slightly more efficient

This reverts commit ae463f366b.

.github/actionlint.yml

@@ -28,7 +28,3 @@ config-variables:
   - DEV_AWS_OIDC_ROLE_MANAGE_BENCHMARK_EC2_VMS_ARN
   - SLACK_ON_CALL_STORAGE_STAGING_STREAM
   - SLACK_CICD_CHANNEL_ID
-  - SLACK_STORAGE_CHANNEL_ID
-  - NEON_DEV_AWS_ACCOUNT_ID
-  - NEON_PROD_AWS_ACCOUNT_ID
-  - AWS_ECR_REGION

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

@@ -19,11 +19,7 @@ inputs:
     default: '[1, 1]'
   # settings below only needed if you want the project to be sharded from the beginning
   shard_split_project:
-    description: 'by default new projects are not shard-split initiailly, but only when shard-split threshold is reached, specify true to explicitly shard-split initially'
-    required: false
-    default: 'false'
-  disable_sharding:
-    description: 'by default new projects use storage controller default policy to shard-split when shard-split threshold is reached, specify true to explicitly disable sharding'
+    description: 'by default new projects are not shard-split, specify true to shard-split'
     required: false
     default: 'false'
   admin_api_key:
@@ -111,21 +107,6 @@ runs:
             -H "Accept: application/json" -H "Content-Type: application/json" -H "Authorization: Bearer ${ADMIN_API_KEY}" \
             -d "{\"new_shard_count\": $SHARD_COUNT, \"new_stripe_size\": $STRIPE_SIZE}"
         fi
-        if [ "${DISABLE_SHARDING}" = "true" ]; then
-          # determine tenant ID
-          TENANT_ID=`${PSQL} ${dsn} -t -A -c "SHOW neon.tenant_id"`
-
-          echo "Explicitly disabling shard-splitting for project ${project_id} with tenant_id ${TENANT_ID}"
-
-          echo "Sending PUT request to https://${API_HOST}/regions/${REGION_ID}/api/v1/admin/storage/proxy/control/v1/tenant/${TENANT_ID}/policy"
-          echo "with body {\"scheduling\": \"Essential\"}"
-
-          # we need an ADMIN API KEY to invoke storage controller API for shard splitting (bash -u above checks that the variable is set)
-          curl -X PUT \
-            "https://${API_HOST}/regions/${REGION_ID}/api/v1/admin/storage/proxy/control/v1/tenant/${TENANT_ID}/policy" \
-            -H "Accept: application/json" -H "Content-Type: application/json" -H "Authorization: Bearer ${ADMIN_API_KEY}" \
-            -d "{\"scheduling\": \"Essential\"}"
-        fi
 
       env:
         API_HOST: ${{ inputs.api_host }}
@@ -135,7 +116,6 @@ runs:
         MIN_CU: ${{ fromJSON(inputs.compute_units)[0] }}
         MAX_CU: ${{ fromJSON(inputs.compute_units)[1] }}
         SHARD_SPLIT_PROJECT: ${{ inputs.shard_split_project }}
-        DISABLE_SHARDING: ${{ inputs.disable_sharding }}
         ADMIN_API_KEY: ${{ inputs.admin_api_key }}
         SHARD_COUNT: ${{ inputs.shard_count }}
         STRIPE_SIZE: ${{ inputs.stripe_size }}

.github/workflows/_push-to-container-registry.yml

@@ -2,7 +2,7 @@ name: Push images to Container Registry
 on:
   workflow_call:
     inputs:
-      # Example: {"docker.io/neondatabase/neon:13196061314":["${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/neon:13196061314","neoneastus2.azurecr.io/neondatabase/neon:13196061314"]}
+      # Example: {"docker.io/neondatabase/neon:13196061314":["369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:13196061314","neoneastus2.azurecr.io/neondatabase/neon:13196061314"]}
       image-map:
         description: JSON map of images, mapping from a source image to an array of target images that should be pushed.
         required: true

.github/workflows/build_and_test.yml

@@ -68,7 +68,7 @@ jobs:
   tag:
     needs: [ check-permissions ]
     runs-on: [ self-hosted, small ]
-    container: ${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/base:pinned
+    container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/base:pinned
     outputs:
       build-tag: ${{steps.build-tag.outputs.tag}}
 
@@ -859,17 +859,14 @@ jobs:
           BRANCH: "${{ github.ref_name }}"
           DEV_ACR: "${{ vars.AZURE_DEV_REGISTRY_NAME }}"
           PROD_ACR: "${{ vars.AZURE_PROD_REGISTRY_NAME }}"
-          DEV_AWS: "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}"
-          PROD_AWS: "${{ vars.NEON_PROD_AWS_ACCOUNT_ID }}"
-          AWS_REGION: "${{ vars.AWS_ECR_REGION }}"
 
   push-neon-image-dev:
     needs: [ generate-image-maps, neon-image ]
     uses: ./.github/workflows/_push-to-container-registry.yml
     with:
       image-map: '${{ needs.generate-image-maps.outputs.neon-dev }}'
-      aws-region: ${{ vars.AWS_ECR_REGION }}
-      aws-account-ids: "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}"
+      aws-region: eu-central-1
+      aws-account-ids: "369495373322"
       azure-client-id: ${{ vars.AZURE_DEV_CLIENT_ID }}
       azure-subscription-id: ${{ vars.AZURE_DEV_SUBSCRIPTION_ID }}
       azure-tenant-id: ${{ vars.AZURE_TENANT_ID }}
@@ -884,8 +881,8 @@ jobs:
     uses: ./.github/workflows/_push-to-container-registry.yml
     with:
       image-map: '${{ needs.generate-image-maps.outputs.compute-dev }}'
-      aws-region: ${{ vars.AWS_ECR_REGION }}
-      aws-account-ids: "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}"
+      aws-region: eu-central-1
+      aws-account-ids: "369495373322"
       azure-client-id: ${{ vars.AZURE_DEV_CLIENT_ID }}
       azure-subscription-id: ${{ vars.AZURE_DEV_SUBSCRIPTION_ID }}
       azure-tenant-id: ${{ vars.AZURE_TENANT_ID }}
@@ -901,8 +898,8 @@ jobs:
     uses: ./.github/workflows/_push-to-container-registry.yml
     with:
       image-map: '${{ needs.generate-image-maps.outputs.neon-prod }}'
-      aws-region: ${{ vars.AWS_ECR_REGION }}
-      aws-account-ids: "${{ vars.NEON_PROD_AWS_ACCOUNT_ID }}"
+      aws-region: eu-central-1
+      aws-account-ids: "093970136003"
       azure-client-id: ${{ vars.AZURE_PROD_CLIENT_ID }}
       azure-subscription-id: ${{ vars.AZURE_PROD_SUBSCRIPTION_ID }}
       azure-tenant-id: ${{ vars.AZURE_TENANT_ID }}
@@ -918,8 +915,8 @@ jobs:
     uses: ./.github/workflows/_push-to-container-registry.yml
     with:
       image-map: '${{ needs.generate-image-maps.outputs.compute-prod }}'
-      aws-region: ${{ vars.AWS_ECR_REGION }}
-      aws-account-ids: "${{ vars.NEON_PROD_AWS_ACCOUNT_ID }}"
+      aws-region: eu-central-1
+      aws-account-ids: "093970136003"
       azure-client-id: ${{ vars.AZURE_PROD_CLIENT_ID }}
       azure-subscription-id: ${{ vars.AZURE_PROD_SUBSCRIPTION_ID }}
       azure-tenant-id: ${{ vars.AZURE_TENANT_ID }}
@@ -1032,7 +1029,7 @@ jobs:
       statuses: write
       contents: write
     runs-on: [ self-hosted, small ]
-    container: ${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/ansible:latest
+    container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/ansible:latest
     steps:
       - uses: actions/checkout@v4
 
@@ -1181,22 +1178,6 @@ jobs:
             exit 1
           fi
 
-  notify-storage-release-deploy-failure:
-    needs: [ deploy ]
-    # We want this to run even if (transitive) dependencies are skipped, because deploy should really be successful on release branch workflow runs.
-    if: github.ref_name == 'release' && needs.deploy.result != 'success' && always()
-    runs-on: ubuntu-22.04
-    steps:
-      - name: Post release-deploy failure to team-storage slack channel
-        uses: slackapi/slack-github-action@v2
-        with:
-          method: chat.postMessage
-          token: ${{ secrets.SLACK_BOT_TOKEN }}
-          payload: |
-            channel: ${{ vars.SLACK_STORAGE_CHANNEL_ID }}
-            text: |
-              🔴 @oncall-storage: deploy job on release branch had unexpected status "${{ needs.deploy.result }}" <${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>.
-
   # 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 ]
@@ -1293,7 +1274,7 @@ jobs:
           done
 
   pin-build-tools-image:
-    needs: [ build-build-tools-image, test-images, build-and-test-locally ]
+    needs: [ build-build-tools-image, push-compute-image-prod, push-neon-image-prod, build-and-test-locally ]
     if: github.ref_name == 'main'
     uses: ./.github/workflows/pin-build-tools-image.yml
     with:

.github/workflows/build_and_test_with_sanitizers.yml

@@ -27,7 +27,7 @@ env:
 jobs:
   tag:
     runs-on: [ self-hosted, small ]
-    container: ${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/base:pinned
+    container: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/base:pinned
     outputs:
       build-tag: ${{steps.build-tag.outputs.tag}}
 

.github/workflows/ingest_benchmark.yml

@@ -32,27 +32,18 @@ jobs:
           - target_project: new_empty_project_stripe_size_2048 
             stripe_size: 2048 # 16 MiB
             postgres_version: 16
-            disable_sharding: false
           - target_project: new_empty_project_stripe_size_32768
             stripe_size: 32768 # 256 MiB # note that this is different from null because using null will shard_split the project only if it reaches the threshold
                                # while here it is sharded from the beginning with a shard size of 256 MiB
-            disable_sharding: false
             postgres_version: 16
           - target_project: new_empty_project
             stripe_size: null # run with neon defaults which will shard split only when reaching the threshold
-            disable_sharding: false
             postgres_version: 16
           - target_project: new_empty_project
             stripe_size: null # run with neon defaults which will shard split only when reaching the threshold
-            disable_sharding: false
             postgres_version: 17
           - target_project: large_existing_project
             stripe_size: null # cannot re-shared or choose different stripe size for existing, already sharded project
-            disable_sharding: false
-            postgres_version: 16
-          - target_project: new_empty_project_unsharded
-            stripe_size: null # run with neon defaults which will shard split only when reaching the threshold
-            disable_sharding: true
             postgres_version: 16
       max-parallel: 1 # we want to run each stripe size sequentially to be able to compare the results
     permissions:
@@ -105,7 +96,6 @@ jobs:
         admin_api_key: ${{ secrets.NEON_STAGING_ADMIN_API_KEY }} 
         shard_count: 8
         stripe_size: ${{ matrix.stripe_size }}
-        disable_sharding: ${{ matrix.disable_sharding }} 
 
     - name: Initialize Neon project
       if: ${{ startsWith(matrix.target_project, 'new_empty_project') }}

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

@@ -33,6 +33,10 @@ concurrency:
 # No permission for GITHUB_TOKEN by default; the **minimal required** set of permissions should be granted in each job.
 permissions: {}
 
+env:
+  FROM_TAG: ${{ inputs.from-tag }}
+  TO_TAG: pinned
+
 jobs:
   check-manifests:
     runs-on: ubuntu-22.04
@@ -42,14 +46,11 @@ jobs:
     steps:
       - name: Check if we really need to pin the image
         id: check-manifests
-        env:
-          FROM_TAG: ${{ inputs.from-tag }}
-          TO_TAG: pinned
         run: |
-          docker manifest inspect "docker.io/neondatabase/build-tools:${FROM_TAG}" > "${FROM_TAG}.json"
-          docker manifest inspect "docker.io/neondatabase/build-tools:${TO_TAG}"   > "${TO_TAG}.json"
+          docker manifest inspect neondatabase/build-tools:${FROM_TAG} > ${FROM_TAG}.json
+          docker manifest inspect neondatabase/build-tools:${TO_TAG}   > ${TO_TAG}.json
 
-          if diff "${FROM_TAG}.json" "${TO_TAG}.json"; then
+          if diff ${FROM_TAG}.json ${TO_TAG}.json; then
             skip=true
           else
             skip=false
@@ -63,34 +64,55 @@ jobs:
     # use format(..) to catch both inputs.force = true AND inputs.force = 'true'
     if: needs.check-manifests.outputs.skip == 'false' || format('{0}', inputs.force) == 'true'
 
-    permissions:
-      id-token: write  # Required for aws/azure login
+    runs-on: ubuntu-22.04
 
-    uses: ./.github/workflows/_push-to-container-registry.yml
-    with:
-      image-map: |
-        {
-          "docker.io/neondatabase/build-tools:${{ inputs.from-tag }}-bullseye": [
-            "docker.io/neondatabase/build-tools:pinned-bullseye",
-            "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/build-tools:pinned-bullseye",
-            "${{ vars.AZURE_DEV_REGISTRY_NAME }}.azurecr.io/neondatabase/build-tools:pinned-bullseye"
-          ],
-          "docker.io/neondatabase/build-tools:${{ inputs.from-tag }}-bookworm": [
-            "docker.io/neondatabase/build-tools:pinned-bookworm",
-            "docker.io/neondatabase/build-tools:pinned",
-            "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/build-tools:pinned-bookworm",
-            "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/build-tools:pinned",
-            "${{ vars.AZURE_DEV_REGISTRY_NAME }}.azurecr.io/neondatabase/build-tools:pinned-bookworm",
-            "${{ vars.AZURE_DEV_REGISTRY_NAME }}.azurecr.io/neondatabase/build-tools:pinned"
-          ]
-        }
-      aws-region: ${{ vars.AWS_ECR_REGION }}
-      aws-account-ids: "${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}"
-      azure-client-id: ${{ vars.AZURE_DEV_CLIENT_ID }}
-      azure-subscription-id: ${{ vars.AZURE_DEV_SUBSCRIPTION_ID }}
-      azure-tenant-id: ${{ vars.AZURE_TENANT_ID }}
-      acr-registry-name: ${{ vars.AZURE_DEV_REGISTRY_NAME }}
-    secrets:
-      aws-role-to-assume: "${{ vars.DEV_AWS_OIDC_ROLE_ARN }}"
-      docker-hub-username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
-      docker-hub-password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
+    permissions:
+      id-token: write # for `azure/login` and aws auth
+
+    steps:
+      - uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
+          password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
+
+      - name: Configure AWS credentials
+        uses: aws-actions/configure-aws-credentials@v4
+        with:
+          aws-region: eu-central-1
+          role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
+          role-duration-seconds: 3600
+
+      - name: Login to Amazon Dev ECR
+        uses: aws-actions/amazon-ecr-login@v2
+
+      - name: Azure login
+        uses: azure/login@6c251865b4e6290e7b78be643ea2d005bc51f69a  # @v2.1.1
+        with:
+          client-id: ${{ secrets.AZURE_DEV_CLIENT_ID }}
+          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
+          subscription-id: ${{ secrets.AZURE_DEV_SUBSCRIPTION_ID }}
+
+      - name: Login to ACR
+        run: |
+          az acr login --name=neoneastus2
+
+      - name: Tag build-tools with `${{ env.TO_TAG }}` in Docker Hub, ECR, and ACR
+        env:
+          DEFAULT_DEBIAN_VERSION: bookworm
+        run: |
+          for debian_version in bullseye bookworm; do
+            tags=()
+
+            tags+=("-t" "neondatabase/build-tools:${TO_TAG}-${debian_version}")
+            tags+=("-t" "369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:${TO_TAG}-${debian_version}")
+            tags+=("-t" "neoneastus2.azurecr.io/neondatabase/build-tools:${TO_TAG}-${debian_version}")
+
+            if [ "${debian_version}" == "${DEFAULT_DEBIAN_VERSION}" ]; then
+              tags+=("-t" "neondatabase/build-tools:${TO_TAG}")
+              tags+=("-t" "369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:${TO_TAG}")
+              tags+=("-t" "neoneastus2.azurecr.io/neondatabase/build-tools:${TO_TAG}")
+            fi
+
+            docker buildx imagetools create "${tags[@]}" \
+                                              neondatabase/build-tools:${FROM_TAG}-${debian_version}
+          done

Cargo.lock

@@ -1316,6 +1316,7 @@ dependencies = [
  "flate2",
  "futures",
  "http 1.1.0",
+ "jsonwebtoken",
  "metrics",
  "nix 0.27.1",
  "notify",
@@ -1325,6 +1326,7 @@ dependencies = [
  "opentelemetry_sdk",
  "postgres",
  "postgres_initdb",
+ "prometheus",
  "regex",
  "remote_storage",
  "reqwest",
@@ -1343,6 +1345,7 @@ dependencies = [
  "tower 0.5.2",
  "tower-http",
  "tracing",
+ "tracing-opentelemetry",
  "tracing-subscriber",
  "tracing-utils",
  "url",
@@ -1874,12 +1877,6 @@ dependencies = [
  "syn 2.0.90",
 ]
 
-[[package]]
-name = "difflib"
-version = "0.4.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6184e33543162437515c2e2b48714794e37845ec9851711914eec9d308f6ebe8"
-
 [[package]]
 name = "digest"
 version = "0.10.7"
@@ -3337,17 +3334,6 @@ dependencies = [
  "wasm-bindgen",
 ]
 
-[[package]]
-name = "json-structural-diff"
-version = "0.2.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e878e36a8a44c158505c2c818abdc1350413ad83dcb774a0459f6a7ef2b65cbf"
-dependencies = [
- "difflib",
- "regex",
- "serde_json",
-]
-
 [[package]]
 name = "jsonwebtoken"
 version = "9.2.0"
@@ -6460,7 +6446,6 @@ dependencies = [
  "humantime",
  "hyper 0.14.30",
  "itertools 0.10.5",
- "json-structural-diff",
  "lasso",
  "measured",
  "metrics",
@@ -6483,7 +6468,6 @@ dependencies = [
  "strum",
  "strum_macros",
  "thiserror 1.0.69",
- "tikv-jemallocator",
  "tokio",
  "tokio-postgres",
  "tokio-postgres-rustls",
@@ -7037,11 +7021,14 @@ dependencies = [
 name = "tokio-postgres2"
 version = "0.1.0"
 dependencies = [
+ "async-trait",
+ "byteorder",
  "bytes",
  "fallible-iterator",
  "futures-util",
  "log",
  "parking_lot 0.12.1",
+ "percent-encoding",
  "phf",
  "pin-project-lite",
  "postgres-protocol2",
@@ -7628,13 +7615,13 @@ dependencies = [
  "hex",
  "hex-literal",
  "humantime",
+ "inferno 0.12.0",
  "jsonwebtoken",
  "metrics",
  "nix 0.27.1",
  "once_cell",
  "pin-project-lite",
  "postgres_connection",
- "pprof",
  "pq_proto",
  "rand 0.8.5",
  "regex",

Cargo.toml

@@ -210,7 +210,6 @@ rustls-native-certs = "0.8"
 x509-parser = "0.16"
 whoami = "1.5.1"
 zerocopy = { version = "0.7", features = ["derive"] }
-json-structural-diff = { version = "0.2.0" }
 
 ## TODO replace this with tracing
 env_logger = "0.10"

build-tools.Dockerfile

@@ -292,7 +292,7 @@ WORKDIR /home/nonroot
 
 # Rust
 # Please keep the version of llvm (installed above) in sync with rust llvm (`rustc --version --verbose | grep LLVM`)
-ENV RUSTC_VERSION=1.85.0
+ENV RUSTC_VERSION=1.84.1
 ENV RUSTUP_HOME="/home/nonroot/.rustup"
 ENV PATH="/home/nonroot/.cargo/bin:${PATH}"
 ARG RUSTFILT_VERSION=0.2.1

compute/compute-node.Dockerfile

@@ -395,22 +395,15 @@ RUN case "${PG_VERSION:?}" in \
     cd plv8-src && \
     if [[ "${PG_VERSION:?}" < "v17" ]]; then patch -p1 < /ext-src/plv8-3.1.10.patch; fi
 
-# Step 1: Build the vendored V8 engine. It doesn't depend on PostgreSQL, so use
-# 'build-deps' as the base. This enables caching and avoids unnecessary rebuilds.
-# (The V8 engine takes a very long time to build)
-FROM build-deps AS plv8-build
+FROM pg-build AS plv8-build
 ARG PG_VERSION
-WORKDIR /ext-src/plv8-src
 RUN apt update && \
     apt install --no-install-recommends --no-install-suggests -y \
     ninja-build python3-dev libncurses5 binutils clang \
     && apt clean && rm -rf /var/lib/apt/lists/*
-COPY --from=plv8-src /ext-src/ /ext-src/
-RUN make DOCKER=1 -j $(getconf _NPROCESSORS_ONLN) v8
 
-# Step 2: Build the PostgreSQL-dependent parts
-COPY --from=pg-build /usr/local/pgsql /usr/local/pgsql
-ENV PATH="/usr/local/pgsql/bin:$PATH"
+COPY --from=plv8-src /ext-src/ /ext-src/
+WORKDIR /ext-src/plv8-src
 RUN \
     # generate and copy upgrade scripts
     make generate_upgrades && \
@@ -1516,73 +1509,6 @@ WORKDIR /ext-src/pg_repack-src
 RUN make -j $(getconf _NPROCESSORS_ONLN) && \
     make -j $(getconf _NPROCESSORS_ONLN) install
 
-
-#########################################################################################
-#
-# Layer "pgaudit"
-# compile pgaudit extension
-#
-#########################################################################################
-
-FROM build-deps AS pgaudit-src
-ARG PG_VERSION
-WORKDIR /ext-src
-RUN case "${PG_VERSION}" in \
-    "v14") \
-    export PGAUDIT_VERSION=1.6.2 \
-    export PGAUDIT_CHECKSUM=1f350d70a0cbf488c0f2b485e3a5c9b11f78ad9e3cbb95ef6904afa1eb3187eb \
-    ;; \
-    "v15") \
-    export PGAUDIT_VERSION=1.7.0 \
-    export PGAUDIT_CHECKSUM=8f4a73e451c88c567e516e6cba7dc1e23bc91686bb6f1f77f8f3126d428a8bd8 \
-    ;; \
-    "v16") \
-    export PGAUDIT_VERSION=16.0 \
-    export PGAUDIT_CHECKSUM=d53ef985f2d0b15ba25c512c4ce967dce07b94fd4422c95bd04c4c1a055fe738 \
-    ;; \
-    "v17") \
-    export PGAUDIT_VERSION=17.0 \
-    export PGAUDIT_CHECKSUM=7d0d08d030275d525f36cd48b38c6455f1023da863385badff0cec44965bfd8c \
-    ;; \
-    *) \
-    echo "pgaudit is not supported on this PostgreSQL version" && exit 1;; \
-    esac && \
-    wget https://github.com/pgaudit/pgaudit/archive/refs/tags/${PGAUDIT_VERSION}.tar.gz -O pgaudit.tar.gz && \
-    echo "${PGAUDIT_CHECKSUM} pgaudit.tar.gz" | sha256sum --check && \
-    mkdir pgaudit-src && cd pgaudit-src && tar xzf ../pgaudit.tar.gz --strip-components=1 -C .
-
-FROM pg-build AS pgaudit-build
-COPY --from=pgaudit-src /ext-src/ /ext-src/
-WORKDIR /ext-src/pgaudit-src
-RUN make install USE_PGXS=1 -j $(getconf _NPROCESSORS_ONLN)
-
-#########################################################################################
-#
-# Layer "pgauditlogtofile"
-# compile pgauditlogtofile extension
-#
-#########################################################################################
-
-FROM build-deps AS pgauditlogtofile-src
-ARG PG_VERSION
-WORKDIR /ext-src
-RUN case "${PG_VERSION}" in \
-    "v14" | "v15" | "v16" | "v17") \
-    export PGAUDITLOGTOFILE_VERSION=v1.6.4 \
-    export PGAUDITLOGTOFILE_CHECKSUM=ef801eb09c26aaa935c0dabd92c81eb9ebe338930daa9674d420a280c6bc2d70 \
-    ;; \
-    *) \
-    echo "pgauditlogtofile is not supported on this PostgreSQL version" && exit 1;; \
-    esac && \
-    wget https://github.com/fmbiete/pgauditlogtofile/archive/refs/tags/${PGAUDITLOGTOFILE_VERSION}.tar.gz -O pgauditlogtofile.tar.gz && \
-    echo "${PGAUDITLOGTOFILE_CHECKSUM} pgauditlogtofile.tar.gz" | sha256sum --check && \
-    mkdir pgauditlogtofile-src && cd pgauditlogtofile-src && tar xzf ../pgauditlogtofile.tar.gz --strip-components=1 -C .
-
-FROM pg-build AS pgauditlogtofile-build
-COPY --from=pgauditlogtofile-src /ext-src/ /ext-src/
-WORKDIR /ext-src/pgauditlogtofile-src
-RUN make install USE_PGXS=1 -j $(getconf _NPROCESSORS_ONLN)
-
 #########################################################################################
 #
 # Layer "neon-ext-build"
@@ -1676,14 +1602,8 @@ COPY --from=pg_anon-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg_ivm-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg_partman-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg_mooncake-build /usr/local/pgsql/ /usr/local/pgsql/
-
-# Disabled temporarily, because it clashed with pg_mooncake. pg_mooncake
-# also depends on libduckdb, but a different version.
-#COPY --from=pg_duckdb-build /usr/local/pgsql/ /usr/local/pgsql/
-
+COPY --from=pg_duckdb-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg_repack-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY --from=pgaudit-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY --from=pgauditlogtofile-build /usr/local/pgsql/ /usr/local/pgsql/
 
