not do anything with extensions at all;
-   v1 -- put them into the same manifest; v2 -- make them separate entities with their own lifecycle.
-
-   **N.B.** This method is intended to be used in CI workflows, and CI/network can be flaky. It's reasonable
-   to assume that we could retry the request several times, even though it's already succeeded. Although it's
-   not a big deal to create several identical releases one-by-one, it's better to avoid it, so the control plane
-   should check if the latest release is identical and just return `304 Not Modified` in this case.
-
-2. `POST /management/api/v2/compute_releases/rollback` to rollback to any previously deployed release. With payload
-   including the revision of the release to rollback to:
-
-   ```json
-   {
-      "revision": 52
-   }
-   ```
-
-   Rollback marks the current release as `superseded` and creates a new release with all the same data as the
-   requested revision, but with a new revision number.
-
-   This rollback API is not strictly needed, as we can just use `infra` repo workflow to deploy any
-   available tag. It's still nice to have for on-call and any urgent matters, for example, if we need
-   to rollback and GitHub is down. It's much easier to specify only the revision number vs. crafting
-   all the necessary data for the new release payload.
-
-### Compute->storage compatibility tests
-
-In order to safely release new compute versions independently from storage, we need to ensure that the currently
-deployed storage is compatible with the new compute version. Currently, we maintain backward compatibility
-in storage, but newer computes may require a newer storage version.
-
-Remote end-to-end (e2e) tests [already accept](https://github.com/neondatabase/cloud/blob/e3468d433e0d73d02b7d7e738d027f509b522408/.github/workflows/testing.yml#L43-L48)
-`storage_image_tag` and `compute_image_tag` as separate inputs. That means that we could reuse e2e tests to ensure
-compatibility between storage and compute:
-
-1. Pick the latest storage release tag and use it as `storage_image_tag`.
-2. Pick a new compute tag built in the current compute release PR and use it as `compute_image_tag`.
-   Here, we should use a temporary ECR image tag, because the final tag will be known only after the release PR is merged.
-3. Trigger e2e tests as usual.
-
-### Release flow
-
-```mermaid
-  sequenceDiagram
-
-  actor oncall as Compute on-call person
-  participant neon as neondatabase/neon
-
-  box private
-    participant cloud as neondatabase/cloud
-    participant exts as neondatabase/build-custom-extensions
-    participant infra as neondatabase/infra
-  end
-
-  box cloud
-    participant preprod as Pre-prod control plane
-    participant prod as Production control plane
-    participant k8s as Compute k8s
-  end
-
-  oncall ->> neon: Open release PR into release-compute
-
-  activate neon
-  neon ->> cloud: CI: trigger e2e compatibility tests
-  activate cloud
-  cloud -->> neon: CI: e2e tests pass
-  deactivate cloud
-  neon ->> neon: CI: pass PR checks, get approvals
-  deactivate neon
-
-  oncall ->> neon: Merge release PR into release-compute
-
-  activate neon
-  neon ->> neon: CI: pass checks, build and push images
-  neon ->> exts: CI: trigger extensions build
-  activate exts
-  exts -->> neon: CI: extensions are ready
-  deactivate exts
-  neon ->> neon: CI: create release tag
-  neon ->> infra: Trigger release workflow using the produced tag
-  deactivate neon
-
-  activate infra
-  infra ->> infra: CI: pass checks
-  infra ->> preprod: Release new compute image to pre-prod automatically <br/> POST /management/api/v2/compute_releases
-  activate preprod
-  preprod -->> infra: 200 OK
-  deactivate preprod
-
-  infra ->> infra: CI: wait for per-region production deploy approvals
-  oncall ->> infra: CI: approve deploys region by region
-  infra ->> k8s: Prewarm new compute image
-  infra ->> prod: POST /management/api/v2/compute_releases
-  activate prod
-  prod -->> infra: 200 OK
-  deactivate prod
-  deactivate infra
-```
-
-## Further work
-
-As briefly mentioned in other sections, eventually, we would like to use more complex deployment strategies.
-For example, we can pass a fraction of the total compute starts that should use the new release. Then we can
-mark the release as `partial` or `canary` and monitor its performance. If everything is fine, we can promote it
-to `deployed` status. If not, we can roll back to the previous one.
-
-## Alternatives
-
-In theory, we can try using Helm as-is:
-
-1. Write a compute Helm chart. That will actually have only some config map, which the control plane can access and read.
-   N.B. We could reuse the control plane chart as well, but then it's not a fully independent release again and even more fuzzy.
-2. The control plane will read it and start using the new compute version for new starts.
-
-Drawbacks:
-
-1. Helm releases work best if the workload is controlled by the Helm chart itself. Then you can have different
-   deployment strategies like rolling update or canary or blue/green deployments. At Neon, the compute starts are controlled
-   by control plane, so it makes it much more tricky.
-2. Releases visibility will suffer, i.e. instead of a nice table in the control plane and Admin UI, we would need to use
-   `helm` cli and/or K8s UIs like K8sLens.
-3. We do not restart all computes shortly after the new version release. This means that for some features and compatibility
-   purpose (see above) control plane may need some auxiliary info from the previous releases.

libs/compute_api/Cargo.toml

@@ -8,6 +8,7 @@ license.workspace = true
 anyhow.workspace = true
 chrono.workspace = true
 serde.workspace = true
+serde_with.workspace = true
 serde_json.workspace = true
 regex.workspace = true
 

libs/consumption_metrics/Cargo.toml

@@ -5,6 +5,9 @@ edition = "2021"
 license = "Apache-2.0"
 
 [dependencies]
+anyhow.workspace = true
 chrono = { workspace = true, features = ["serde"] }
 rand.workspace = true
 serde.workspace = true
+serde_with.workspace = true
+utils.workspace = true

libs/consumption_metrics/src/lib.rs

@@ -5,7 +5,7 @@ use chrono::{DateTime, Utc};
 use rand::Rng;
 use serde::{Deserialize, Serialize};
 
-#[derive(Serialize, Deserialize, Debug, Clone, Copy, Eq, PartialEq, Ord, PartialOrd)]
+#[derive(Serialize, serde::Deserialize, Debug, Clone, Copy, Eq, PartialEq, Ord, PartialOrd)]
 #[serde(tag = "type")]
 pub enum EventType {
     #[serde(rename = "absolute")]
@@ -107,7 +107,7 @@ pub const CHUNK_SIZE: usize = 1000;
 
 // Just a wrapper around a slice of events
 // to serialize it as `{"events" : [ ] }
-#[derive(serde::Serialize, Deserialize)]
+#[derive(serde::Serialize, serde::Deserialize)]
 pub struct EventChunk<'a, T: Clone> {
     pub events: std::borrow::Cow<'a, [T]>,
 }

libs/desim/Cargo.toml

@@ -12,4 +12,5 @@ bytes.workspace = true
 utils.workspace = true
 parking_lot.workspace = true
 hex.workspace = true
+scopeguard.workspace = true
 smallvec = { workspace = true, features = ["write"] }

libs/pageserver_api/src/config.rs

@@ -104,6 +104,9 @@ pub struct ConfigToml {
     pub image_compression: ImageCompressionAlgorithm,
     pub ephemeral_bytes_per_memory_kb: usize,
     pub l0_flush: Option<crate::models::L0FlushConfig>,
+    #[serde(skip_serializing)]
+    // TODO(https://github.com/neondatabase/neon/issues/8184): remove after this field is removed from all pageserver.toml's
+    pub compact_level0_phase1_value_access: serde::de::IgnoredAny,
     pub virtual_file_direct_io: crate::models::virtual_file::DirectIoMode,
     pub io_buffer_alignment: usize,
 }
@@ -170,6 +173,40 @@ impl Default for EvictionOrder {
     }
 }
 
+#[derive(
+    Eq,
+    PartialEq,
+    Debug,
+    Copy,
+    Clone,
+    strum_macros::EnumString,
+    strum_macros::Display,
+    serde_with::DeserializeFromStr,
+    serde_with::SerializeDisplay,
+)]
+#[strum(serialize_all = "kebab-case")]
+pub enum GetVectoredImpl {
+    Sequential,
+    Vectored,
+}
+
+#[derive(
+    Eq,
+    PartialEq,
+    Debug,
+    Copy,
+    Clone,
+    strum_macros::EnumString,
+    strum_macros::Display,
+    serde_with::DeserializeFromStr,
+    serde_with::SerializeDisplay,
+)]
+#[strum(serialize_all = "kebab-case")]
+pub enum GetImpl {
+    Legacy,
+    Vectored,
+}
+
 #[derive(Copy, Clone, Debug, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
 #[serde(transparent)]
 pub struct MaxVectoredReadBytes(pub NonZeroUsize);
@@ -301,6 +338,8 @@ pub mod defaults {
     pub const DEFAULT_IMAGE_COMPRESSION: ImageCompressionAlgorithm =
         ImageCompressionAlgorithm::Zstd { level: Some(1) };
 
+    pub const DEFAULT_VALIDATE_VECTORED_GET: bool = false;
+
     pub const DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB: usize = 0;
 
     pub const DEFAULT_IO_BUFFER_ALIGNMENT: usize = 512;
@@ -337,10 +376,7 @@ impl Default for ConfigToml {
 
             concurrent_tenant_warmup: (NonZeroUsize::new(DEFAULT_CONCURRENT_TENANT_WARMUP)
                 .expect("Invalid default constant")),
-            concurrent_tenant_size_logical_size_queries: NonZeroUsize::new(
+            concurrent_tenant_size_logical_size_queries: NonZeroUsize::new(1).unwrap(),
-                DEFAULT_CONCURRENT_TENANT_SIZE_LOGICAL_SIZE_QUERIES,
-            )
-            .unwrap(),
             metric_collection_interval: (humantime::parse_duration(
                 DEFAULT_METRIC_COLLECTION_INTERVAL,
             )
@@ -381,6 +417,7 @@ impl Default for ConfigToml {
             image_compression: (DEFAULT_IMAGE_COMPRESSION),
             ephemeral_bytes_per_memory_kb: (DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB),
             l0_flush: None,
+            compact_level0_phase1_value_access: Default::default(),
             virtual_file_direct_io: crate::models::virtual_file::DirectIoMode::default(),
 
             io_buffer_alignment: DEFAULT_IO_BUFFER_ALIGNMENT,
@@ -430,6 +467,8 @@ pub mod tenant_conf_defaults {
     // By default ingest enough WAL for two new L0 layers before checking if new image
     // image layers should be created.
     pub const DEFAULT_IMAGE_LAYER_CREATION_CHECK_THRESHOLD: u8 = 2;
+
+    pub const DEFAULT_INGEST_BATCH_SIZE: u64 = 100;
 }
 
 impl Default for TenantConfigToml {

libs/pageserver_api/src/controller_api.rs

@@ -1,5 +1,4 @@
 use std::collections::{HashMap, HashSet};
-use std::fmt::Display;
 use std::str::FromStr;
 use std::time::{Duration, Instant};
 
@@ -58,7 +57,7 @@ pub struct NodeRegisterRequest {
     pub listen_http_addr: String,
     pub listen_http_port: u16,
 
-    pub availability_zone_id: AvailabilityZone,
+    pub availability_zone_id: String,
 }
 
 #[derive(Serialize, Deserialize)]
@@ -75,19 +74,10 @@ pub struct TenantPolicyRequest {
     pub scheduling: Option<ShardSchedulingPolicy>,
 }
 
-#[derive(Clone, Serialize, Deserialize, PartialEq, Eq, Hash)]
-pub struct AvailabilityZone(pub String);
-
-impl Display for AvailabilityZone {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(f, "{}", self.0)
-    }
-}
-
 #[derive(Serialize, Deserialize)]
 pub struct ShardsPreferredAzsRequest {
     #[serde(flatten)]
-    pub preferred_az_ids: HashMap<TenantShardId, AvailabilityZone>,
+    pub preferred_az_ids: HashMap<TenantShardId, String>,
 }
 
 #[derive(Serialize, Deserialize)]

libs/pageserver_api/src/key.rs

@@ -1,8 +1,8 @@
 use anyhow::{bail, Result};
 use byteorder::{ByteOrder, BE};
 use postgres_ffi::relfile_utils::{FSM_FORKNUM, VISIBILITYMAP_FORKNUM};
-use postgres_ffi::Oid;
 use postgres_ffi::RepOriginId;
+use postgres_ffi::{Oid, TransactionId};
 use serde::{Deserialize, Serialize};
 use std::{fmt, ops::Range};
 
@@ -350,17 +350,7 @@ impl Key {
 // 02 00000000 00000000 00000000 00   00000000
 //
 // TwoPhaseFile:
-//
+// 02 00000000 00000000 00000000 00   XID
-// 02 00000000 00000000 00XXXXXX XX   XXXXXXXX
-//
-//                        \______XID_________/
-//
-// The 64-bit XID is stored a little awkwardly in field6, field5 and
-// field4. PostgreSQL v16 and below only stored a 32-bit XID, which
-// fit completely in field6, but starting with PostgreSQL v17, a full
-// 64-bit XID is used. Most pageserver code that accesses
-// TwoPhaseFiles now deals with 64-bit XIDs even on v16, the high bits
-// are just unused.
 //
 // ControlFile:
 // 03 00000000 00000000 00000000 00   00000000
@@ -592,36 +582,35 @@ pub const TWOPHASEDIR_KEY: Key = Key {
 };
 
 #[inline(always)]
-pub fn twophase_file_key(xid: u64) -> Key {
+pub fn twophase_file_key(xid: TransactionId) -> Key {
     Key {
         field1: 0x02,
         field2: 0,
         field3: 0,
-        field4: ((xid & 0xFFFFFF0000000000) >> 40) as u32,
+        field4: 0,
-        field5: ((xid & 0x000000FF00000000) >> 32) as u8,
+        field5: 0,
-        field6: (xid & 0x00000000FFFFFFFF) as u32,
+        field6: xid,
     }
 }
 
 #[inline(always)]
-pub fn twophase_key_range(xid: u64) -> Range<Key> {
+pub fn twophase_key_range(xid: TransactionId) -> Range<Key> {
-    // 64-bit XIDs really should not overflow
     let (next_xid, overflowed) = xid.overflowing_add(1);
 
     Key {
         field1: 0x02,
         field2: 0,
         field3: 0,
-        field4: ((xid & 0xFFFFFF0000000000) >> 40) as u32,
+        field4: 0,
-        field5: ((xid & 0x000000FF00000000) >> 32) as u8,
+        field5: 0,
-        field6: (xid & 0x00000000FFFFFFFF) as u32,
+        field6: xid,
     }..Key {
         field1: 0x02,
         field2: 0,
-        field3: u32::from(overflowed),
+        field3: 0,
-        field4: ((next_xid & 0xFFFFFF0000000000) >> 40) as u32,
+        field4: 0,
-        field5: ((next_xid & 0x000000FF00000000) >> 32) as u8,
+        field5: u8::from(overflowed),
-        field6: (next_xid & 0x00000000FFFFFFFF) as u32,
+        field6: next_xid,
     }
 }
 

libs/pageserver_api/src/models.rs

@@ -37,11 +37,14 @@ use bytes::{Buf, BufMut, Bytes, BytesMut};
 /// ```mermaid
 /// stateDiagram-v2
 ///
+///     [*] --> Loading: spawn_load()
 ///     [*] --> Attaching: spawn_attach()
 ///
+///     Loading --> Activating: activate()
 ///     Attaching --> Activating: activate()
 ///     Activating --> Active: infallible
 ///
+///     Loading --> Broken: load() failure
 ///     Attaching --> Broken: attach() failure
 ///
 ///     Active --> Stopping: set_stopping(), part of shutdown & detach
@@ -65,6 +68,10 @@ use bytes::{Buf, BufMut, Bytes, BytesMut};
 )]
 #[serde(tag = "slug", content = "data")]
 pub enum TenantState {
+    /// This tenant is being loaded from local disk.
+    ///
+    /// `set_stopping()` and `set_broken()` do not work in this state and wait for it to pass.
+    Loading,
     /// This tenant is being attached to the pageserver.
     ///
     /// `set_stopping()` and `set_broken()` do not work in this state and wait for it to pass.
@@ -114,6 +121,8 @@ impl TenantState {
             // But, our attach task might still be fetching the remote timelines, etc.
             // So, return `Maybe` while Attaching, making Console wait for the attach task to finish.
             Self::Attaching | Self::Activating(ActivatingFrom::Attaching) => Maybe,
+            // tenant mgr startup distinguishes attaching from loading via marker file.
+            Self::Loading | Self::Activating(ActivatingFrom::Loading) => Attached,
             // We only reach Active after successful load / attach.
             // So, call atttachment status Attached.
             Self::Active => Attached,
@@ -182,11 +191,10 @@ impl LsnLease {
 }
 
 /// The only [`TenantState`] variants we could be `TenantState::Activating` from.
-///
-/// XXX: We used to have more variants here, but now it's just one, which makes this rather
-/// useless. Remove, once we've checked that there's no client code left that looks at this.
 #[derive(Clone, Copy, Debug, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
 pub enum ActivatingFrom {
+    /// Arrived to [`TenantState::Activating`] from [`TenantState::Loading`]
+    Loading,
     /// Arrived to [`TenantState::Activating`] from [`TenantState::Attaching`]
     Attaching,
 }
@@ -487,7 +495,7 @@ pub struct CompactionAlgorithmSettings {
     pub kind: CompactionAlgorithm,
 }
 
-#[derive(Debug, PartialEq, Eq, Clone, Deserialize, Serialize)]
+#[derive(Debug, PartialEq, Eq, Clone, serde::Deserialize, serde::Serialize)]
 #[serde(tag = "mode", rename_all = "kebab-case", deny_unknown_fields)]
 pub enum L0FlushConfig {
     #[serde(rename_all = "snake_case")]
@@ -1554,8 +1562,11 @@ mod tests {
 
     #[test]
     fn tenantstatus_activating_serde() {
-        let states = [TenantState::Activating(ActivatingFrom::Attaching)];
+        let states = [
-        let expected = "[{\"slug\":\"Activating\",\"data\":\"Attaching\"}]";
+            TenantState::Activating(ActivatingFrom::Loading),
+            TenantState::Activating(ActivatingFrom::Attaching),
+        ];
+        let expected = "[{\"slug\":\"Activating\",\"data\":\"Loading\"},{\"slug\":\"Activating\",\"data\":\"Attaching\"}]";
 
         let actual = serde_json::to_string(&states).unwrap();
 