 #########################################################################################
 #
@@ -1855,20 +1775,14 @@ COPY --from=pg_semver-src /ext-src/ /ext-src/
 COPY --from=pg_ivm-src /ext-src/ /ext-src/
 COPY --from=pg_partman-src /ext-src/ /ext-src/
 #COPY --from=pg_mooncake-src /ext-src/ /ext-src/
-COPY --from=pg_repack-src /ext-src/ /ext-src/
-COPY --from=pg_repack-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY compute/patches/pg_repack.patch /ext-src
-RUN cd /ext-src/pg_repack-src && patch -p1 </ext-src/pg_repack.patch && rm -f /ext-src/pg_repack.patch
+#COPY --from=pg_repack-src /ext-src/ /ext-src/
 
 COPY --chmod=755 docker-compose/run-tests.sh /run-tests.sh
-RUN apt-get update && apt-get install -y libtap-parser-sourcehandler-pgtap-perl\
-   && apt clean && rm -rf /ext-src/*.tar.gz /var/lib/apt/lists/*
 ENV PATH=/usr/local/pgsql/bin:$PATH
 ENV PGHOST=compute
 ENV PGPORT=55433
 ENV PGUSER=cloud_admin
 ENV PGDATABASE=postgres
-ENV PG_VERSION=${PG_VERSION:?}
 
 #########################################################################################
 #

compute/patches/pg_repack.patch

@@ -1,72 +0,0 @@
-diff --git a/regress/Makefile b/regress/Makefile
-index bf6edcb..89b4c7f 100644
---- a/regress/Makefile
-+++ b/regress/Makefile
-@@ -17,7 +17,7 @@ INTVERSION := $(shell echo $$(($$(echo $(VERSION).0 | sed 's/\([[:digit:]]\{1,\}
- # Test suite
- #
- 
--REGRESS := init-extension repack-setup repack-run error-on-invalid-idx no-error-on-invalid-idx after-schema repack-check nosuper tablespace get_order_by trigger
-+REGRESS := init-extension repack-setup repack-run error-on-invalid-idx no-error-on-invalid-idx after-schema repack-check nosuper get_order_by trigger
- 
- USE_PGXS = 1	# use pgxs if not in contrib directory
- PGXS := $(shell $(PG_CONFIG) --pgxs)
-diff --git a/regress/expected/nosuper.out b/regress/expected/nosuper.out
-index 8d0a94e..63b68bf 100644
---- a/regress/expected/nosuper.out
-+++ b/regress/expected/nosuper.out
-@@ -4,22 +4,22 @@
- SET client_min_messages = error;
- DROP ROLE IF EXISTS nosuper;
- SET client_min_messages = warning;
--CREATE ROLE nosuper WITH LOGIN;
-+CREATE ROLE nosuper WITH LOGIN PASSWORD 'NoSuPeRpAsSwOrD';
- -- => OK
- \! pg_repack --dbname=contrib_regression --table=tbl_cluster --no-superuser-check
- INFO: repacking table "public.tbl_cluster"
- -- => ERROR
--\! pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper
-+\! PGPASSWORD=NoSuPeRpAsSwOrD pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper
- ERROR: pg_repack failed with error: You must be a superuser to use pg_repack
- -- => ERROR
--\! pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
-+\! PGPASSWORD=NoSuPeRpAsSwOrD pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
- ERROR: pg_repack failed with error: ERROR:  permission denied for schema repack
- LINE 1: select repack.version(), repack.version_sql()
-                ^
- GRANT ALL ON ALL TABLES IN SCHEMA repack TO nosuper;
- GRANT USAGE ON SCHEMA repack TO nosuper;
- -- => ERROR
--\! pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
-+\! PGPASSWORD=NoSuPeRpAsSwOrD pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
- INFO: repacking table "public.tbl_cluster"
- ERROR: query failed: ERROR:  current transaction is aborted, commands ignored until end of transaction block
- DETAIL: query was: RESET lock_timeout
-diff --git a/regress/sql/nosuper.sql b/regress/sql/nosuper.sql
-index 072f0fa..dbe60f8 100644
---- a/regress/sql/nosuper.sql
-+++ b/regress/sql/nosuper.sql
-@@ -4,19 +4,19 @@
- SET client_min_messages = error;
- DROP ROLE IF EXISTS nosuper;
- SET client_min_messages = warning;
--CREATE ROLE nosuper WITH LOGIN;
-+CREATE ROLE nosuper WITH LOGIN PASSWORD 'NoSuPeRpAsSwOrD';
- -- => OK
- \! pg_repack --dbname=contrib_regression --table=tbl_cluster --no-superuser-check
- -- => ERROR
--\! pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper
-+\! PGPASSWORD=NoSuPeRpAsSwOrD pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper
- -- => ERROR
--\! pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
-+\! PGPASSWORD=NoSuPeRpAsSwOrD pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
- 
- GRANT ALL ON ALL TABLES IN SCHEMA repack TO nosuper;
- GRANT USAGE ON SCHEMA repack TO nosuper;
- 
- -- => ERROR
--\! pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
-+\! PGPASSWORD=NoSuPeRpAsSwOrD pg_repack --dbname=contrib_regression --table=tbl_cluster --username=nosuper --no-superuser-check
- 
- REVOKE ALL ON ALL TABLES IN SCHEMA repack FROM nosuper;
- REVOKE USAGE ON SCHEMA repack FROM nosuper;

compute_tools/Cargo.toml

@@ -25,6 +25,7 @@ fail.workspace = true
 flate2.workspace = true
 futures.workspace = true
 http.workspace = true
+jsonwebtoken.workspace = true
 metrics.workspace = true
 nix.workspace = true
 notify.workspace = true
@@ -47,11 +48,13 @@ tokio-postgres.workspace = true
 tokio-util.workspace = true
 tokio-stream.workspace = true
 tracing.workspace = true
+tracing-opentelemetry.workspace = true
 tracing-subscriber.workspace = true
 tracing-utils.workspace = true
 thiserror.workspace = true
 url.workspace = true
 uuid.workspace = true
+prometheus.workspace = true
 walkdir.workspace = true
 
 postgres_initdb.workspace = true

compute_tools/src/bin/compute_ctl.rs

@@ -41,6 +41,7 @@ use std::process::exit;
 use std::str::FromStr;
 use std::sync::atomic::Ordering;
 use std::sync::{mpsc, Arc, Condvar, Mutex, RwLock};
+use std::time::SystemTime;
 use std::{thread, time::Duration};
 
 use anyhow::{Context, Result};
@@ -85,6 +86,19 @@ fn parse_remote_ext_config(arg: &str) -> Result<String> {
     }
 }
 
+/// Generate a compute ID if one is not supplied. This exists to keep forward
+/// compatibility tests working, but will be removed in a future iteration.
+fn generate_compute_id() -> String {
+    let now = SystemTime::now();
+
+    format!(
+        "compute-{}",
+        now.duration_since(SystemTime::UNIX_EPOCH)
+            .unwrap()
+            .as_secs()
+    )
+}
+
 #[derive(Parser)]
 #[command(rename_all = "kebab-case")]
 struct Cli {
@@ -98,13 +112,16 @@ struct Cli {
     /// outside the compute will talk to the compute through this port. Keep
     /// the previous name for this argument around for a smoother release
     /// with the control plane.
-    #[arg(long, default_value_t = 3080)]
+    ///
+    /// TODO: Remove the alias after the control plane release which teaches the
+    /// control plane about the renamed argument.
+    #[arg(long, alias = "http-port", default_value_t = 3080)]
     pub external_http_port: u16,
 
-    /// The port to bind the internal listening HTTP server to. Clients include
+    /// The port to bind the internal listening HTTP server to. Clients like
     /// the neon extension (for installing remote extensions) and local_proxy.
-    #[arg(long, default_value_t = 3081)]
-    pub internal_http_port: u16,
+    #[arg(long)]
+    pub internal_http_port: Option<u16>,
 
     #[arg(short = 'D', long, value_name = "DATADIR")]
     pub pgdata: String,
@@ -139,7 +156,7 @@ struct Cli {
     #[arg(short = 'S', long, group = "spec-path")]
     pub spec_path: Option<OsString>,
 
-    #[arg(short = 'i', long, group = "compute-id")]
+    #[arg(short = 'i', long, group = "compute-id", default_value = generate_compute_id())]
     pub compute_id: String,
 
     #[arg(short = 'p', long, conflicts_with_all = ["spec", "spec-path"], value_name = "CONTROL_PLANE_API_BASE_URL")]
@@ -342,7 +359,7 @@ fn wait_spec(
         pgbin: cli.pgbin.clone(),
         pgversion: get_pg_version_string(&cli.pgbin),
         external_http_port: cli.external_http_port,
-        internal_http_port: cli.internal_http_port,
+        internal_http_port: cli.internal_http_port.unwrap_or(cli.external_http_port + 1),
         live_config_allowed,
         state: Mutex::new(new_state),
         state_changed: Condvar::new(),
@@ -366,7 +383,7 @@ fn wait_spec(
 
     // The internal HTTP server could be launched later, but there isn't much
     // sense in waiting.
-    Server::Internal(cli.internal_http_port).launch(&compute);
+    Server::Internal(cli.internal_http_port.unwrap_or(cli.external_http_port + 1)).launch(&compute);
 
     if !spec_set {
         // No spec provided, hang waiting for it.

compute_tools/src/bin/fast_import.rs

@@ -361,14 +361,6 @@ async fn run_dump_restore(
             // how we run it
             .env_clear()
             .env("LD_LIBRARY_PATH", &pg_lib_dir)
-            .env(
-                "ASAN_OPTIONS",
-                std::env::var("ASAN_OPTIONS").unwrap_or_default(),
-            )
-            .env(
-                "UBSAN_OPTIONS",
-                std::env::var("UBSAN_OPTIONS").unwrap_or_default(),
-            )
             .kill_on_drop(true)
             .stdout(std::process::Stdio::piped())
             .stderr(std::process::Stdio::piped())
@@ -402,14 +394,6 @@ async fn run_dump_restore(
             // how we run it
             .env_clear()
             .env("LD_LIBRARY_PATH", &pg_lib_dir)
-            .env(
-                "ASAN_OPTIONS",
-                std::env::var("ASAN_OPTIONS").unwrap_or_default(),
-            )
-            .env(
-                "UBSAN_OPTIONS",
-                std::env::var("UBSAN_OPTIONS").unwrap_or_default(),
-            )
             .kill_on_drop(true)
             .stdout(std::process::Stdio::piped())
             .stderr(std::process::Stdio::piped())

compute_tools/src/sql/drop_subscriptions.sql

@@ -2,7 +2,6 @@ DO $$
 DECLARE
     subname TEXT;
 BEGIN
-    LOCK TABLE pg_subscription IN ACCESS EXCLUSIVE MODE;
     FOR subname IN SELECT pg_subscription.subname FROM pg_subscription WHERE subdbid = (SELECT oid FROM pg_database WHERE datname = {datname_str}) LOOP
         EXECUTE format('ALTER SUBSCRIPTION %I DISABLE;', subname);
         EXECUTE format('ALTER SUBSCRIPTION %I SET (slot_name = NONE);', subname);

control_plane/src/endpoint.rs

@@ -46,8 +46,6 @@ use std::process::Command;
 use std::str::FromStr;
 use std::sync::Arc;
 use std::time::Duration;
-use std::time::SystemTime;
-use std::time::UNIX_EPOCH;
 
 use anyhow::{anyhow, bail, Context, Result};
 use compute_api::requests::ConfigurationRequest;
@@ -61,7 +59,6 @@ use nix::sys::signal::Signal;
 use pageserver_api::shard::ShardStripeSize;
 use reqwest::header::CONTENT_TYPE;
 use serde::{Deserialize, Serialize};
-use tracing::debug;
 use url::Host;
 use utils::id::{NodeId, TenantId, TimelineId};
 
@@ -84,10 +81,8 @@ pub struct EndpointConf {
     internal_http_port: u16,
     pg_version: u32,
     skip_pg_catalog_updates: bool,
-    reconfigure_concurrency: usize,
     drop_subscriptions_before_start: bool,
     features: Vec<ComputeFeature>,
-    cluster: Option<Cluster>,
 }
 
 //
@@ -184,9 +179,7 @@ impl ComputeControlPlane {
             // we also skip catalog updates in the cloud.
             skip_pg_catalog_updates,
             drop_subscriptions_before_start,
-            reconfigure_concurrency: 1,
             features: vec![],
-            cluster: None,
         });
 
         ep.create_endpoint_dir()?;
@@ -203,9 +196,7 @@ impl ComputeControlPlane {
                 pg_version,
                 skip_pg_catalog_updates,
                 drop_subscriptions_before_start,
-                reconfigure_concurrency: 1,
                 features: vec![],
-                cluster: None,
             })?,
         )?;
         std::fs::write(
@@ -270,11 +261,8 @@ pub struct Endpoint {
     skip_pg_catalog_updates: bool,
 
     drop_subscriptions_before_start: bool,
-    reconfigure_concurrency: usize,
     // Feature flags
     features: Vec<ComputeFeature>,
-    // Cluster settings
-    cluster: Option<Cluster>,
 }
 
 #[derive(PartialEq, Eq)]
@@ -314,8 +302,6 @@ impl Endpoint {
         let conf: EndpointConf =
             serde_json::from_slice(&std::fs::read(entry.path().join("endpoint.json"))?)?;
 
-        debug!("serialized endpoint conf: {:?}", conf);
-
         Ok(Endpoint {
             pg_address: SocketAddr::new(IpAddr::from(Ipv4Addr::LOCALHOST), conf.pg_port),
             external_http_address: SocketAddr::new(
@@ -333,10 +319,8 @@ impl Endpoint {
             tenant_id: conf.tenant_id,
             pg_version: conf.pg_version,
             skip_pg_catalog_updates: conf.skip_pg_catalog_updates,
-            reconfigure_concurrency: conf.reconfigure_concurrency,
             drop_subscriptions_before_start: conf.drop_subscriptions_before_start,
             features: conf.features,
-            cluster: conf.cluster,
         })
     }
 
@@ -623,7 +607,7 @@ impl Endpoint {
         };
 
         // Create spec file
-        let mut spec = ComputeSpec {
+        let spec = ComputeSpec {
             skip_pg_catalog_updates: self.skip_pg_catalog_updates,
             format_version: 1.0,
             operation_uuid: None,
@@ -656,7 +640,7 @@ impl Endpoint {
                     Vec::new()
                 },
                 settings: None,
-                postgresql_conf: Some(postgresql_conf.clone()),
+                postgresql_conf: Some(postgresql_conf),
             },
             delta_operations: None,
             tenant_id: Some(self.tenant_id),
@@ -669,35 +653,9 @@ impl Endpoint {
             pgbouncer_settings: None,
             shard_stripe_size: Some(shard_stripe_size),
             local_proxy_config: None,
-            reconfigure_concurrency: self.reconfigure_concurrency,
+            reconfigure_concurrency: 1,
             drop_subscriptions_before_start: self.drop_subscriptions_before_start,
         };
-
-        // this strange code is needed to support respec() in tests
-        if self.cluster.is_some() {
-            debug!("Cluster is already set in the endpoint spec, using it");
-            spec.cluster = self.cluster.clone().unwrap();
-
-            debug!("spec.cluster {:?}", spec.cluster);
-
-            // fill missing fields again
-            if create_test_user {
-                spec.cluster.roles.push(Role {
-                    name: PgIdent::from_str("test").unwrap(),
-                    encrypted_password: None,
-                    options: None,
-                });
-                spec.cluster.databases.push(Database {
-                    name: PgIdent::from_str("neondb").unwrap(),
-                    owner: PgIdent::from_str("test").unwrap(),
-                    options: None,
-                    restrict_conn: false,
-                    invalid: false,
-                });
-            }
-            spec.cluster.postgresql_conf = Some(postgresql_conf);
-        }
-
         let spec_path = self.endpoint_path().join("spec.json");
         std::fs::write(spec_path, serde_json::to_string_pretty(&spec)?)?;
 
@@ -715,14 +673,18 @@ impl Endpoint {
             println!("Also at '{}'", conn_str);
         }
         let mut cmd = Command::new(self.env.neon_distrib_dir.join("compute_ctl"));
+        //cmd.args([
+        //    "--external-http-port",
+        //    &self.external_http_address.port().to_string(),
+        //])
+        //.args([
+        //    "--internal-http-port",
+        //    &self.internal_http_address.port().to_string(),
+        //])
         cmd.args([
-            "--external-http-port",
+            "--http-port",
             &self.external_http_address.port().to_string(),
         ])
-        .args([
-            "--internal-http-port",
-            &self.internal_http_address.port().to_string(),
-        ])
         .args(["--pgdata", self.pgdata().to_str().unwrap()])
         .args(["--connstr", &conn_str])
         .args([
@@ -739,16 +701,20 @@ impl Endpoint {
         ])
         // TODO: It would be nice if we generated compute IDs with the same
         // algorithm as the real control plane.
-        .args([
-            "--compute-id",
-            &format!(
-                "compute-{}",
-                SystemTime::now()
-                    .duration_since(UNIX_EPOCH)
-                    .unwrap()
-                    .as_secs()
-            ),
-        ])
+        //
+        // TODO: Add this back when
+        // https://github.com/neondatabase/neon/pull/10747 is merged.
+        //
+        //.args([
+        //    "--compute-id",
+        //    &format!(
+        //        "compute-{}",
+        //        SystemTime::now()
+        //            .duration_since(UNIX_EPOCH)
+        //            .unwrap()
+        //            .as_secs()
+        //    ),
+        //])
         .stdin(std::process::Stdio::null())
         .stderr(logfile.try_clone()?)
         .stdout(logfile);

control_plane/src/pageserver.rs

@@ -335,21 +335,13 @@ impl PageServerNode {
                 .map(|x| x.parse::<u64>())
                 .transpose()
                 .context("Failed to parse 'checkpoint_distance' as an integer")?,
-            checkpoint_timeout: settings
-                .remove("checkpoint_timeout")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'checkpoint_timeout' as duration")?,
+            checkpoint_timeout: settings.remove("checkpoint_timeout").map(|x| x.to_string()),
             compaction_target_size: settings
                 .remove("compaction_target_size")
                 .map(|x| x.parse::<u64>())
                 .transpose()
                 .context("Failed to parse 'compaction_target_size' as an integer")?,
-            compaction_period: settings
-                .remove("compaction_period")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'compaction_period' as duration")?,
+            compaction_period: settings.remove("compaction_period").map(|x| x.to_string()),
             compaction_threshold: settings
                 .remove("compaction_threshold")
                 .map(|x| x.parse::<usize>())
@@ -395,10 +387,7 @@ impl PageServerNode {
                 .map(|x| x.parse::<u64>())
                 .transpose()
                 .context("Failed to parse 'gc_horizon' as an integer")?,
-            gc_period: settings.remove("gc_period")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'gc_period' as duration")?,
+            gc_period: settings.remove("gc_period").map(|x| x.to_string()),
             image_creation_threshold: settings
                 .remove("image_creation_threshold")
                 .map(|x| x.parse::<usize>())
@@ -414,20 +403,13 @@ impl PageServerNode {
                 .map(|x| x.parse::<usize>())
                 .transpose()
                 .context("Failed to parse 'image_creation_preempt_threshold' as integer")?,
-            pitr_interval: settings.remove("pitr_interval")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'pitr_interval' as duration")?,
+            pitr_interval: settings.remove("pitr_interval").map(|x| x.to_string()),
             walreceiver_connect_timeout: settings
                 .remove("walreceiver_connect_timeout")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'walreceiver_connect_timeout' as duration")?,
+                .map(|x| x.to_string()),
             lagging_wal_timeout: settings
                 .remove("lagging_wal_timeout")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'lagging_wal_timeout' as duration")?,
+                .map(|x| x.to_string()),
             max_lsn_wal_lag: settings
                 .remove("max_lsn_wal_lag")
                 .map(|x| x.parse::<NonZeroU64>())
@@ -445,14 +427,8 @@ impl PageServerNode {
                 .context("Failed to parse 'min_resident_size_override' as integer")?,
             evictions_low_residence_duration_metric_threshold: settings
                 .remove("evictions_low_residence_duration_metric_threshold")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'evictions_low_residence_duration_metric_threshold' as duration")?,
-            heatmap_period: settings
-                .remove("heatmap_period")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'heatmap_period' as duration")?,
+                .map(|x| x.to_string()),
+            heatmap_period: settings.remove("heatmap_period").map(|x| x.to_string()),
             lazy_slru_download: settings
                 .remove("lazy_slru_download")
                 .map(|x| x.parse::<bool>())
@@ -463,15 +439,10 @@ impl PageServerNode {
                 .map(serde_json::from_str)
                 .transpose()
                 .context("parse `timeline_get_throttle` from json")?,
-            lsn_lease_length: settings.remove("lsn_lease_length")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'lsn_lease_length' as duration")?,
+            lsn_lease_length: settings.remove("lsn_lease_length").map(|x| x.to_string()),
             lsn_lease_length_for_ts: settings
                 .remove("lsn_lease_length_for_ts")
-                .map(humantime::parse_duration)
-                .transpose()
-                .context("Failed to parse 'lsn_lease_length_for_ts' as duration")?,
+                .map(|x| x.to_string()),
             timeline_offloading: settings
                 .remove("timeline_offloading")
                 .map(|x| x.parse::<bool>())

control_plane/storcon_cli/src/main.rs

@@ -22,7 +22,7 @@ use pageserver_api::{
 };
 use pageserver_client::mgmt_api::{self};
 use reqwest::{Method, StatusCode, Url};
-use utils::id::{NodeId, TenantId, TimelineId};
+use utils::id::{NodeId, TenantId};
 
 use pageserver_api::controller_api::{
     NodeConfigureRequest, NodeRegisterRequest, NodeSchedulingPolicy, PlacementPolicy,
@@ -47,9 +47,6 @@ enum Command {
         listen_http_addr: String,
         #[arg(long)]
         listen_http_port: u16,
-        #[arg(long)]
-        listen_https_port: Option<u16>,
-
         #[arg(long)]
         availability_zone_id: String,
     },
@@ -242,19 +239,6 @@ enum Command {
         #[arg(long)]
         scheduling_policy: SkSchedulingPolicyArg,
     },
-    /// Downloads any missing heatmap layers for all shard for a given timeline
-    DownloadHeatmapLayers {
-        /// Tenant ID or tenant shard ID. When an unsharded tenant ID is specified,
-        /// the operation is performed on all shards. When a sharded tenant ID is
-        /// specified, the operation is only performed on the specified shard.
-        #[arg(long)]
-        tenant_shard_id: TenantShardId,
-        #[arg(long)]
-        timeline_id: TimelineId,
-        /// Optional: Maximum download concurrency (default is 16)
-        #[arg(long)]
-        concurrency: Option<usize>,
-    },
 }
 
 #[derive(Parser)]
@@ -397,7 +381,6 @@ async fn main() -> anyhow::Result<()> {
             listen_pg_port,
             listen_http_addr,
             listen_http_port,
-            listen_https_port,
             availability_zone_id,
         } => {
             storcon_client
@@ -410,7 +393,6 @@ async fn main() -> anyhow::Result<()> {
                         listen_pg_port,
                         listen_http_addr,
                         listen_http_port,
-                        listen_https_port,
                         availability_zone_id: AvailabilityZone(availability_zone_id),
                     }),
                 )
@@ -959,7 +941,7 @@ async fn main() -> anyhow::Result<()> {
                                 threshold: threshold.into(),
                             },
                         )),
-                        heatmap_period: Some(Duration::from_secs(300)),
+                        heatmap_period: Some("300s".to_string()),
                         ..Default::default()
                     },
                 })
@@ -1265,24 +1247,6 @@ async fn main() -> anyhow::Result<()> {
                 String::from(scheduling_policy)
             );
         }
-        Command::DownloadHeatmapLayers {
-            tenant_shard_id,
-            timeline_id,
-            concurrency,
-        } => {
-            let mut path = format!(
-                "/v1/tenant/{}/timeline/{}/download_heatmap_layers",
-                tenant_shard_id, timeline_id,
-            );
-
-            if let Some(c) = concurrency {
-                path = format!("{path}?concurrency={c}");
-            }
-
-            storcon_client
-                .dispatch::<(), ()>(Method::POST, path, None)
-                .await?;
-        }
     }
 
     Ok(())

docker-compose/compute_wrapper/shell/compute.sh

@@ -77,5 +77,4 @@ echo "Start compute node"
 /usr/local/bin/compute_ctl --pgdata /var/db/postgres/compute \
      -C "postgresql://cloud_admin@localhost:55433/postgres"  \
      -b /usr/local/bin/postgres                              \
-     --compute-id "compute-$RANDOM"                          \
      -S ${SPEC_FILE}

docker-compose/docker_compose_test.sh

@@ -81,8 +81,15 @@ for pg_version in ${TEST_VERSION_ONLY-14 15 16 17}; do
             [ $EXT_SUCCESS -eq 0 ] && FAILED=$(tail -1 testout.txt | awk '{for(i=1;i<=NF;i++){print "/ext-src/"$i;}}')
             [ $CONTRIB_SUCCESS -eq 0 ] && CONTRIB_FAILED=$(tail -1 testout_contrib.txt | awk '{for(i=0;i<=NF;i++){print "/postgres/contrib/"$i;}}')
             for d in $FAILED $CONTRIB_FAILED; do
-                docker exec $TEST_CONTAINER_NAME bash -c 'for file in $(find '"$d"' -name regression.diffs -o -name regression.out); do cat $file; done' || [ $? -eq 1 ]
+                dn="$(basename $d)"
+                rm -rf $dn
+                mkdir $dn
+                docker cp $TEST_CONTAINER_NAME:$d/regression.diffs $dn || [ $? -eq 1 ]
+                docker cp $TEST_CONTAINER_NAME:$d/regression.out $dn || [ $? -eq 1 ]
+                cat $dn/regression.out $dn/regression.diffs || true
+                rm -rf $dn
             done
+        rm -rf $FAILED
         exit 1
         fi
     fi

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

@@ -1,5 +0,0 @@
-#!/bin/sh
-set -ex
-cd "$(dirname ${0})"
-PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
-${PG_REGRESS} --use-existing --inputdir=./regress --bindir='/usr/local/pgsql/bin' --dbname=contrib_regression repack-setup repack-run error-on-invalid-idx no-error-on-invalid-idx after-schema repack-check nosuper get_order_by trigger

docker-compose/ext-src/pg_semver-src/test-upgrade-17.patch

@@ -1,24 +0,0 @@
-diff --git a/test/sql/base.sql b/test/sql/base.sql
-index 53adb30..2eed91b 100644
---- a/test/sql/base.sql
-+++ b/test/sql/base.sql
-@@ -2,7 +2,6 @@
- BEGIN;
- 
- \i test/pgtap-core.sql
--CREATE EXTENSION semver;
- 
- SELECT plan(334);
- --SELECT * FROM no_plan();
-diff --git a/test/sql/corpus.sql b/test/sql/corpus.sql
-index c0fe98e..39cdd2e 100644
---- a/test/sql/corpus.sql
-+++ b/test/sql/corpus.sql
-@@ -4,7 +4,6 @@ BEGIN;
- -- Test the SemVer corpus from https://regex101.com/r/Ly7O1x/3/.
- 
- \i test/pgtap-core.sql
--CREATE EXTENSION semver;
- 
- SELECT plan(76);
- --SELECT * FROM no_plan();

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

@@ -1,7 +1,6 @@
 #!/bin/sh
 set -ex
 cd "$(dirname ${0})"
-patch -p1 <test-upgrade-${PG_VERSION}.patch
-psql -d contrib_regression -c "DROP EXTENSION IF EXISTS pgtap"
+patch -p1 <test-upgrade.patch
 PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
 ${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin'    --inputdir=test --dbname=contrib_regression base corpus

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

@@ -1,16 +1,3 @@
-diff --git a/Makefile b/Makefile
-index f255fe6..0a0fa65 100644
---- a/Makefile
-+++ b/Makefile
-@@ -346,7 +346,7 @@ test: test-serial test-parallel
- TB_DIR = test/build
- GENERATED_SCHEDULE_DEPS = $(TB_DIR)/all_tests $(TB_DIR)/exclude_tests
- REGRESS = --schedule $(TB_DIR)/run.sch # Set this again just to be safe
--REGRESS_OPTS = --inputdir=test --max-connections=$(PARALLEL_CONN) --schedule $(SETUP_SCH) $(REGRESS_CONF)
-+REGRESS_OPTS = --use-existing --dbname=pgtap_regression --inputdir=test --max-connections=$(PARALLEL_CONN) --schedule $(SETUP_SCH) $(REGRESS_CONF)
- SETUP_SCH = test/schedule/main.sch # schedule to use for test setup; this can be forcibly changed by some targets!
- IGNORE_TESTS = $(notdir $(EXCLUDE_TEST_FILES:.sql=))
- PARALLEL_TESTS = $(filter-out $(IGNORE_TESTS),$(filter-out $(SERIAL_TESTS),$(ALL_TESTS)))
 diff --git a/test/schedule/create.sql b/test/schedule/create.sql
 index ba355ed..7e250f5 100644
 --- a/test/schedule/create.sql

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

@@ -2,4 +2,5 @@
 set -ex
 cd "$(dirname ${0})"
 patch -p1 <test-upgrade.patch
-make installcheck
+PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin'    --inputdir=test --max-connections=86 --schedule test/schedule/main.sch   --schedule test/build/run.sch --dbname contrib_regression --use-existing

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

@@ -2,5 +2,4 @@
 set -ex
 cd "$(dirname ${0})"
 PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
-REGRESS="$(make -n installcheck | awk '{print substr($0,index($0,"init-extension")+15);}')"
-${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin'  --use-existing --dbname=contrib_regression ${REGRESS}
+${PG_REGRESS} --inputdir=./ --bindir='/usr/local/pgsql/bin'  --use-existing --dbname=contrib_regression plv8 plv8-errors scalar_args inline json startup_pre startup varparam json_conv jsonb_conv window guc es6 arraybuffer composites currentresource startup_perms bytea find_function_perms memory_limits reset show array_spread regression dialect bigint procedure

docker-compose/test_extensions_upgrade.sh

@@ -43,8 +43,7 @@ EXTENSIONS='[
 {"extname": "semver", "extdir": "pg_semver-src"},
 {"extname": "pg_ivm", "extdir": "pg_ivm-src"},
 {"extname": "pgjwt", "extdir": "pgjwt-src"},
-{"extname": "pgtap", "extdir": "pgtap-src"},
-{"extname": "pg_repack", "extdir": "pg_repack-src"}
+{"extname": "pgtap", "extdir": "pgtap-src"}
 ]'
 EXTNAMES=$(echo ${EXTENSIONS} | jq -r '.[].extname' | paste -sd ' ' -)
 TAG=${NEWTAG} docker compose --profile test-extensions up --quiet-pull --build -d
@@ -60,8 +59,6 @@ wait_for_ready
 docker compose cp  ext-src neon-test-extensions:/
 docker compose exec neon-test-extensions psql -c "DROP DATABASE IF EXISTS contrib_regression"
 docker compose exec neon-test-extensions psql -c "CREATE DATABASE contrib_regression"
-docker compose exec neon-test-extensions psql -c "CREATE DATABASE pgtap_regression"
-docker compose exec neon-test-extensions psql -d pgtap_regression -c "CREATE EXTENSION pgtap"
 create_extensions "${EXTNAMES}"
 if [ "${FORCE_ALL_UPGRADE_TESTS:-false}" = true ]; then
   exts="${EXTNAMES}"

libs/compute_api/src/spec.rs

@@ -252,7 +252,7 @@ pub enum ComputeMode {
     Replica,
 }
 
-#[derive(Clone, Debug, Default, Deserialize, Serialize, PartialEq, Eq)]
+#[derive(Clone, Debug, Default, Deserialize, Serialize)]
 pub struct Cluster {
     pub cluster_id: Option<String>,
     pub name: Option<String>,
@@ -283,7 +283,7 @@ pub struct DeltaOp {
 
 /// Rust representation of Postgres role info with only those fields
 /// that matter for us.
-#[derive(Clone, Debug, Deserialize, Serialize, PartialEq, Eq)]
+#[derive(Clone, Debug, Deserialize, Serialize)]
 pub struct Role {
     pub name: PgIdent,
     pub encrypted_password: Option<String>,
@@ -292,7 +292,7 @@ pub struct Role {
 
 /// Rust representation of Postgres database info with only those fields
 /// that matter for us.
-#[derive(Clone, Debug, Deserialize, Serialize, PartialEq, Eq)]
+#[derive(Clone, Debug, Deserialize, Serialize)]
 pub struct Database {
     pub name: PgIdent,
     pub owner: PgIdent,
@@ -308,7 +308,7 @@ pub struct Database {
 /// Common type representing both SQL statement params with or without value,
 /// like `LOGIN` or `OWNER username` in the `CREATE/ALTER ROLE`, and config
 /// options like `wal_level = logical`.
-#[derive(Clone, Debug, Deserialize, Serialize, PartialEq, Eq)]
+#[derive(Clone, Debug, Deserialize, Serialize)]
 pub struct GenericOption {
     pub name: String,
     pub value: Option<String>,

libs/http-utils/src/pprof.rs

@@ -2,6 +2,7 @@ use anyhow::bail;
 use flate2::write::{GzDecoder, GzEncoder};
 use flate2::Compression;
 use itertools::Itertools as _;
+use once_cell::sync::Lazy;
 use pprof::protos::{Function, Line, Location, Message as _, Profile};
 use regex::Regex;
 
@@ -57,30 +58,38 @@ pub fn symbolize(mut profile: Profile) -> anyhow::Result<Profile> {
 
         // Resolve the line and function for each location.
         backtrace::resolve(loc.address as *mut c_void, |symbol| {
-            let Some(symbol_name) = symbol.name() else {
+            let Some(symname) = symbol.name() else {
                 return;
             };
+            let mut name = symname.to_string();
 
-            let function_name = format!("{symbol_name:#}");
-            let functions_len = functions.len();
-            let function_id = functions
-                .entry(function_name)
-                .or_insert_with_key(|function_name| {
-                    let function_id = functions_len as u64 + 1;
-                    let system_name = String::from_utf8_lossy(symbol_name.as_bytes());
+            // Strip the Rust monomorphization suffix from the symbol name.
+            static SUFFIX_REGEX: Lazy<Regex> =
+                Lazy::new(|| Regex::new("::h[0-9a-f]{16}$").expect("invalid regex"));
+            if let Some(m) = SUFFIX_REGEX.find(&name) {
+                name.truncate(m.start());
+            }
+
+            let function_id = match functions.get(&name) {
+                Some(function) => function.id,
+                None => {
+                    let id = functions.len() as u64 + 1;
+                    let system_name = String::from_utf8_lossy(symname.as_bytes());
                     let filename = symbol
                         .filename()
                         .map(|path| path.to_string_lossy())
                         .unwrap_or(Cow::Borrowed(""));
-                    Function {
-                        id: function_id,
-                        name: string_id(function_name),
+                    let function = Function {
+                        id,
+                        name: string_id(&name),
                         system_name: string_id(&system_name),
                         filename: string_id(&filename),
                         ..Default::default()
-                    }
-                })
-                .id;
+                    };
+                    functions.insert(name, function);
+                    id
+                }
+            };
             loc.line.push(Line {
                 function_id,
                 line: symbol.lineno().unwrap_or(0) as i64,

libs/pageserver_api/src/config.rs

@@ -122,8 +122,6 @@ pub struct ConfigToml {
     pub page_service_pipelining: PageServicePipeliningConfig,
     pub get_vectored_concurrent_io: GetVectoredConcurrentIo,
     pub enable_read_path_debugging: Option<bool>,
-    #[serde(skip_serializing_if = "Option::is_none")]
-    pub validate_wal_contiguity: Option<bool>,
 }
 
 #[derive(Debug, Clone, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
@@ -523,7 +521,6 @@ impl Default for ConfigToml {
             } else {
                 None
             },
-            validate_wal_contiguity: None,
         }
     }
 }
@@ -547,11 +544,10 @@ pub mod tenant_conf_defaults {
     pub const DEFAULT_COMPACTION_PERIOD: &str = "20 s";
     pub const DEFAULT_COMPACTION_THRESHOLD: usize = 10;
 
-    // This value needs to be tuned to avoid OOM. We have 3/4*CPUs threads for L0 compaction, that's
-    // 3/4*16=9 on most of our pageservers. Compacting 20 layers requires about 1 GB memory (could
-    // be reduced later by optimizing L0 hole calculation to avoid loading all keys into memory). So
-    // with this config, we can get a maximum peak compaction usage of 9 GB.
-    pub const DEFAULT_COMPACTION_UPPER_LIMIT: usize = 20;
+    // This value needs to be tuned to avoid OOM. We have 3/4 of the total CPU threads to do background works, that's 16*3/4=9 on
+    // most of our pageservers. Compaction ~50 layers requires about 2GB memory (could be reduced later by optimizing L0 hole
+    // calculation to avoid loading all keys into the memory). So with this config, we can get a maximum peak compaction usage of 18GB.
+    pub const DEFAULT_COMPACTION_UPPER_LIMIT: usize = 50;
     pub const DEFAULT_COMPACTION_L0_FIRST: bool = false;
     pub const DEFAULT_COMPACTION_L0_SEMAPHORE: bool = true;
 

libs/pageserver_api/src/controller_api.rs

@@ -57,7 +57,6 @@ pub struct NodeRegisterRequest {
 
     pub listen_http_addr: String,
     pub listen_http_port: u16,
-    pub listen_https_port: Option<u16>,
 
     pub availability_zone_id: AvailabilityZone,
 }
@@ -106,7 +105,6 @@ pub struct TenantLocateResponseShard {
 
     pub listen_http_addr: String,
     pub listen_http_port: u16,
-    pub listen_https_port: Option<u16>,
 }
 
 #[derive(Serialize, Deserialize)]
@@ -150,7 +148,6 @@ pub struct NodeDescribeResponse {
 
     pub listen_http_addr: String,
     pub listen_http_port: u16,
-    pub listen_https_port: Option<u16>,
 
     pub listen_pg_addr: String,
     pub listen_pg_port: u16,

libs/pageserver_api/src/models.rs

@@ -526,13 +526,9 @@ pub struct TenantConfigPatch {
 #[derive(Serialize, Deserialize, Debug, Default, Clone, Eq, PartialEq)]
 pub struct TenantConfig {
     pub checkpoint_distance: Option<u64>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub checkpoint_timeout: Option<Duration>,
+    pub checkpoint_timeout: Option<String>,
     pub compaction_target_size: Option<u64>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub compaction_period: Option<Duration>,
+    pub compaction_period: Option<String>,
     pub compaction_threshold: Option<usize>,
     pub compaction_upper_limit: Option<usize>,
     // defer parsing compaction_algorithm, like eviction_policy
@@ -543,38 +539,22 @@ pub struct TenantConfig {
     pub l0_flush_stall_threshold: Option<usize>,
     pub l0_flush_wait_upload: Option<bool>,
     pub gc_horizon: Option<u64>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub gc_period: Option<Duration>,
+    pub gc_period: Option<String>,
     pub image_creation_threshold: Option<usize>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub pitr_interval: Option<Duration>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub walreceiver_connect_timeout: Option<Duration>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub lagging_wal_timeout: Option<Duration>,
+    pub pitr_interval: Option<String>,
+    pub walreceiver_connect_timeout: Option<String>,
+    pub lagging_wal_timeout: Option<String>,
     pub max_lsn_wal_lag: Option<NonZeroU64>,
     pub eviction_policy: Option<EvictionPolicy>,
     pub min_resident_size_override: Option<u64>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub evictions_low_residence_duration_metric_threshold: Option<Duration>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub heatmap_period: Option<Duration>,
+    pub evictions_low_residence_duration_metric_threshold: Option<String>,
+    pub heatmap_period: Option<String>,
     pub lazy_slru_download: Option<bool>,
     pub timeline_get_throttle: Option<ThrottleConfig>,
     pub image_layer_creation_check_threshold: Option<u8>,
     pub image_creation_preempt_threshold: Option<usize>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub lsn_lease_length: Option<Duration>,
-    #[serde(default)]
-    #[serde(with = "humantime_serde")]
-    pub lsn_lease_length_for_ts: Option<Duration>,
+    pub lsn_lease_length: Option<String>,
+    pub lsn_lease_length_for_ts: Option<String>,
     pub timeline_offloading: Option<bool>,
     pub wal_receiver_protocol_override: Option<PostgresClientProtocol>,
     pub rel_size_v2_enabled: Option<bool>,
@@ -584,10 +564,7 @@ pub struct TenantConfig {
 }
 
 impl TenantConfig {
-    pub fn apply_patch(
-        self,
-        patch: TenantConfigPatch,
-    ) -> Result<TenantConfig, humantime::DurationError> {
+    pub fn apply_patch(self, patch: TenantConfigPatch) -> TenantConfig {
         let Self {
             mut checkpoint_distance,
             mut checkpoint_timeout,
@@ -627,17 +604,11 @@ impl TenantConfig {
         } = self;
 
         patch.checkpoint_distance.apply(&mut checkpoint_distance);
-        patch
-            .checkpoint_timeout
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut checkpoint_timeout);
+        patch.checkpoint_timeout.apply(&mut checkpoint_timeout);
         patch
             .compaction_target_size
             .apply(&mut compaction_target_size);
-        patch
-            .compaction_period
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut compaction_period);
+        patch.compaction_period.apply(&mut compaction_period);
         patch.compaction_threshold.apply(&mut compaction_threshold);
         patch
             .compaction_upper_limit
@@ -655,25 +626,15 @@ impl TenantConfig {
             .apply(&mut l0_flush_stall_threshold);
         patch.l0_flush_wait_upload.apply(&mut l0_flush_wait_upload);
         patch.gc_horizon.apply(&mut gc_horizon);
-        patch
-            .gc_period
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut gc_period);
+        patch.gc_period.apply(&mut gc_period);
         patch
             .image_creation_threshold
             .apply(&mut image_creation_threshold);
-        patch
-            .pitr_interval
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut pitr_interval);
+        patch.pitr_interval.apply(&mut pitr_interval);
         patch
             .walreceiver_connect_timeout
-            .map(|v| humantime::parse_duration(&v))?
             .apply(&mut walreceiver_connect_timeout);
-        patch
-            .lagging_wal_timeout
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut lagging_wal_timeout);
+        patch.lagging_wal_timeout.apply(&mut lagging_wal_timeout);
         patch.max_lsn_wal_lag.apply(&mut max_lsn_wal_lag);
         patch.eviction_policy.apply(&mut eviction_policy);
         patch
@@ -681,12 +642,8 @@ impl TenantConfig {
             .apply(&mut min_resident_size_override);
         patch
             .evictions_low_residence_duration_metric_threshold
-            .map(|v| humantime::parse_duration(&v))?
             .apply(&mut evictions_low_residence_duration_metric_threshold);
-        patch
-            .heatmap_period
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut heatmap_period);
+        patch.heatmap_period.apply(&mut heatmap_period);
         patch.lazy_slru_download.apply(&mut lazy_slru_download);
         patch
             .timeline_get_throttle
@@ -697,13 +654,9 @@ impl TenantConfig {
         patch
             .image_creation_preempt_threshold
             .apply(&mut image_creation_preempt_threshold);
-        patch
-            .lsn_lease_length
-            .map(|v| humantime::parse_duration(&v))?
-            .apply(&mut lsn_lease_length);
+        patch.lsn_lease_length.apply(&mut lsn_lease_length);
         patch
             .lsn_lease_length_for_ts
-            .map(|v| humantime::parse_duration(&v))?
             .apply(&mut lsn_lease_length_for_ts);
         patch.timeline_offloading.apply(&mut timeline_offloading);
         patch
@@ -720,7 +673,7 @@ impl TenantConfig {
             .gc_compaction_ratio_percent
             .apply(&mut gc_compaction_ratio_percent);
 
-        Ok(Self {
+        Self {
             checkpoint_distance,
             checkpoint_timeout,
             compaction_target_size,
@@ -756,7 +709,7 @@ impl TenantConfig {
             gc_compaction_enabled,
             gc_compaction_initial_threshold_kb,
             gc_compaction_ratio_percent,
-        })
+        }
     }
 }
 
@@ -1127,7 +1080,8 @@ pub struct TenantInfo {
 
     /// Opaque explanation if gc is being blocked.
     ///
-    /// Only looked up for the individual tenant detail, not the listing.
+    /// Only looked up for the individual tenant detail, not the listing. This is purely for
+    /// debugging, not included in openapi.
     #[serde(skip_serializing_if = "Option::is_none")]
     pub gc_blocking: Option<String>,
 }
@@ -2550,7 +2504,7 @@ mod tests {
             ..base.clone()
         };
 
-        let patched = base.apply_patch(decoded.config).unwrap();
+        let patched = base.apply_patch(decoded.config);
 
         assert_eq!(patched, expected);
     }

libs/postgres_ffi/src/lib.rs

@@ -278,7 +278,7 @@ pub fn generate_pg_control(
     checkpoint_bytes: &[u8],
     lsn: Lsn,
     pg_version: u32,
-) -> anyhow::Result<(Bytes, u64, bool)> {
+) -> anyhow::Result<(Bytes, u64)> {
     dispatch_pgversion!(
         pg_version,
         pgv::xlog_utils::generate_pg_control(pg_control_bytes, checkpoint_bytes, lsn),

libs/postgres_ffi/src/xlog_utils.rs

@@ -124,59 +124,23 @@ pub fn normalize_lsn(lsn: Lsn, seg_sz: usize) -> Lsn {
     }
 }
 