@@ -1570,7 +1581,13 @@ mod tests {
     fn tenantstatus_activating_strum() {
         // tests added, because we use these for metrics
         let examples = [
+            (line!(), TenantState::Loading, "Loading"),
             (line!(), TenantState::Attaching, "Attaching"),
+            (
+                line!(),
+                TenantState::Activating(ActivatingFrom::Loading),
+                "Activating",
+            ),
             (
                 line!(),
                 TenantState::Activating(ActivatingFrom::Attaching),

libs/postgres_backend/Cargo.toml

@@ -5,8 +5,10 @@ edition.workspace = true
 license.workspace = true
 
 [dependencies]
+async-trait.workspace = true
 anyhow.workspace = true
 bytes.workspace = true
+futures.workspace = true
 rustls.workspace = true
 serde.workspace = true
 thiserror.workspace = true

libs/postgres_backend/src/lib.rs

@@ -280,6 +280,16 @@ pub struct PostgresBackend<IO> {
 
 pub type PostgresBackendTCP = PostgresBackend<tokio::net::TcpStream>;
 
+pub fn query_from_cstring(query_string: Bytes) -> Vec<u8> {
+    let mut query_string = query_string.to_vec();
+    if let Some(ch) = query_string.last() {
+        if *ch == 0 {
+            query_string.pop();
+        }
+    }
+    query_string
+}
+
 /// Cast a byte slice to a string slice, dropping null terminator if there's one.
 fn cstr_to_str(bytes: &[u8]) -> anyhow::Result<&str> {
     let without_null = bytes.strip_suffix(&[0]).unwrap_or(bytes);

libs/postgres_ffi/Cargo.toml

@@ -5,10 +5,13 @@ edition.workspace = true
 license.workspace = true
 
 [dependencies]
+rand.workspace = true
 regex.workspace = true
 bytes.workspace = true
+byteorder.workspace = true
 anyhow.workspace = true
 crc32c.workspace = true
+hex.workspace = true
 once_cell.workspace = true
 log.workspace = true
 memoffset.workspace = true

libs/postgres_ffi/build.rs

@@ -56,7 +56,7 @@ fn main() -> anyhow::Result<()> {
         PathBuf::from("pg_install")
     };
 
-    for pg_version in &["v14", "v15", "v16", "v17"] {
+    for pg_version in &["v14", "v15", "v16"] {
         let mut pg_install_dir_versioned = pg_install_dir.join(pg_version);
         if pg_install_dir_versioned.is_relative() {
             let cwd = env::current_dir().context("Failed to get current_dir")?;

libs/postgres_ffi/src/lib.rs

@@ -57,7 +57,6 @@ macro_rules! for_all_postgres_versions {
         $macro!(v14);
         $macro!(v15);
         $macro!(v16);
-        $macro!(v17);
     };
 }
 
@@ -92,7 +91,6 @@ macro_rules! dispatch_pgversion {
                 14 : v14,
                 15 : v15,
                 16 : v16,
-                17 : v17,
             ]
         )
     };
@@ -123,7 +121,6 @@ macro_rules! enum_pgversion_dispatch {
                 V14 : v14,
                 V15 : v15,
                 V16 : v16,
-                V17 : v17,
             ]
         )
     };
@@ -153,7 +150,6 @@ macro_rules! enum_pgversion {
                 V14 : v14,
                 V15 : v15,
                 V16 : v16,
-                V17 : v17,
             ]
         }
     };
@@ -166,7 +162,6 @@ macro_rules! enum_pgversion {
                 V14 : v14,
                 V15 : v15,
                 V16 : v16,
-                V17 : v17,
             ]
         }
     };

libs/postgres_ffi/src/pg_constants.rs

@@ -9,8 +9,8 @@
 //! comments on them.
 //!
 
-use crate::PageHeaderData;
 use crate::BLCKSZ;
+use crate::{PageHeaderData, XLogRecord};
 
 //
 // From pg_tablespace_d.h
@@ -152,9 +152,6 @@ pub const XLH_UPDATE_OLD_ALL_VISIBLE_CLEARED: u8 = (1 << 0) as u8;
 pub const XLH_UPDATE_NEW_ALL_VISIBLE_CLEARED: u8 = (1 << 1) as u8;
 pub const XLH_DELETE_ALL_VISIBLE_CLEARED: u8 = (1 << 0) as u8;
 
-// From heapam_xlog.h
-pub const XLOG_HEAP2_REWRITE: u8 = 0x00;
-
 // From replication/message.h
 pub const XLOG_LOGICAL_MESSAGE: u8 = 0x00;
 
@@ -194,6 +191,8 @@ pub const XLR_RMGR_INFO_MASK: u8 = 0xF0;
 pub const XLOG_TBLSPC_CREATE: u8 = 0x00;
 pub const XLOG_TBLSPC_DROP: u8 = 0x10;
 
+pub const SIZEOF_XLOGRECORD: u32 = size_of::<XLogRecord>() as u32;
+
 //
 // from xlogrecord.h
 //
@@ -217,21 +216,18 @@ pub const BKPIMAGE_HAS_HOLE: u8 = 0x01; /* page image has "hole" */
 /* From transam.h */
 pub const FIRST_NORMAL_TRANSACTION_ID: u32 = 3;
 pub const INVALID_TRANSACTION_ID: u32 = 0;
+pub const FIRST_BOOTSTRAP_OBJECT_ID: u32 = 12000;
+pub const FIRST_NORMAL_OBJECT_ID: u32 = 16384;
 
-/* pg_control.h */
 pub const XLOG_CHECKPOINT_SHUTDOWN: u8 = 0x00;
 pub const XLOG_CHECKPOINT_ONLINE: u8 = 0x10;
-pub const XLOG_PARAMETER_CHANGE: u8 = 0x60;
+pub const XLP_FIRST_IS_CONTRECORD: u16 = 0x0001;
-pub const XLOG_END_OF_RECOVERY: u8 = 0x90;
+pub const XLP_LONG_HEADER: u16 = 0x0002;
 
 /* From xlog.h */
 pub const XLOG_REPLORIGIN_SET: u8 = 0x00;
 pub const XLOG_REPLORIGIN_DROP: u8 = 0x10;
 
-/* xlog_internal.h */
-pub const XLP_FIRST_IS_CONTRECORD: u16 = 0x0001;
-pub const XLP_LONG_HEADER: u16 = 0x0002;
-
 /* From replication/slot.h */
 pub const REPL_SLOT_ON_DISK_OFFSETOF_RESTART_LSN: usize = 4*4  /* offset of `slotdata` in ReplicationSlotOnDisk  */
    + 64 /* NameData */  + 4*4;
@@ -249,6 +245,33 @@ pub const VM_HEAPBLOCKS_PER_PAGE: u32 =
 /* From origin.c */
 pub const REPLICATION_STATE_MAGIC: u32 = 0x1257DADE;
 
+// List of subdirectories inside pgdata.
+// Copied from src/bin/initdb/initdb.c
+pub const PGDATA_SUBDIRS: [&str; 22] = [
+    "global",
+    "pg_wal/archive_status",
+    "pg_commit_ts",
+    "pg_dynshmem",
+    "pg_notify",
+    "pg_serial",
+    "pg_snapshots",
+    "pg_subtrans",
+    "pg_twophase",
+    "pg_multixact",
+    "pg_multixact/members",
+    "pg_multixact/offsets",
+    "base",
+    "base/1",
+    "pg_replslot",
+    "pg_tblspc",
+    "pg_stat",
+    "pg_stat_tmp",
+    "pg_xact",
+    "pg_logical",
+    "pg_logical/snapshots",
+    "pg_logical/mappings",
+];
+
 // Don't include postgresql.conf as it is inconvenient on node start:
 // we need postgresql.conf before basebackup to synchronize safekeepers
 // so no point in overwriting it during backup restore. Rest of the files

libs/postgres_ffi/src/pg_constants_v14.rs

@@ -5,33 +5,6 @@ pub const BKPIMAGE_IS_COMPRESSED: u8 = 0x02; /* page image is compressed */
 pub const BKPIMAGE_APPLY: u8 = 0x04; /* page image should be restored during replay */
 pub const SIZEOF_RELMAPFILE: usize = 512; /* sizeof(RelMapFile) in relmapper.c */
 
-// List of subdirectories inside pgdata.
-// Copied from src/bin/initdb/initdb.c
-pub const PGDATA_SUBDIRS: [&str; 22] = [
-    "global",
-    "pg_wal/archive_status",
-    "pg_commit_ts",
-    "pg_dynshmem",
-    "pg_notify",
-    "pg_serial",
-    "pg_snapshots",
-    "pg_subtrans",
-    "pg_twophase",
-    "pg_multixact",
-    "pg_multixact/members",
-    "pg_multixact/offsets",
-    "base",
-    "base/1",
-    "pg_replslot",
-    "pg_tblspc",
-    "pg_stat",
-    "pg_stat_tmp",
-    "pg_xact",
-    "pg_logical",
-    "pg_logical/snapshots",
-    "pg_logical/mappings",
-];
-
 pub fn bkpimg_is_compressed(bimg_info: u8) -> bool {
     (bimg_info & BKPIMAGE_IS_COMPRESSED) != 0
 }

libs/postgres_ffi/src/pg_constants_v15.rs

@@ -11,8 +11,6 @@ pub const BKPIMAGE_COMPRESS_ZSTD: u8 = 0x10; /* page image is compressed */
 
 pub const SIZEOF_RELMAPFILE: usize = 512; /* sizeof(RelMapFile) in relmapper.c */
 
-pub use super::super::v14::bindings::PGDATA_SUBDIRS;
-
 pub fn bkpimg_is_compressed(bimg_info: u8) -> bool {
     const ANY_COMPRESS_FLAG: u8 = BKPIMAGE_COMPRESS_PGLZ | BKPIMAGE_COMPRESS_LZ4 | BKPIMAGE_COMPRESS_ZSTD;
 

libs/postgres_ffi/src/pg_constants_v16.rs

@@ -11,8 +11,6 @@ pub const BKPIMAGE_COMPRESS_ZSTD: u8 = 0x10; /* page image is compressed */
 
 pub const SIZEOF_RELMAPFILE: usize = 524; /* sizeof(RelMapFile) in relmapper.c */
 
-pub use super::super::v14::bindings::PGDATA_SUBDIRS;
-
 pub fn bkpimg_is_compressed(bimg_info: u8) -> bool {
     const ANY_COMPRESS_FLAG: u8 = BKPIMAGE_COMPRESS_PGLZ | BKPIMAGE_COMPRESS_LZ4 | BKPIMAGE_COMPRESS_ZSTD;
 

libs/postgres_ffi/src/pg_constants_v17.rs

@@ -1,55 +0,0 @@
-pub const XACT_XINFO_HAS_DROPPED_STATS: u32 = 1u32 << 8;
-
-pub const XLOG_DBASE_CREATE_FILE_COPY: u8 = 0x00;
-pub const XLOG_DBASE_CREATE_WAL_LOG: u8 = 0x10;
-pub const XLOG_DBASE_DROP: u8 = 0x20;
-
-pub const BKPIMAGE_APPLY: u8 = 0x02; /* page image should be restored during replay */
-pub const BKPIMAGE_COMPRESS_PGLZ: u8 = 0x04; /* page image is compressed */
-pub const BKPIMAGE_COMPRESS_LZ4: u8 = 0x08; /* page image is compressed */
-pub const BKPIMAGE_COMPRESS_ZSTD: u8 = 0x10; /* page image is compressed */
-
-pub const SIZEOF_RELMAPFILE: usize = 524; /* sizeof(RelMapFile) in relmapper.c */
-
-// List of subdirectories inside pgdata.
-// Copied from src/bin/initdb/initdb.c
-pub const PGDATA_SUBDIRS: [&str; 23] = [
-    "global",
-    "pg_wal/archive_status",
-    "pg_wal/summaries",
-    "pg_commit_ts",
-    "pg_dynshmem",
-    "pg_notify",
-    "pg_serial",
-    "pg_snapshots",
-    "pg_subtrans",
-    "pg_twophase",
-    "pg_multixact",
-    "pg_multixact/members",
-    "pg_multixact/offsets",
-    "base",
-    "base/1",
-    "pg_replslot",
-    "pg_tblspc",
-    "pg_stat",
-    "pg_stat_tmp",
-    "pg_xact",
-    "pg_logical",
-    "pg_logical/snapshots",
-    "pg_logical/mappings",
-];
-
-pub fn bkpimg_is_compressed(bimg_info: u8) -> bool {
-    const ANY_COMPRESS_FLAG: u8 = BKPIMAGE_COMPRESS_PGLZ | BKPIMAGE_COMPRESS_LZ4 | BKPIMAGE_COMPRESS_ZSTD;
-
-    (bimg_info & ANY_COMPRESS_FLAG) != 0
-}
-
-
-pub const XLOG_HEAP2_PRUNE_ON_ACCESS: u8 = 0x10;
-pub const XLOG_HEAP2_PRUNE_VACUUM_SCAN: u8 = 0x20;
-pub const XLOG_HEAP2_PRUNE_VACUUM_CLEANUP: u8 = 0x30;
-
-
-pub const XLOG_OVERWRITE_CONTRECORD: u8 = 0xD0;
-pub const XLOG_CHECKPOINT_REDO: u8 = 0xE0;

libs/postgres_ffi/src/xlog_utils.rs

@@ -26,12 +26,11 @@ use bytes::{Buf, Bytes};
 use log::*;
 
 use serde::Serialize;
-use std::ffi::OsStr;
 use std::fs::File;
 use std::io::prelude::*;
 use std::io::ErrorKind;
 use std::io::SeekFrom;
-use std::path::Path;
+use std::path::{Path, PathBuf};
 use std::time::SystemTime;
 use utils::bin_ser::DeserializeError;
 use utils::bin_ser::SerializeError;
@@ -79,34 +78,19 @@ pub fn XLogFileName(tli: TimeLineID, logSegNo: XLogSegNo, wal_segsz_bytes: usize
     )
 }
 
-pub fn XLogFromFileName(
+pub fn XLogFromFileName(fname: &str, wal_seg_size: usize) -> (XLogSegNo, TimeLineID) {
-    fname: &OsStr,
+    let tli = u32::from_str_radix(&fname[0..8], 16).unwrap();
-    wal_seg_size: usize,
+    let log = u32::from_str_radix(&fname[8..16], 16).unwrap() as XLogSegNo;
-) -> anyhow::Result<(XLogSegNo, TimeLineID)> {
+    let seg = u32::from_str_radix(&fname[16..24], 16).unwrap() as XLogSegNo;
-    if let Some(fname_str) = fname.to_str() {
+    (log * XLogSegmentsPerXLogId(wal_seg_size) + seg, tli)
-        let tli = u32::from_str_radix(&fname_str[0..8], 16)?;
-        let log = u32::from_str_radix(&fname_str[8..16], 16)? as XLogSegNo;
-        let seg = u32::from_str_radix(&fname_str[16..24], 16)? as XLogSegNo;
-        Ok((log * XLogSegmentsPerXLogId(wal_seg_size) + seg, tli))
-    } else {
-        anyhow::bail!("non-ut8 filename: {:?}", fname);
-    }
 }
 
-pub fn IsXLogFileName(fname: &OsStr) -> bool {
+pub fn IsXLogFileName(fname: &str) -> bool {
-    if let Some(fname) = fname.to_str() {
+    return fname.len() == XLOG_FNAME_LEN && fname.chars().all(|c| c.is_ascii_hexdigit());
-        fname.len() == XLOG_FNAME_LEN && fname.chars().all(|c| c.is_ascii_hexdigit())
-    } else {
-        false
-    }
 }
 
-pub fn IsPartialXLogFileName(fname: &OsStr) -> bool {
+pub fn IsPartialXLogFileName(fname: &str) -> bool {
-    if let Some(fname) = fname.to_str() {
+    fname.ends_with(".partial") && IsXLogFileName(&fname[0..fname.len() - 8])
-        fname.ends_with(".partial") && IsXLogFileName(OsStr::new(&fname[0..fname.len() - 8]))
-    } else {
-        false
-    }
 }
 
 /// If LSN points to the beginning of the page, then shift it to first record,
@@ -276,6 +260,13 @@ fn open_wal_segment(seg_file_path: &Path) -> anyhow::Result<Option<File>> {
     }
 }
 
+pub fn main() {
+    let mut data_dir = PathBuf::new();
+    data_dir.push(".");
+    let wal_end = find_end_of_wal(&data_dir, WAL_SEGMENT_SIZE, Lsn(0)).unwrap();
+    println!("wal_end={:?}", wal_end);
+}
+
 impl XLogRecord {
     pub fn from_slice(buf: &[u8]) -> Result<XLogRecord, DeserializeError> {
         use utils::bin_ser::LeSer;

libs/postgres_ffi/wal_craft/Cargo.toml

@@ -9,6 +9,7 @@ anyhow.workspace = true
 clap.workspace = true
 env_logger.workspace = true
 log.workspace = true
+once_cell.workspace = true
 postgres.workspace = true
 postgres_ffi.workspace = true
 camino-tempfile.workspace = true

libs/postgres_ffi/wal_craft/src/lib.rs

@@ -7,7 +7,6 @@ use postgres_ffi::{WAL_SEGMENT_SIZE, XLOG_BLCKSZ};
 use postgres_ffi::{
     XLOG_SIZE_OF_XLOG_LONG_PHD, XLOG_SIZE_OF_XLOG_RECORD, XLOG_SIZE_OF_XLOG_SHORT_PHD,
 };
-use std::ffi::OsStr;
 use std::path::{Path, PathBuf};
 use std::process::Command;
 use std::time::{Duration, Instant};
@@ -27,6 +26,7 @@ macro_rules! xlog_utils_test {
 
 postgres_ffi::for_all_postgres_versions! { xlog_utils_test }
 
+#[derive(Debug, Clone, PartialEq, Eq)]
 pub struct Conf {
     pub pg_version: u32,
     pub pg_distrib_dir: PathBuf,
@@ -53,7 +53,7 @@ impl Conf {
 