-/// Generate a pg_control file, for a basebackup for starting up Postgres at the given LSN
-///
-/// 'pg_control_bytes' and 'checkpoint_bytes' are the contents of those keys persisted in
-/// the pageserver. They use the same format as the PostgreSQL control file and the
-/// checkpoint record, but see walingest.rs for how exactly they are kept up to date.
-/// 'lsn' is the LSN at which we're starting up.
-///
-/// Returns:
-/// - pg_control file contents
-/// - system_identifier, extracted from the persisted information
-/// - true, if we're starting up from a "clean shutdown", i.e. if there was a shutdown
-///   checkpoint at the given LSN
 pub fn generate_pg_control(
     pg_control_bytes: &[u8],
     checkpoint_bytes: &[u8],
     lsn: Lsn,
-) -> anyhow::Result<(Bytes, u64, bool)> {
+) -> anyhow::Result<(Bytes, u64)> {
     let mut pg_control = ControlFileData::decode(pg_control_bytes)?;
     let mut checkpoint = CheckPoint::decode(checkpoint_bytes)?;
 
     // Generate new pg_control needed for bootstrap
-    //
-    // NB: In the checkpoint struct that we persist in the pageserver, we have a different
-    // convention for the 'redo' field than in PostgreSQL: On a shutdown checkpoint,
-    // 'redo' points the *end* of the checkpoint WAL record. On PostgreSQL, it points to
-    // the beginning. Furthermore, on an online checkpoint, 'redo' is set to 0.
-    //
-    // We didn't always have this convention however, and old persisted records will have
-    // old REDO values that point to some old LSN.
-    //
-    // The upshot is that if 'redo' is equal to the "current" LSN, there was a shutdown
-    // checkpoint record at that point in WAL, with no new WAL records after it. That case
-    // can be treated as starting from a clean shutdown. All other cases are treated as
-    // non-clean shutdown. In Neon, we don't do WAL replay at startup in either case, so
-    // that distinction doesn't matter very much. As of this writing, it only affects
-    // whether the persisted pg_stats information can be used or not.
-    //
-    // In the Checkpoint struct in the returned pg_control file, the redo pointer is
-    // always set to the LSN we're starting at, to hint that no WAL replay is required.
-    // (There's some neon-specific code in Postgres startup to make that work, though.
-    // Just setting the redo pointer is not sufficient.)
-    let was_shutdown = Lsn(checkpoint.redo) == lsn;
     checkpoint.redo = normalize_lsn(lsn, WAL_SEGMENT_SIZE).0;
 
-    // We use DBState_DB_SHUTDOWNED even if it was not a clean shutdown.  The
-    // neon-specific code at postgres startup ignores the state stored in the control
-    // file, similar to archive recovery in standalone PostgreSQL. Similarly, the
-    // checkPoint pointer is ignored, so just set it to 0.
+    //save new values in pg_control
     pg_control.checkPoint = 0;
     pg_control.checkPointCopy = checkpoint;
     pg_control.state = DBState_DB_SHUTDOWNED;
 
-    Ok((pg_control.encode(), pg_control.system_identifier, was_shutdown))
+    Ok((pg_control.encode(), pg_control.system_identifier))
 }
 
 pub fn get_current_timestamp() -> TimestampTz {

libs/proxy/tokio-postgres2/Cargo.toml

@@ -5,15 +5,18 @@ edition = "2021"
 license = "MIT/Apache-2.0"
 
 [dependencies]
+async-trait.workspace = true
 bytes.workspace = true
+byteorder.workspace = true
 fallible-iterator.workspace = true
 futures-util = { workspace = true, features = ["sink"] }
 log = "0.4"
 parking_lot.workspace = true
+percent-encoding = "2.0"
 pin-project-lite.workspace = true
 phf = "0.11"
 postgres-protocol2 = { path = "../postgres-protocol2" }
 postgres-types2 = { path = "../postgres-types2" }
 tokio = { workspace = true, features = ["io-util", "time", "net"] }
 tokio-util = { workspace = true, features = ["codec"] }
-serde = { workspace = true, features = ["derive"] }
+serde = { workspace = true, features = ["derive"] }

libs/safekeeper_api/src/membership.rs

@@ -9,43 +9,13 @@ use anyhow::bail;
 use serde::{Deserialize, Serialize};
 use utils::id::NodeId;
 
-/// 1 is the first valid generation, 0 is used as
-/// a placeholder before we fully migrate to generations.
-pub const INVALID_GENERATION: SafekeeperGeneration = SafekeeperGeneration::new(0);
-pub const INITIAL_GENERATION: SafekeeperGeneration = SafekeeperGeneration::new(1);
-
 /// Number uniquely identifying safekeeper configuration.
 /// Note: it is a part of sk control file.
-///
-/// Like tenant generations, but for safekeepers.
-#[derive(Debug, Copy, Clone, PartialEq, Eq, PartialOrd, Ord, Serialize, Deserialize)]
-pub struct SafekeeperGeneration(u32);
-
-impl SafekeeperGeneration {
-    pub const fn new(v: u32) -> Self {
-        Self(v)
-    }
-
-    #[track_caller]
-    pub fn previous(&self) -> Option<Self> {
-        Some(Self(self.0.checked_sub(1)?))
-    }
-
-    #[track_caller]
-    pub fn next(&self) -> Self {
-        Self(self.0 + 1)
-    }
-
-    pub fn into_inner(self) -> u32 {
-        self.0
-    }
-}
-
-impl Display for SafekeeperGeneration {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
-        write!(f, "{}", self.0)
-    }
-}
+pub type Generation = u32;
+/// 1 is the first valid generation, 0 is used as
+/// a placeholder before we fully migrate to generations.
+pub const INVALID_GENERATION: Generation = 0;
+pub const INITIAL_GENERATION: Generation = 1;
 
 /// Membership is defined by ids so e.g. walproposer uses them to figure out
 /// quorums, but we also carry host and port to give wp idea where to connect.
@@ -119,7 +89,7 @@ impl Display for MemberSet {
 #[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
 pub struct Configuration {
     /// Unique id.
-    pub generation: SafekeeperGeneration,
+    pub generation: Generation,
     /// Current members of the configuration.
     pub members: MemberSet,
     /// Some means it is a joint conf.

libs/safekeeper_api/src/models.rs

@@ -282,18 +282,3 @@ pub struct TimelineTermBumpResponse {
 pub struct SafekeeperUtilization {
     pub timeline_count: u64,
 }
-
-/// pull_timeline request body.
-#[derive(Debug, Deserialize, Serialize)]
-pub struct PullTimelineRequest {
-    pub tenant_id: TenantId,
-    pub timeline_id: TimelineId,
-    pub http_hosts: Vec<String>,
-}
-
-#[derive(Debug, Serialize, Deserialize)]
-pub struct PullTimelineResponse {
-    // Donor safekeeper host
-    pub safekeeper_host: String,
-    // TODO: add more fields?
-}

libs/utils/Cargo.toml

@@ -24,10 +24,11 @@ diatomic-waker.workspace = true
 git-version.workspace = true
 hex = { workspace = true, features = ["serde"] }
 humantime.workspace = true
+inferno.workspace = true
 fail.workspace = true
 futures = { workspace = true }
 jsonwebtoken.workspace = true
-nix = { workspace = true, features = ["ioctl"] }
+nix = {workspace = true, features = [ "ioctl" ] }
 once_cell.workspace = true
 pin-project-lite.workspace = true
 regex.workspace = true
@@ -61,7 +62,6 @@ bytes.workspace = true
 criterion.workspace = true
 hex-literal.workspace = true
 camino-tempfile.workspace = true
-pprof.workspace = true
 serde_assert.workspace = true
 tokio = { workspace = true, features = ["test-util"] }
 

libs/utils/benches/README.md

@@ -1,26 +0,0 @@
-## Utils Benchmarks
-
-To run benchmarks:
-
-```sh
-# All benchmarks.
-cargo bench --package utils
-
-# Specific file.
-cargo bench --package utils --bench benchmarks
-
-# Specific benchmark.
-cargo bench --package utils --bench benchmarks warn_slow/enabled=true
-
-# List available benchmarks.
-cargo bench --package utils --benches -- --list
-
-# Generate flamegraph profiles using pprof-rs, profiling for 10 seconds.
-# Output in target/criterion/*/profile/flamegraph.svg.
-cargo bench --package utils --bench benchmarks warn_slow/enabled=true --profile-time 10
-```
-
-Additional charts and statistics are available in `target/criterion/report/index.html`.
-
-Benchmarks are automatically compared against the previous run. To compare against other runs, see
-`--baseline` and `--save-baseline`.

libs/utils/benches/benchmarks.rs

@@ -1,18 +1,5 @@
-use std::time::Duration;
-
-use criterion::{criterion_group, criterion_main, Bencher, Criterion};
-use pprof::criterion::{Output, PProfProfiler};
+use criterion::{criterion_group, criterion_main, Criterion};
 use utils::id;
-use utils::logging::warn_slow;
-
-// Register benchmarks with Criterion.
-criterion_group!(
-    name = benches;
-    config = Criterion::default().with_profiler(PProfProfiler::new(100, Output::Flamegraph(None)));
-    targets = bench_id_stringify,
-    bench_warn_slow,
-);
-criterion_main!(benches);
 
 pub fn bench_id_stringify(c: &mut Criterion) {
     // Can only use public methods.
@@ -29,31 +16,5 @@ pub fn bench_id_stringify(c: &mut Criterion) {
     });
 }
 
-pub fn bench_warn_slow(c: &mut Criterion) {
-    for enabled in [false, true] {
-        c.bench_function(&format!("warn_slow/enabled={enabled}"), |b| {
-            run_bench(b, enabled).unwrap()
-        });
-    }
-
-    // The actual benchmark.
-    fn run_bench(b: &mut Bencher, enabled: bool) -> anyhow::Result<()> {
-        const THRESHOLD: Duration = Duration::from_secs(1);
-
-        // Use a multi-threaded runtime to avoid thread parking overhead when yielding.
-        let runtime = tokio::runtime::Builder::new_multi_thread()
-            .enable_all()
-            .build()?;
-
-        // Test both with and without warn_slow, since we're essentially measuring Tokio scheduling
-        // performance too. Use a simple noop future that yields once, to avoid any scheduler fast
-        // paths for a ready future.
-        if enabled {
-            b.iter(|| runtime.block_on(warn_slow("ready", THRESHOLD, tokio::task::yield_now())));
-        } else {
-            b.iter(|| runtime.block_on(tokio::task::yield_now()));
-        }
-
-        Ok(())
-    }
-}
+criterion_group!(benches, bench_id_stringify);
+criterion_main!(benches);

libs/utils/src/bin_ser.rs

@@ -286,11 +286,6 @@ mod tests {
     const SHORT2_ENC_LE: &[u8] = &[8, 0, 0, 3, 7];
     const SHORT2_ENC_LE_TRAILING: &[u8] = &[8, 0, 0, 3, 7, 0xff, 0xff, 0xff];
 
-    #[derive(Debug, PartialEq, Eq, Serialize, Deserialize)]
-    struct NewTypeStruct(u32);
-    const NT1: NewTypeStruct = NewTypeStruct(414243);
-    const NT1_INNER: u32 = 414243;
-
     #[derive(Debug, PartialEq, Eq, Serialize, Deserialize)]
     pub struct LongMsg {
         pub tag: u8,
@@ -413,42 +408,4 @@ mod tests {
         let msg2 = LongMsg::des(&encoded).unwrap();
         assert_eq!(msg, msg2);
     }
-
-    #[test]
-    /// Ensure that newtype wrappers around u32 don't change the serialization format
-    fn be_nt() {
-        use super::BeSer;
-
-        assert_eq!(NT1.serialized_size().unwrap(), 4);
-
-        let msg = NT1;
-
-        let encoded = msg.ser().unwrap();
-        let expected = hex_literal::hex!("0006 5223");
-        assert_eq!(encoded, expected);
-
-        assert_eq!(encoded, NT1_INNER.ser().unwrap());
-
-        let msg2 = NewTypeStruct::des(&encoded).unwrap();
-        assert_eq!(msg, msg2);
-    }
-
-    #[test]
-    /// Ensure that newtype wrappers around u32 don't change the serialization format
-    fn le_nt() {
-        use super::LeSer;
-
-        assert_eq!(NT1.serialized_size().unwrap(), 4);
-
-        let msg = NT1;
-
-        let encoded = msg.ser().unwrap();
-        let expected = hex_literal::hex!("2352 0600");
-        assert_eq!(encoded, expected);
-
-        assert_eq!(encoded, NT1_INNER.ser().unwrap());
-
-        let msg2 = NewTypeStruct::des(&encoded).unwrap();
-        assert_eq!(msg, msg2);
-    }
 }

libs/utils/src/logging.rs

@@ -1,13 +1,9 @@
-use std::future::Future;
 use std::str::FromStr;
-use std::time::Duration;
 
 use anyhow::Context;
 use metrics::{IntCounter, IntCounterVec};
 use once_cell::sync::Lazy;
 use strum_macros::{EnumString, VariantNames};
-use tokio::time::Instant;
-use tracing::warn;
 
 /// Logs a critical error, similarly to `tracing::error!`. This will:
 ///
@@ -322,41 +318,6 @@ impl std::fmt::Debug for SecretString {
     }
 }
 
-/// Logs a periodic warning if a future is slow to complete.
-///
-/// This is performance-sensitive as it's used on the GetPage read path.
-#[inline]
-pub async fn warn_slow<O>(name: &str, threshold: Duration, f: impl Future<Output = O>) -> O {
-    // TODO: we unfortunately have to pin the future on the heap, since GetPage futures are huge and
-    // won't fit on the stack.
-    let mut f = Box::pin(f);
-
-    let started = Instant::now();
-    let mut attempt = 1;
-
-    loop {
-        // NB: use timeout_at() instead of timeout() to avoid an extra clock reading in the common
-        // case where the timeout doesn't fire.
-        let deadline = started + attempt * threshold;
-        if let Ok(output) = tokio::time::timeout_at(deadline, &mut f).await {
-            // NB: we check if we exceeded the threshold even if the timeout never fired, because
-            // scheduling or execution delays may cause the future to succeed even if it exceeds the
-            // timeout. This costs an extra unconditional clock reading, but seems worth it to avoid
-            // false negatives.
-            let elapsed = started.elapsed();
-            if elapsed >= threshold {
-                warn!("slow {name} completed after {:.3}s", elapsed.as_secs_f64());
-            }
-            return output;
-        }
-
-        let elapsed = started.elapsed().as_secs_f64();
-        warn!("slow {name} still running after {elapsed:.3}s",);
-
-        attempt += 1;
-    }
-}
-
 #[cfg(test)]
 mod tests {
     use metrics::{core::Opts, IntCounterVec};

libs/utils/src/shard.rs

@@ -117,10 +117,6 @@ impl TenantShardId {
         )
     }
 
-    pub fn range(&self) -> RangeInclusive<Self> {
-        RangeInclusive::new(*self, *self)
-    }
-
     pub fn shard_slug(&self) -> impl std::fmt::Display + '_ {
         ShardSlug(self)
     }

libs/wal_decoder/proto/interpreted_wal.proto

@@ -5,7 +5,6 @@ package interpreted_wal;
 message InterpretedWalRecords {
   repeated InterpretedWalRecord records = 1;
   optional uint64 next_record_lsn = 2;
-  optional uint64 raw_wal_start_lsn = 3;
 }
 
 message InterpretedWalRecord {

libs/wal_decoder/src/models.rs

@@ -60,11 +60,7 @@ pub struct InterpretedWalRecords {
     pub records: Vec<InterpretedWalRecord>,
     // Start LSN of the next record after the batch.
     // Note that said record may not belong to the current shard.
-    pub next_record_lsn: Lsn,
-    // Inclusive start LSN of the PG WAL from which the interpreted
-    // WAL records were extracted. Note that this is not necessarily the
-    // start LSN of the first interpreted record in the batch.
-    pub raw_wal_start_lsn: Option<Lsn>,
+    pub next_record_lsn: Option<Lsn>,
 }
 
 /// An interpreted Postgres WAL record, ready to be handled by the pageserver

libs/wal_decoder/src/wire_format.rs

@@ -167,8 +167,7 @@ impl TryFrom<InterpretedWalRecords> for proto::InterpretedWalRecords {
             .collect::<Result<Vec<_>, _>>()?;
         Ok(proto::InterpretedWalRecords {
             records,
-            next_record_lsn: Some(value.next_record_lsn.0),
-            raw_wal_start_lsn: value.raw_wal_start_lsn.map(|l| l.0),
+            next_record_lsn: value.next_record_lsn.map(|l| l.0),
         })
     }
 }
@@ -255,11 +254,7 @@ impl TryFrom<proto::InterpretedWalRecords> for InterpretedWalRecords {
 
         Ok(InterpretedWalRecords {
             records,
-            next_record_lsn: value
-                .next_record_lsn
-                .map(Lsn::from)
-                .expect("Always provided"),
-            raw_wal_start_lsn: value.raw_wal_start_lsn.map(Lsn::from),
+            next_record_lsn: value.next_record_lsn.map(Lsn::from),
         })
     }
 }

pageserver/client/src/mgmt_api.rs

@@ -477,26 +477,6 @@ impl Client {
         self.request(Method::POST, &uri, ()).await.map(|_| ())
     }
 
-    pub async fn timeline_download_heatmap_layers(
-        &self,
-        tenant_shard_id: TenantShardId,
-        timeline_id: TimelineId,
-        concurrency: Option<usize>,
-    ) -> Result<()> {
-        let mut path = reqwest::Url::parse(&format!(
-            "{}/v1/tenant/{}/timeline/{}/download_heatmap_layers",
-            self.mgmt_api_endpoint, tenant_shard_id, timeline_id
-        ))
-        .expect("Cannot build URL");
-
-        if let Some(concurrency) = concurrency {
-            path.query_pairs_mut()
-                .append_pair("concurrency", &format!("{}", concurrency));
-        }
-
-        self.request(Method::POST, path, ()).await.map(|_| ())
-    }
-
     pub async fn tenant_reset(&self, tenant_shard_id: TenantShardId) -> Result<()> {
         let uri = format!(
             "{}/v1/tenant/{}/reset",

pageserver/ctl/src/key.rs

@@ -345,7 +345,6 @@ impl AuxFileV2 {
                 AuxFileV2::Recognized("pg_logical/replorigin_checkpoint", hash)
             }
             (2, 1) => AuxFileV2::Recognized("pg_replslot/", hash),
-            (3, 1) => AuxFileV2::Recognized("pg_stat/pgstat.stat", hash),
             (1, 0xff) => AuxFileV2::OtherWithPrefix("pg_logical/", hash),
             (0xff, 0xff) => AuxFileV2::Other(hash),
             _ => return None,

pageserver/src/aux_file.rs

@@ -39,7 +39,6 @@ fn aux_hash_to_metadata_key(dir_level1: u8, dir_level2: u8, data: &[u8]) -> Key
 
 const AUX_DIR_PG_LOGICAL: u8 = 0x01;
 const AUX_DIR_PG_REPLSLOT: u8 = 0x02;
-const AUX_DIR_PG_STAT: u8 = 0x03;
 const AUX_DIR_PG_UNKNOWN: u8 = 0xFF;
 
 /// Encode the aux file into a fixed-size key.
@@ -54,7 +53,6 @@ const AUX_DIR_PG_UNKNOWN: u8 = 0xFF;
 /// * pg_logical/replorigin_checkpoint -> 0x0103
 /// * pg_logical/others -> 0x01FF
 /// * pg_replslot/ -> 0x0201
-/// * pg_stat/pgstat.stat -> 0x0301
 /// * others -> 0xFFFF
 ///
 /// If you add new AUX files to this function, please also add a test case to `test_encoding_portable`.
@@ -77,8 +75,6 @@ pub fn encode_aux_file_key(path: &str) -> Key {
         aux_hash_to_metadata_key(AUX_DIR_PG_LOGICAL, 0xFF, fname.as_bytes())
     } else if let Some(fname) = path.strip_prefix("pg_replslot/") {
         aux_hash_to_metadata_key(AUX_DIR_PG_REPLSLOT, 0x01, fname.as_bytes())
-    } else if let Some(fname) = path.strip_prefix("pg_stat/") {
-        aux_hash_to_metadata_key(AUX_DIR_PG_STAT, 0x01, fname.as_bytes())
     } else {
         if cfg!(debug_assertions) {
             warn!(

pageserver/src/basebackup.rs

@@ -264,31 +264,6 @@ where
     async fn send_tarball(mut self) -> Result<(), BasebackupError> {
         // TODO include checksum
 
-        // Construct the pg_control file from the persisted checkpoint and pg_control
-        // information. But we only add this to the tarball at the end, so that if the
-        // writing is interrupted half-way through, the resulting incomplete tarball will
-        // be missing the pg_control file, which prevents PostgreSQL from starting up on
-        // it. With proper error handling, you should never try to start up from an
-        // incomplete basebackup in the first place, of course, but this is a nice little
-        // extra safety measure.
-        let checkpoint_bytes = self
-            .timeline
-            .get_checkpoint(self.lsn, self.ctx)
-            .await
-            .context("failed to get checkpoint bytes")?;
-        let pg_control_bytes = self
-            .timeline
-            .get_control_file(self.lsn, self.ctx)
-            .await
-            .context("failed to get control bytes")?;
-        let (pg_control_bytes, system_identifier, was_shutdown) =
-            postgres_ffi::generate_pg_control(
-                &pg_control_bytes,
-                &checkpoint_bytes,
-                self.lsn,
-                self.timeline.pg_version,
-            )?;
-
         let lazy_slru_download = self.timeline.get_lazy_slru_download() && !self.full_backup;
 
         let pgversion = self.timeline.pg_version;
@@ -426,10 +401,6 @@ where
                 // In future we will not generate AUX record for "pg_logical/replorigin_checkpoint" at all,
                 // but now we should handle (skip) it for backward compatibility.
                 continue;
-            } else if path == "pg_stat/pgstat.stat" && !was_shutdown {
-                // Drop statistic in case of abnormal termination, i.e. if we're not starting from the exact LSN
-                // of a shutdown checkpoint.
-                continue;
             }
             let header = new_tar_header(&path, content.len() as u64)?;
             self.ar
@@ -491,9 +462,8 @@ where
             )))
         });
 
-        // Last, add the pg_control file and bootstrap WAL segment.
-        self.add_pgcontrol_file(pg_control_bytes, system_identifier)
-            .await?;
+        // Generate pg_control and bootstrap WAL segment.
+        self.add_pgcontrol_file().await?;
         self.ar
             .finish()
             .await
@@ -701,11 +671,7 @@ where
     // Add generated pg_control file and bootstrap WAL segment.
     // Also send zenith.signal file with extra bootstrap data.
     //
-    async fn add_pgcontrol_file(
-        &mut self,
-        pg_control_bytes: Bytes,
-        system_identifier: u64,
-    ) -> Result<(), BasebackupError> {
+    async fn add_pgcontrol_file(&mut self) -> Result<(), BasebackupError> {
         // add zenith.signal file
         let mut zenith_signal = String::new();
         if self.prev_record_lsn == Lsn(0) {
@@ -728,6 +694,24 @@ where
             .await
             .map_err(|e| BasebackupError::Client(e, "add_pgcontrol_file,zenith.signal"))?;
 
+        let checkpoint_bytes = self
+            .timeline
+            .get_checkpoint(self.lsn, self.ctx)
+            .await
+            .context("failed to get checkpoint bytes")?;
+        let pg_control_bytes = self
+            .timeline
+            .get_control_file(self.lsn, self.ctx)
+            .await
+            .context("failed get control bytes")?;
+
+        let (pg_control_bytes, system_identifier) = postgres_ffi::generate_pg_control(
+            &pg_control_bytes,
+            &checkpoint_bytes,
+            self.lsn,
+            self.timeline.pg_version,
+        )?;
+
         //send pg_control
         let header = new_tar_header("global/pg_control", pg_control_bytes.len() as u64)?;
         self.ar

pageserver/src/bin/pageserver.rs

@@ -134,7 +134,6 @@ fn main() -> anyhow::Result<()> {
     info!(?conf.virtual_file_io_engine, "starting with virtual_file IO engine");
     info!(?conf.virtual_file_io_mode, "starting with virtual_file IO mode");
     info!(?conf.wal_receiver_protocol, "starting with WAL receiver protocol");
-    info!(?conf.validate_wal_contiguity, "starting with WAL contiguity validation");
     info!(?conf.page_service_pipelining, "starting with page service pipelining config");
     info!(?conf.get_vectored_concurrent_io, "starting with get_vectored IO concurrency config");
 

pageserver/src/config.rs

@@ -197,10 +197,6 @@ pub struct PageServerConf {
     /// Enable read path debugging. If enabled, read key errors will print a backtrace of the layer
     /// files read.
     pub enable_read_path_debugging: bool,
-
-    /// Interpreted protocol feature: if enabled, validate that the logical WAL received from
-    /// safekeepers does not have gaps.
-    pub validate_wal_contiguity: bool,
 }
 
 /// Token for authentication to safekeepers
@@ -364,7 +360,6 @@ impl PageServerConf {
             page_service_pipelining,
             get_vectored_concurrent_io,
             enable_read_path_debugging,
-            validate_wal_contiguity,
         } = config_toml;
 
         let mut conf = PageServerConf {
@@ -451,7 +446,6 @@ impl PageServerConf {
             virtual_file_io_mode: virtual_file_io_mode.unwrap_or(virtual_file::IoMode::preferred()),
             no_sync: no_sync.unwrap_or(false),
             enable_read_path_debugging: enable_read_path_debugging.unwrap_or(false),
-            validate_wal_contiguity: validate_wal_contiguity.unwrap_or(false),
         };
 
         // ------------------------------------------------------------

pageserver/src/context.rs

@@ -98,7 +98,6 @@ pub struct RequestContext {
     download_behavior: DownloadBehavior,
     access_stats_behavior: AccessStatsBehavior,
     page_content_kind: PageContentKind,
-    read_path_debug: bool,
 }
 
 /// The kind of access to the page cache.
@@ -156,7 +155,6 @@ impl RequestContextBuilder {
                 download_behavior: DownloadBehavior::Download,
                 access_stats_behavior: AccessStatsBehavior::Update,
                 page_content_kind: PageContentKind::Unknown,
-                read_path_debug: false,
             },
         }
     }
@@ -170,7 +168,6 @@ impl RequestContextBuilder {
                 download_behavior: original.download_behavior,
                 access_stats_behavior: original.access_stats_behavior,
                 page_content_kind: original.page_content_kind,
-                read_path_debug: original.read_path_debug,
             },
         }
     }
@@ -194,11 +191,6 @@ impl RequestContextBuilder {
         self
     }
 
-    pub(crate) fn read_path_debug(mut self, b: bool) -> Self {
-        self.inner.read_path_debug = b;
-        self
-    }
-
     pub fn build(self) -> RequestContext {
         self.inner
     }
@@ -299,8 +291,4 @@ impl RequestContext {
     pub(crate) fn page_content_kind(&self) -> PageContentKind {
         self.page_content_kind
     }
-
-    pub(crate) fn read_path_debug(&self) -> bool {
-        self.read_path_debug
-    }
 }

pageserver/src/controller_upcall_client.rs

@@ -173,7 +173,6 @@ impl ControlPlaneGenerationsApi for ControllerUpcallClient {
                         listen_pg_port: m.postgres_port,
                         listen_http_addr: m.http_host,
                         listen_http_port: m.http_port,
-                        listen_https_port: None, // TODO: Support https.
                         availability_zone_id: az_id.expect("Checked above"),
                     })
                 }

pageserver/src/http/openapi_spec.yml

@@ -824,38 +824,6 @@ paths:
               schema:
                 $ref: "#/components/schemas/TenantConfigResponse"
 
-  /v1/tenant/{tenant_shard_id}/timeline/{timeline_id}/download_heatmap_layers:
-    parameters:
-      - name: tenant_shard_id
-        in: path
-        required: true
-        schema:
-          type: string
-      - name: timeline_id
-        in: path
-        required: true
-        schema:
-          type: string
-      - name: concurrency
-        description: Maximum number of concurrent downloads (capped at remote storage concurrency)
-        in: query
-        required: false
-        schema:
-          type: integer
-    post:
-      description: |
-        Download all layers in the specified timeline's heatmap. The `tenant_shard_id` parameter
-        may be used to target all shards of a tenant when the unsharded form is used, or a specific
-        tenant shard with the sharded form.
-      responses:
-        "200":
-          description: Success
-    delete:
-      description: Stop any on-going background downloads of heatmap layers for the specified timeline.
-      responses:
-        "200":
-          description: Success
-
   /v1/utilization:
     get:
       description: |
@@ -914,8 +882,6 @@ components:
               properties:
                 reason:
                   type: string
-        gc_blocking:
-          type: string
 
     TenantCreateRequest:
       allOf:
@@ -1117,9 +1083,6 @@ components:
         min_readable_lsn:
           type: string
           format: hex
-        latest_gc_cutoff_lsn:
-          type: string
-          format: hex
         applied_gc_cutoff_lsn:
           type: string
           format: hex

pageserver/src/http/routes.rs

@@ -68,7 +68,6 @@ use tokio_util::sync::CancellationToken;
 use tracing::*;
 
 use crate::config::PageServerConf;
-use crate::context::RequestContextBuilder;
 use crate::context::{DownloadBehavior, RequestContext};
 use crate::deletion_queue::DeletionQueueClient;
 use crate::pgdatadir_mapping::LsnForTimestamp;
@@ -1464,59 +1463,6 @@ async fn timeline_layer_scan_disposable_keys(
     )
 }
 
-async fn timeline_download_heatmap_layers_handler(
-    request: Request<Body>,
-    _cancel: CancellationToken,
-) -> Result<Response<Body>, ApiError> {
-    // Only used in the case where remote storage is not configured.
-    const DEFAULT_MAX_CONCURRENCY: usize = 100;
-    // A conservative default.
-    const DEFAULT_CONCURRENCY: usize = 16;
-
-    let tenant_shard_id: TenantShardId = parse_request_param(&request, "tenant_shard_id")?;
-    let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
-
-    let desired_concurrency =
-        parse_query_param(&request, "concurrency")?.unwrap_or(DEFAULT_CONCURRENCY);
-
-    check_permission(&request, Some(tenant_shard_id.tenant_id))?;
-
-    let state = get_state(&request);
-    let timeline =
-        active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
-            .await?;
-
-    let max_concurrency = get_config(&request)
-        .remote_storage_config
-        .as_ref()
-        .map(|c| c.concurrency_limit())
-        .unwrap_or(DEFAULT_MAX_CONCURRENCY);
-    let concurrency = std::cmp::min(max_concurrency, desired_concurrency);
-
-    timeline.start_heatmap_layers_download(concurrency).await?;
-
-    json_response(StatusCode::ACCEPTED, ())
-}
-
-async fn timeline_shutdown_download_heatmap_layers_handler(
-    request: Request<Body>,
-    _cancel: CancellationToken,
-) -> Result<Response<Body>, ApiError> {
-    let tenant_shard_id: TenantShardId = parse_request_param(&request, "tenant_shard_id")?;
-    let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
-
-    check_permission(&request, Some(tenant_shard_id.tenant_id))?;
-
-    let state = get_state(&request);
-    let timeline =
-        active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
-            .await?;
-
-    timeline.stop_and_drain_heatmap_layers_download().await;
-
-    json_response(StatusCode::OK, ())
-}
-
 async fn layer_download_handler(
     request: Request<Body>,
     _cancel: CancellationToken,
@@ -2395,7 +2341,6 @@ async fn timeline_checkpoint_handler(
                     match e {
                         CompactionError::ShuttingDown => ApiError::ShuttingDown,
                         CompactionError::Offload(e) => ApiError::InternalServerError(anyhow::anyhow!(e)),
-                        CompactionError::CollectKeySpaceError(e) => ApiError::InternalServerError(anyhow::anyhow!(e)),
                         CompactionError::Other(e) => ApiError::InternalServerError(e)
                     }
                 )?;
@@ -2573,30 +2518,14 @@ async fn deletion_queue_flush(
     }
 }
 
-async fn getpage_at_lsn_handler(
-    request: Request<Body>,
-    cancel: CancellationToken,
-) -> Result<Response<Body>, ApiError> {
-    getpage_at_lsn_handler_inner(false, request, cancel).await
-}
-
-async fn touchpage_at_lsn_handler(
-    request: Request<Body>,
-    cancel: CancellationToken,
-) -> Result<Response<Body>, ApiError> {
-    getpage_at_lsn_handler_inner(true, request, cancel).await
-}
-
 /// Try if `GetPage@Lsn` is successful, useful for manual debugging.
-async fn getpage_at_lsn_handler_inner(
-    touch: bool,
+async fn getpage_at_lsn_handler(
     request: Request<Body>,
     _cancel: CancellationToken,
 ) -> Result<Response<Body>, ApiError> {
     let tenant_shard_id: TenantShardId = parse_request_param(&request, "tenant_shard_id")?;
     let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
-    // Require pageserver admin permission for this API instead of only tenant-level token.
-    check_permission(&request, None)?;
+    check_permission(&request, Some(tenant_shard_id.tenant_id))?;
     let state = get_state(&request);
 
     struct Key(pageserver_api::key::Key);
@@ -2611,29 +2540,22 @@ async fn getpage_at_lsn_handler_inner(
 
     let key: Key = parse_query_param(&request, "key")?
         .ok_or_else(|| ApiError::BadRequest(anyhow!("missing 'key' query parameter")))?;
-    let lsn: Option<Lsn> = parse_query_param(&request, "lsn")?;
+    let lsn: Lsn = parse_query_param(&request, "lsn")?
+        .ok_or_else(|| ApiError::BadRequest(anyhow!("missing 'lsn' query parameter")))?;
 
     async {
         let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
-        // Enable read path debugging
-        let ctx = RequestContextBuilder::extend(&ctx).read_path_debug(true).build();
         let timeline = active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id).await?;
 
-        // Use last_record_lsn if no lsn is provided
-        let lsn = lsn.unwrap_or_else(|| timeline.get_last_record_lsn());
         let page = timeline.get(key.0, lsn, &ctx).await?;
 
-        if touch {
-            json_response(StatusCode::OK, ())
-        } else {
-            Result::<_, ApiError>::Ok(
-                Response::builder()
-                    .status(StatusCode::OK)
-                    .header(header::CONTENT_TYPE, "application/octet-stream")
-                    .body(hyper::Body::from(page))
-                    .unwrap(),
-            )
-        }
+        Result::<_, ApiError>::Ok(
+            Response::builder()
+                .status(StatusCode::OK)
+                .header(header::CONTENT_TYPE, "application/octet-stream")
+                .body(hyper::Body::from(page))
+                .unwrap(),
+        )
     }
     .instrument(info_span!("timeline_get", tenant_id = %tenant_shard_id.tenant_id, shard_id = %tenant_shard_id.shard_slug(), %timeline_id))
     .await
@@ -3704,14 +3626,6 @@ pub fn make_router(
             "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/layer",
             |r| api_handler(r, layer_map_info_handler),
         )
-        .post(
-            "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/download_heatmap_layers",
-            |r| api_handler(r, timeline_download_heatmap_layers_handler),
-        )
-        .delete(
-            "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/download_heatmap_layers",
-            |r| api_handler(r, timeline_shutdown_download_heatmap_layers_handler),
-        )
         .get(
             "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/layer/:layer_file_name",
             |r| api_handler(r, layer_download_handler),
@@ -3768,10 +3682,6 @@ pub fn make_router(
             "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/getpage",
             |r| testing_api_handler("getpage@lsn", r, getpage_at_lsn_handler),
         )
-        .get(
-            "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/touchpage",
-            |r| api_handler(r, touchpage_at_lsn_handler),
-        )
         .get(
             "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/keyspace",
             |r| api_handler(r, timeline_collect_keyspace),

pageserver/src/page_service.rs

@@ -34,13 +34,11 @@ use std::str::FromStr;
 use std::sync::Arc;
 use std::time::SystemTime;
 use std::time::{Duration, Instant};
-use strum_macros::IntoStaticStr;
 use tokio::io::{AsyncRead, AsyncWrite};
 use tokio::io::{AsyncWriteExt, BufWriter};
 use tokio::task::JoinHandle;
 use tokio_util::sync::CancellationToken;
 use tracing::*;
-use utils::logging::warn_slow;
 use utils::sync::gate::{Gate, GateGuard};
 use utils::sync::spsc_fold;
 use utils::{
@@ -83,9 +81,6 @@ use std::os::fd::AsRawFd;
 /// NB: this is a different value than [`crate::http::routes::ACTIVE_TENANT_TIMEOUT`].
 const ACTIVE_TENANT_TIMEOUT: Duration = Duration::from_millis(30000);
 
-/// Threshold at which to log a warning about slow GetPage requests.
-const WARN_SLOW_GETPAGE_THRESHOLD: Duration = Duration::from_secs(30);
-
 ///////////////////////////////////////////////////////////////////////////////
 
 pub struct Listener {
@@ -599,7 +594,6 @@ struct BatchedTestRequest {
 /// NB: we only hold [`timeline::handle::WeakHandle`] inside this enum,
 /// so that we don't keep the [`Timeline::gate`] open while the batch
 /// is being built up inside the [`spsc_fold`] (pagestream pipelining).
-#[derive(IntoStaticStr)]
 enum BatchedFeMessage {
     Exists {
         span: Span,
@@ -644,10 +638,6 @@ enum BatchedFeMessage {
 }
 
 impl BatchedFeMessage {
-    fn as_static_str(&self) -> &'static str {
-        self.into()
-    }
-
     fn observe_execution_start(&mut self, at: Instant) {
         match self {
             BatchedFeMessage::Exists { timer, .. }
@@ -1473,20 +1463,17 @@ impl PageServerHandler {
                 }
             };
 
-            let result = warn_slow(
-                msg.as_static_str(),
-                WARN_SLOW_GETPAGE_THRESHOLD,
-                self.pagesteam_handle_batched_message(
+            let err = self
+                .pagesteam_handle_batched_message(
                     pgb_writer,
                     msg,
                     io_concurrency.clone(),
                     &cancel,
                     protocol_version,
                     ctx,
-                ),
-            )
-            .await;
-            match result {
+                )
+                .await;
+            match err {
                 Ok(()) => {}
                 Err(e) => break e,
             }
@@ -1649,17 +1636,13 @@ impl PageServerHandler {
                             return Err(e);
                         }
                     };
-                    warn_slow(
-                        batch.as_static_str(),
-                        WARN_SLOW_GETPAGE_THRESHOLD,
-                        self.pagesteam_handle_batched_message(
-                            pgb_writer,
-                            batch,
-                            io_concurrency.clone(),
-                            &cancel,
-                            protocol_version,
-                            &ctx,
-                        ),
+                    self.pagesteam_handle_batched_message(
+                        pgb_writer,
+                        batch,
+                        io_concurrency.clone(),
+                        &cancel,
+                        protocol_version,
+                        &ctx,
                     )
                     .await?;
                 }
@@ -1816,13 +1799,6 @@ impl PageServerHandler {
                 .as_millis()
                 .to_string()
         });
-
-        info!(
-            "acquired lease for {} until {}",
-            lsn,
-            valid_until_str.as_deref().unwrap_or("<unknown>")
-        );
-
         let bytes = valid_until_str.as_ref().map(|x| x.as_bytes());
 
         pgb.write_message_noflush(&BeMessage::RowDescription(&[RowDescriptor::text_col(

pageserver/src/pgdatadir_mapping.rs

@@ -45,7 +45,7 @@ use std::ops::ControlFlow;
 use std::ops::Range;
 use strum::IntoEnumIterator;
 use tokio_util::sync::CancellationToken;
-use tracing::{debug, info, trace, warn};
+use tracing::{debug, trace, warn};
 use utils::bin_ser::DeserializeError;
 use utils::pausable_failpoint;
 use utils::{bin_ser::BeSer, lsn::Lsn};
@@ -2264,13 +2264,6 @@ impl DatadirModification<'_> {
                 self.tline.aux_file_size_estimator.on_add(content.len());
                 new_files.push((path, content));
             }
-            // Compute may request delete of old version of pgstat AUX file if new one exceeds size limit.
-            // Compute doesn't know if previous version of this file exists or not, so
-            // attempt to delete non-existing file can cause this message.
-            // To avoid false alarms, log it as info rather than warning.
-            (None, true) if path.starts_with("pg_stat/") => {
-                info!("removing non-existing pg_stat file: {}", path)
-            }
             (None, true) => warn!("removing non-existing aux file: {}", path),
         }
         let new_val = aux_file::encode_file_value(&new_files)?;

pageserver/src/tenant.rs

@@ -3101,9 +3101,6 @@ impl Tenant {
                 if let Some(queue) = queue {
                     outcome = queue
                         .iteration(cancel, ctx, &self.gc_block, &timeline)
-                        .instrument(
-                            info_span!("gc_compact_timeline", timeline_id = %timeline.timeline_id),
-                        )
                         .await?;
                 }
             }
@@ -3150,12 +3147,6 @@ impl Tenant {
             // Offload failures don't trip the circuit breaker, since they're cheap to retry and
             // shouldn't block compaction.
             CompactionError::Offload(_) => {}
-            CompactionError::CollectKeySpaceError(err) => {
-                self.compaction_circuit_breaker
-                    .lock()
-                    .unwrap()
-                    .fail(&CIRCUIT_BREAKERS_BROKEN, err);
-            }
             CompactionError::Other(err) => {
                 self.compaction_circuit_breaker
                     .lock()
@@ -7855,6 +7846,18 @@ mod tests {
             }
 
             tline.freeze_and_flush().await?;
+            // Force layers to L1
+            tline
+                .compact(
+                    &cancel,
+                    {
+                        let mut flags = EnumSet::new();
+                        flags.insert(CompactFlags::ForceL0Compaction);
+                        flags
+                    },
+                    &ctx,
+                )
+                .await?;
 
             if iter % 5 == 0 {
                 let (_, before_delta_file_accessed) =
@@ -7867,6 +7870,7 @@ mod tests {
                             let mut flags = EnumSet::new();
                             flags.insert(CompactFlags::ForceImageLayerCreation);
                             flags.insert(CompactFlags::ForceRepartition);
+                            flags.insert(CompactFlags::ForceL0Compaction);
                             flags
                         },
                         &ctx,
@@ -8313,6 +8317,8 @@ mod tests {
 
         let cancel = CancellationToken::new();
 
+        // Image layer creation happens on the disk_consistent_lsn so we need to force set it now.
+        tline.force_set_disk_consistent_lsn(Lsn(0x40));
         tline
             .compact(
                 &cancel,
@@ -8326,8 +8332,7 @@ mod tests {
             )
             .await
             .unwrap();
-
-        // Image layers are created at last_record_lsn
+        // Image layers are created at repartition LSN
         let images = tline
             .inspect_image_layers(Lsn(0x40), &ctx, io_concurrency.clone())
             .await

pageserver/src/tenant/config.rs

@@ -693,15 +693,16 @@ impl TryFrom<&'_ models::TenantConfig> for TenantConfOpt {
 /// This is a conversion from our internal tenant config object to the one used
 /// in external APIs.
 impl From<TenantConfOpt> for models::TenantConfig {
-    // TODO(vlad): These are now the same, but they have different serialization logic.
-    // Can we merge them?
     fn from(value: TenantConfOpt) -> Self {
+        fn humantime(d: Duration) -> String {
+            format!("{}s", d.as_secs())
+        }
         Self {
             checkpoint_distance: value.checkpoint_distance,
-            checkpoint_timeout: value.checkpoint_timeout,
+            checkpoint_timeout: value.checkpoint_timeout.map(humantime),
             compaction_algorithm: value.compaction_algorithm,
             compaction_target_size: value.compaction_target_size,
-            compaction_period: value.compaction_period,
+            compaction_period: value.compaction_period.map(humantime),
             compaction_threshold: value.compaction_threshold,
             compaction_upper_limit: value.compaction_upper_limit,
             compaction_l0_first: value.compaction_l0_first,
@@ -710,23 +711,24 @@ impl From<TenantConfOpt> for models::TenantConfig {
             l0_flush_stall_threshold: value.l0_flush_stall_threshold,
             l0_flush_wait_upload: value.l0_flush_wait_upload,
             gc_horizon: value.gc_horizon,
-            gc_period: value.gc_period,
+            gc_period: value.gc_period.map(humantime),
             image_creation_threshold: value.image_creation_threshold,
-            pitr_interval: value.pitr_interval,
-            walreceiver_connect_timeout: value.walreceiver_connect_timeout,
-            lagging_wal_timeout: value.lagging_wal_timeout,
+            pitr_interval: value.pitr_interval.map(humantime),
+            walreceiver_connect_timeout: value.walreceiver_connect_timeout.map(humantime),
+            lagging_wal_timeout: value.lagging_wal_timeout.map(humantime),
             max_lsn_wal_lag: value.max_lsn_wal_lag,
             eviction_policy: value.eviction_policy,
             min_resident_size_override: value.min_resident_size_override,
             evictions_low_residence_duration_metric_threshold: value
-                .evictions_low_residence_duration_metric_threshold,
-            heatmap_period: value.heatmap_period,
+                .evictions_low_residence_duration_metric_threshold
+                .map(humantime),
+            heatmap_period: value.heatmap_period.map(humantime),
             lazy_slru_download: value.lazy_slru_download,
             timeline_get_throttle: value.timeline_get_throttle,
             image_layer_creation_check_threshold: value.image_layer_creation_check_threshold,
             image_creation_preempt_threshold: value.image_creation_preempt_threshold,
-            lsn_lease_length: value.lsn_lease_length,
-            lsn_lease_length_for_ts: value.lsn_lease_length_for_ts,
+            lsn_lease_length: value.lsn_lease_length.map(humantime),
+            lsn_lease_length_for_ts: value.lsn_lease_length_for_ts.map(humantime),
             timeline_offloading: value.timeline_offloading,
             wal_receiver_protocol_override: value.wal_receiver_protocol_override,
             rel_size_v2_enabled: value.rel_size_v2_enabled,
@@ -758,10 +760,29 @@ mod tests {
         assert_eq!(small_conf, serde_json::from_str(&json_form).unwrap());
     }
 