         #[allow(clippy::manual_range_patterns)]
         match self.pg_version {
-            14 | 15 | 16 | 17 => Ok(path.join(format!("v{}", self.pg_version))),
+            14 | 15 | 16 => Ok(path.join(format!("v{}", self.pg_version))),
             _ => bail!("Unsupported postgres version: {}", self.pg_version),
         }
     }
@@ -136,8 +136,8 @@ impl Conf {
 
     pub fn pg_waldump(
         &self,
-        first_segment_name: &OsStr,
+        first_segment_name: &str,
-        last_segment_name: &OsStr,
+        last_segment_name: &str,
     ) -> anyhow::Result<std::process::Output> {
         let first_segment_file = self.datadir.join(first_segment_name);
         let last_segment_file = self.datadir.join(last_segment_name);

libs/postgres_ffi/wal_craft/src/xlog_utils_test.rs

@@ -4,7 +4,6 @@ use super::*;
 use crate::{error, info};
 use regex::Regex;
 use std::cmp::min;
-use std::ffi::OsStr;
 use std::fs::{self, File};
 use std::io::Write;
 use std::{env, str::FromStr};
@@ -55,7 +54,7 @@ fn test_end_of_wal<C: crate::Crafter>(test_name: &str) {
         .wal_dir()
         .read_dir()
         .unwrap()
-        .map(|f| f.unwrap().file_name())
+        .map(|f| f.unwrap().file_name().into_string().unwrap())
         .filter(|fname| IsXLogFileName(fname))
         .max()
         .unwrap();
@@ -71,11 +70,11 @@ fn test_end_of_wal<C: crate::Crafter>(test_name: &str) {
             start_lsn
         );
         for file in fs::read_dir(cfg.wal_dir()).unwrap().flatten() {
-            let fname = file.file_name();
+            let fname = file.file_name().into_string().unwrap();
             if !IsXLogFileName(&fname) {
                 continue;
             }
-            let (segno, _) = XLogFromFileName(&fname, WAL_SEGMENT_SIZE).unwrap();
+            let (segno, _) = XLogFromFileName(&fname, WAL_SEGMENT_SIZE);
             let seg_start_lsn = XLogSegNoOffsetToRecPtr(segno, 0, WAL_SEGMENT_SIZE);
             if seg_start_lsn > u64::from(*start_lsn) {
                 continue;
@@ -94,10 +93,10 @@ fn test_end_of_wal<C: crate::Crafter>(test_name: &str) {
     }
 }
 
-fn find_pg_waldump_end_of_wal(cfg: &crate::Conf, last_segment: &OsStr) -> Lsn {
+fn find_pg_waldump_end_of_wal(cfg: &crate::Conf, last_segment: &str) -> Lsn {
     // Get the actual end of WAL by pg_waldump
     let waldump_output = cfg
-        .pg_waldump(OsStr::new("000000010000000000000001"), last_segment)
+        .pg_waldump("000000010000000000000001", last_segment)
         .unwrap()
         .stderr;
     let waldump_output = std::str::from_utf8(&waldump_output).unwrap();
@@ -118,7 +117,7 @@ fn find_pg_waldump_end_of_wal(cfg: &crate::Conf, last_segment: &OsStr) -> Lsn {
 
 fn check_end_of_wal(
     cfg: &crate::Conf,
-    last_segment: &OsStr,
+    last_segment: &str,
     start_lsn: Lsn,
     expected_end_of_wal: Lsn,
 ) {
@@ -133,8 +132,7 @@ fn check_end_of_wal(
     // Rename file to partial to actually find last valid lsn, then rename it back.
     fs::rename(
         cfg.wal_dir().join(last_segment),
-        cfg.wal_dir()
+        cfg.wal_dir().join(format!("{}.partial", last_segment)),
-            .join(format!("{}.partial", last_segment.to_str().unwrap())),
     )
     .unwrap();
     let wal_end = find_end_of_wal(&cfg.wal_dir(), WAL_SEGMENT_SIZE, start_lsn).unwrap();
@@ -144,8 +142,7 @@ fn check_end_of_wal(
     );
     assert_eq!(wal_end, expected_end_of_wal);
     fs::rename(
-        cfg.wal_dir()
+        cfg.wal_dir().join(format!("{}.partial", last_segment)),
-            .join(format!("{}.partial", last_segment.to_str().unwrap())),
         cfg.wal_dir().join(last_segment),
     )
     .unwrap();

libs/pq_proto/Cargo.toml

@@ -8,8 +8,10 @@ license.workspace = true
 bytes.workspace = true
 byteorder.workspace = true
 itertools.workspace = true
+pin-project-lite.workspace = true
 postgres-protocol.workspace = true
 rand.workspace = true
 tokio = { workspace = true, features = ["io-util"] }
+tracing.workspace = true
 thiserror.workspace = true
 serde.workspace = true

libs/remote_storage/Cargo.toml

@@ -13,11 +13,14 @@ aws-smithy-async.workspace = true
 aws-smithy-types.workspace = true
 aws-config.workspace = true
 aws-sdk-s3.workspace = true
+aws-credential-types.workspace = true
 bytes.workspace = true
 camino = { workspace = true, features = ["serde1"] }
+humantime.workspace = true
 humantime-serde.workspace = true
 hyper = { workspace = true, features = ["stream"] }
 futures.workspace = true
+rand.workspace = true
 serde.workspace = true
 serde_json.workspace = true
 tokio = { workspace = true, features = ["sync", "fs", "io-util"] }

libs/remote_storage/src/lib.rs

@@ -127,6 +127,10 @@ impl RemotePath {
         &self.0
     }
 
+    pub fn extension(&self) -> Option<&str> {
+        self.0.extension()
+    }
+
     pub fn strip_prefix(&self, p: &RemotePath) -> Result<&Utf8Path, std::path::StripPrefixError> {
         self.0.strip_prefix(&p.0)
     }

libs/safekeeper_api/Cargo.toml

@@ -6,5 +6,6 @@ license.workspace = true
 
 [dependencies]
 serde.workspace = true
+serde_with.workspace = true
 const_format.workspace = true
 utils.workspace = true

libs/tracing-utils/Cargo.toml

@@ -9,9 +9,8 @@ hyper.workspace = true
 opentelemetry = { workspace = true, features=["rt-tokio"] }
 opentelemetry-otlp = { workspace = true, default-features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
 opentelemetry-semantic-conventions.workspace = true
+reqwest = { workspace = true, default-features = false, features = ["rustls-tls"] }
 tokio = { workspace = true, features = ["rt", "rt-multi-thread"] }
 tracing.workspace = true
 tracing-opentelemetry.workspace = true
-
+tracing-subscriber.workspace = true
-[dev-dependencies]
-tracing-subscriber.workspace = true    # For examples in docs

libs/utils/Cargo.toml

@@ -19,7 +19,6 @@ bincode.workspace = true
 bytes.workspace = true
 camino.workspace = true
 chrono.workspace = true
-git-version.workspace = true
 hex = { workspace = true, features = ["serde"] }
 humantime.workspace = true
 hyper = { workspace = true, features = ["full"] }
@@ -43,6 +42,7 @@ tracing.workspace = true
 tracing-error.workspace = true
 tracing-subscriber = { workspace = true, features = ["json", "registry"] }
 rand.workspace = true
+serde_with.workspace = true
 strum.workspace = true
 strum_macros.workspace = true
 url.workspace = true

libs/utils/src/accum.rs

@@ -0,0 +1,33 @@
+/// A helper to "accumulate" a value similar to `Iterator::reduce`, but lets you
+/// feed the accumulated values by calling the 'accum' function, instead of having an
+/// iterator.
+///
+/// For example, to calculate the smallest value among some integers:
+///
+/// ```
+/// use utils::accum::Accum;
+///
+/// let values = [1, 2, 3];
+///
+/// let mut min_value: Accum<u32> = Accum(None);
+/// for new_value in &values {
+///     min_value.accum(std::cmp::min, *new_value);
+/// }
+///
+/// assert_eq!(min_value.0.unwrap(), 1);
+/// ```
+pub struct Accum<T>(pub Option<T>);
+impl<T: Copy> Accum<T> {
+    pub fn accum<F>(&mut self, func: F, new_value: T)
+    where
+        F: FnOnce(T, T) -> T,
+    {
+        // If there is no previous value, just store the new value.
+        // Otherwise call the function to decide which one to keep.
+        self.0 = Some(if let Some(accum) = self.0 {
+            func(accum, new_value)
+        } else {
+            new_value
+        });
+    }
+}

libs/utils/src/http/error.rs

@@ -82,7 +82,7 @@ impl ApiError {
                 StatusCode::INTERNAL_SERVER_ERROR,
             ),
             ApiError::InternalServerError(err) => HttpErrorBody::response_from_msg_and_status(
-                format!("{err:#}"), // use alternative formatting so that we give the cause without backtrace
+                err.to_string(),
                 StatusCode::INTERNAL_SERVER_ERROR,
             ),
         }

libs/utils/src/id.rs

@@ -88,6 +88,12 @@ impl<'de> Deserialize<'de> for Id {
 }
 
 impl Id {
+    pub fn get_from_buf(buf: &mut impl bytes::Buf) -> Id {
+        let mut arr = [0u8; 16];
+        buf.copy_to_slice(&mut arr);
+        Id::from(arr)
+    }
+
     pub fn from_slice(src: &[u8]) -> Result<Id, IdError> {
         if src.len() != 16 {
             return Err(IdError::SliceParseError(src.len()));
@@ -173,6 +179,10 @@ impl fmt::Debug for Id {
 macro_rules! id_newtype {
     ($t:ident) => {
         impl $t {
+            pub fn get_from_buf(buf: &mut impl bytes::Buf) -> $t {
+                $t(Id::get_from_buf(buf))
+            }
+
             pub fn from_slice(src: &[u8]) -> Result<$t, IdError> {
                 Ok($t(Id::from_slice(src)?))
             }

libs/utils/src/leaky_bucket.rs

@@ -21,13 +21,7 @@
 //!
 //! Another explaination can be found here: <https://brandur.org/rate-limiting>
 
-use std::{
+use std::{sync::Mutex, time::Duration};
-    sync::{
-        atomic::{AtomicU64, Ordering},
-        Mutex,
-    },
-    time::Duration,
-};
 
 use tokio::{sync::Notify, time::Instant};
 
@@ -134,7 +128,6 @@ impl LeakyBucketState {
 
 pub struct RateLimiter {
     pub config: LeakyBucketConfig,
-    pub sleep_counter: AtomicU64,
     pub state: Mutex<LeakyBucketState>,
     /// a queue to provide this fair ordering.
     pub queue: Notify,
@@ -151,7 +144,6 @@ impl Drop for Requeue<'_> {
 impl RateLimiter {
     pub fn with_initial_tokens(config: LeakyBucketConfig, initial_tokens: f64) -> Self {
         RateLimiter {
-            sleep_counter: AtomicU64::new(0),
             state: Mutex::new(LeakyBucketState::with_initial_tokens(
                 &config,
                 initial_tokens,
@@ -171,16 +163,15 @@ impl RateLimiter {
 
     /// returns true if we did throttle
     pub async fn acquire(&self, count: usize) -> bool {
-        let start = tokio::time::Instant::now();
+        let mut throttled = false;
 
-        let start_count = self.sleep_counter.load(Ordering::Acquire);
+        let start = tokio::time::Instant::now();
-        let mut end_count = start_count;
 
         // wait until we are the first in the queue
         let mut notified = std::pin::pin!(self.queue.notified());
         if !notified.as_mut().enable() {
+            throttled = true;
             notified.await;
-            end_count = self.sleep_counter.load(Ordering::Acquire);
         }
 
         // notify the next waiter in the queue when we are done.
@@ -193,22 +184,9 @@ impl RateLimiter {
                 .unwrap()
                 .add_tokens(&self.config, start, count as f64);
             match res {
-                Ok(()) => return end_count > start_count,
+                Ok(()) => return throttled,
                 Err(ready_at) => {
-                    struct Increment<'a>(&'a AtomicU64);
+                    throttled = true;
-
-                    impl Drop for Increment<'_> {
-                        fn drop(&mut self) {
-                            self.0.fetch_add(1, Ordering::AcqRel);
-                        }
-                    }
-
-                    // increment the counter after we finish sleeping (or cancel this task).
-                    // this ensures that tasks that have already started the acquire will observe
-                    // the new sleep count when they are allowed to resume on the notify.
-                    let _inc = Increment(&self.sleep_counter);
-                    end_count += 1;
-
                     tokio::time::sleep_until(ready_at).await;
                 }
             }

libs/utils/src/lib.rs

@@ -43,9 +43,16 @@ pub mod logging;
 pub mod lock_file;
 pub mod pid_file;
 
+// Misc
+pub mod accum;
+pub mod shutdown;
+
 // Utility for binding TcpListeners with proper socket options.
 pub mod tcp_listener;
 
+// Utility for putting a raw file descriptor into non-blocking mode
+pub mod nonblock;
+
 // Default signal handling
 pub mod sentry_init;
 pub mod signals;
@@ -92,10 +99,6 @@ pub mod toml_edit_ext;
 
 pub mod circuit_breaker;
 
-// Re-export used in macro. Avoids adding git-version as dep in target crates.
-#[doc(hidden)]
-pub use git_version;
-
 /// This is a shortcut to embed git sha into binaries and avoid copying the same build script to all packages
 ///
 /// we have several cases:
@@ -135,7 +138,7 @@ macro_rules! project_git_version {
     ($const_identifier:ident) => {
         // this should try GIT_VERSION first only then git_version::git_version!
         const $const_identifier: &::core::primitive::str = {
-            const __COMMIT_FROM_GIT: &::core::primitive::str = $crate::git_version::git_version! {
+            const __COMMIT_FROM_GIT: &::core::primitive::str = git_version::git_version! {
                 prefix = "",
                 fallback = "unknown",
                 args = ["--abbrev=40", "--always", "--dirty=-modified"] // always use full sha

libs/utils/src/lsn.rs

@@ -1,5 +1,6 @@
 #![warn(missing_docs)]
 
+use camino::Utf8Path;
 use serde::{de::Visitor, Deserialize, Serialize};
 use std::fmt;
 use std::ops::{Add, AddAssign};
@@ -144,6 +145,14 @@ impl Lsn {
         i128::from(self.0) - i128::from(other)
     }
 
+    /// Parse an LSN from a filename in the form `0000000000000000`
+    pub fn from_filename<F>(filename: F) -> Result<Self, LsnParseError>
+    where
+        F: AsRef<Utf8Path>,
+    {
+        Lsn::from_hex(filename.as_ref().as_str())
+    }
+
     /// Parse an LSN from a string in the form `0000000000000000`
     pub fn from_hex<S>(s: S) -> Result<Self, LsnParseError>
     where

libs/utils/src/nonblock.rs

@@ -0,0 +1,17 @@
+use nix::fcntl::{fcntl, OFlag, F_GETFL, F_SETFL};
+use std::os::unix::io::RawFd;
+
+/// Put a file descriptor into non-blocking mode
+pub fn set_nonblock(fd: RawFd) -> Result<(), std::io::Error> {
+    let bits = fcntl(fd, F_GETFL)?;
+
+    // If F_GETFL returns some unknown bits, they should be valid
+    // for passing back to F_SETFL, too. If we left them out, the F_SETFL
+    // would effectively clear them, which is not what we want.
+    let mut flags = OFlag::from_bits_retain(bits);
+    flags |= OFlag::O_NONBLOCK;
+
+    fcntl(fd, F_SETFL(flags))?;
+
+    Ok(())
+}

libs/utils/src/shutdown.rs

@@ -0,0 +1,7 @@
+/// Immediately terminate the calling process without calling
+/// atexit callbacks, C runtime destructors etc. We mainly use
+/// this to protect coverage data from concurrent writes.
+pub fn exit_now(code: u8) -> ! {
+    // SAFETY: exiting is safe, the ffi is not safe
+    unsafe { nix::libc::_exit(code as _) };
+}

libs/utils/src/vec_map.rs

@@ -120,6 +120,32 @@ impl<K: Ord, V> VecMap<K, V> {
         Ok((None, delta_size))
     }
 
+    /// Split the map into two.
+    ///
+    /// The left map contains everything before `cutoff` (exclusive).
+    /// Right map contains `cutoff` and everything after (inclusive).
+    pub fn split_at(&self, cutoff: &K) -> (Self, Self)
+    where
+        K: Clone,
+        V: Clone,
+    {
+        let split_idx = self
+            .data
+            .binary_search_by_key(&cutoff, extract_key)
+            .unwrap_or_else(std::convert::identity);
+
+        (
+            VecMap {
+                data: self.data[..split_idx].to_vec(),
+                ordering: self.ordering,
+            },
+            VecMap {
+                data: self.data[split_idx..].to_vec(),
+                ordering: self.ordering,
+            },
+        )
+    }
+
     /// Move items from `other` to the end of `self`, leaving `other` empty.
     /// If the `other` ordering is different from `self` ordering
     /// `ExtendOrderingError` error will be returned.

libs/vm_monitor/Cargo.toml

@@ -15,11 +15,13 @@ anyhow.workspace = true
 axum.workspace = true
 clap.workspace = true
 futures.workspace = true
+inotify.workspace = true
 serde.workspace = true
 serde_json.workspace = true
 sysinfo.workspace = true
 tokio = { workspace = true, features = ["rt-multi-thread"] }
 tokio-postgres.workspace = true
+tokio-stream.workspace = true
 tokio-util.workspace = true
 tracing.workspace = true
 tracing-subscriber.workspace = true

libs/walproposer/build.rs

@@ -5,8 +5,6 @@ use std::{env, path::PathBuf, process::Command};
 
 use anyhow::{anyhow, Context};
 
-const WALPROPOSER_PG_VERSION: &str = "v17";
-
 fn main() -> anyhow::Result<()> {
     // Tell cargo to invalidate the built crate whenever the wrapper changes
     println!("cargo:rerun-if-changed=bindgen_deps.h");
@@ -38,10 +36,7 @@ fn main() -> anyhow::Result<()> {
     // Rebuild crate when libwalproposer.a changes
     println!("cargo:rerun-if-changed={walproposer_lib_search_str}/libwalproposer.a");
 