+    #[test]
+    fn test_try_from_models_tenant_config_err() {
+        let tenant_config = models::TenantConfig {
+            lagging_wal_timeout: Some("5a".to_string()),
+            ..TenantConfig::default()
+        };
+
+        let tenant_conf_opt = TenantConfOpt::try_from(&tenant_config);
+
+        assert!(
+            tenant_conf_opt.is_err(),
+            "Suceeded to convert TenantConfig to TenantConfOpt"
+        );
+
+        let expected_error_str =
+            "lagging_wal_timeout: invalid value: string \"5a\", expected a duration";
+        assert_eq!(tenant_conf_opt.unwrap_err().to_string(), expected_error_str);
+    }
+
     #[test]
     fn test_try_from_models_tenant_config_success() {
         let tenant_config = models::TenantConfig {
-            lagging_wal_timeout: Some(Duration::from_secs(5)),
+            lagging_wal_timeout: Some("5s".to_string()),
             ..TenantConfig::default()
         };
 

pageserver/src/tenant/layer_map.rs

@@ -570,8 +570,12 @@ impl LayerMap {
         self.historic.iter()
     }
 
+    pub fn riter_historic_layers(&self) -> impl '_ + Iterator<Item = Arc<PersistentLayerDesc>> {
+        self.historic.riter()
+    }
+
     /// Get a ref counted pointer for the first in memory layer that matches the provided predicate.
-    pub fn find_in_memory_layer<Pred>(&self, mut pred: Pred) -> Option<Arc<InMemoryLayer>>
+    pub(crate) fn find_in_memory_layer<Pred>(&self, mut pred: Pred) -> Option<Arc<InMemoryLayer>>
     where
         Pred: FnMut(&Arc<InMemoryLayer>) -> bool,
     {
@@ -900,6 +904,24 @@ impl LayerMap {
         Ok(())
     }
 
+    /// Efficiency: this is a single btreemap walk to the end of the map in the common case where
+    /// we are queried for image layers after the start of an ephemeral layer.  In the general case
+    /// where we are called with some arbitrary LSN, this function is O(N) -- so don't use it like that.
+    pub(crate) fn get_newest_image_after(&self, lsn: Lsn) -> Option<Arc<PersistentLayerDesc>> {
+        // TODO: an efficient equivalent, this is a crude placeholder
+        for layer in self.riter_historic_layers() {
+            if !layer.is_delta() && layer.image_layer_lsn() >= lsn {
+                return Some(layer);
+            }
+
+            if layer.lsn_range.start < lsn {
+                // We are past the layers that could possibly intersect with the requested bound
+                break;
+            }
+        }
+        None
+    }
+
     /// `read_points` represent the tip of a timeline and any branch points, i.e. the places
     /// where we expect to serve reads.
     ///

pageserver/src/tenant/layer_map/historic_layer_coverage.rs

@@ -509,6 +509,18 @@ impl<Value: Clone> BufferedHistoricLayerCoverage<Value> {
         self.layers.values().cloned()
     }
 
+    /// Iterate all the layers in reverse order (newest LSNs first)
+    pub fn riter(&self) -> impl '_ + Iterator<Item = Value> {
+        // NOTE we can actually perform this without rebuilding,
+        //      but it's not necessary for now.
+        if !self.buffer.is_empty() {
+            panic!("rebuild pls")
+        }
+
+        // TODO: is cloned() really needed?
+        self.layers.values().rev().cloned()
+    }
+
     /// Return a reference to a queryable map, assuming all updates
     /// have already been processed using self.rebuild()
     pub fn get(&self) -> anyhow::Result<&HistoricLayerCoverage<Value>> {

pageserver/src/tenant/storage_layer/delta_layer.rs

@@ -51,7 +51,8 @@ use camino::{Utf8Path, Utf8PathBuf};
 use futures::StreamExt;
 use itertools::Itertools;
 use pageserver_api::config::MaxVectoredReadBytes;
-use pageserver_api::key::{Key, DBDIR_KEY, KEY_SIZE};
+use pageserver_api::key::DBDIR_KEY;
+use pageserver_api::key::{Key, KEY_SIZE};
 use pageserver_api::keyspace::KeySpace;
 use pageserver_api::models::ImageCompressionAlgorithm;
 use pageserver_api::shard::TenantShardId;
@@ -966,10 +967,7 @@ impl DeltaLayerInner {
                 .as_slice()
                 .iter()
                 .filter_map(|(_, blob_meta)| {
-                    if blob_meta.key.is_rel_dir_key()
-                        || blob_meta.key == DBDIR_KEY
-                        || blob_meta.key.is_aux_file_key()
-                    {
+                    if blob_meta.key.is_rel_dir_key() || blob_meta.key == DBDIR_KEY {
                         // The size of values for these keys is unbounded and can
                         // grow very large in pathological cases.
                         None

pageserver/src/tenant/storage_layer/image_layer.rs

@@ -48,7 +48,8 @@ use camino::{Utf8Path, Utf8PathBuf};
 use hex;
 use itertools::Itertools;
 use pageserver_api::config::MaxVectoredReadBytes;
-use pageserver_api::key::{Key, DBDIR_KEY, KEY_SIZE};
+use pageserver_api::key::DBDIR_KEY;
+use pageserver_api::key::{Key, KEY_SIZE};
 use pageserver_api::keyspace::KeySpace;
 use pageserver_api::shard::{ShardIdentity, TenantShardId};
 use pageserver_api::value::Value;
@@ -602,10 +603,7 @@ impl ImageLayerInner {
                     .as_slice()
                     .iter()
                     .filter_map(|(_, blob_meta)| {
-                        if blob_meta.key.is_rel_dir_key()
-                            || blob_meta.key == DBDIR_KEY
-                            || blob_meta.key.is_aux_file_key()
-                        {
+                        if blob_meta.key.is_rel_dir_key() || blob_meta.key == DBDIR_KEY {
                             // The size of values for these keys is unbounded and can
                             // grow very large in pathological cases.
                             None

pageserver/src/tenant/tasks.rs

@@ -287,7 +287,6 @@ fn log_compaction_error(
     sleep_duration: Duration,
     task_cancelled: bool,
 ) {
-    use crate::pgdatadir_mapping::CollectKeySpaceError;
     use crate::tenant::upload_queue::NotInitialized;
     use crate::tenant::PageReconstructError;
     use CompactionError::*;
@@ -295,8 +294,6 @@ fn log_compaction_error(
     let level = match err {
         ShuttingDown => return,
         Offload(_) => Level::ERROR,
-        CollectKeySpaceError(CollectKeySpaceError::Cancelled) => Level::INFO,
-        CollectKeySpaceError(_) => Level::ERROR,
         _ if task_cancelled => Level::INFO,
         Other(err) => {
             let root_cause = err.root_cause();

pageserver/src/tenant/timeline.rs

@@ -4,7 +4,6 @@ pub mod delete;
 pub(crate) mod detach_ancestor;
 mod eviction_task;
 pub(crate) mod handle;
-mod heatmap_layers_downloader;
 pub(crate) mod import_pgdata;
 mod init;
 pub mod layer_manager;
@@ -22,7 +21,6 @@ use chrono::{DateTime, Utc};
 use compaction::CompactionOutcome;
 use enumset::EnumSet;
 use fail::fail_point;
-use futures::FutureExt;
 use futures::{stream::FuturesUnordered, StreamExt};
 use handle::ShardTimelineId;
 use layer_manager::Shutdown;
@@ -469,10 +467,6 @@ pub struct Timeline {
     pub(crate) page_trace: ArcSwapOption<Sender<PageTraceEvent>>,
 
     previous_heatmap: ArcSwapOption<PreviousHeatmap>,
-
-    /// May host a background Tokio task which downloads all the layers from the current
-    /// heatmap on demand.
-    heatmap_layers_downloader: Mutex<Option<heatmap_layers_downloader::HeatmapLayersDownloader>>,
 }
 
 pub(crate) enum PreviousHeatmap {
@@ -1299,7 +1293,7 @@ impl Timeline {
         reconstruct_state: &mut ValuesReconstructState,
         ctx: &RequestContext,
     ) -> Result<BTreeMap<Key, Result<Bytes, PageReconstructError>>, GetVectoredError> {
-        let read_path = if self.conf.enable_read_path_debugging || ctx.read_path_debug() {
+        let read_path = if self.conf.enable_read_path_debugging {
             Some(ReadPath::new(keyspace.clone(), lsn))
         } else {
             None
@@ -1882,7 +1876,7 @@ impl Timeline {
         // Signal compaction failure to avoid L0 flush stalls when it's broken.
         match result {
             Ok(_) => self.compaction_failed.store(false, AtomicOrdering::Relaxed),
-            Err(CompactionError::Other(_)) | Err(CompactionError::CollectKeySpaceError(_)) => {
+            Err(CompactionError::Other(_)) => {
                 self.compaction_failed.store(true, AtomicOrdering::Relaxed)
             }
             // Don't change the current value on offload failure or shutdown. We don't want to
@@ -2045,11 +2039,6 @@ impl Timeline {
         tracing::debug!("Cancelling CancellationToken");
         self.cancel.cancel();
 
-        // If we have a background task downloading heatmap layers stop it.
-        // The background downloads are sensitive to timeline cancellation (done above),
-        // so the drain will be immediate.
-        self.stop_and_drain_heatmap_layers_download().await;
-
         // Ensure Prevent new page service requests from starting.
         self.handles.shutdown();
 
@@ -2763,8 +2752,6 @@ impl Timeline {
                 page_trace: Default::default(),
 
                 previous_heatmap: ArcSwapOption::from_pointee(previous_heatmap),
-
-                heatmap_layers_downloader: Mutex::new(None),
             };
 
             result.repartition_threshold =
@@ -2874,7 +2861,6 @@ impl Timeline {
                 auth_token: crate::config::SAFEKEEPER_AUTH_TOKEN.get().cloned(),
                 availability_zone: self.conf.availability_zone.clone(),
                 ingest_batch_size: self.conf.ingest_batch_size,
-                validate_wal_contiguity: self.conf.validate_wal_contiguity,
             },
             broker_client,
             ctx,
@@ -3789,6 +3775,8 @@ impl Timeline {
         let mut completed_keyspace = KeySpace::default();
         let mut image_covered_keyspace = KeySpaceRandomAccum::new();
 
+        let mut in_memory_layers_considered = Vec::new();
+
         // Prevent GC from progressing while visiting the current timeline.
         // If we are GC-ing because a new image layer was added while traversing
         // the timeline, then it will remove layers that are required for fulfilling
@@ -3824,12 +3812,34 @@ impl Timeline {
 
                 let in_memory_layer = layers.find_in_memory_layer(|l| {
                     let start_lsn = l.get_lsn_range().start;
-                    cont_lsn > start_lsn
+                    !in_memory_layers_considered.contains(&start_lsn) && cont_lsn > start_lsn
                 });
 
                 match in_memory_layer {
                     Some(l) => {
-                        let lsn_range = l.get_lsn_range().start..cont_lsn;
+                        in_memory_layers_considered.push(l.get_lsn_range().start);
+
+                        // Search for image layers that overlap with the in-memory layer: this is rare but permitted, and
+                        // we must bound the `lsn_range` of this layer to avoid skipping past the image layer.
+                        // TODO: a narrower search that only hits on image layers matching `unmapped_keyspace`
+                        let lsn_range = if let Some(image) =
+                            layers.get_newest_image_after(l.get_lsn_range().start)
+                        {
+                            // Note that this does not guarantee serving a read from an image layer, just that we will
+                            // not skip considering thge image layer in our Fringe.  We can still end up doing walredo work
+                            // in spite of the presence of an image layer, if the inmemory layers we visit contain enough
+                            // information to fully construct a page.  For example:
+                            //  - ephemeral layer contains I1, D1, D2, <LSN X>
+                            //  - image layer at LSN X contains image equal to I2
+                            //  - we will end up doing a walredo of I1+D1+D2, rather than reading from the image layer
+                            //
+                            //  This is not a problem for correctness, and is rare enough that the wasted time doing walredo
+                            //  doesn't matter.
+                            image.get_lsn_range().start + 1..cont_lsn
+                        } else {
+                            l.get_lsn_range().start..cont_lsn
+                        };
+
                         fringe.update(
                             ReadableLayer::InMemoryLayer(l),
                             unmapped_keyspace.clone(),
@@ -4606,10 +4616,7 @@ impl Timeline {
             ));
         }
 
-        let (dense_ks, sparse_ks) = self
-            .collect_keyspace(lsn, ctx)
-            .await
-            .map_err(CompactionError::CollectKeySpaceError)?;
+        let (dense_ks, sparse_ks) = self.collect_keyspace(lsn, ctx).await?;
         let dense_partitioning = dense_ks.partition(&self.shard_identity, partition_size);
         let sparse_partitioning = SparseKeyPartitioning {
             parts: vec![sparse_ks],
@@ -5130,26 +5137,20 @@ impl Timeline {
                     // image layer generation taking too long time and blocking L0 compaction. So in this
                     // mode, we also inspect the current number of L0 layers and skip image layer generation
                     // if there are too many of them.
+                    let num_of_l0_layers = {
+                        let layers = self.layers.read().await;
+                        layers.layer_map()?.level0_deltas().len()
+                    };
                     let image_preempt_threshold = self.get_image_creation_preempt_threshold()
                         * self.get_compaction_threshold();
-                    // TODO: currently we do not respect `get_image_creation_preempt_threshold` and always yield
-                    // when there is a single timeline with more than L0 threshold L0 layers. As long as the
-                    // `get_image_creation_preempt_threshold` is set to a value greater than 0, we will yield for L0 compaction.
-                    if image_preempt_threshold != 0 {
-                        let should_yield = self
-                            .l0_compaction_trigger
-                            .notified()
-                            .now_or_never()
-                            .is_some();
-                        if should_yield {
-                            tracing::info!(
-                                "preempt image layer generation at {lsn} when processing partition {}..{}: too many L0 layers",
-                                partition.start().unwrap(), partition.end().unwrap()
-                            );
-                            last_partition_processed = Some(partition.clone());
-                            all_generated = false;
-                            break;
-                        }
+                    if image_preempt_threshold != 0 && num_of_l0_layers >= image_preempt_threshold {
+                        tracing::info!(
+                        "preempt image layer generation at {lsn} when processing partition {}..{}: too many L0 layers {}",
+                        partition.start().unwrap(), partition.end().unwrap(), num_of_l0_layers
+                    );
+                        last_partition_processed = Some(partition.clone());
+                        all_generated = false;
+                        break;
                     }
                 }
             }
@@ -5178,16 +5179,14 @@ impl Timeline {
             .map(|l| l.metadata().file_size)
             .sum::<u64>();
 
-        if !image_layers.is_empty() {
-            info!(
-                "created {} image layers ({} bytes) in {}s, processed {} out of {} partitions",
-                image_layers.len(),
-                total_layer_size,
-                duration.as_secs_f64(),
-                partition_processed,
-                total_partitions
-            );
-        }
+        info!(
+            "created {} image layers ({} bytes) in {}s, processed {} out of {} partitions",
+            image_layers.len(),
+            total_layer_size,
+            duration.as_secs_f64(),
+            partition_processed,
+            total_partitions
+        );
 
         Ok((
             image_layers,
@@ -5330,8 +5329,6 @@ pub(crate) enum CompactionError {
     #[error("Failed to offload timeline: {0}")]
     Offload(OffloadError),
     /// Compaction cannot be done right now; page reconstruction and so on.
-    #[error("Failed to collect keyspace: {0}")]
-    CollectKeySpaceError(CollectKeySpaceError),
     #[error(transparent)]
     Other(anyhow::Error),
 }
@@ -5345,6 +5342,12 @@ impl From<OffloadError> for CompactionError {
     }
 }
 
+impl CompactionError {
+    pub fn is_cancelled(&self) -> bool {
+        matches!(self, CompactionError::ShuttingDown)
+    }
+}
+
 impl From<CollectKeySpaceError> for CompactionError {
     fn from(err: CollectKeySpaceError) -> Self {
         match err {
@@ -6609,7 +6612,7 @@ impl TimelineWriter<'_> {
 
         if let Some(wait_threshold) = wait_threshold {
             if l0_count >= wait_threshold {
-                debug!("layer roll waiting for flush due to compaction backpressure at {l0_count} L0 layers");
+                info!("layer roll waiting for flush due to compaction backpressure at {l0_count} L0 layers");
                 self.tl.wait_flush_completion(flush_id).await?;
             }
         }

pageserver/src/tenant/timeline/compaction.rs

@@ -11,8 +11,7 @@ use std::sync::Arc;
 use super::layer_manager::LayerManager;
 use super::{
     CompactFlags, CompactOptions, CreateImageLayersError, DurationRecorder, GetVectoredError,
-    ImageLayerCreationMode, LastImageLayerCreationStatus, PageReconstructError, RecordedDuration,
-    Timeline,
+    ImageLayerCreationMode, LastImageLayerCreationStatus, RecordedDuration, Timeline,
 };
 
 use anyhow::{anyhow, bail, Context};
@@ -26,13 +25,12 @@ use pageserver_api::models::CompactInfoResponse;
 use pageserver_api::shard::{ShardCount, ShardIdentity, TenantShardId};
 use serde::Serialize;
 use tokio_util::sync::CancellationToken;
-use tracing::{debug, error, info, info_span, trace, warn, Instrument};
+use tracing::{debug, info, info_span, trace, warn, Instrument};
 use utils::critical;
 use utils::id::TimelineId;
 
 use crate::context::{AccessStatsBehavior, RequestContext, RequestContextBuilder};
 use crate::page_cache;
-use crate::pgdatadir_mapping::CollectKeySpaceError;
 use crate::statvfs::Statvfs;
 use crate::tenant::checks::check_valid_layermap;
 use crate::tenant::gc_block::GcBlock;
@@ -303,12 +301,18 @@ impl GcCompactionQueue {
                         let mut guard = self.inner.lock().unwrap();
                         guard.gc_guards.insert(id, gc_guard);
                     }
-                    let _ = timeline.compact_with_options(cancel, options, ctx).await?;
+                    let _ = timeline
+                        .compact_with_options(cancel, options, ctx)
+                        .instrument(info_span!("scheduled_compact_timeline", %timeline.timeline_id))
+                        .await?;
                     self.notify_and_unblock(id);
                 }
             }
             GcCompactionQueueItem::SubCompactionJob(options) => {
-                let _ = timeline.compact_with_options(cancel, options, ctx).await?;
+                let _ = timeline
+                    .compact_with_options(cancel, options, ctx)
+                    .instrument(info_span!("scheduled_compact_timeline", %timeline.timeline_id))
+                    .await?;
             }
             GcCompactionQueueItem::Notify(id) => {
                 self.notify_and_unblock(id);
@@ -688,6 +692,21 @@ impl Timeline {
 
         // Define partitioning schema if needed
 
+        let l0_l1_boundary_lsn = {
+            // We do the repartition on the L0-L1 boundary. All data below the boundary
+            // are compacted by L0 with low read amplification, thus making the `repartition`
+            // function run fast.
+            let guard = self.layers.read().await;
+            let l0_min_lsn = guard
+                .layer_map()?
+                .level0_deltas()
+                .iter()
+                .map(|l| l.get_lsn_range().start)
+                .min()
+                .unwrap_or(self.get_disk_consistent_lsn());
+            l0_min_lsn.max(self.get_ancestor_lsn())
+        };
+
         // 1. L0 Compact
         let l0_outcome = {
             let timer = self.metrics.compact_time_histo.start_timer();
@@ -714,87 +733,86 @@ impl Timeline {
             return Ok(CompactionOutcome::YieldForL0);
         }
 
-        // 2. Repartition and create image layers if necessary
-        match self
-            .repartition(
-                self.get_last_record_lsn(),
-                self.get_compaction_target_size(),
-                options.flags,
-                ctx,
-            )
-            .await
-        {
-            Ok(((dense_partitioning, sparse_partitioning), lsn)) => {
-                // Disables access_stats updates, so that the files we read remain candidates for eviction after we're done with them
-                let image_ctx = RequestContextBuilder::extend(ctx)
-                    .access_stats_behavior(AccessStatsBehavior::Skip)
-                    .build();
+        if l0_l1_boundary_lsn < self.partitioning.read().1 {
+            // We never go backwards when repartition and create image layers.
+            info!("skipping image layer generation because repartition LSN is greater than L0-L1 boundary LSN.");
+        } else {
+            // 2. Repartition and create image layers if necessary
+            match self
+                .repartition(
+                    l0_l1_boundary_lsn,
+                    self.get_compaction_target_size(),
+                    options.flags,
+                    ctx,
+                )
+                .await
+            {
+                Ok(((dense_partitioning, sparse_partitioning), lsn)) => {
+                    // Disables access_stats updates, so that the files we read remain candidates for eviction after we're done with them
+                    let image_ctx = RequestContextBuilder::extend(ctx)
+                        .access_stats_behavior(AccessStatsBehavior::Skip)
+                        .build();
 
-                let mut partitioning = dense_partitioning;
-                partitioning
-                    .parts
-                    .extend(sparse_partitioning.into_dense().parts);
+                    let mut partitioning = dense_partitioning;
+                    partitioning
+                        .parts
+                        .extend(sparse_partitioning.into_dense().parts);
 
-                // 3. Create new image layers for partitions that have been modified "enough".
-                let (image_layers, outcome) = self
-                    .create_image_layers(
-                        &partitioning,
-                        lsn,
-                        if options
-                            .flags
-                            .contains(CompactFlags::ForceImageLayerCreation)
-                        {
-                            ImageLayerCreationMode::Force
-                        } else {
-                            ImageLayerCreationMode::Try
-                        },
-                        &image_ctx,
-                        self.last_image_layer_creation_status
-                            .load()
-                            .as_ref()
-                            .clone(),
-                        !options.flags.contains(CompactFlags::NoYield),
-                    )
-                    .await
-                    .inspect_err(|err| {
-                        if let CreateImageLayersError::GetVectoredError(
-                            GetVectoredError::MissingKey(_),
-                        ) = err
-                        {
-                            critical!("missing key during compaction: {err:?}");
-                        }
-                    })?;
+                    // 3. Create new image layers for partitions that have been modified "enough".
+                    let (image_layers, outcome) = self
+                        .create_image_layers(
+                            &partitioning,
+                            lsn,
+                            if options
+                                .flags
+                                .contains(CompactFlags::ForceImageLayerCreation)
+                            {
+                                ImageLayerCreationMode::Force
+                            } else {
+                                ImageLayerCreationMode::Try
+                            },
+                            &image_ctx,
+                            self.last_image_layer_creation_status
+                                .load()
+                                .as_ref()
+                                .clone(),
+                            !options.flags.contains(CompactFlags::NoYield),
+                        )
+                        .await
+                        .inspect_err(|err| {
+                            if let CreateImageLayersError::GetVectoredError(
+                                GetVectoredError::MissingKey(_),
+                            ) = err
+                            {
+                                critical!("missing key during compaction: {err:?}");
+                            }
+                        })?;
 
-                self.last_image_layer_creation_status
-                    .store(Arc::new(outcome.clone()));
+                    self.last_image_layer_creation_status
+                        .store(Arc::new(outcome.clone()));
 
-                self.upload_new_image_layers(image_layers)?;
-                if let LastImageLayerCreationStatus::Incomplete { .. } = outcome {
-                    // Yield and do not do any other kind of compaction.
-                    info!("skipping shard ancestor compaction due to pending image layer generation tasks (preempted by L0 compaction).");
-                    return Ok(CompactionOutcome::YieldForL0);
+                    self.upload_new_image_layers(image_layers)?;
+                    if let LastImageLayerCreationStatus::Incomplete { .. } = outcome {
+                        // Yield and do not do any other kind of compaction.
+                        info!("skipping shard ancestor compaction due to pending image layer generation tasks (preempted by L0 compaction).");
+                        return Ok(CompactionOutcome::YieldForL0);
+                    }
                 }
-            }
-
-            // Suppress errors when cancelled.
-            Err(_) if self.cancel.is_cancelled() => {}
-            Err(CompactionError::ShuttingDown) => {}
-
-            // Alert on critical errors that indicate data corruption.
-            Err(
-                err @ CompactionError::CollectKeySpaceError(
-                    CollectKeySpaceError::Decode(_)
-                    | CollectKeySpaceError::PageRead(
-                        PageReconstructError::MissingKey(_) | PageReconstructError::WalRedo(_),
-                    ),
-                ),
-            ) => critical!("could not compact, repartitioning keyspace failed: {err:?}"),
-
-            // Log other errors. No partitioning? This is normal, if the timeline was just created
-            // as an empty timeline. Also in unit tests, when we use the timeline as a simple
-            // key-value store, ignoring the datadir layout. Log the error but continue.
-            Err(err) => error!("could not compact, repartitioning keyspace failed: {err:?}"),
-        };
+                Err(err) => {
+                    // no partitioning? This is normal, if the timeline was just created
+                    // as an empty timeline. Also in unit tests, when we use the timeline
+                    // as a simple key-value store, ignoring the datadir layout. Log the
+                    // error but continue.
+                    //
+                    // Suppress error when it's due to cancellation
+                    if !self.cancel.is_cancelled() && !err.is_cancelled() {
+                        tracing::error!(
+                            "could not compact, repartitioning keyspace failed: {err:?}"
+                        );
+                    }
+                }
+            };
+        }
 
         let partition_count = self.partitioning.read().0 .0.parts.len();
 