-    let pg_config_bin = pg_install_abs
+    let pg_config_bin = pg_install_abs.join("v16").join("bin").join("pg_config");
-        .join(WALPROPOSER_PG_VERSION)
-        .join("bin")
-        .join("pg_config");
     let inc_server_path: String = if pg_config_bin.exists() {
         let output = Command::new(pg_config_bin)
             .arg("--includedir-server")
@@ -58,7 +53,7 @@ fn main() -> anyhow::Result<()> {
             .into()
     } else {
         let server_path = pg_install_abs
-            .join(WALPROPOSER_PG_VERSION)
+            .join("v16")
             .join("include")
             .join("postgresql")
             .join("server")

pageserver/Cargo.toml

@@ -15,6 +15,7 @@ anyhow.workspace = true
 arc-swap.workspace = true
 async-compression.workspace = true
 async-stream.workspace = true
+async-trait.workspace = true
 bit_field.workspace = true
 byteorder.workspace = true
 bytes.workspace = true
@@ -22,11 +23,15 @@ camino.workspace = true
 camino-tempfile.workspace = true
 chrono = { workspace = true, features = ["serde"] }
 clap = { workspace = true, features = ["string"] }
+const_format.workspace = true
 consumption_metrics.workspace = true
 crc32c.workspace = true
+crossbeam-utils.workspace = true
 either.workspace = true
+flate2.workspace = true
 fail.workspace = true
 futures.workspace = true
+git-version.workspace = true
 hex.workspace = true
 humantime.workspace = true
 humantime-serde.workspace = true
@@ -52,6 +57,10 @@ serde.workspace = true
 serde_json = { workspace = true, features = ["raw_value"] }
 serde_path_to_error.workspace = true
 serde_with.workspace = true
+signal-hook.workspace = true
+smallvec = { workspace = true, features = ["write"] }
+svg_fmt.workspace = true
+sync_wrapper.workspace = true
 sysinfo.workspace = true
 tokio-tar.workspace = true
 thiserror.workspace = true
@@ -64,6 +73,7 @@ tokio-stream.workspace = true
 tokio-util.workspace = true
 toml_edit = { workspace = true, features = [ "serde" ] }
 tracing.workspace = true
+twox-hash.workspace = true
 url.workspace = true
 walkdir.workspace = true
 metrics.workspace = true

pageserver/benches/bench_walredo.rs

@@ -1,7 +1,7 @@
 //! Quantify a single walredo manager's throughput under N concurrent callers.
 //!
 //! The benchmark implementation ([`bench_impl`]) is parametrized by
-//! - `redo_work` => an async closure that takes a `PostgresRedoManager` and performs one redo
+//! - `redo_work` => [`Request::short_request`] or [`Request::medium_request`]
 //! - `n_redos` => number of times the benchmark shell execute the `redo_work`
 //! - `nclients` => number of clients (more on this shortly).
 //!
@@ -10,7 +10,7 @@
 //! Each task executes the `redo_work` `n_redos/nclients` times.
 //!
 //! We exercise the following combinations:
-//! - `redo_work = ping / short / medium``
+//! - `redo_work = short / medium``
 //! - `nclients = [1, 2, 4, 8, 16, 32, 64, 128]`
 //!
 //! We let `criterion` determine the `n_redos` using `iter_custom`.
@@ -27,43 +27,33 @@
 //!
 //! # Reference Numbers
 //!
-//! 2024-09-18 on im4gn.2xlarge
+//! 2024-04-15 on i3en.3xlarge
 //!
 //! ```text
-//! ping/1                  time:   [21.789 µs 21.918 µs 22.078 µs]
+//! short/1           time:   [24.584 µs 24.737 µs 24.922 µs]
-//! ping/2                  time:   [27.686 µs 27.812 µs 27.970 µs]
+//! short/2           time:   [33.479 µs 33.660 µs 33.888 µs]
-//! ping/4                  time:   [35.468 µs 35.671 µs 35.926 µs]
+//! short/4           time:   [42.713 µs 43.046 µs 43.440 µs]
-//! ping/8                  time:   [59.682 µs 59.987 µs 60.363 µs]
+//! short/8           time:   [71.814 µs 72.478 µs 73.240 µs]
-//! ping/16                 time:   [101.79 µs 102.37 µs 103.08 µs]
+//! short/16          time:   [132.73 µs 134.45 µs 136.22 µs]
-//! ping/32                 time:   [184.18 µs 185.15 µs 186.36 µs]
+//! short/32          time:   [258.31 µs 260.73 µs 263.27 µs]
-//! ping/64                 time:   [349.86 µs 351.45 µs 353.47 µs]
+//! short/64          time:   [511.61 µs 514.44 µs 517.51 µs]
-//! ping/128                time:   [684.53 µs 687.98 µs 692.17 µs]
+//! short/128         time:   [992.64 µs 998.23 µs 1.0042 ms]
-//! short/1                 time:   [31.833 µs 32.126 µs 32.428 µs]
+//! medium/1          time:   [110.11 µs 110.50 µs 110.96 µs]
-//! short/2                 time:   [35.558 µs 35.756 µs 35.992 µs]
+//! medium/2          time:   [153.06 µs 153.85 µs 154.99 µs]
-//! short/4                 time:   [44.850 µs 45.138 µs 45.484 µs]
+//! medium/4          time:   [317.51 µs 319.92 µs 322.85 µs]
-//! short/8                 time:   [65.985 µs 66.379 µs 66.853 µs]
+//! medium/8          time:   [638.30 µs 644.68 µs 652.12 µs]
-//! short/16                time:   [127.06 µs 127.90 µs 128.87 µs]
+//! medium/16         time:   [1.2651 ms 1.2773 ms 1.2914 ms]
-//! short/32                time:   [252.98 µs 254.70 µs 256.73 µs]
+//! medium/32         time:   [2.5117 ms 2.5410 ms 2.5720 ms]
-//! short/64                time:   [497.13 µs 499.86 µs 503.26 µs]
+//! medium/64         time:   [4.8088 ms 4.8555 ms 4.9047 ms]
-//! short/128               time:   [987.46 µs 993.45 µs 1.0004 ms]
+//! medium/128        time:   [8.8311 ms 8.9849 ms 9.1263 ms]
-//! medium/1                time:   [137.91 µs 138.55 µs 139.35 µs]
-//! medium/2                time:   [192.00 µs 192.91 µs 194.07 µs]
-//! medium/4                time:   [389.62 µs 391.55 µs 394.01 µs]
-//! medium/8                time:   [776.80 µs 780.33 µs 784.77 µs]
-//! medium/16               time:   [1.5323 ms 1.5383 ms 1.5459 ms]
-//! medium/32               time:   [3.0120 ms 3.0226 ms 3.0350 ms]
-//! medium/64               time:   [5.7405 ms 5.7787 ms 5.8166 ms]
-//! medium/128              time:   [10.412 ms 10.574 ms 10.718 ms]
 //! ```
 
 use anyhow::Context;
 use bytes::{Buf, Bytes};
 use criterion::{BenchmarkId, Criterion};
-use once_cell::sync::Lazy;
 use pageserver::{config::PageServerConf, walrecord::NeonWalRecord, walredo::PostgresRedoManager};
 use pageserver_api::{key::Key, shard::TenantShardId};
 use std::{
-    future::Future,
     sync::Arc,
     time::{Duration, Instant},
 };
@@ -71,59 +61,40 @@ use tokio::{sync::Barrier, task::JoinSet};
 use utils::{id::TenantId, lsn::Lsn};
 
 fn bench(c: &mut Criterion) {
-    macro_rules! bench_group {
+    {
-        ($name:expr, $redo_work:expr) => {{
+        let nclients = [1, 2, 4, 8, 16, 32, 64, 128];
-            let name: &str = $name;
+        for nclients in nclients {
-            let nclients = [1, 2, 4, 8, 16, 32, 64, 128];
+            let mut group = c.benchmark_group("short");
-            for nclients in nclients {
+            group.bench_with_input(
-                let mut group = c.benchmark_group(name);
+                BenchmarkId::from_parameter(nclients),
-                group.bench_with_input(
+                &nclients,
-                    BenchmarkId::from_parameter(nclients),
+                |b, nclients| {
-                    &nclients,
+                    let redo_work = Arc::new(Request::short_input());
-                    |b, nclients| {
+                    b.iter_custom(|iters| bench_impl(Arc::clone(&redo_work), iters, *nclients));
-                        b.iter_custom(|iters| bench_impl($redo_work, iters, *nclients));
+                },
-                    },
+            );
-                );
+        }
-            }
+    }
-        }};
+    {
+        let nclients = [1, 2, 4, 8, 16, 32, 64, 128];
+        for nclients in nclients {
+            let mut group = c.benchmark_group("medium");
+            group.bench_with_input(
+                BenchmarkId::from_parameter(nclients),
+                &nclients,
+                |b, nclients| {
+                    let redo_work = Arc::new(Request::medium_input());
+                    b.iter_custom(|iters| bench_impl(Arc::clone(&redo_work), iters, *nclients));
+                },
+            );
+        }
     }
-    //
-    // benchmark the protocol implementation
-    //
-    let pg_version = 14;
-    bench_group!(
-        "ping",
-        Arc::new(move |mgr: Arc<PostgresRedoManager>| async move {
-            let _: () = mgr.ping(pg_version).await.unwrap();
-        })
-    );
-    //
-    // benchmarks with actual record redo
-    //
-    let make_redo_work = |req: &'static Request| {
-        Arc::new(move |mgr: Arc<PostgresRedoManager>| async move {
-            let page = req.execute(&mgr).await.unwrap();
-            assert_eq!(page.remaining(), 8192);
-        })
-    };
-    bench_group!("short", {
-        static REQUEST: Lazy<Request> = Lazy::new(Request::short_input);
-        make_redo_work(&REQUEST)
-    });
-    bench_group!("medium", {
-        static REQUEST: Lazy<Request> = Lazy::new(Request::medium_input);
-        make_redo_work(&REQUEST)
-    });
 }
 criterion::criterion_group!(benches, bench);
 criterion::criterion_main!(benches);
 
 // Returns the sum of each client's wall-clock time spent executing their share of the n_redos.
-fn bench_impl<F, Fut>(redo_work: Arc<F>, n_redos: u64, nclients: u64) -> Duration
+fn bench_impl(redo_work: Arc<Request>, n_redos: u64, nclients: u64) -> Duration {
-where
-    F: Fn(Arc<PostgresRedoManager>) -> Fut + Send + Sync + 'static,
-    Fut: Future<Output = ()> + Send + 'static,
-{
     let repo_dir = camino_tempfile::tempdir_in(env!("CARGO_TARGET_TMPDIR")).unwrap();
 
     let conf = PageServerConf::dummy_conf(repo_dir.path().to_path_buf());
@@ -164,20 +135,17 @@ where
     })
 }
 
-async fn client<F, Fut>(
+async fn client(
     mgr: Arc<PostgresRedoManager>,
     start: Arc<Barrier>,
-    redo_work: Arc<F>,
+    redo_work: Arc<Request>,
     n_redos: u64,
-) -> Duration
+) -> Duration {
-where
-    F: Fn(Arc<PostgresRedoManager>) -> Fut + Send + Sync + 'static,
-    Fut: Future<Output = ()> + Send + 'static,
-{
     start.wait().await;
     let start = Instant::now();
     for _ in 0..n_redos {
-        redo_work(Arc::clone(&mgr)).await;
+        let page = redo_work.execute(&mgr).await.unwrap();
+        assert_eq!(page.remaining(), 8192);
         // The real pageserver will rarely if ever do 2 walredos in a row without
         // yielding to the executor.
         tokio::task::yield_now().await;

pageserver/client/src/mgmt_api.rs

@@ -432,7 +432,7 @@ impl Client {
             self.mgmt_api_endpoint
         );
 
-        self.request(Method::PUT, &uri, req)
+        self.request(Method::POST, &uri, req)
             .await?
             .json()
             .await

pageserver/compaction/Cargo.toml

@@ -9,18 +9,41 @@ default = []
 
 [dependencies]
 anyhow.workspace = true
+async-compression.workspace = true
 async-stream.workspace = true
+byteorder.workspace = true
+bytes.workspace = true
+chrono = { workspace = true, features = ["serde"] }
 clap = { workspace = true, features = ["string"] }
+const_format.workspace = true
+consumption_metrics.workspace = true
+crossbeam-utils.workspace = true
+either.workspace = true
+flate2.workspace = true
+fail.workspace = true
 futures.workspace = true
+git-version.workspace = true
+hex.workspace = true
+humantime.workspace = true
+humantime-serde.workspace = true
 itertools.workspace = true
 once_cell.workspace = true
 pageserver_api.workspace = true
 pin-project-lite.workspace = true
 rand.workspace = true
+smallvec = { workspace = true, features = ["write"] }
 svg_fmt.workspace = true
+sync_wrapper.workspace = true
+thiserror.workspace = true
 tokio = { workspace = true, features = ["process", "sync", "fs", "rt", "io-util", "time"] }
+tokio-io-timeout.workspace = true
+tokio-util.workspace = true
 tracing.workspace = true
+tracing-error.workspace = true
 tracing-subscriber.workspace = true
+url.workspace = true
+walkdir.workspace = true
+metrics.workspace = true
 utils.workspace = true
 workspace_hack.workspace = true
 

pageserver/ctl/Cargo.toml

@@ -8,8 +8,10 @@ license.workspace = true
 
 [dependencies]
 anyhow.workspace = true
+bytes.workspace = true
 camino.workspace = true
 clap = { workspace = true, features = ["string"] }
+git-version.workspace = true
 humantime.workspace = true
 pageserver = { path = ".." }
 pageserver_api.workspace = true
@@ -22,4 +24,5 @@ toml_edit.workspace = true
 utils.workspace = true
 svg_fmt.workspace = true
 workspace_hack.workspace = true
+serde.workspace = true
 serde_json.workspace = true

pageserver/ctl/src/layer_map_analyzer.rs

@@ -79,24 +79,16 @@ pub(crate) fn parse_filename(name: &str) -> Option<LayerFile> {
         return None;
     }
     let keys: Vec<&str> = split[0].split('-').collect();
-    let lsn_and_opt_generation: Vec<&str> = split[1].split('v').collect();
+    let mut lsns: Vec<&str> = split[1].split('-').collect();
-    let lsns: Vec<&str> = lsn_and_opt_generation[0].split('-').collect();
+    let is_delta = if lsns.len() == 1 {
-    let the_lsns: [&str; 2];
+        lsns.push(lsns[0]);
-
-    /*
-     * Generations add a -vX-XXXXXX postfix, which causes issues when we try to
-     * parse 'vX' as an LSN.
-     */
-    let is_delta = if lsns.len() == 1 || lsns[1].is_empty() {
-        the_lsns = [lsns[0], lsns[0]];
         false
     } else {
-        the_lsns = [lsns[0], lsns[1]];
         true
     };
 
     let key_range = Key::from_hex(keys[0]).unwrap()..Key::from_hex(keys[1]).unwrap();
-    let lsn_range = Lsn::from_hex(the_lsns[0]).unwrap()..Lsn::from_hex(the_lsns[1]).unwrap();
+    let lsn_range = Lsn::from_hex(lsns[0]).unwrap()..Lsn::from_hex(lsns[1]).unwrap();
     let holes = Vec::new();
     Some(LayerFile {
         key_range,

pageserver/src/basebackup.rs

@@ -30,8 +30,9 @@ use pageserver_api::reltag::{RelTag, SlruKind};
 
 use postgres_ffi::dispatch_pgversion;
 use postgres_ffi::pg_constants::{DEFAULTTABLESPACE_OID, GLOBALTABLESPACE_OID};
-use postgres_ffi::pg_constants::{PGDATA_SPECIAL_FILES, PG_HBA};
+use postgres_ffi::pg_constants::{PGDATA_SPECIAL_FILES, PGDATA_SUBDIRS, PG_HBA};
 use postgres_ffi::relfile_utils::{INIT_FORKNUM, MAIN_FORKNUM};
+use postgres_ffi::TransactionId;
 use postgres_ffi::XLogFileName;
 use postgres_ffi::PG_TLI;
 use postgres_ffi::{BLCKSZ, RELSEG_SIZE, WAL_SEGMENT_SIZE};
@@ -254,11 +255,8 @@ where
 
         let lazy_slru_download = self.timeline.get_lazy_slru_download() && !self.full_backup;
 
-        let pgversion = self.timeline.pg_version;
-        let subdirs = dispatch_pgversion!(pgversion, &pgv::bindings::PGDATA_SUBDIRS[..]);
-
         // Create pgdata subdirs structure
-        for dir in subdirs.iter() {
+        for dir in PGDATA_SUBDIRS.iter() {
             let header = new_tar_header_dir(dir)?;
             self.ar
                 .append(&header, &mut io::empty())
@@ -608,7 +606,7 @@ where
     //
     // Extract twophase state files
     //
-    async fn add_twophase_file(&mut self, xid: u64) -> Result<(), BasebackupError> {
+    async fn add_twophase_file(&mut self, xid: TransactionId) -> Result<(), BasebackupError> {
         let img = self
             .timeline
             .get_twophase_file(xid, self.lsn, self.ctx)
@@ -619,11 +617,7 @@ where
         buf.extend_from_slice(&img[..]);
         let crc = crc32c::crc32c(&img[..]);
         buf.put_u32_le(crc);
-        let path = if self.timeline.pg_version < 17 {
+        let path = format!("pg_twophase/{:>08X}", xid);
-            format!("pg_twophase/{:>08X}", xid)
-        } else {
-            format!("pg_twophase/{:>016X}", xid)
-        };
         let header = new_tar_header(&path, buf.len() as u64)?;
         self.ar
             .append(&header, &buf[..])

pageserver/src/config.rs

@@ -13,6 +13,7 @@ use pageserver_api::{
 use remote_storage::{RemotePath, RemoteStorageConfig};
 use std::env;
 use storage_broker::Uri;
+use utils::crashsafe::path_with_suffix_extension;
 use utils::logging::SecretString;
 
 use once_cell::sync::OnceCell;
@@ -32,7 +33,7 @@ use crate::tenant::storage_layer::inmemory_layer::IndexEntry;
 use crate::tenant::{TENANTS_SEGMENT_NAME, TIMELINES_SEGMENT_NAME};
 use crate::virtual_file;
 use crate::virtual_file::io_engine;
-use crate::{TENANT_HEATMAP_BASENAME, TENANT_LOCATION_CONFIG_NAME};
+use crate::{TENANT_HEATMAP_BASENAME, TENANT_LOCATION_CONFIG_NAME, TIMELINE_DELETE_MARK_SUFFIX};
 
 /// Global state of pageserver.
 ///
@@ -256,6 +257,17 @@ impl PageServerConf {
             .join(timeline_id.to_string())
     }
 
+    pub(crate) fn timeline_delete_mark_file_path(
+        &self,
+        tenant_shard_id: TenantShardId,
+        timeline_id: TimelineId,
+    ) -> Utf8PathBuf {
+        path_with_suffix_extension(
+            self.timeline_path(&tenant_shard_id, &timeline_id),
+            TIMELINE_DELETE_MARK_SUFFIX,
+        )
+    }
+
     /// Turns storage remote path of a file into its local path.
     pub fn local_path(&self, remote_path: &RemotePath) -> Utf8PathBuf {
         remote_path.with_base(&self.workdir)
@@ -269,7 +281,7 @@ impl PageServerConf {
 
         #[allow(clippy::manual_range_patterns)]
         match pg_version {
-            14 | 15 | 16 | 17 => Ok(path.join(format!("v{pg_version}"))),
+            14 | 15 | 16 => Ok(path.join(format!("v{pg_version}"))),
             _ => bail!("Unsupported postgres version: {}", pg_version),
         }
     }
@@ -324,6 +336,7 @@ impl PageServerConf {
             max_vectored_read_bytes,
             image_compression,
             ephemeral_bytes_per_memory_kb,
+            compact_level0_phase1_value_access: _,
             l0_flush,
             virtual_file_direct_io,
             concurrent_tenant_warmup,
@@ -478,6 +491,11 @@ pub struct ConfigurableSemaphore {
 }
 
 impl ConfigurableSemaphore {
+    pub const DEFAULT_INITIAL: NonZeroUsize = match NonZeroUsize::new(1) {
+        Some(x) => x,
+        None => panic!("const unwrap is not yet stable"),
+    };
+
     /// Initializse using a non-zero amount of permits.
     ///
     /// Require a non-zero initial permits, because using permits == 0 is a crude way to disable a
@@ -498,6 +516,12 @@ impl ConfigurableSemaphore {
     }
 }
 
+impl Default for ConfigurableSemaphore {
+    fn default() -> Self {
+        Self::new(Self::DEFAULT_INITIAL)
+    }
+}
+
 impl PartialEq for ConfigurableSemaphore {
     fn eq(&self, other: &Self) -> bool {
         // the number of permits can be increased at runtime, so we cannot really fulfill the
@@ -534,6 +558,16 @@ mod tests {
             .expect("parse_and_validate");
     }
 
+    #[test]
+    fn test_compactl0_phase1_access_mode_is_ignored_silently() {
+        let input = indoc::indoc! {r#"
+            [compact_level0_phase1_value_access]
+            mode = "streaming-kmerge"
+            validate = "key-lsn-value"
+        "#};
+        toml_edit::de::from_str::<pageserver_api::config::ConfigToml>(input).unwrap();
+    }
+
     /// If there's a typo in the pageserver config, we'd rather catch that typo
     /// and fail pageserver startup than silently ignoring the typo, leaving whoever
     /// made it in the believe that their config change is effective.

pageserver/src/consumption_metrics.rs

@@ -178,7 +178,7 @@ async fn collect_metrics(
                 )
                 .await;
                 if let Err(e) = res {
-                    tracing::error!("failed to upload to remote storage: {e:#}");
+                    tracing::error!("failed to upload to S3: {e:#}");
                 }
             }
         };

pageserver/src/control_plane_client.rs

@@ -2,7 +2,7 @@ use std::collections::HashMap;
 
 use futures::Future;
 use pageserver_api::{
-    controller_api::{AvailabilityZone, NodeRegisterRequest},
+    controller_api::NodeRegisterRequest,
     shard::TenantShardId,
     upcall_api::{
         ReAttachRequest, ReAttachResponse, ReAttachResponseTenant, ValidateRequest,
@@ -148,10 +148,10 @@ impl ControlPlaneGenerationsApi for ControlPlaneClient {
                             .and_then(|jv| jv.as_str().map(|str| str.to_owned()));
 
                         match az_id_from_metadata {
-                            Some(az_id) => Some(AvailabilityZone(az_id)),
+                            Some(az_id) => Some(az_id),
                             None => {
                                 tracing::warn!("metadata.json does not contain an 'availability_zone_id' field");
-                                conf.availability_zone.clone().map(AvailabilityZone)
+                                conf.availability_zone.clone()
                             }
                         }
                     };

pageserver/src/http/routes.rs

@@ -2955,7 +2955,7 @@ pub fn make_router(
             "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/preserve_initdb_archive",
             |r| api_handler(r, timeline_preserve_initdb_handler),
         )
-        .put(
+        .post(
             "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/archival_config",
             |r| api_handler(r, timeline_archival_config_handler),
         )

pageserver/src/import_datadir.rs

@@ -580,11 +580,9 @@ async fn import_file(
         import_slru(modification, slru, file_path, reader, len, ctx).await?;
         debug!("imported multixact members slru");
     } else if file_path.starts_with("pg_twophase") {
-        let bytes = read_all_bytes(reader).await?;
+        let xid = u32::from_str_radix(file_name.as_ref(), 16)?;
 
-        // In PostgreSQL v17, this is a 64-bit FullTransactionid. In previous versions,
+        let bytes = read_all_bytes(reader).await?;
-        // it's a 32-bit TransactionId, which fits in u64 anyway.
-        let xid = u64::from_str_radix(file_name.as_ref(), 16)?;
         modification
             .put_twophase_file(xid, Bytes::copy_from_slice(&bytes[..]), ctx)
             .await?;

pageserver/src/metrics.rs

@@ -1177,10 +1177,10 @@ pub(crate) mod virtual_file_io_engine {
 }
 
 struct GlobalAndPerTimelineHistogramTimer<'a, 'c> {
-    global_latency_histo: &'a Histogram,
+    global_metric: &'a Histogram,
 
     // Optional because not all op types are tracked per-timeline
-    per_timeline_latency_histo: Option<&'a Histogram>,
+    timeline_metric: Option<&'a Histogram>,
 
     ctx: &'c RequestContext,
     start: std::time::Instant,
@@ -1212,10 +1212,9 @@ impl<'a, 'c> Drop for GlobalAndPerTimelineHistogramTimer<'a, 'c> {
                 elapsed
             }
         };
-        self.global_latency_histo
+        self.global_metric.observe(ex_throttled.as_secs_f64());
-            .observe(ex_throttled.as_secs_f64());
+        if let Some(timeline_metric) = self.timeline_metric {
-        if let Some(per_timeline_getpage_histo) = self.per_timeline_latency_histo {
+            timeline_metric.observe(ex_throttled.as_secs_f64());
-            per_timeline_getpage_histo.observe(ex_throttled.as_secs_f64());
         }
     }
 }
@@ -1241,32 +1240,10 @@ pub enum SmgrQueryType {
 
 #[derive(Debug)]
 pub(crate) struct SmgrQueryTimePerTimeline {
-    global_started: [IntCounter; SmgrQueryType::COUNT],
+    global_metrics: [Histogram; SmgrQueryType::COUNT],
-    global_latency: [Histogram; SmgrQueryType::COUNT],
+    per_timeline_getpage: Histogram,
-    per_timeline_getpage_started: IntCounter,
-    per_timeline_getpage_latency: Histogram,
 }
 
-static SMGR_QUERY_STARTED_GLOBAL: Lazy<IntCounterVec> = Lazy::new(|| {
-    register_int_counter_vec!(
-        // it's a counter, but, name is prepared to extend it to a histogram of queue depth
-        "pageserver_smgr_query_started_global_count",
-        "Number of smgr queries started, aggregated by query type.",
-        &["smgr_query_type"],
-    )
-    .expect("failed to define a metric")
-});
-
-static SMGR_QUERY_STARTED_PER_TENANT_TIMELINE: Lazy<IntCounterVec> = Lazy::new(|| {
-    register_int_counter_vec!(
-        // it's a counter, but, name is prepared to extend it to a histogram of queue depth
-        "pageserver_smgr_query_started_count",
-        "Number of smgr queries started, aggregated by query type and tenant/timeline.",
-        &["smgr_query_type", "tenant_id", "shard_id", "timeline_id"],
-    )
-    .expect("failed to define a metric")
-});
-
 static SMGR_QUERY_TIME_PER_TENANT_TIMELINE: Lazy<HistogramVec> = Lazy::new(|| {
     register_histogram_vec!(
         "pageserver_smgr_query_seconds",
@@ -1342,20 +1319,14 @@ impl SmgrQueryTimePerTimeline {
         let tenant_id = tenant_shard_id.tenant_id.to_string();
         let shard_slug = format!("{}", tenant_shard_id.shard_slug());
         let timeline_id = timeline_id.to_string();
-        let global_started = std::array::from_fn(|i| {
+        let global_metrics = std::array::from_fn(|i| {
-            let op = SmgrQueryType::from_repr(i).unwrap();
-            SMGR_QUERY_STARTED_GLOBAL
-                .get_metric_with_label_values(&[op.into()])
-                .unwrap()
-        });
-        let global_latency = std::array::from_fn(|i| {
             let op = SmgrQueryType::from_repr(i).unwrap();
             SMGR_QUERY_TIME_GLOBAL
                 .get_metric_with_label_values(&[op.into()])
                 .unwrap()
         });
 
-        let per_timeline_getpage_started = SMGR_QUERY_STARTED_PER_TENANT_TIMELINE
+        let per_timeline_getpage = SMGR_QUERY_TIME_PER_TENANT_TIMELINE
             .get_metric_with_label_values(&[
                 SmgrQueryType::GetPageAtLsn.into(),
                 &tenant_id,
@@ -1363,32 +1334,18 @@ impl SmgrQueryTimePerTimeline {
                 &timeline_id,
             ])
             .unwrap();
-        let per_timeline_getpage_latency = SMGR_QUERY_TIME_PER_TENANT_TIMELINE
-            .get_metric_with_label_values(&[
-                SmgrQueryType::GetPageAtLsn.into(),
-                &tenant_id,
-                &shard_slug,
-                &timeline_id,
-            ])
-            .unwrap();
-
         Self {
-            global_started,
+            global_metrics,
-            global_latency,
+            per_timeline_getpage,
-            per_timeline_getpage_latency,
-            per_timeline_getpage_started,
         }
     }
     pub(crate) fn start_timer<'c: 'a, 'a>(
         &'a self,
         op: SmgrQueryType,
         ctx: &'c RequestContext,
-    ) -> Option<impl Drop + 'a> {
+    ) -> Option<impl Drop + '_> {
+        let global_metric = &self.global_metrics[op as usize];
         let start = Instant::now();
-
-        self.global_started[op as usize].inc();
-
-        // We subtract time spent throttled from the observed latency.
         match ctx.micros_spent_throttled.open() {
             Ok(()) => (),
             Err(error) => {
@@ -1407,16 +1364,15 @@ impl SmgrQueryTimePerTimeline {
             }
         }
 
-        let per_timeline_latency_histo = if matches!(op, SmgrQueryType::GetPageAtLsn) {
+        let timeline_metric = if matches!(op, SmgrQueryType::GetPageAtLsn) {
-            self.per_timeline_getpage_started.inc();
+            Some(&self.per_timeline_getpage)
-            Some(&self.per_timeline_getpage_latency)
         } else {
             None
         };
 
         Some(GlobalAndPerTimelineHistogramTimer {
-            global_latency_histo: &self.global_latency[op as usize],
+            global_metric,
-            per_timeline_latency_histo,
+            timeline_metric,
             ctx,
             start,
             op,
@@ -1467,12 +1423,9 @@ mod smgr_query_time_tests {
             let get_counts = || {
                 let global: u64 = ops
                     .iter()
-                    .map(|op| metrics.global_latency[*op as usize].get_sample_count())
+                    .map(|op| metrics.global_metrics[*op as usize].get_sample_count())
                     .sum();
-                (
+                (global, metrics.per_timeline_getpage.get_sample_count())
-                    global,
-                    metrics.per_timeline_getpage_latency.get_sample_count(),
-                )
             };
 
             let (pre_global, pre_per_tenant_timeline) = get_counts();
@@ -1534,7 +1487,7 @@ impl BasebackupQueryTime {
     pub(crate) fn start_recording<'c: 'a, 'a>(
         &'a self,
         ctx: &'c RequestContext,
-    ) -> BasebackupQueryTimeOngoingRecording<'a, 'a> {
+    ) -> BasebackupQueryTimeOngoingRecording<'_, '_> {
         let start = Instant::now();
         match ctx.micros_spent_throttled.open() {
             Ok(()) => (),
@@ -1824,7 +1777,7 @@ pub(crate) static SECONDARY_MODE: Lazy<SecondaryModeMetrics> = Lazy::new(|| {
     .expect("failed to define a metric"),
     upload_heatmap_duration: register_histogram!(
         "pageserver_secondary_upload_heatmap_duration",
-        "Time to build and upload a heatmap, including any waiting inside the remote storage client"
+        "Time to build and upload a heatmap, including any waiting inside the S3 client"
     )
     .expect("failed to define a metric"),
     download_heatmap: register_int_counter!(
@@ -2623,12 +2576,6 @@ impl TimelineMetrics {
             let _ = STORAGE_IO_SIZE.remove_label_values(&[op, tenant_id, shard_id, timeline_id]);
         }
 
-        let _ = SMGR_QUERY_STARTED_PER_TENANT_TIMELINE.remove_label_values(&[
-            SmgrQueryType::GetPageAtLsn.into(),
-            tenant_id,
-            shard_id,
-            timeline_id,
-        ]);
         let _ = SMGR_QUERY_TIME_PER_TENANT_TIMELINE.remove_label_values(&[
             SmgrQueryType::GetPageAtLsn.into(),
             tenant_id,
@@ -2645,8 +2592,6 @@ pub(crate) fn remove_tenant_metrics(tenant_shard_id: &TenantShardId) {
         let _ = TENANT_SYNTHETIC_SIZE_METRIC.remove_label_values(&[&tid]);
     }
 
-    tenant_throttling::remove_tenant_metrics(tenant_shard_id);
-
     // we leave the BROKEN_TENANTS_SET entry if any
 }
 
@@ -3110,173 +3055,41 @@ pub mod tokio_epoll_uring {
 pub(crate) mod tenant_throttling {
     use metrics::{register_int_counter_vec, IntCounter};
     use once_cell::sync::Lazy;
-    use utils::shard::TenantShardId;
 
     use crate::tenant::{self, throttle::Metric};
 
-    struct GlobalAndPerTenantIntCounter {
-        global: IntCounter,
-        per_tenant: IntCounter,
-    }
-
-    impl GlobalAndPerTenantIntCounter {
-        #[inline(always)]
-        pub(crate) fn inc(&self) {
-            self.inc_by(1)
-        }
-        #[inline(always)]
-        pub(crate) fn inc_by(&self, n: u64) {
-            self.global.inc_by(n);
-            self.per_tenant.inc_by(n);
-        }
-    }
-
     pub(crate) struct TimelineGet {
-        count_accounted_start: GlobalAndPerTenantIntCounter,
+        wait_time: IntCounter,
-        count_accounted_finish: GlobalAndPerTenantIntCounter,
+        count: IntCounter,
-        wait_time: GlobalAndPerTenantIntCounter,
-        count_throttled: GlobalAndPerTenantIntCounter,
     }
 
-    static COUNT_ACCOUNTED_START: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
+    pub(crate) static TIMELINE_GET: Lazy<TimelineGet> = Lazy::new(|| {
-        register_int_counter_vec!(
+        static WAIT_USECS: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-            "pageserver_tenant_throttling_count_accounted_start_global",
+            register_int_counter_vec!(
-            "Count of tenant throttling starts, by kind of throttle.",
-            &["kind"]
-        )
-        .unwrap()
-    });
-    static COUNT_ACCOUNTED_START_PER_TENANT: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-        register_int_counter_vec!(
-            "pageserver_tenant_throttling_count_accounted_start",
-            "Count of tenant throttling starts, by kind of throttle.",
-            &["kind", "tenant_id", "shard_id"]
-        )
-        .unwrap()
-    });
-    static COUNT_ACCOUNTED_FINISH: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-        register_int_counter_vec!(
-            "pageserver_tenant_throttling_count_accounted_finish_global",
-            "Count of tenant throttling finishes, by kind of throttle.",
-            &["kind"]
-        )
-        .unwrap()
-    });
-    static COUNT_ACCOUNTED_FINISH_PER_TENANT: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-        register_int_counter_vec!(
-            "pageserver_tenant_throttling_count_accounted_finish",
-            "Count of tenant throttling finishes, by kind of throttle.",
-            &["kind", "tenant_id", "shard_id"]
-        )
-        .unwrap()
-    });
-    static WAIT_USECS: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-        register_int_counter_vec!(
             "pageserver_tenant_throttling_wait_usecs_sum_global",
-            "Sum of microseconds that spent waiting throttle by kind of throttle.",
+            "Sum of microseconds that tenants spent waiting for a tenant throttle of a given kind.",
             &["kind"]
         )
-        .unwrap()
+            .unwrap()
-    });
+        });
-    static WAIT_USECS_PER_TENANT: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
+
-        register_int_counter_vec!(
+        static WAIT_COUNT: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-            "pageserver_tenant_throttling_wait_usecs_sum",
+            register_int_counter_vec!(
-            "Sum of microseconds that spent waiting throttle by kind of throttle.",
+                "pageserver_tenant_throttling_count_global",
-            &["kind", "tenant_id", "shard_id"]
+                "Count of tenant throttlings, by kind of throttle.",
-        )
+                &["kind"]
-        .unwrap()
+            )
+            .unwrap()
+        });
+
+        let kind = "timeline_get";
+        TimelineGet {
+            wait_time: WAIT_USECS.with_label_values(&[kind]),
+            count: WAIT_COUNT.with_label_values(&[kind]),
+        }
     });
 
-    static WAIT_COUNT: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
+    impl Metric for &'static TimelineGet {
-        register_int_counter_vec!(
-            "pageserver_tenant_throttling_count_global",
-            "Count of tenant throttlings, by kind of throttle.",
-            &["kind"]
-        )
-        .unwrap()
-    });
-    static WAIT_COUNT_PER_TENANT: Lazy<metrics::IntCounterVec> = Lazy::new(|| {
-        register_int_counter_vec!(
-            "pageserver_tenant_throttling_count",
-            "Count of tenant throttlings, by kind of throttle.",
-            &["kind", "tenant_id", "shard_id"]
-        )
-        .unwrap()
-    });
-
-    const KIND: &str = "timeline_get";
-
-    impl TimelineGet {
-        pub(crate) fn new(tenant_shard_id: &TenantShardId) -> Self {
-            let per_tenant_label_values = &[
-                KIND,
-                &tenant_shard_id.tenant_id.to_string(),
-                &tenant_shard_id.shard_slug().to_string(),
-            ];
-            TimelineGet {
-                count_accounted_start: {
-                    GlobalAndPerTenantIntCounter {
-                        global: COUNT_ACCOUNTED_START.with_label_values(&[KIND]),
-                        per_tenant: COUNT_ACCOUNTED_START_PER_TENANT
-                            .with_label_values(per_tenant_label_values),
-                    }
-                },
-                count_accounted_finish: {
-                    GlobalAndPerTenantIntCounter {
-                        global: COUNT_ACCOUNTED_FINISH.with_label_values(&[KIND]),
-                        per_tenant: COUNT_ACCOUNTED_FINISH_PER_TENANT
-                            .with_label_values(per_tenant_label_values),
-                    }
-                },
-                wait_time: {
-                    GlobalAndPerTenantIntCounter {
-                        global: WAIT_USECS.with_label_values(&[KIND]),
-                        per_tenant: WAIT_USECS_PER_TENANT
-                            .with_label_values(per_tenant_label_values),
-                    }
-                },
-                count_throttled: {
-                    GlobalAndPerTenantIntCounter {
-                        global: WAIT_COUNT.with_label_values(&[KIND]),
-                        per_tenant: WAIT_COUNT_PER_TENANT
-                            .with_label_values(per_tenant_label_values),
-                    }
-                },
-            }
-        }
-    }
-
-    pub(crate) fn preinitialize_global_metrics() {
-        Lazy::force(&COUNT_ACCOUNTED_START);
-        Lazy::force(&COUNT_ACCOUNTED_FINISH);
-        Lazy::force(&WAIT_USECS);
-        Lazy::force(&WAIT_COUNT);
-    }
-
-    pub(crate) fn remove_tenant_metrics(tenant_shard_id: &TenantShardId) {
-        for m in &[
-            &COUNT_ACCOUNTED_START_PER_TENANT,
-            &COUNT_ACCOUNTED_FINISH_PER_TENANT,
-            &WAIT_USECS_PER_TENANT,
-            &WAIT_COUNT_PER_TENANT,
-        ] {
-            let _ = m.remove_label_values(&[
-                KIND,
-                &tenant_shard_id.tenant_id.to_string(),
-                &tenant_shard_id.shard_slug().to_string(),
-            ]);
-        }
-    }
-
-    impl Metric for TimelineGet {
-        #[inline(always)]
-        fn accounting_start(&self) {
-            self.count_accounted_start.inc();
-        }
-        #[inline(always)]
-        fn accounting_finish(&self) {
-            self.count_accounted_finish.inc();
-        }
         #[inline(always)]
         fn observe_throttling(
             &self,
@@ -3284,7 +3097,7 @@ pub(crate) mod tenant_throttling {
         ) {
             let val = u64::try_from(wait_time.as_micros()).unwrap();
             self.wait_time.inc_by(val);
-            self.count_throttled.inc();
+            self.count.inc();
         }
     }
 }
@@ -3414,14 +3227,11 @@ pub fn preinitialize_metrics() {
     }
 