@@ -2212,7 +2230,7 @@ impl Timeline {
         let sub_compaction_max_job_size_mb =
             sub_compaction_max_job_size_mb.unwrap_or(GC_COMPACT_MAX_SIZE_MB);
 
-        let mut compact_jobs = Vec::<GcCompactJob>::new();
+        let mut compact_jobs = Vec::new();
         // For now, we simply use the key partitioning information; we should do a more fine-grained partitioning
         // by estimating the amount of files read for a compaction job. We should also partition on LSN.
         let ((dense_ks, sparse_ks), _) = self.partitioning.read().as_ref().clone();
@@ -2299,25 +2317,16 @@ impl Timeline {
                 } else {
                     end
                 };
-                if total_size == 0 && !compact_jobs.is_empty() {
-                    info!(
-                        "splitting compaction job: {}..{}, estimated_size={}, extending the previous job",
-                        start, end, total_size
-                    );
-                    compact_jobs.last_mut().unwrap().compact_key_range.end = end;
-                    current_start = Some(end);
-                } else {
-                    info!(
-                        "splitting compaction job: {}..{}, estimated_size={}",
-                        start, end, total_size
-                    );
-                    compact_jobs.push(GcCompactJob {
-                        dry_run: job.dry_run,
-                        compact_key_range: start..end,
-                        compact_lsn_range: job.compact_lsn_range.start..compact_below_lsn,
-                    });
-                    current_start = Some(end);
-                }
+                info!(
+                    "splitting compaction job: {}..{}, estimated_size={}",
+                    start, end, total_size
+                );
+                compact_jobs.push(GcCompactJob {
+                    dry_run: job.dry_run,
+                    compact_key_range: start..end,
+                    compact_lsn_range: job.compact_lsn_range.start..compact_below_lsn,
+                });
+                current_start = Some(end);
             }
         }
         Ok(compact_jobs)

pageserver/src/tenant/timeline/heatmap_layers_downloader.rs

@@ -1,162 +0,0 @@
-//! Timeline utility module to hydrate everything from the current heatmap.
-//!
-//! Provides utilities to spawn and abort a background task where the downloads happen.
-//! See /v1/tenant/:tenant_shard_id/timeline/:timeline_id/download_heatmap_layers.
-
-use futures::StreamExt;
-use http_utils::error::ApiError;
-use std::sync::{Arc, Mutex};
-use tokio_util::sync::CancellationToken;
-use utils::sync::gate::Gate;
-
-use super::Timeline;
-
-// This status is not strictly necessary now, but gives us a nice place
-// to store progress information if we ever wish to expose it.
-pub(super) enum HeatmapLayersDownloadStatus {
-    InProgress,
-    Complete,
-}
-
-pub(super) struct HeatmapLayersDownloader {
-    handle: tokio::task::JoinHandle<()>,
-    status: Arc<Mutex<HeatmapLayersDownloadStatus>>,
-    cancel: CancellationToken,
-    downloads_guard: Arc<Gate>,
-}
-
-impl HeatmapLayersDownloader {
-    fn new(
-        timeline: Arc<Timeline>,
-        concurrency: usize,
-    ) -> Result<HeatmapLayersDownloader, ApiError> {
-        let tl_guard = timeline.gate.enter().map_err(|_| ApiError::Cancelled)?;
-
-        let cancel = timeline.cancel.child_token();
-        let downloads_guard = Arc::new(Gate::default());
-
-        let status = Arc::new(Mutex::new(HeatmapLayersDownloadStatus::InProgress));
-
-        let handle = tokio::task::spawn({
-            let status = status.clone();
-            let downloads_guard = downloads_guard.clone();
-            let cancel = cancel.clone();
-
-            async move {
-                let _guard = tl_guard;
-
-                scopeguard::defer! {
-                    *status.lock().unwrap() = HeatmapLayersDownloadStatus::Complete;
-                }
-
-                let Some(heatmap) = timeline.generate_heatmap().await else {
-                    tracing::info!("Heatmap layers download failed to generate heatmap");
-                    return;
-                };
-
-                tracing::info!(
-                    resident_size=%timeline.resident_physical_size(),
-                    heatmap_layers=%heatmap.layers.len(),
-                    "Starting heatmap layers download"
-                );
-
-                let stream = futures::stream::iter(heatmap.layers.into_iter().filter_map(
-                    |layer| {
-                        let tl = timeline.clone();
-                        let dl_guard = match downloads_guard.enter() {
-                            Ok(g) => g,
-                            Err(_) => {
-                                // [`Self::shutdown`] was called. Don't spawn any more downloads.
-                                return None;
-                            }
-                        };
-
-                        Some(async move {
-                            let _dl_guard = dl_guard;
-
-                            let res = tl.download_layer(&layer.name).await;
-                            if let Err(err) = res {
-                                if !err.is_cancelled() {
-                                    tracing::warn!(layer=%layer.name,"Failed to download heatmap layer: {err}")
-                                }
-                            }
-                        })
-                    }
-                )).buffered(concurrency);
-
-                tokio::select! {
-                    _ = stream.collect::<()>() => {
-                        tracing::info!(
-                            resident_size=%timeline.resident_physical_size(),
-                            "Heatmap layers download completed"
-                        );
-                    },
-                    _ = cancel.cancelled() => {
-                        tracing::info!("Heatmap layers download cancelled");
-                    }
-                }
-            }
-        });
-
-        Ok(Self {
-            status,
-            handle,
-            cancel,
-            downloads_guard,
-        })
-    }
-
-    fn is_complete(&self) -> bool {
-        matches!(
-            *self.status.lock().unwrap(),
-            HeatmapLayersDownloadStatus::Complete
-        )
-    }
-
-    /// Drive any in-progress downloads to completion and stop spawning any new ones.
-    ///
-    /// This has two callers and they behave differently
-    /// 1. [`Timeline::shutdown`]: the drain will be immediate since downloads themselves
-    ///    are sensitive to timeline cancellation.
-    ///
-    /// 2. Endpoint handler in [`crate::http::routes`]: the drain will wait for any in-progress
-    ///    downloads to complete.
-    async fn stop_and_drain(self) {
-        // Counterintuitive: close the guard before cancelling.
-        // Something needs to poll the already created download futures to completion.
-        // If we cancel first, then the underlying task exits and we lost
-        // the poller.
-        self.downloads_guard.close().await;
-        self.cancel.cancel();
-        if let Err(err) = self.handle.await {
-            tracing::warn!("Failed to join heatmap layer downloader task: {err}");
-        }
-    }
-}
-
-impl Timeline {
-    pub(crate) async fn start_heatmap_layers_download(
-        self: &Arc<Self>,
-        concurrency: usize,
-    ) -> Result<(), ApiError> {
-        let mut locked = self.heatmap_layers_downloader.lock().unwrap();
-        if locked.as_ref().map(|dl| dl.is_complete()).unwrap_or(true) {
-            let dl = HeatmapLayersDownloader::new(self.clone(), concurrency)?;
-            *locked = Some(dl);
-            Ok(())
-        } else {
-            Err(ApiError::Conflict("Already running".to_string()))
-        }
-    }
-
-    pub(crate) async fn stop_and_drain_heatmap_layers_download(&self) {
-        // This can race with the start of a new downloader and lead to a situation
-        // where one donloader is shutting down and another one is in-flight.
-        // The only impact is that we'd end up using more remote storage semaphore
-        // units than expected.
-        let downloader = self.heatmap_layers_downloader.lock().unwrap().take();
-        if let Some(dl) = downloader {
-            dl.stop_and_drain().await;
-        }
-    }
-}

pageserver/src/tenant/timeline/walreceiver.rs

@@ -56,7 +56,6 @@ pub struct WalReceiverConf {
     pub auth_token: Option<Arc<String>>,
     pub availability_zone: Option<String>,
     pub ingest_batch_size: u64,
-    pub validate_wal_contiguity: bool,
 }
 
 pub struct WalReceiver {

pageserver/src/tenant/timeline/walreceiver/connection_manager.rs

@@ -537,7 +537,6 @@ impl ConnectionManagerState {
         let connect_timeout = self.conf.wal_connect_timeout;
         let ingest_batch_size = self.conf.ingest_batch_size;
         let protocol = self.conf.protocol;
-        let validate_wal_contiguity = self.conf.validate_wal_contiguity;
         let timeline = Arc::clone(&self.timeline);
         let ctx = ctx.detached_child(
             TaskKind::WalReceiverConnectionHandler,
@@ -559,7 +558,6 @@ impl ConnectionManagerState {
                     ctx,
                     node_id,
                     ingest_batch_size,
-                    validate_wal_contiguity,
                 )
                 .await;
 
@@ -1565,7 +1563,6 @@ mod tests {
                 auth_token: None,
                 availability_zone: None,
                 ingest_batch_size: 1,
-                validate_wal_contiguity: false,
             },
             wal_connection: None,
             wal_stream_candidates: HashMap::new(),

pageserver/src/tenant/timeline/walreceiver/walreceiver_connection.rs

@@ -120,7 +120,6 @@ pub(super) async fn handle_walreceiver_connection(
     ctx: RequestContext,
     safekeeper_node: NodeId,
     ingest_batch_size: u64,
-    validate_wal_contiguity: bool,
 ) -> Result<(), WalReceiverError> {
     debug_assert_current_span_has_tenant_and_timeline_id();
 
@@ -275,7 +274,6 @@ pub(super) async fn handle_walreceiver_connection(
         } => Some((format, compression)),
     };
 
-    let mut expected_wal_start = startpoint;
     while let Some(replication_message) = {
         select! {
             _ = cancellation.cancelled() => {
@@ -342,49 +340,13 @@ pub(super) async fn handle_walreceiver_connection(
                     )
                     })?;
 
-                // Guard against WAL gaps. If the start LSN of the PG WAL section
-                // from which the interpreted records were extracted, doesn't match
-                // the end of the previous batch (or the starting point for the first batch),
-                // then kill this WAL receiver connection and start a new one.
-                if validate_wal_contiguity {
-                    if let Some(raw_wal_start_lsn) = batch.raw_wal_start_lsn {
-                        match raw_wal_start_lsn.cmp(&expected_wal_start) {
-                            std::cmp::Ordering::Greater => {
-                                let msg = format!(
-                                    "Gap in streamed WAL: [{}, {})",
-                                    expected_wal_start, raw_wal_start_lsn
-                                );
-                                critical!("{msg}");
-                                return Err(WalReceiverError::Other(anyhow!(msg)));
-                            }
-                            std::cmp::Ordering::Less => {
-                                // Other shards are reading WAL behind us.
-                                // This is valid, but check that we received records
-                                // that we haven't seen before.
-                                if let Some(first_rec) = batch.records.first() {
-                                    if first_rec.next_record_lsn < last_rec_lsn {
-                                        let msg = format!(
-                                            "Received record with next_record_lsn multiple times ({} < {})",
-                                            first_rec.next_record_lsn, expected_wal_start
-                                        );
-                                        critical!("{msg}");
-                                        return Err(WalReceiverError::Other(anyhow!(msg)));
-                                    }
-                                }
-                            }
-                            std::cmp::Ordering::Equal => {}
-                        }
-                    }
-                }
-
                 let InterpretedWalRecords {
                     records,
                     next_record_lsn,
-                    raw_wal_start_lsn: _,
                 } = batch;
 
                 tracing::debug!(
-                    "Received WAL up to {} with next_record_lsn={}",
+                    "Received WAL up to {} with next_record_lsn={:?}",
                     streaming_lsn,
                     next_record_lsn
                 );
@@ -461,11 +423,12 @@ pub(super) async fn handle_walreceiver_connection(
                 // need to advance last record LSN on all shards. If we've not ingested the latest
                 // record, then set the LSN of the modification past it. This way all shards
                 // advance their last record LSN at the same time.
-                let needs_last_record_lsn_advance = if next_record_lsn > modification.get_lsn() {
-                    modification.set_lsn(next_record_lsn).unwrap();
-                    true
-                } else {
-                    false
+                let needs_last_record_lsn_advance = match next_record_lsn {
+                    Some(lsn) if lsn > modification.get_lsn() => {
+                        modification.set_lsn(lsn).unwrap();
+                        true
+                    }
+                    _ => false,
                 };
 
                 if uncommitted_records > 0 || needs_last_record_lsn_advance {
@@ -483,8 +446,9 @@ pub(super) async fn handle_walreceiver_connection(
                     timeline.get_last_record_lsn()
                 );
 
-                last_rec_lsn = next_record_lsn;
-                expected_wal_start = streaming_lsn;
+                if let Some(lsn) = next_record_lsn {
+                    last_rec_lsn = lsn;
+                }
 
                 Some(streaming_lsn)
             }

pageserver/src/walingest.rs

@@ -1180,50 +1180,6 @@ impl WalIngest {
                 } else {
                     cp.oldestActiveXid = xlog_checkpoint.oldestActiveXid;
                 }
-                // NB: We abuse the Checkpoint.redo field:
-                //
-                // - In PostgreSQL, the Checkpoint struct doesn't store the information
-                //   of whether this is an online checkpoint or a shutdown checkpoint. It's
-                //   stored in the XLOG info field of the WAL record, shutdown checkpoints
-                //   use record type XLOG_CHECKPOINT_SHUTDOWN and online checkpoints use
-                //   XLOG_CHECKPOINT_ONLINE. We don't store the original WAL record headers
-                //   in the pageserver, however.
-                //
-                // - In PostgreSQL, the Checkpoint.redo field stores the *start* of the
-                //   checkpoint record, if it's a shutdown checkpoint. But when we are
-                //   starting from a shutdown checkpoint, the basebackup LSN is the *end*
-                //   of the shutdown checkpoint WAL record. That makes it difficult to
-                //   correctly detect whether we're starting from a shutdown record or
-                //   not.
-                //
-                // To address both of those issues, we store 0 in the redo field if it's
-                // an online checkpoint record, and the record's *end* LSN if it's a
-                // shutdown checkpoint. We don't need the original redo pointer in neon,
-                // because we don't perform WAL replay at startup anyway, so we can get
-                // away with abusing the redo field like this.
-                //
-                // XXX: Ideally, we would persist the extra information in a more
-                // explicit format, rather than repurpose the fields of the Postgres
-                // struct like this. However, we already have persisted data like this,
-                // so we need to maintain backwards compatibility.
-                //
-                // NB: We didn't originally have this convention, so there are still old
-                // persisted records that didn't do this. Before, we didn't update the
-                // persisted redo field at all. That means that old records have a bogus
-                // redo pointer that points to some old value, from the checkpoint record
-                // that was originally imported from the data directory. If it was a
-                // project created in Neon, that means it points to the first checkpoint
-                // after initdb. That's OK for our purposes: all such old checkpoints are
-                // treated as old online checkpoints when the basebackup is created.
-                cp.redo = if info == pg_constants::XLOG_CHECKPOINT_SHUTDOWN {
-                    // Store the *end* LSN of the checkpoint record. Or to be precise,
-                    // the start LSN of the *next* record, i.e. if the record ends
-                    // exactly at page boundary, the redo LSN points to just after the
-                    // page header on the next page.
-                    lsn.into()
-                } else {
-                    Lsn::INVALID.into()
-                };
 
                 // Write a new checkpoint key-value pair on every checkpoint record, even
                 // if nothing really changed. Not strictly required, but it seems nice to

pageserver/src/walredo/process.rs

@@ -136,9 +136,7 @@ impl WalRedoProcess {
                         Ok(0) => break Ok(()), // eof
                         Ok(num_bytes) => {
                             let output = String::from_utf8_lossy(&buf[..num_bytes]);
-                            if !output.contains("LOG:") {
-                               error!(%output, "received output");
-                            }
+                            error!(%output, "received output");
                         }
                         Err(e) => {
                             break Err(e);

pgxn/neon/neon.c

@@ -56,7 +56,6 @@ uint32		WAIT_EVENT_NEON_LFC_MAINTENANCE;
 uint32		WAIT_EVENT_NEON_LFC_READ;
 uint32		WAIT_EVENT_NEON_LFC_TRUNCATE;
 uint32		WAIT_EVENT_NEON_LFC_WRITE;
-uint32		WAIT_EVENT_NEON_LFC_CV_WAIT;
 uint32		WAIT_EVENT_NEON_PS_STARTING;
 uint32		WAIT_EVENT_NEON_PS_CONFIGURING;
 uint32		WAIT_EVENT_NEON_PS_SEND;
@@ -539,7 +538,6 @@ neon_shmem_startup_hook(void)
 	WAIT_EVENT_NEON_LFC_READ = WaitEventExtensionNew("Neon/FileCache_Read");
 	WAIT_EVENT_NEON_LFC_TRUNCATE = WaitEventExtensionNew("Neon/FileCache_Truncate");
 	WAIT_EVENT_NEON_LFC_WRITE = WaitEventExtensionNew("Neon/FileCache_Write");
-	WAIT_EVENT_NEON_LFC_CV_WAIT = WaitEventExtensionNew("Neon/FileCache_CvWait");
 	WAIT_EVENT_NEON_PS_STARTING = WaitEventExtensionNew("Neon/PS_Starting");
 	WAIT_EVENT_NEON_PS_CONFIGURING = WaitEventExtensionNew("Neon/PS_Configuring");
 	WAIT_EVENT_NEON_PS_SEND = WaitEventExtensionNew("Neon/PS_SendIO");

pgxn/neon/neon.h

@@ -28,7 +28,6 @@ extern uint32		WAIT_EVENT_NEON_LFC_MAINTENANCE;
 extern uint32		WAIT_EVENT_NEON_LFC_READ;
 extern uint32		WAIT_EVENT_NEON_LFC_TRUNCATE;
 extern uint32		WAIT_EVENT_NEON_LFC_WRITE;
-extern uint32		WAIT_EVENT_NEON_LFC_CV_WAIT;
 extern uint32		WAIT_EVENT_NEON_PS_STARTING;
 extern uint32		WAIT_EVENT_NEON_PS_CONFIGURING;
 extern uint32		WAIT_EVENT_NEON_PS_SEND;
@@ -39,7 +38,6 @@ extern uint32		WAIT_EVENT_NEON_WAL_DL;
 #define WAIT_EVENT_NEON_LFC_READ		WAIT_EVENT_BUFFILE_READ
 #define WAIT_EVENT_NEON_LFC_TRUNCATE	WAIT_EVENT_BUFFILE_TRUNCATE
 #define WAIT_EVENT_NEON_LFC_WRITE		WAIT_EVENT_BUFFILE_WRITE
-#define WAIT_EVENT_NEON_LFC_CV_WAIT 	WAIT_EVENT_BUFFILE_READ
 #define WAIT_EVENT_NEON_PS_STARTING		PG_WAIT_EXTENSION
 #define WAIT_EVENT_NEON_PS_CONFIGURING	PG_WAIT_EXTENSION
 #define WAIT_EVENT_NEON_PS_SEND			PG_WAIT_EXTENSION

pgxn/neon/pagestore_client.h

@@ -233,7 +233,6 @@ extern char *neon_timeline;
 extern char *neon_tenant;
 extern int32 max_cluster_size;
 extern int  neon_protocol_version;
-extern bool lfc_store_prefetch_result;
 
 extern shardno_t get_shard_number(BufferTag* tag);
 
@@ -302,16 +301,14 @@ extern bool lfc_cache_contains(NRelFileInfo rinfo, ForkNumber forkNum,
 							   BlockNumber blkno);
 extern int lfc_cache_containsv(NRelFileInfo rinfo, ForkNumber forkNum,
 							   BlockNumber blkno, int nblocks, bits8 *bitmap);
+extern void lfc_evict(NRelFileInfo rinfo, ForkNumber forkNum, BlockNumber blkno);
 extern void lfc_init(void);
-extern bool lfc_prefetch(NRelFileInfo rinfo, ForkNumber forknum, BlockNumber blkno,
-						 const void* buffer, XLogRecPtr lsn);
-
 
 static inline bool
 lfc_read(NRelFileInfo rinfo, ForkNumber forkNum, BlockNumber blkno,
 		 void *buffer)
 {
-	bits8		rv = 1;
+	bits8		rv = 0;
 	return lfc_readv_select(rinfo, forkNum, blkno, &buffer, 1, &rv) == 1;
 }
 

pgxn/neon/pagestore_smgr.c

@@ -162,7 +162,7 @@ static uint32 local_request_counter;
  * UNUSED ------> REQUESTED --> RECEIVED
  *   ^         :      |            |
  *   |         :      v            |
- *   |         : TAG_REMAINS       |
+ *   |         : TAG_UNUSED        |
  *   |         :      |            |
  *   +----------------+------------+
  *             :
@@ -181,7 +181,7 @@ typedef enum PrefetchStatus
 /* must fit in uint8; bits 0x1 are used */
 typedef enum {
 	PRFSF_NONE	= 0x0,
-	PRFSF_LFC	= 0x1  /* received prefetch result is stored in LFC */
+	PRFSF_SEQ	= 0x1,
 } PrefetchRequestFlags;
 
 typedef struct PrefetchRequest
@@ -305,7 +305,7 @@ GetLastWrittenLSNv(NRelFileInfo relfilenode, ForkNumber forknum,
 static void
 neon_get_request_lsns(NRelFileInfo rinfo, ForkNumber forknum,
 					  BlockNumber blkno, neon_request_lsns *output,
-					  BlockNumber nblocks);
+					  BlockNumber nblocks, const bits8 *mask);
 static bool neon_prefetch_response_usable(neon_request_lsns *request_lsns,
 										  PrefetchRequest *slot);
 
@@ -363,7 +363,6 @@ compact_prefetch_buffers(void)
 		target_slot->buftag = source_slot->buftag;
 		target_slot->shard_no = source_slot->shard_no;
 		target_slot->status = source_slot->status;
-		target_slot->flags = source_slot->flags;
 		target_slot->response = source_slot->response;
 		target_slot->reqid = source_slot->reqid;
 		target_slot->request_lsns = source_slot->request_lsns;
@@ -453,18 +452,6 @@ prefetch_pump_state(void)
 		/* update slot state */
 		slot->status = PRFS_RECEIVED;
 		slot->response = response;
-
-		if (response->tag == T_NeonGetPageResponse && !(slot->flags & PRFSF_LFC) && lfc_store_prefetch_result)
-		{
-			/*
-			 * Store prefetched result in LFC (please read comments to lfc_prefetch
-			 * explaining why it can be done without holding shared buffer lock
-			 */
-			if (lfc_prefetch(BufTagGetNRelFileInfo(slot->buftag), slot->buftag.forkNum, slot->buftag.blockNum, ((NeonGetPageResponse*)response)->page, slot->request_lsns.not_modified_since))
-			{
-				slot->flags |= PRFSF_LFC;
-			}
-		}
 	}
 }
 