     // countervecs
-    [
+    [&BACKGROUND_LOOP_PERIOD_OVERRUN_COUNT]
-        &BACKGROUND_LOOP_PERIOD_OVERRUN_COUNT,
+        .into_iter()
-        &SMGR_QUERY_STARTED_GLOBAL,
+        .for_each(|c| {
-    ]
+            Lazy::force(c);
-    .into_iter()
+        });
-    .for_each(|c| {
-        Lazy::force(c);
-    });
 
     // gauges
     WALRECEIVER_ACTIVE_MANAGERS.get();
@@ -3443,8 +3253,7 @@ pub fn preinitialize_metrics() {
 
     // Custom
     Lazy::force(&RECONSTRUCT_TIME);
+    Lazy::force(&tenant_throttling::TIMELINE_GET);
     Lazy::force(&BASEBACKUP_QUERY_TIME);
     Lazy::force(&COMPUTE_COMMANDS_COUNTERS);
-
-    tenant_throttling::preinitialize_global_metrics();
 }

pageserver/src/pgdatadir_mapping.rs

@@ -633,7 +633,7 @@ impl Timeline {
 
     pub(crate) async fn get_twophase_file(
         &self,
-        xid: u64,
+        xid: TransactionId,
         lsn: Lsn,
         ctx: &RequestContext,
     ) -> Result<Bytes, PageReconstructError> {
@@ -646,19 +646,11 @@ impl Timeline {
         &self,
         lsn: Lsn,
         ctx: &RequestContext,
-    ) -> Result<HashSet<u64>, PageReconstructError> {
+    ) -> Result<HashSet<TransactionId>, PageReconstructError> {
         // fetch directory entry
         let buf = self.get(TWOPHASEDIR_KEY, lsn, ctx).await?;
 
-        if self.pg_version >= 17 {
+        Ok(TwoPhaseDirectory::des(&buf)?.xids)
-            Ok(TwoPhaseDirectoryV17::des(&buf)?.xids)
-        } else {
-            Ok(TwoPhaseDirectory::des(&buf)?
-                .xids
-                .iter()
-                .map(|x| u64::from(*x))
-                .collect())
-        }
     }
 
     pub(crate) async fn get_control_file(
@@ -840,36 +832,6 @@ impl Timeline {
         Ok(total_size * BLCKSZ as u64)
     }
 
-    /// Get a KeySpace that covers all the Keys that are in use at AND below the given LSN. This is only used
-    /// for gc-compaction.
-    ///
-    /// gc-compaction cannot use the same `collect_keyspace` function as the legacy compaction because it
-    /// processes data at multiple LSNs and needs to be aware of the fact that some key ranges might need to
-    /// be kept only for a specific range of LSN.
-    ///
-    /// Consider the case that the user created branches at LSN 10 and 20, where the user created a table A at
-    /// LSN 10 and dropped that table at LSN 20. `collect_keyspace` at LSN 10 will return the key range
-    /// corresponding to that table, while LSN 20 won't. The keyspace info at a single LSN is not enough to
-    /// determine which keys to retain/drop for gc-compaction.
-    ///
-    /// For now, it only drops AUX-v1 keys. But in the future, the function will be extended to return the keyspace
-    /// to be retained for each of the branch LSN.
-    ///
-    /// The return value is (dense keyspace, sparse keyspace).
-    pub(crate) async fn collect_gc_compaction_keyspace(
-        &self,
-    ) -> Result<(KeySpace, SparseKeySpace), CollectKeySpaceError> {
-        let metadata_key_begin = Key::metadata_key_range().start;
-        let aux_v1_key = AUX_FILES_KEY;
-        let dense_keyspace = KeySpace {
-            ranges: vec![Key::MIN..aux_v1_key, aux_v1_key.next()..metadata_key_begin],
-        };
-        Ok((
-            dense_keyspace,
-            SparseKeySpace(KeySpace::single(Key::metadata_key_range())),
-        ))
-    }
-
     ///
     /// Get a KeySpace that covers all the Keys that are in use at the given LSN.
     /// Anything that's not listed maybe removed from the underlying storage (from
@@ -940,13 +902,9 @@ impl Timeline {
 
         // Then pg_twophase
         result.add_key(TWOPHASEDIR_KEY);
-
+        let buf = self.get(TWOPHASEDIR_KEY, lsn, ctx).await?;
-        let mut xids: Vec<u64> = self
+        let twophase_dir = TwoPhaseDirectory::des(&buf)?;
-            .list_twophase_files(lsn, ctx)
+        let mut xids: Vec<TransactionId> = twophase_dir.xids.iter().cloned().collect();
-            .await?
-            .iter()
-            .cloned()
-            .collect();
         xids.sort_unstable();
         for xid in xids {
             result.add_key(twophase_file_key(xid));
@@ -1169,15 +1127,9 @@ impl<'a> DatadirModification<'a> {
         // Create AuxFilesDirectory
         self.init_aux_dir()?;
 
-        let buf = if self.tline.pg_version >= 17 {
+        let buf = TwoPhaseDirectory::ser(&TwoPhaseDirectory {
-            TwoPhaseDirectoryV17::ser(&TwoPhaseDirectoryV17 {
+            xids: HashSet::new(),
-                xids: HashSet::new(),
+        })?;
-            })
-        } else {
-            TwoPhaseDirectory::ser(&TwoPhaseDirectory {
-                xids: HashSet::new(),
-            })
-        }?;
         self.pending_directory_entries
             .push((DirectoryKind::TwoPhase, 0));
         self.put(TWOPHASEDIR_KEY, Value::Image(buf.into()));
@@ -1369,31 +1321,22 @@ impl<'a> DatadirModification<'a> {
 
     pub async fn put_twophase_file(
         &mut self,
-        xid: u64,
+        xid: TransactionId,
         img: Bytes,
         ctx: &RequestContext,
     ) -> anyhow::Result<()> {
         // Add it to the directory entry
-        let dirbuf = self.get(TWOPHASEDIR_KEY, ctx).await?;
+        let buf = self.get(TWOPHASEDIR_KEY, ctx).await?;
-        let newdirbuf = if self.tline.pg_version >= 17 {
+        let mut dir = TwoPhaseDirectory::des(&buf)?;
-            let mut dir = TwoPhaseDirectoryV17::des(&dirbuf)?;
+        if !dir.xids.insert(xid) {
-            if !dir.xids.insert(xid) {
+            anyhow::bail!("twophase file for xid {} already exists", xid);
-                anyhow::bail!("twophase file for xid {} already exists", xid);
+        }
-            }
+        self.pending_directory_entries
-            self.pending_directory_entries
+            .push((DirectoryKind::TwoPhase, dir.xids.len()));
-                .push((DirectoryKind::TwoPhase, dir.xids.len()));
+        self.put(
-            Bytes::from(TwoPhaseDirectoryV17::ser(&dir)?)
+            TWOPHASEDIR_KEY,
-        } else {
+            Value::Image(Bytes::from(TwoPhaseDirectory::ser(&dir)?)),
-            let xid = xid as u32;
+        );
-            let mut dir = TwoPhaseDirectory::des(&dirbuf)?;
-            if !dir.xids.insert(xid) {
-                anyhow::bail!("twophase file for xid {} already exists", xid);
-            }
-            self.pending_directory_entries
-                .push((DirectoryKind::TwoPhase, dir.xids.len()));
-            Bytes::from(TwoPhaseDirectory::ser(&dir)?)
-        };
-        self.put(TWOPHASEDIR_KEY, Value::Image(newdirbuf));
 
         self.put(twophase_file_key(xid), Value::Image(img));
         Ok(())
@@ -1696,32 +1639,22 @@ impl<'a> DatadirModification<'a> {
     /// This method is used for marking truncated SLRU files
     pub async fn drop_twophase_file(
         &mut self,
-        xid: u64,
+        xid: TransactionId,
         ctx: &RequestContext,
     ) -> anyhow::Result<()> {
         // Remove it from the directory entry
         let buf = self.get(TWOPHASEDIR_KEY, ctx).await?;
-        let newdirbuf = if self.tline.pg_version >= 17 {
+        let mut dir = TwoPhaseDirectory::des(&buf)?;
-            let mut dir = TwoPhaseDirectoryV17::des(&buf)?;
 
-            if !dir.xids.remove(&xid) {
+        if !dir.xids.remove(&xid) {
-                warn!("twophase file for xid {} does not exist", xid);
+            warn!("twophase file for xid {} does not exist", xid);
-            }
+        }
-            self.pending_directory_entries
+        self.pending_directory_entries
-                .push((DirectoryKind::TwoPhase, dir.xids.len()));
+            .push((DirectoryKind::TwoPhase, dir.xids.len()));
-            Bytes::from(TwoPhaseDirectoryV17::ser(&dir)?)
+        self.put(
-        } else {
+            TWOPHASEDIR_KEY,
-            let xid: u32 = u32::try_from(xid)?;
+            Value::Image(Bytes::from(TwoPhaseDirectory::ser(&dir)?)),
-            let mut dir = TwoPhaseDirectory::des(&buf)?;
+        );
-
-            if !dir.xids.remove(&xid) {
-                warn!("twophase file for xid {} does not exist", xid);
-            }
-            self.pending_directory_entries
-                .push((DirectoryKind::TwoPhase, dir.xids.len()));
-            Bytes::from(TwoPhaseDirectory::ser(&dir)?)
-        };
-        self.put(TWOPHASEDIR_KEY, Value::Image(newdirbuf));
 
         // Delete it
         self.delete(twophase_key_range(xid));
@@ -2191,21 +2124,11 @@ struct DbDirectory {
     dbdirs: HashMap<(Oid, Oid), bool>,
 }
 
-// The format of TwoPhaseDirectory changed in PostgreSQL v17, because the filenames of
-// pg_twophase files was expanded from 32-bit XIDs to 64-bit XIDs.  Previously, the files
-// were named like "pg_twophase/000002E5", now they're like
-// "pg_twophsae/0000000A000002E4".
-
 #[derive(Debug, Serialize, Deserialize)]
 struct TwoPhaseDirectory {
     xids: HashSet<TransactionId>,
 }
 
-#[derive(Debug, Serialize, Deserialize)]
-struct TwoPhaseDirectoryV17 {
-    xids: HashSet<u64>,
-}
-
 #[derive(Debug, Serialize, Deserialize, Default)]
 struct RelDirectory {
     // Set of relations that exist. (relfilenode, forknum)

pageserver/src/tenant.rs

@@ -18,6 +18,7 @@ use camino::Utf8Path;
 use camino::Utf8PathBuf;
 use enumset::EnumSet;
 use futures::stream::FuturesUnordered;
+use futures::FutureExt;
 use futures::StreamExt;
 use pageserver_api::models;
 use pageserver_api::models::AuxFilePolicy;
@@ -33,7 +34,6 @@ use remote_storage::GenericRemoteStorage;
 use remote_storage::TimeoutOrCancel;
 use std::collections::BTreeMap;
 use std::fmt;
-use std::future::Future;
 use std::sync::Weak;
 use std::time::SystemTime;
 use storage_broker::BrokerClientChannel;
@@ -140,7 +140,6 @@ pub mod metadata;
 pub mod remote_timeline_client;
 pub mod storage_layer;
 
-pub mod checks;
 pub mod config;
 pub mod mgr;
 pub mod secondary;
@@ -302,7 +301,7 @@ pub struct Tenant {
     /// Throttle applied at the top of [`Timeline::get`].
     /// All [`Tenant::timelines`] of a given [`Tenant`] instance share the same [`throttle::Throttle`] instance.
     pub(crate) timeline_get_throttle:
-        Arc<throttle::Throttle<crate::metrics::tenant_throttling::TimelineGet>>,
+        Arc<throttle::Throttle<&'static crate::metrics::tenant_throttling::TimelineGet>>,
 
     /// An ongoing timeline detach concurrency limiter.
     ///
@@ -1031,9 +1030,13 @@ impl Tenant {
         }
 
         Ok(TenantPreload {
-            timelines: self
+            timelines: Self::load_timeline_metadata(
-                .load_timelines_metadata(remote_timeline_ids, remote_storage, cancel)
+                self,
-                .await?,
+                remote_timeline_ids,
+                remote_storage,
+                cancel,
+            )
+            .await?,
         })
     }
 
@@ -1299,7 +1302,7 @@ impl Tenant {
         .await
     }
 
-    async fn load_timelines_metadata(
+    async fn load_timeline_metadata(
         self: &Arc<Tenant>,
         timeline_ids: HashSet<TimelineId>,
         remote_storage: &GenericRemoteStorage,
@@ -1307,10 +1310,33 @@ impl Tenant {
     ) -> anyhow::Result<HashMap<TimelineId, TimelinePreload>> {
         let mut part_downloads = JoinSet::new();
         for timeline_id in timeline_ids {
+            let client = RemoteTimelineClient::new(
+                remote_storage.clone(),
+                self.deletion_queue_client.clone(),
+                self.conf,
+                self.tenant_shard_id,
+                timeline_id,
+                self.generation,
+            );
             let cancel_clone = cancel.clone();
             part_downloads.spawn(
-                self.load_timeline_metadata(timeline_id, remote_storage.clone(), cancel_clone)
+                async move {
-                    .instrument(info_span!("download_index_part", %timeline_id)),
+                    debug!("starting index part download");
+
+                    let index_part = client.download_index_file(&cancel_clone).await;
+
+                    debug!("finished index part download");
+
+                    Result::<_, anyhow::Error>::Ok(TimelinePreload {
+                        client,
+                        timeline_id,
+                        index_part,
+                    })
+                }
+                .map(move |res| {
+                    res.with_context(|| format!("download index part for timeline {timeline_id}"))
+                })
+                .instrument(info_span!("download_index_part", %timeline_id)),
             );
         }
 
@@ -1321,7 +1347,8 @@ impl Tenant {
                 next = part_downloads.join_next() => {
                     match next {
                         Some(result) => {
-                            let preload = result.context("join preload task")?;
+                            let preload_result = result.context("join preload task")?;
+                            let preload = preload_result?;
                             timeline_preloads.insert(preload.timeline_id, preload);
                         },
                         None => {
@@ -1338,36 +1365,6 @@ impl Tenant {
         Ok(timeline_preloads)
     }
 
-    fn load_timeline_metadata(
-        self: &Arc<Tenant>,
-        timeline_id: TimelineId,
-        remote_storage: GenericRemoteStorage,
-        cancel: CancellationToken,
-    ) -> impl Future<Output = TimelinePreload> {
-        let client = RemoteTimelineClient::new(
-            remote_storage.clone(),
-            self.deletion_queue_client.clone(),
-            self.conf,
-            self.tenant_shard_id,
-            timeline_id,
-            self.generation,
-        );
-        async move {
-            debug_assert_current_span_has_tenant_and_timeline_id();
-            debug!("starting index part download");
-
-            let index_part = client.download_index_file(&cancel).await;
-
-            debug!("finished index part download");
-
-            TimelinePreload {
-                client,
-                timeline_id,
-                index_part,
-            }
-        }
-    }
-
     pub(crate) async fn apply_timeline_archival_config(
         &self,
         timeline_id: TimelineId,
@@ -1576,9 +1573,6 @@ impl Tenant {
         image_layer_desc: Vec<(Lsn, Vec<(pageserver_api::key::Key, bytes::Bytes)>)>,
         end_lsn: Lsn,
     ) -> anyhow::Result<Arc<Timeline>> {
-        use checks::check_valid_layermap;
-        use itertools::Itertools;
-
         let tline = self
             .create_test_timeline(new_timeline_id, initdb_lsn, pg_version, ctx)
             .await?;
@@ -1593,18 +1587,6 @@ impl Tenant {
                 .force_create_image_layer(lsn, images, Some(initdb_lsn), ctx)
                 .await?;
         }
-        let layer_names = tline
-            .layers
-            .read()
-            .await
-            .layer_map()
-            .unwrap()
-            .iter_historic_layers()
-            .map(|layer| layer.layer_name())
-            .collect_vec();
-        if let Some(err) = check_valid_layermap(&layer_names) {
-            bail!("invalid layermap: {err}");
-        }
         Ok(tline)
     }
 
@@ -1968,6 +1950,9 @@ impl Tenant {
                 TenantState::Activating(_) | TenantState::Active | TenantState::Broken { .. } | TenantState::Stopping { .. } => {
                     panic!("caller is responsible for calling activate() only on Loading / Attaching tenants, got {state:?}", state = current_state);
                 }
+                TenantState::Loading => {
+                    *current_state = TenantState::Activating(ActivatingFrom::Loading);
+                }
                 TenantState::Attaching => {
                     *current_state = TenantState::Activating(ActivatingFrom::Attaching);
                 }
@@ -2148,7 +2133,7 @@ impl Tenant {
     async fn set_stopping(
         &self,
         progress: completion::Barrier,
-        _allow_transition_from_loading: bool,
+        allow_transition_from_loading: bool,
         allow_transition_from_attaching: bool,
     ) -> Result<(), SetStoppingError> {
         let mut rx = self.state.subscribe();
@@ -2163,6 +2148,7 @@ impl Tenant {
                 );
                 false
             }
+            TenantState::Loading => allow_transition_from_loading,
             TenantState::Active | TenantState::Broken { .. } | TenantState::Stopping { .. } => true,
         })
         .await
@@ -2181,6 +2167,13 @@ impl Tenant {
                 *current_state = TenantState::Stopping { progress };
                 true
             }
+            TenantState::Loading => {
+                if !allow_transition_from_loading {
+                    unreachable!("3we ensured above that we're done with activation, and, there is no re-activation")
+                };
+                *current_state = TenantState::Stopping { progress };
+                true
+            }
             TenantState::Active => {
                 // FIXME: due to time-of-check vs time-of-use issues, it can happen that new timelines
                 // are created after the transition to Stopping. That's harmless, as the Timelines
@@ -2236,7 +2229,7 @@ impl Tenant {
         // The load & attach routines own the tenant state until it has reached `Active`.
         // So, wait until it's done.
         rx.wait_for(|state| match state {
-            TenantState::Activating(_) | TenantState::Attaching => {
+            TenantState::Activating(_) | TenantState::Loading | TenantState::Attaching => {
                 info!(
                     "waiting for {} to turn Active|Broken|Stopping",
                     <&'static str>::from(state)
@@ -2256,7 +2249,7 @@ impl Tenant {
         let reason = reason.to_string();
         self.state.send_modify(|current_state| {
             match *current_state {
-                TenantState::Activating(_) | TenantState::Attaching => {
+                TenantState::Activating(_) | TenantState::Loading | TenantState::Attaching => {
                     unreachable!("we ensured above that we're done with activation, and, there is no re-activation")
                 }
                 TenantState::Active => {
@@ -2300,7 +2293,7 @@ impl Tenant {
         loop {
             let current_state = receiver.borrow_and_update().clone();
             match current_state {
-                TenantState::Attaching | TenantState::Activating(_) => {
+                TenantState::Loading | TenantState::Attaching | TenantState::Activating(_) => {
                     // in these states, there's a chance that we can reach ::Active
                     self.activate_now();
                     match timeout_cancellable(timeout, &self.cancel, receiver.changed()).await {
@@ -2822,7 +2815,7 @@ impl Tenant {
             gate: Gate::default(),
             timeline_get_throttle: Arc::new(throttle::Throttle::new(
                 Tenant::get_timeline_get_throttle_config(conf, &attached_conf.tenant_conf),
-                crate::metrics::tenant_throttling::TimelineGet::new(&tenant_shard_id),
+                &crate::metrics::tenant_throttling::TIMELINE_GET,
             )),
             tenant_conf: Arc::new(ArcSwap::from_pointee(attached_conf)),
             ongoing_timeline_detach: std::sync::Mutex::default(),
@@ -3204,9 +3197,6 @@ impl Tenant {
         image_layer_desc: Vec<(Lsn, Vec<(pageserver_api::key::Key, bytes::Bytes)>)>,
         end_lsn: Lsn,
     ) -> anyhow::Result<Arc<Timeline>> {
-        use checks::check_valid_layermap;
-        use itertools::Itertools;
-
         let tline = self
             .branch_timeline_test(src_timeline, dst_id, ancestor_lsn, ctx)
             .await?;
@@ -3227,18 +3217,6 @@ impl Tenant {
                 .force_create_image_layer(lsn, images, Some(ancestor_lsn), ctx)
                 .await?;
         }
-        let layer_names = tline
-            .layers
-            .read()
-            .await
-            .layer_map()
-            .unwrap()
-            .iter_historic_layers()
-            .map(|layer| layer.layer_name())
-            .collect_vec();
-        if let Some(err) = check_valid_layermap(&layer_names) {
-            bail!("invalid layermap: {err}");
-        }
         Ok(tline)
     }
 
@@ -3616,7 +3594,7 @@ impl Tenant {
         start_lsn: Lsn,
         ancestor: Option<Arc<Timeline>>,
         last_aux_file_policy: Option<AuxFilePolicy>,
-    ) -> anyhow::Result<UninitializedTimeline<'a>> {
+    ) -> anyhow::Result<UninitializedTimeline> {
         let tenant_shard_id = self.tenant_shard_id;
 
         let resources = self.build_timeline_resources(new_timeline_id);
@@ -4133,7 +4111,7 @@ pub(crate) mod harness {
             let walredo_mgr = Arc::new(WalRedoManager::from(TestRedoManager));
 
             let tenant = Arc::new(Tenant::new(
-                TenantState::Attaching,
+                TenantState::Loading,
                 self.conf,
                 AttachedTenantConf::try_from(LocationConf::attached_single(
                     TenantConfOpt::from(self.tenant_conf.clone()),
@@ -4186,18 +4164,9 @@ pub(crate) mod harness {
             let records_neon = records.iter().all(|r| apply_neon::can_apply_in_neon(&r.1));
             if records_neon {
                 // For Neon wal records, we can decode without spawning postgres, so do so.
-                let mut page = match (base_img, records.first()) {
+                let base_img = base_img.expect("Neon WAL redo requires base image").1;
-                    (Some((_lsn, img)), _) => {
+                let mut page = BytesMut::new();
-                        let mut page = BytesMut::new();
+                page.extend_from_slice(&base_img);
-                        page.extend_from_slice(&img);
-                        page
-                    }
-                    (_, Some((_lsn, rec))) if rec.will_init() => BytesMut::new(),
-                    _ => {
-                        panic!("Neon WAL redo requires base image or will init record");
-                    }
-                };
-
                 for (record_lsn, record) in records {
                     apply_neon::apply_in_neon(&record, record_lsn, key, &mut page)?;
                 }
@@ -8501,135 +8470,4 @@ mod tests {
 
         Ok(())
     }
-
-    // Regression test for https://github.com/neondatabase/neon/issues/9012
-    // Create an image arrangement where we have to read at different LSN ranges
-    // from a delta layer. This is achieved by overlapping an image layer on top of
-    // a delta layer. Like so:
-    //
-    //     A      B
-    // +----------------+ -> delta_layer
-    // |                |                           ^ lsn
-    // |       =========|-> nested_image_layer      |
-    // |       C        |                           |
-    // +----------------+                           |
-    // ======== -> baseline_image_layer             +-------> key
-    //
-    //
-    // When querying the key range [A, B) we need to read at different LSN ranges
-    // for [A, C) and [C, B). This test checks that the described edge case is handled correctly.
-    #[tokio::test]
-    async fn test_vectored_read_with_nested_image_layer() -> anyhow::Result<()> {
-        let harness = TenantHarness::create("test_vectored_read_with_nested_image_layer").await?;
-        let (tenant, ctx) = harness.load().await;
-
-        let will_init_keys = [2, 6];
-        fn get_key(id: u32) -> Key {
-            let mut key = Key::from_hex("110000000033333333444444445500000000").unwrap();
-            key.field6 = id;
-            key
-        }
-
-        let mut expected_key_values = HashMap::new();
-
-        let baseline_image_layer_lsn = Lsn(0x10);
-        let mut baseline_img_layer = Vec::new();
-        for i in 0..5 {
-            let key = get_key(i);
-            let value = format!("value {i}@{baseline_image_layer_lsn}");
-
-            let removed = expected_key_values.insert(key, value.clone());
-            assert!(removed.is_none());
-
-            baseline_img_layer.push((key, Bytes::from(value)));
-        }
-
-        let nested_image_layer_lsn = Lsn(0x50);
-        let mut nested_img_layer = Vec::new();
-        for i in 5..10 {
-            let key = get_key(i);
-            let value = format!("value {i}@{nested_image_layer_lsn}");
-
-            let removed = expected_key_values.insert(key, value.clone());
-            assert!(removed.is_none());
-
-            nested_img_layer.push((key, Bytes::from(value)));
-        }
-
-        let mut delta_layer_spec = Vec::default();
-        let delta_layer_start_lsn = Lsn(0x20);
-        let mut delta_layer_end_lsn = delta_layer_start_lsn;
-
-        for i in 0..10 {
-            let key = get_key(i);
-            let key_in_nested = nested_img_layer
-                .iter()
-                .any(|(key_with_img, _)| *key_with_img == key);
-            let lsn = {
-                if key_in_nested {
-                    Lsn(nested_image_layer_lsn.0 + 0x10)
-                } else {
-                    delta_layer_start_lsn
-                }
-            };
-
-            let will_init = will_init_keys.contains(&i);
-            if will_init {
-                delta_layer_spec.push((key, lsn, Value::WalRecord(NeonWalRecord::wal_init())));
-
-                expected_key_values.insert(key, "".to_string());
-            } else {
-                let delta = format!("@{lsn}");
-                delta_layer_spec.push((
-                    key,
-                    lsn,
-                    Value::WalRecord(NeonWalRecord::wal_append(&delta)),
-                ));
-
-                expected_key_values
-                    .get_mut(&key)
-                    .expect("An image exists for each key")
-                    .push_str(delta.as_str());
-            }
-            delta_layer_end_lsn = std::cmp::max(delta_layer_start_lsn, lsn);
-        }
-
-        delta_layer_end_lsn = Lsn(delta_layer_end_lsn.0 + 1);
-
-        assert!(
-            nested_image_layer_lsn > delta_layer_start_lsn
-                && nested_image_layer_lsn < delta_layer_end_lsn
-        );
-
-        let tline = tenant
-            .create_test_timeline_with_layers(
-                TIMELINE_ID,
-                baseline_image_layer_lsn,
-                DEFAULT_PG_VERSION,
-                &ctx,
-                vec![DeltaLayerTestDesc::new_with_inferred_key_range(
-                    delta_layer_start_lsn..delta_layer_end_lsn,
-                    delta_layer_spec,
-                )], // delta layers
-                vec![
-                    (baseline_image_layer_lsn, baseline_img_layer),
-                    (nested_image_layer_lsn, nested_img_layer),
-                ], // image layers
-                delta_layer_end_lsn,
-            )
-            .await?;
-
-        let keyspace = KeySpace::single(get_key(0)..get_key(10));
-        let results = tline
-            .get_vectored(keyspace, delta_layer_end_lsn, &ctx)
-            .await
-            .expect("No vectored errors");
-        for (key, res) in results {
-            let value = res.expect("No key errors");
-            let expected_value = expected_key_values.remove(&key).expect("No unknown keys");
-            assert_eq!(value, Bytes::from(expected_value));
-        }
-
-        Ok(())
-    }
 }

pageserver/src/tenant/checks.rs

@@ -1,55 +0,0 @@
-use std::collections::BTreeSet;
-
-use itertools::Itertools;
-
-use super::storage_layer::LayerName;
-
-/// Checks whether a layer map is valid (i.e., is a valid result of the current compaction algorithm if nothing goes wrong).
-/// The function checks if we can split the LSN range of a delta layer only at the LSNs of the delta layers. For example,
-///
-/// ```plain
-/// |       |                 |       |
-/// |   1   |    |   2   |    |   3   |
-/// |       |    |       |    |       |
-/// ```
-///
-/// This is not a valid layer map because the LSN range of layer 1 intersects with the LSN range of layer 2. 1 and 2 should have
-/// the same LSN range.
-///
-/// The exception is that when layer 2 only contains a single key, it could be split over the LSN range. For example,
-///
-/// ```plain
-/// |       |    |   2   |    |       |
-/// |   1   |    |-------|    |   3   |
-/// |       |    |   4   |    |       |
-///
-/// If layer 2 and 4 contain the same single key, this is also a valid layer map.
-pub fn check_valid_layermap(metadata: &[LayerName]) -> Option<String> {
-    let mut lsn_split_point = BTreeSet::new(); // TODO: use a better data structure (range tree / range set?)
-    let mut all_delta_layers = Vec::new();
-    for name in metadata {
-        if let LayerName::Delta(layer) = name {
-            if layer.key_range.start.next() != layer.key_range.end {
-                all_delta_layers.push(layer.clone());
-            }
-        }
-    }
-    for layer in &all_delta_layers {
-        let lsn_range = &layer.lsn_range;
-        lsn_split_point.insert(lsn_range.start);
-        lsn_split_point.insert(lsn_range.end);
-    }
-    for layer in &all_delta_layers {
-        let lsn_range = layer.lsn_range.clone();
-        let intersects = lsn_split_point.range(lsn_range).collect_vec();
-        if intersects.len() > 1 {
-            let err = format!(
-                "layer violates the layer map LSN split assumption: layer {} intersects with LSN [{}]",
-                layer,
-                intersects.into_iter().map(|lsn| lsn.to_string()).join(", ")
-            );
-            return Some(err);
-        }
-    }
-    None
-}

pageserver/src/tenant/gc_block.rs

@@ -1,29 +1,11 @@
-use std::{collections::HashMap, time::Duration};
+use std::collections::HashMap;
 
-use super::remote_timeline_client::index::GcBlockingReason;
-use tokio::time::Instant;
 use utils::id::TimelineId;
 
-type TimelinesBlocked = HashMap<TimelineId, enumset::EnumSet<GcBlockingReason>>;
+use super::remote_timeline_client::index::GcBlockingReason;
 
-#[derive(Default)]
+type Storage = HashMap<TimelineId, enumset::EnumSet<GcBlockingReason>>;
-struct Storage {
-    timelines_blocked: TimelinesBlocked,
-    /// The deadline before which we are blocked from GC so that
-    /// leases have a chance to be renewed.
-    lsn_lease_deadline: Option<Instant>,
-}
 
-impl Storage {
-    fn is_blocked_by_lsn_lease_deadline(&self) -> bool {
-        self.lsn_lease_deadline
-            .map(|d| Instant::now() < d)
-            .unwrap_or(false)
-    }
-}
-
-/// GcBlock provides persistent (per-timeline) gc blocking and facilitates transient time based gc
-/// blocking.
 #[derive(Default)]
 pub(crate) struct GcBlock {
     /// The timelines which have current reasons to block gc.
@@ -31,12 +13,6 @@ pub(crate) struct GcBlock {
     /// LOCK ORDER: this is held locked while scheduling the next index_part update. This is done
     /// to keep the this field up to date with RemoteTimelineClient `upload_queue.dirty`.
     reasons: std::sync::Mutex<Storage>,
-
-    /// GC background task or manually run `Tenant::gc_iteration` holds a lock on this.
-    ///
-    /// Do not add any more features taking and forbidding taking this lock. It should be
-    /// `tokio::sync::Notify`, but that is rarely used. On the other side, [`GcBlock::insert`]
-    /// synchronizes with gc attempts by locking and unlocking this mutex.
     blocking: tokio::sync::Mutex<()>,
 }
 
@@ -66,20 +42,6 @@ impl GcBlock {
         }
     }
 
-    /// Sets a deadline before which we cannot proceed to GC due to lsn lease.
-    ///
-    /// We do this as the leases mapping are not persisted to disk. By delaying GC by lease
-    /// length, we guarantee that all the leases we granted before will have a chance to renew
-    /// when we run GC for the first time after restart / transition from AttachedMulti to AttachedSingle.
-    pub(super) fn set_lsn_lease_deadline(&self, lsn_lease_length: Duration) {
-        let deadline = Instant::now() + lsn_lease_length;
-        let mut g = self.reasons.lock().unwrap();
-        g.lsn_lease_deadline = Some(deadline);
-    }
-
-    /// Describe the current gc blocking reasons.
-    ///
-    /// TODO: make this json serializable.
     pub(crate) fn summary(&self) -> Option<BlockingReasons> {
         let g = self.reasons.lock().unwrap();
 
@@ -102,7 +64,7 @@ impl GcBlock {
     ) -> anyhow::Result<bool> {
         let (added, uploaded) = {
             let mut g = self.reasons.lock().unwrap();
-            let set = g.timelines_blocked.entry(timeline.timeline_id).or_default();
+            let set = g.entry(timeline.timeline_id).or_default();
             let added = set.insert(reason);
 
             // LOCK ORDER: intentionally hold the lock, see self.reasons.
@@ -133,7 +95,7 @@ impl GcBlock {
 
         let (remaining_blocks, uploaded) = {
             let mut g = self.reasons.lock().unwrap();
-            match g.timelines_blocked.entry(timeline.timeline_id) {
+            match g.entry(timeline.timeline_id) {
                 Entry::Occupied(mut oe) => {
                     let set = oe.get_mut();
                     set.remove(reason);
@@ -147,7 +109,7 @@ impl GcBlock {
                 }
             }
 
-            let remaining_blocks = g.timelines_blocked.len();
+            let remaining_blocks = g.len();
 
             // LOCK ORDER: intentionally hold the lock while scheduling; see self.reasons
             let uploaded = timeline
@@ -172,11 +134,11 @@ impl GcBlock {
     pub(crate) fn before_delete(&self, timeline: &super::Timeline) {
         let unblocked = {
             let mut g = self.reasons.lock().unwrap();
-            if g.timelines_blocked.is_empty() {
+            if g.is_empty() {
                 return;
             }
 
-            g.timelines_blocked.remove(&timeline.timeline_id);
+            g.remove(&timeline.timeline_id);
 
             BlockingReasons::clean_and_summarize(g).is_none()
         };
@@ -187,11 +149,10 @@ impl GcBlock {
     }
 
     /// Initialize with the non-deleted timelines of this tenant.
-    pub(crate) fn set_scanned(&self, scanned: TimelinesBlocked) {
+    pub(crate) fn set_scanned(&self, scanned: Storage) {
         let mut g = self.reasons.lock().unwrap();
-        assert!(g.timelines_blocked.is_empty());
+        assert!(g.is_empty());
-        g.timelines_blocked
+        g.extend(scanned.into_iter().filter(|(_, v)| !v.is_empty()));
-            .extend(scanned.into_iter().filter(|(_, v)| !v.is_empty()));
 
         if let Some(reasons) = BlockingReasons::clean_and_summarize(g) {
             tracing::info!(summary=?reasons, "initialized with gc blocked");
@@ -205,7 +166,6 @@ pub(super) struct Guard<'a> {
 
 #[derive(Debug)]
 pub(crate) struct BlockingReasons {
-    tenant_blocked_by_lsn_lease_deadline: bool,
     timelines: usize,
     reasons: enumset::EnumSet<GcBlockingReason>,
 }
@@ -214,8 +174,8 @@ impl std::fmt::Display for BlockingReasons {
     fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
         write!(
             f,
-            "tenant_blocked_by_lsn_lease_deadline: {}, {} timelines block for {:?}",
+            "{} timelines block for {:?}",
-            self.tenant_blocked_by_lsn_lease_deadline, self.timelines, self.reasons
+            self.timelines, self.reasons
         )
     }
 }
@@ -223,15 +183,13 @@ impl std::fmt::Display for BlockingReasons {
 impl BlockingReasons {
     fn clean_and_summarize(mut g: std::sync::MutexGuard<'_, Storage>) -> Option<Self> {
         let mut reasons = enumset::EnumSet::empty();
-        g.timelines_blocked.retain(|_key, value| {
+        g.retain(|_key, value| {
             reasons = reasons.union(*value);
             !value.is_empty()
         });
-        let blocked_by_lsn_lease_deadline = g.is_blocked_by_lsn_lease_deadline();
+        if !g.is_empty() {
-        if !g.timelines_blocked.is_empty() || blocked_by_lsn_lease_deadline {
             Some(BlockingReasons {
-                tenant_blocked_by_lsn_lease_deadline: blocked_by_lsn_lease_deadline,
+                timelines: g.len(),
-                timelines: g.timelines_blocked.len(),
                 reasons,
             })
         } else {
@@ -240,17 +198,14 @@ impl BlockingReasons {
     }
 
     fn summarize(g: &std::sync::MutexGuard<'_, Storage>) -> Option<Self> {
-        let blocked_by_lsn_lease_deadline = g.is_blocked_by_lsn_lease_deadline();
+        if g.is_empty() {
-        if g.timelines_blocked.is_empty() && !blocked_by_lsn_lease_deadline {
             None
         } else {
             let reasons = g
-                .timelines_blocked
                 .values()
                 .fold(enumset::EnumSet::empty(), |acc, next| acc.union(*next));
             Some(BlockingReasons {
-                tenant_blocked_by_lsn_lease_deadline: blocked_by_lsn_lease_deadline,
+                timelines: g.len(),
-                timelines: g.timelines_blocked.len(),
                 reasons,
             })
         }

pageserver/src/tenant/mgr.rs

@@ -949,12 +949,6 @@ impl TenantManager {
                 (LocationMode::Attached(attach_conf), Some(TenantSlot::Attached(tenant))) => {
                     match attach_conf.generation.cmp(&tenant.generation) {
                         Ordering::Equal => {
-                            if attach_conf.attach_mode == AttachmentMode::Single {
-                                tenant
-                                    .gc_block
-                                    .set_lsn_lease_deadline(tenant.get_lsn_lease_length());
-                            }
-
                             // A transition from Attached to Attached in the same generation, we may
                             // take our fast path and just provide the updated configuration
                             // to the tenant.

pageserver/src/tenant/storage_layer.rs

@@ -1,13 +1,13 @@
 //! Common traits and structs for layers
 
 pub mod delta_layer;
-pub mod filter_iterator;
 pub mod image_layer;
 pub mod inmemory_layer;
 pub(crate) mod layer;
 mod layer_desc;
 mod layer_name;
 pub mod merge_iterator;
+
 pub mod split_writer;
 
 use crate::context::{AccessStatsBehavior, RequestContext};
@@ -276,16 +276,6 @@ pub(crate) enum LayerId {
     InMemoryLayerId(InMemoryLayerFileId),
 }
 
-/// Uniquely identify a layer visit by the layer
-/// and LSN floor (or start LSN) of the reads.
-/// The layer itself is not enough since we may
-/// have different LSN lower bounds for delta layer reads.
-#[derive(Debug, PartialEq, Eq, Clone, Hash)]
-struct LayerToVisitId {
-    layer_id: LayerId,
-    lsn_floor: Lsn,
-}
-
 /// Layer wrapper for the read path. Note that it is valid
 /// to use these layers even after external operations have
 /// been performed on them (compaction, freeze, etc.).
@@ -297,9 +287,9 @@ pub(crate) enum ReadableLayer {
 
 /// A partial description of a read to be done.
 #[derive(Debug, Clone)]
-struct LayerVisit {
+struct ReadDesc {
     /// An id used to resolve the readable layer within the fringe
-    layer_to_visit_id: LayerToVisitId,
+    layer_id: LayerId,
     /// Lsn range for the read, used for selecting the next read
     lsn_range: Range<Lsn>,
 }
@@ -313,12 +303,12 @@ struct LayerVisit {
 /// a two layer indexing scheme.
 #[derive(Debug)]
 pub(crate) struct LayerFringe {
-    planned_visits_by_lsn: BinaryHeap<LayerVisit>,
+    planned_reads_by_lsn: BinaryHeap<ReadDesc>,
-    visit_reads: HashMap<LayerToVisitId, LayerVisitReads>,
+    layers: HashMap<LayerId, LayerKeyspace>,
 }
 
 #[derive(Debug)]
-struct LayerVisitReads {
+struct LayerKeyspace {
     layer: ReadableLayer,
     target_keyspace: KeySpaceRandomAccum,
 }
@@ -326,23 +316,23 @@ struct LayerVisitReads {
 impl LayerFringe {
     pub(crate) fn new() -> Self {
         LayerFringe {
-            planned_visits_by_lsn: BinaryHeap::new(),
+            planned_reads_by_lsn: BinaryHeap::new(),
-            visit_reads: HashMap::new(),
+            layers: HashMap::new(),
         }
     }
 
     pub(crate) fn next_layer(&mut self) -> Option<(ReadableLayer, KeySpace, Range<Lsn>)> {
-        let read_desc = match self.planned_visits_by_lsn.pop() {
+        let read_desc = match self.planned_reads_by_lsn.pop() {
             Some(desc) => desc,
             None => return None,
         };
 
-        let removed = self.visit_reads.remove_entry(&read_desc.layer_to_visit_id);
+        let removed = self.layers.remove_entry(&read_desc.layer_id);
 
         match removed {
             Some((
                 _,
-                LayerVisitReads {
+                LayerKeyspace {
                     layer,
                     mut target_keyspace,
                 },
@@ -361,24 +351,20 @@ impl LayerFringe {
         keyspace: KeySpace,
         lsn_range: Range<Lsn>,
     ) {
-        let layer_to_visit_id = LayerToVisitId {
+        let layer_id = layer.id();
-            layer_id: layer.id(),
+        let entry = self.layers.entry(layer_id.clone());
-            lsn_floor: lsn_range.start,
-        };
-
-        let entry = self.visit_reads.entry(layer_to_visit_id.clone());
         match entry {
             Entry::Occupied(mut entry) => {
                 entry.get_mut().target_keyspace.add_keyspace(keyspace);
             }
             Entry::Vacant(entry) => {
-                self.planned_visits_by_lsn.push(LayerVisit {
+                self.planned_reads_by_lsn.push(ReadDesc {
                     lsn_range,
-                    layer_to_visit_id: layer_to_visit_id.clone(),
+                    layer_id: layer_id.clone(),
                 });
                 let mut accum = KeySpaceRandomAccum::new();
                 accum.add_keyspace(keyspace);
-                entry.insert(LayerVisitReads {
+                entry.insert(LayerKeyspace {
                     layer,
                     target_keyspace: accum,
                 });
@@ -393,7 +379,7 @@ impl Default for LayerFringe {
     }
 }
 
-impl Ord for LayerVisit {
+impl Ord for ReadDesc {
     fn cmp(&self, other: &Self) -> Ordering {
         let ord = self.lsn_range.end.cmp(&other.lsn_range.end);
         if ord == std::cmp::Ordering::Equal {
@@ -404,19 +390,19 @@ impl Ord for LayerVisit {
     }
 }
 
-impl PartialOrd for LayerVisit {
+impl PartialOrd for ReadDesc {
     fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
         Some(self.cmp(other))
     }
 }
 
-impl PartialEq for LayerVisit {
+impl PartialEq for ReadDesc {
     fn eq(&self, other: &Self) -> bool {
         self.lsn_range == other.lsn_range
     }
 }
 
-impl Eq for LayerVisit {}
+impl Eq for ReadDesc {}
 
 impl ReadableLayer {
     pub(crate) fn id(&self) -> LayerId {

pageserver/src/tenant/storage_layer/delta_layer.rs

@@ -39,7 +39,7 @@ use crate::tenant::disk_btree::{
 use crate::tenant::storage_layer::layer::S3_UPLOAD_LIMIT;
 use crate::tenant::timeline::GetVectoredError;
 use crate::tenant::vectored_blob_io::{
-    BlobFlag, BufView, StreamingVectoredReadPlanner, VectoredBlobReader, VectoredRead,
+    BlobFlag, StreamingVectoredReadPlanner, VectoredBlobReader, VectoredRead,
     VectoredReadCoalesceMode, VectoredReadPlanner,
 };
 use crate::tenant::PageReconstructError;
@@ -1021,30 +1021,13 @@ impl DeltaLayerInner {
                     continue;
                 }
             };
-            let view = BufView::new_slice(&blobs_buf.buf);
+
             for meta in blobs_buf.blobs.iter().rev() {
                 if Some(meta.meta.key) == ignore_key_with_err {
                     continue;
                 }
-                let blob_read = meta.read(&view).await;
-                let blob_read = match blob_read {
-                    Ok(buf) => buf,
-                    Err(e) => {
-                        reconstruct_state.on_key_error(
-                            meta.meta.key,
-                            PageReconstructError::Other(anyhow!(e).context(format!(
-                                "Failed to decompress blob from virtual file {}",
-                                self.file.path,
-                            ))),
-                        );
-
-                        ignore_key_with_err = Some(meta.meta.key);
-                        continue;
-                    }
-                };
-
-                let value = Value::des(&blob_read);
 
+                let value = Value::des(&blobs_buf.buf[meta.start..meta.end]);
                 let value = match value {
                     Ok(v) => v,
                     Err(e) => {
@@ -1260,21 +1243,21 @@ impl DeltaLayerInner {
                 buf.reserve(read.size());
                 let res = reader.read_blobs(&read, buf, ctx).await?;
 
-                let view = BufView::new_slice(&res.buf);
-
                 for blob in res.blobs {
                     let key = blob.meta.key;
                     let lsn = blob.meta.lsn;
-
+                    let data = &res.buf[blob.start..blob.end];
-                    let data = blob.read(&view).await?;
 
                     #[cfg(debug_assertions)]
-                    Value::des(&data)
+                    Value::des(data)
                         .with_context(|| {
                             format!(
-                                "blob failed to deserialize for {}: {:?}",
+                                "blob failed to deserialize for {}@{}, {}..{}: {:?}",
-                                blob,
+                                blob.meta.key,
-                                utils::Hex(&data)
+                                blob.meta.lsn,
+                                blob.start,
+                                blob.end,
+                                utils::Hex(data)
                             )
                         })
                         .unwrap();
@@ -1282,15 +1265,15 @@ impl DeltaLayerInner {
                     // is it an image or will_init walrecord?
                     // FIXME: this could be handled by threading the BlobRef to the
                     // VectoredReadBuilder
-                    let will_init = crate::repository::ValueBytes::will_init(&data)
+                    let will_init = crate::repository::ValueBytes::will_init(data)
                         .inspect_err(|_e| {
                             #[cfg(feature = "testing")]
-                            tracing::error!(data=?utils::Hex(&data), err=?_e, %key, %lsn, "failed to parse will_init out of serialized value");
+                            tracing::error!(data=?utils::Hex(data), err=?_e, %key, %lsn, "failed to parse will_init out of serialized value");
                         })
                         .unwrap_or(false);
 
                     per_blob_copy.clear();
-                    per_blob_copy.extend_from_slice(&data);
+                    per_blob_copy.extend_from_slice(data);
 
                     let (tmp, res) = writer
                         .put_value_bytes(
@@ -1555,11 +1538,8 @@ impl<'a> DeltaLayerIterator<'a> {
             .read_blobs(&plan, buf, self.ctx)
             .await?;
         let frozen_buf = blobs_buf.buf.freeze();
-        let view = BufView::new_bytes(frozen_buf);
         for meta in blobs_buf.blobs.iter() {
-            let blob_read = meta.read(&view).await?;
+            let value = Value::des(&frozen_buf[meta.start..meta.end])?;
-            let value = Value::des(&blob_read)?;
-
             next_batch.push_back((meta.meta.key, meta.meta.lsn, value));
         }
         self.key_values_batch = next_batch;
@@ -1936,13 +1916,9 @@ pub(crate) mod test {
                 let blobs_buf = vectored_blob_reader
                     .read_blobs(&read, buf.take().expect("Should have a buffer"), &ctx)
                     .await?;
-                let view = BufView::new_slice(&blobs_buf.buf);
                 for meta in blobs_buf.blobs.iter() {
-                    let value = meta.read(&view).await?;
+                    let value = &blobs_buf.buf[meta.start..meta.end];
-                    assert_eq!(
+                    assert_eq!(value, entries_meta.index[&(meta.meta.key, meta.meta.lsn)]);
-                        &value[..],
-                        &entries_meta.index[&(meta.meta.key, meta.meta.lsn)]
-                    );
                 }
 
                 buf = Some(blobs_buf.buf);

pageserver/src/tenant/storage_layer/filter_iterator.rs

@@ -1,205 +0,0 @@
-use std::ops::Range;
-
-use anyhow::bail;
-use pageserver_api::{
-    key::Key,
-    keyspace::{KeySpace, SparseKeySpace},
-};
-use utils::lsn::Lsn;
-
-use crate::repository::Value;
-
-use super::merge_iterator::MergeIterator;
-
-/// A filter iterator over merge iterators (and can be easily extended to other types of iterators).
-///
-/// The iterator will skip any keys not included in the keyspace filter. In other words, the keyspace filter contains the keys
-/// to be retained.
-pub struct FilterIterator<'a> {
-    inner: MergeIterator<'a>,
-    retain_key_filters: Vec<Range<Key>>,
-    current_filter_idx: usize,
-}
-
-impl<'a> FilterIterator<'a> {
-    pub fn create(
-        inner: MergeIterator<'a>,
-        dense_keyspace: KeySpace,
-        sparse_keyspace: SparseKeySpace,
-    ) -> anyhow::Result<Self> {
-        let mut retain_key_filters = Vec::new();
-        retain_key_filters.extend(dense_keyspace.ranges);
-        retain_key_filters.extend(sparse_keyspace.0.ranges);
-        retain_key_filters.sort_by(|a, b| a.start.cmp(&b.start));
-        // Verify key filters are non-overlapping and sorted
-        for window in retain_key_filters.windows(2) {
-            if window[0].end > window[1].start {
-                bail!(
-                    "Key filters are overlapping: {:?} and {:?}",
-                    window[0],
-                    window[1]
-                );
-            }
-        }
-        Ok(Self {
-            inner,
-            retain_key_filters,
-            current_filter_idx: 0,
-        })
-    }
-
-    pub async fn next(&mut self) -> anyhow::Result<Option<(Key, Lsn, Value)>> {
-        while let Some(item) = self.inner.next().await? {
-            while self.current_filter_idx < self.retain_key_filters.len()
-                && item.0 >= self.retain_key_filters[self.current_filter_idx].end
-            {
-                // [filter region]    [filter region]     [filter region]
-                //                                     ^ item
-                //                    ^ current filter
-                self.current_filter_idx += 1;
-                // [filter region]    [filter region]     [filter region]
-                //                                     ^ item
-                //                                        ^ current filter
-            }
-            if self.current_filter_idx >= self.retain_key_filters.len() {
-                // We already exhausted all filters, so we should return now
-                // [filter region] [filter region] [filter region]
-                //                                                    ^ item
-                //                                                 ^ current filter (nothing)
-                return Ok(None);
-            }
-            if self.retain_key_filters[self.current_filter_idx].contains(&item.0) {
-                // [filter region]    [filter region]     [filter region]
-                //                                              ^ item
-                //                                        ^ current filter
-                return Ok(Some(item));
-            }
-            // If the key is not contained in the key retaining filters, continue to the next item.
-            // [filter region]    [filter region]     [filter region]
-            //                                     ^ item
-            //                                        ^ current filter
-        }
-        Ok(None)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    use itertools::Itertools;
-    use pageserver_api::key::Key;
-    use utils::lsn::Lsn;
-
-    use crate::{
-        tenant::{
-            harness::{TenantHarness, TIMELINE_ID},
-            storage_layer::delta_layer::test::produce_delta_layer,
-        },
-        DEFAULT_PG_VERSION,
-    };
-
-    async fn assert_filter_iter_equal(
-        filter_iter: &mut FilterIterator<'_>,
-        expect: &[(Key, Lsn, Value)],
-    ) {
-        let mut expect_iter = expect.iter();
-        loop {
-            let o1 = filter_iter.next().await.unwrap();
-            let o2 = expect_iter.next();
-            assert_eq!(o1.is_some(), o2.is_some());
-            if o1.is_none() && o2.is_none() {
-                break;
-            }
-            let (k1, l1, v1) = o1.unwrap();
-            let (k2, l2, v2) = o2.unwrap();
-            assert_eq!(&k1, k2);
-            assert_eq!(l1, *l2);
-            assert_eq!(&v1, v2);
-        }
-    }
-
-    #[tokio::test]
-    async fn filter_keyspace_iterator() {
-        use crate::repository::Value;
-        use bytes::Bytes;
-
-        let harness = TenantHarness::create("filter_iterator_filter_keyspace_iterator")
-            .await
-            .unwrap();
-        let (tenant, ctx) = harness.load().await;
-
-        let tline = tenant
-            .create_test_timeline(TIMELINE_ID, Lsn(0x10), DEFAULT_PG_VERSION, &ctx)
-            .await
-            .unwrap();
-
-        fn get_key(id: u32) -> Key {
-            let mut key = Key::from_hex("000000000033333333444444445500000000").unwrap();
-            key.field6 = id;
-            key
-        }
-        const N: usize = 100;
-        let test_deltas1 = (0..N)
-            .map(|idx| {
-                (
-                    get_key(idx as u32),
-                    Lsn(0x20 * ((idx as u64) % 10 + 1)),
-                    Value::Image(Bytes::from(format!("img{idx:05}"))),
-                )
-            })
-            .collect_vec();
-        let resident_layer_1 = produce_delta_layer(&tenant, &tline, test_deltas1.clone(), &ctx)
-            .await
-            .unwrap();
-
-        let merge_iter = MergeIterator::create(
-            &[resident_layer_1.get_as_delta(&ctx).await.unwrap()],
-            &[],
-            &ctx,
-        );
-
-        let mut filter_iter = FilterIterator::create(
-            merge_iter,
-            KeySpace {
-                ranges: vec![
-                    get_key(5)..get_key(10),
-                    get_key(20)..get_key(30),
-                    get_key(90)..get_key(110),
-                    get_key(1000)..get_key(2000),
-                ],
-            },
-            SparseKeySpace(KeySpace::default()),
-        )
-        .unwrap();
-        let mut result = Vec::new();
-        result.extend(test_deltas1[5..10].iter().cloned());
-        result.extend(test_deltas1[20..30].iter().cloned());
-        result.extend(test_deltas1[90..100].iter().cloned());
-        assert_filter_iter_equal(&mut filter_iter, &result).await;
-
-        let merge_iter = MergeIterator::create(
-            &[resident_layer_1.get_as_delta(&ctx).await.unwrap()],
-            &[],
-            &ctx,
-        );
-
-        let mut filter_iter = FilterIterator::create(
-            merge_iter,
-            KeySpace {
-                ranges: vec![
-                    get_key(0)..get_key(10),
-                    get_key(20)..get_key(30),
-                    get_key(90)..get_key(95),
-                ],
-            },
-            SparseKeySpace(KeySpace::default()),
-        )
-        .unwrap();
-        let mut result = Vec::new();
-        result.extend(test_deltas1[0..10].iter().cloned());
-        result.extend(test_deltas1[20..30].iter().cloned());
-        result.extend(test_deltas1[90..95].iter().cloned());
-        assert_filter_iter_equal(&mut filter_iter, &result).await;
-    }
-}