@@ -487,7 +474,8 @@ readahead_buffer_resize(int newsize, void *extra)
 	 */
 	if (MyPState->n_requests_inflight > newsize)
 	{
-		prefetch_wait_for(MyPState->ring_unused - newsize - 1);
+		Assert(MyPState->ring_unused >= MyPState->n_requests_inflight - newsize);
+		prefetch_wait_for(MyPState->ring_unused - (MyPState->n_requests_inflight - newsize));
 		Assert(MyPState->n_requests_inflight <= newsize);
 	}
 
@@ -726,18 +714,6 @@ prefetch_read(PrefetchRequest *slot)
 		/* update slot state */
 		slot->status = PRFS_RECEIVED;
 		slot->response = response;
-
-		if (response->tag == T_NeonGetPageResponse && !(slot->flags & PRFSF_LFC) && lfc_store_prefetch_result)
-		{
-			/*
-			 * Store prefetched result in LFC (please read comments to lfc_prefetch
-			 * explaining why it can be done without holding shared buffer lock
-			 */
-			if (lfc_prefetch(BufTagGetNRelFileInfo(buftag), buftag.forkNum, buftag.blockNum, ((NeonGetPageResponse*)response)->page, slot->request_lsns.not_modified_since))
-			{
-				slot->flags |= PRFSF_LFC;
-			}
-		}
 		return true;
 	}
 	else
@@ -889,7 +865,7 @@ prefetch_do_request(PrefetchRequest *slot, neon_request_lsns *force_request_lsns
 	else
 		neon_get_request_lsns(BufTagGetNRelFileInfo(slot->buftag),
 							  slot->buftag.forkNum, slot->buftag.blockNum,
-							  &slot->request_lsns, 1);
+							  &slot->request_lsns, 1, NULL);
 	request.hdr.lsn = slot->request_lsns.request_lsn;
 	request.hdr.not_modified_since = slot->request_lsns.not_modified_since;
 
@@ -915,73 +891,6 @@ prefetch_do_request(PrefetchRequest *slot, neon_request_lsns *force_request_lsns
 	Assert(!found);
 }
 
-/*
- * Lookup of already received prefetch requests. Only already received responses matching required LSNs are accepted.
- * Present pages are marked in "mask" bitmap and total number of such pages is returned.
- */
-static int
-prefetch_lookupv(NRelFileInfo rinfo, ForkNumber forknum, BlockNumber blocknum, neon_request_lsns *lsns,
-				 BlockNumber nblocks, void **buffers, bits8 *mask)
-{
-	int hits = 0;
-	PrefetchRequest hashkey;
-
-	/*
-	 * Use an intermediate PrefetchRequest struct as the hash key to ensure
-	 * correct alignment and that the padding bytes are cleared.
-	 */
-	memset(&hashkey.buftag, 0, sizeof(BufferTag));
-	CopyNRelFileInfoToBufTag(hashkey.buftag, rinfo);
-	hashkey.buftag.forkNum = forknum;
-
-	for (int i = 0; i < nblocks; i++)
-	{
-		PrfHashEntry *entry;
-
-		hashkey.buftag.blockNum = blocknum + i;
-		entry = prfh_lookup(MyPState->prf_hash, &hashkey);
-
-		if (entry != NULL)
-		{
-			PrefetchRequest *slot = entry->slot;
-			uint64 ring_index = slot->my_ring_index;
-			Assert(slot == GetPrfSlot(ring_index));
-
-			Assert(slot->status != PRFS_UNUSED);
-			Assert(MyPState->ring_last <= ring_index &&
-				   ring_index < MyPState->ring_unused);
-			Assert(BufferTagsEqual(&slot->buftag, &hashkey.buftag));
-
-			if (slot->status != PRFS_RECEIVED)
-				continue;
-
-			/*
-			 * If the caller specified a request LSN to use, only accept
-			 * prefetch responses that satisfy that request.
-			 */
-			if (!neon_prefetch_response_usable(&lsns[i], slot))
-				continue;
-
-			memcpy(buffers[i], ((NeonGetPageResponse*)slot->response)->page, BLCKSZ);
-			prefetch_set_unused(ring_index);
-			BITMAP_SET(mask, i);
-
-			hits += 1;
-		}
-	}
-	pgBufferUsage.prefetch.hits += hits;
-	return hits;
-}
-
-#if PG_MAJORVERSION_NUM < 17
-static bool
-prefetch_lookup(NRelFileInfo rinfo, ForkNumber forkNum, BlockNumber blkn, neon_request_lsns *lsns, void *buffer)
-{
-	bits8 present = 0;
-	return prefetch_lookupv(rinfo, forkNum, blkn, lsns, 1, &buffer, &present) != 0;
-}
-#endif
-
 /*
  * prefetch_register_bufferv() - register and prefetch buffers
  *
@@ -1105,6 +1014,8 @@ Retry:
 					/* The buffered request is good enough, return that index */
 					if (is_prefetch)
 						pgBufferUsage.prefetch.duplicates++;
+					else
+						pgBufferUsage.prefetch.hits++;
 					continue;
 				}
 			}
@@ -1206,7 +1117,6 @@ Retry:
 		slot->buftag = hashkey.buftag;
 		slot->shard_no = get_shard_number(&tag);
 		slot->my_ring_index = ring_index;
-		slot->flags = 0;
 
 		min_ring_index = Min(min_ring_index, ring_index);
 
@@ -2147,7 +2057,8 @@ GetLastWrittenLSNv(NRelFileInfo relfilenode, ForkNumber forknum,
  */
 static void
 neon_get_request_lsns(NRelFileInfo rinfo, ForkNumber forknum, BlockNumber blkno,
-					  neon_request_lsns *output, BlockNumber nblocks)
+					  neon_request_lsns *output, BlockNumber nblocks,
+					  const bits8 *mask)
 {
 	XLogRecPtr	last_written_lsns[PG_IOV_MAX];
 
@@ -2235,6 +2146,9 @@ neon_get_request_lsns(NRelFileInfo rinfo, ForkNumber forknum, BlockNumber blkno,
 			neon_request_lsns *result = &output[i];
 			XLogRecPtr	last_written_lsn = last_written_lsns[i];
 
+			if (PointerIsValid(mask) && !BITMAP_ISSET(mask, i))
+				continue;
+
 			if (last_written_lsn > replay_lsn)
 			{
 				/* GetCurrentReplayRecPtr was introduced in v15 */
@@ -2277,6 +2191,8 @@ neon_get_request_lsns(NRelFileInfo rinfo, ForkNumber forknum, BlockNumber blkno,
 			neon_request_lsns *result = &output[i];
 			XLogRecPtr	last_written_lsn = last_written_lsns[i];
 
+			if (PointerIsValid(mask) && !BITMAP_ISSET(mask, i))
+				continue;
 			/*
 			 * Use the latest LSN that was evicted from the buffer cache as the
 			 * 'not_modified_since' hint. Any pages modified by later WAL records
@@ -2498,7 +2414,7 @@ neon_exists(SMgrRelation reln, ForkNumber forkNum)
 	}
 
 	neon_get_request_lsns(InfoFromSMgrRel(reln), forkNum,
-						  REL_METADATA_PSEUDO_BLOCKNO, &request_lsns, 1);
+						  REL_METADATA_PSEUDO_BLOCKNO, &request_lsns, 1, NULL);
 	{
 		NeonExistsRequest request = {
 			.hdr.tag = T_NeonExistsRequest,
@@ -2917,7 +2833,8 @@ neon_prefetch(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 	while (nblocks > 0)
 	{
 		int		iterblocks = Min(nblocks, PG_IOV_MAX);
-		bits8	lfc_present[PG_IOV_MAX / 8] = {0};
+		bits8		lfc_present[PG_IOV_MAX / 8];
+		memset(lfc_present, 0, sizeof(lfc_present));
 
 		if (lfc_cache_containsv(InfoFromSMgrRel(reln), forknum, blocknum,
 								iterblocks, lfc_present) == iterblocks)
@@ -2928,13 +2845,12 @@ neon_prefetch(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 		}
 
 		tag.blockNum = blocknum;
-
+		
 		for (int i = 0; i < PG_IOV_MAX / 8; i++)
 			lfc_present[i] = ~(lfc_present[i]);
 
 		ring_index = prefetch_register_bufferv(tag, NULL, iterblocks,
 											   lfc_present, true);
-
 		nblocks -= iterblocks;
 		blocknum += iterblocks;
 
@@ -3190,8 +3106,7 @@ Retry:
 					}
 				}
 				memcpy(buffer, getpage_resp->page, BLCKSZ);
-				if (!lfc_store_prefetch_result)
-					lfc_write(rinfo, forkNum, blockno, buffer);
+				lfc_write(rinfo, forkNum, blockno, buffer);
 				break;
 			}
 			case T_NeonErrorResponse:
@@ -3276,17 +3191,6 @@ neon_read(SMgrRelation reln, ForkNumber forkNum, BlockNumber blkno, void *buffer
 			neon_log(ERROR, "unknown relpersistence '%c'", reln->smgr_relpersistence);
 	}
 
-	/* Try to read PS results if they are available */
-	prefetch_pump_state();
-
-	neon_get_request_lsns(InfoFromSMgrRel(reln), forkNum, blkno, &request_lsns, 1);
-
-	if (prefetch_lookup(InfoFromSMgrRel(reln), forkNum, blkno, &request_lsns, buffer))
-	{
-		/* Prefetch hit */
-		return;
-	}
-
 	/* Try to read from local file cache */
 	if (lfc_read(InfoFromSMgrRel(reln), forkNum, blkno, buffer))
 	{
@@ -3294,11 +3198,9 @@ neon_read(SMgrRelation reln, ForkNumber forkNum, BlockNumber blkno, void *buffer
 		return;
 	}
 
+	neon_get_request_lsns(InfoFromSMgrRel(reln), forkNum, blkno, &request_lsns, 1, NULL);
 	neon_read_at_lsn(InfoFromSMgrRel(reln), forkNum, blkno, request_lsns, buffer);
 
-	/*
-	 * Try to receive prefetch results once again just to make sure we don't leave the smgr code while the OS might still have buffered bytes.
-	 */
 	prefetch_pump_state();
 
 #ifdef DEBUG_COMPARE_LOCAL
@@ -3379,14 +3281,11 @@ neon_read(SMgrRelation reln, ForkNumber forkNum, BlockNumber blkno, void *buffer
 #if PG_MAJORVERSION_NUM >= 17
 static void
 neon_readv(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
-		   void **buffers, BlockNumber nblocks)
+		void **buffers, BlockNumber nblocks)
 {
-	bits8		prefetch_hits[PG_IOV_MAX / 8] = {0};
-	bits8		lfc_hits[PG_IOV_MAX / 8];
 	bits8		read[PG_IOV_MAX / 8];
 	neon_request_lsns request_lsns[PG_IOV_MAX];
 	int			lfc_result;
-	int			prefetch_result;
 
 	switch (reln->smgr_relpersistence)
 	{
@@ -3409,52 +3308,38 @@ neon_readv(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 		neon_log(ERROR, "Read request too large: %d is larger than max %d",
 				 nblocks, PG_IOV_MAX);
 
-	/* Try to read PS results if they are available */
-	prefetch_pump_state();
-
-	neon_get_request_lsns(InfoFromSMgrRel(reln), forknum, blocknum,
-						  request_lsns, nblocks);
-
-
-	prefetch_result = prefetch_lookupv(InfoFromSMgrRel(reln), forknum, blocknum, request_lsns, nblocks, buffers, prefetch_hits);
-
-	if (prefetch_result == nblocks)
-		return;
-
-	/* invert the result: exclude prefetched blocks */
-	for (int i = 0; i < PG_IOV_MAX / 8; i++)
-		lfc_hits[i] = ~prefetch_hits[i];
+	memset(read, 0, sizeof(read));
 
 	/* Try to read from local file cache */
 	lfc_result = lfc_readv_select(InfoFromSMgrRel(reln), forknum, blocknum, buffers,
-								  nblocks, lfc_hits);
+								  nblocks, read);
 
 	if (lfc_result > 0)
 		MyNeonCounters->file_cache_hits_total += lfc_result;
 
 	/* Read all blocks from LFC, so we're done */
-	if (prefetch_result + lfc_result == nblocks)
+	if (lfc_result == nblocks)
 		return;
 
-	if (lfc_result <= 0)
+	if (lfc_result == -1)
 	{
 		/* can't use the LFC result, so read all blocks from PS */
 		for (int i = 0; i < PG_IOV_MAX / 8; i++)
-			read[i] = ~prefetch_hits[i];
+			read[i] = 0xFF;
 	}
 	else
 	{
 		/* invert the result: exclude blocks read from lfc */
 		for (int i = 0; i < PG_IOV_MAX / 8; i++)
-			read[i] = ~(prefetch_hits[i] | lfc_hits[i]);
+			read[i] = ~(read[i]);
 	}
 
+	neon_get_request_lsns(InfoFromSMgrRel(reln), forknum, blocknum,
+						  request_lsns, nblocks, read);
+
 	neon_read_at_lsnv(InfoFromSMgrRel(reln), forknum, blocknum, request_lsns,
 					  buffers, nblocks, read);
 
-	/*
-	 * Try to receive prefetch results once again just to make sure we don't leave the smgr code while the OS might still have buffered bytes.
-	 */
 	prefetch_pump_state();
 
 #ifdef DEBUG_COMPARE_LOCAL
@@ -3726,7 +3611,7 @@ neon_nblocks(SMgrRelation reln, ForkNumber forknum)
 	}
 
 	neon_get_request_lsns(InfoFromSMgrRel(reln), forknum,
-						  REL_METADATA_PSEUDO_BLOCKNO, &request_lsns, 1);
+						  REL_METADATA_PSEUDO_BLOCKNO, &request_lsns, 1, NULL);
 
 	{
 		NeonNblocksRequest request = {
@@ -3811,7 +3696,7 @@ neon_dbsize(Oid dbNode)
 	NRelFileInfo dummy_node = {0};
 
 	neon_get_request_lsns(dummy_node, MAIN_FORKNUM,
-						  REL_METADATA_PSEUDO_BLOCKNO, &request_lsns, 1);
+						  REL_METADATA_PSEUDO_BLOCKNO, &request_lsns, 1, NULL);
 
 	{
 		NeonDbSizeRequest request = {
@@ -4546,12 +4431,7 @@ neon_redo_read_buffer_filter(XLogReaderState *record, uint8 block_id)
 	if (no_redo_needed)
 	{
 		SetLastWrittenLSNForBlock(end_recptr, rinfo, forknum, blkno);
-		/*
-		 * Redo changes if page exists in LFC.
-		 * We should perform this check after assigning LwLSN to prevent
-		 * prefetching of some older version of the page by some other backend.
-		 */
-		no_redo_needed = !lfc_cache_contains(rinfo, forknum, blkno);
+		lfc_evict(rinfo, forknum, blkno);
 	}
 
 	LWLockRelease(partitionLock);

pgxn/neon_walredo/inmem_smgr.c

@@ -32,8 +32,8 @@
 
 #include "inmem_smgr.h"
 
-/* Size of the in-memory smgr: XLR_MAX_BLOCK_ID is 32, but we can update up to 3 forks for each block */
-#define MAX_PAGES 100
+/* Size of the in-memory smgr */
+#define MAX_PAGES 64
 
 /* If more than WARN_PAGES are used, print a warning in the log */
 #define WARN_PAGES 32
@@ -285,12 +285,12 @@ inmem_write(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
 		 * WARN_PAGES, print a warning so that we get alerted and get to
 		 * investigate why we're accessing so many buffers.
 		 */
-		if (used_pages >= WARN_PAGES)
-			ereport(WARNING, (errmsg("inmem_write() called for %u/%u/%u.%u blk %u: used_pages %u",
-								   RelFileInfoFmt(InfoFromSMgrRel(reln)),
-								   forknum,
-								   blocknum,
-								   used_pages), errbacktrace()));
+		elog(used_pages >= WARN_PAGES ? WARNING : DEBUG1,
+			 "inmem_write() called for %u/%u/%u.%u blk %u: used_pages %u",
+			 RelFileInfoFmt(InfoFromSMgrRel(reln)),
+			 forknum,
+			 blocknum,
+			 used_pages);
 		if (used_pages == MAX_PAGES)
 			elog(ERROR, "Inmem storage overflow");
 

pgxn/neon_walredo/walredoproc.c

@@ -142,7 +142,7 @@ static BufferTag target_redo_tag;
 
 static XLogReaderState *reader_state;
 
-#define TRACE DEBUG1
+#define TRACE LOG
 
 #ifdef HAVE_LIBSECCOMP
 
@@ -194,7 +194,6 @@ static PgSeccompRule allowed_syscalls[] =
 	 * is stored in MyProcPid anyway.
 	 */
 	PG_SCMP_ALLOW(getpid),
-	PG_SCMP_ALLOW(futex), /* needed for errbacktrace */
 
 	/* Enable those for a proper shutdown. */
 #if 0
@@ -254,7 +253,7 @@ WalRedoMain(int argc, char *argv[])
 	 * which is super strange but that's not something we can solve
 	 * for here. ¯\_(-_-)_/¯
 	 */
-	SetConfigOption("log_min_messages", "WARNING", PGC_SUSET, PGC_S_OVERRIDE);
+	SetConfigOption("log_min_messages", "FATAL", PGC_SUSET, PGC_S_OVERRIDE);
 	SetConfigOption("client_min_messages", "ERROR", PGC_SUSET,
 					PGC_S_OVERRIDE);
 
@@ -759,11 +758,6 @@ BeginRedoForBlock(StringInfo input_message)
 	{
 		reln->smgr_cached_nblocks[forknum] = blknum + 1;
 	}
-	if (target_redo_tag.forkNum == MAIN_FORKNUM)
-	{
-		reln->smgr_cached_nblocks[FSM_FORKNUM] = MaxBlockNumber;
-		reln->smgr_cached_nblocks[VISIBILITYMAP_FORKNUM] = MaxBlockNumber;
-	}
 }
 
 /*
@@ -1059,9 +1053,6 @@ GetPage(StringInfo input_message)
 	DropRelationAllLocalBuffers(rinfo);
 	wal_redo_buffer = InvalidBuffer;
 
-	/* Remove relation from SMGR relastion cache */
-	AtEOXact_SMgr();
-
 	elog(TRACE, "Page sent back for block %u", blknum);
 }
 

proxy/src/serverless/local_conn_pool.rs

@@ -279,12 +279,9 @@ impl ClientInnerCommon<postgres_client::Client> {
             local_data.jti += 1;
             let token = resign_jwt(&local_data.key, payload, local_data.jti)?;
 
-            // discard all cannot run in a transaction. must be executed alone.
-            self.inner.batch_execute("discard all").await?;
-
             // initiates the auth session
             // this is safe from query injections as the jwt format free of any escape characters.
-            let query = format!("select auth.jwt_session_init('{token}')");
+            let query = format!("discard all; select auth.jwt_session_init('{token}')");
             self.inner.batch_execute(&query).await?;
 
             let pid = self.inner.get_process_id();

rust-toolchain.toml

@@ -1,5 +1,5 @@
 [toolchain]
-channel = "1.85.0"
+channel = "1.84.1"
 profile = "default"
 # The default profile includes rustc, rust-std, cargo, rust-docs, rustfmt and clippy.
 # https://rust-lang.github.io/rustup/concepts/profiles.html

safekeeper/client/src/mgmt_api.rs

@@ -5,10 +5,7 @@
 
 use http_utils::error::HttpErrorBody;
 use reqwest::{IntoUrl, Method, StatusCode};
-use safekeeper_api::models::{
-    PullTimelineRequest, PullTimelineResponse, SafekeeperUtilization, TimelineCreateRequest,
-    TimelineStatus,
-};
+use safekeeper_api::models::{SafekeeperUtilization, TimelineCreateRequest, TimelineStatus};
 use std::error::Error as _;
 use utils::{
     id::{NodeId, TenantId, TimelineId},
@@ -91,12 +88,6 @@ impl Client {
         resp.json().await.map_err(Error::ReceiveBody)
     }
 
-    pub async fn pull_timeline(&self, req: &PullTimelineRequest) -> Result<PullTimelineResponse> {
-        let uri = format!("{}/v1/pull_timeline", self.mgmt_api_endpoint);
-        let resp = self.post(&uri, req).await?;
-        resp.json().await.map_err(Error::ReceiveBody)
-    }
-
     pub async fn delete_timeline(
         &self,
         tenant_id: TenantId,
@@ -137,7 +128,7 @@ impl Client {
     }
 
     pub async fn utilization(&self) -> Result<SafekeeperUtilization> {
-        let uri = format!("{}/v1/utilization", self.mgmt_api_endpoint);
+        let uri = format!("{}/v1/utilization/", self.mgmt_api_endpoint);
         let resp = self.get(&uri).await?;
         resp.json().await.map_err(Error::ReceiveBody)
     }

safekeeper/src/control_file.rs

@@ -235,7 +235,7 @@ impl Storage for FileStorage {
 #[cfg(test)]
 mod test {
     use super::*;
-    use safekeeper_api::membership::{Configuration, MemberSet, SafekeeperGeneration};
+    use safekeeper_api::membership::{Configuration, MemberSet};
     use tokio::fs;
     use utils::lsn::Lsn;
 
@@ -246,7 +246,7 @@ mod test {
         let tempdir = camino_tempfile::tempdir()?;
         let mut state = TimelinePersistentState::empty();
         state.mconf = Configuration {
-            generation: SafekeeperGeneration::new(42),
+            generation: 42,
             members: MemberSet::empty(),
             new_members: None,
         };

safekeeper/src/http/routes.rs

@@ -2,7 +2,6 @@ use http_utils::failpoints::failpoints_handler;
 use hyper::{Body, Request, Response, StatusCode};
 use safekeeper_api::models;
 use safekeeper_api::models::AcceptorStateStatus;
-use safekeeper_api::models::PullTimelineRequest;
 use safekeeper_api::models::SafekeeperStatus;
 use safekeeper_api::models::TermSwitchApiEntry;
 use safekeeper_api::models::TimelineStatus;
@@ -231,7 +230,7 @@ async fn timeline_delete_handler(mut request: Request<Body>) -> Result<Response<
 async fn timeline_pull_handler(mut request: Request<Body>) -> Result<Response<Body>, ApiError> {
     check_permission(&request, None)?;
 
-    let data: PullTimelineRequest = json_request(&mut request).await?;
+    let data: pull_timeline::Request = json_request(&mut request).await?;
     let conf = get_conf(&request);
     let global_timelines = get_global_timelines(&request);
 

safekeeper/src/pull_timeline.rs

@@ -4,13 +4,10 @@ use camino::Utf8PathBuf;
 use chrono::{DateTime, Utc};
 use futures::{SinkExt, StreamExt, TryStreamExt};
 use postgres_ffi::{XLogFileName, XLogSegNo, PG_TLI};
-use safekeeper_api::{
-    models::{PullTimelineRequest, PullTimelineResponse, TimelineStatus},
-    Term,
-};
+use safekeeper_api::{models::TimelineStatus, Term};
 use safekeeper_client::mgmt_api;
 use safekeeper_client::mgmt_api::Client;
-use serde::Deserialize;
+use serde::{Deserialize, Serialize};
 use std::{
     cmp::min,
     io::{self, ErrorKind},
@@ -36,7 +33,7 @@ use crate::{
 };
 use utils::{
     crashsafe::fsync_async_opt,
-    id::{NodeId, TenantTimelineId},
+    id::{NodeId, TenantId, TenantTimelineId, TimelineId},
     logging::SecretString,
     lsn::Lsn,
     pausable_failpoint,
@@ -381,6 +378,21 @@ impl WalResidentTimeline {
     }
 }
 
+/// pull_timeline request body.
+#[derive(Debug, Deserialize)]
+pub struct Request {
+    pub tenant_id: TenantId,
+    pub timeline_id: TimelineId,
+    pub http_hosts: Vec<String>,
+}
+
+#[derive(Debug, Serialize)]
+pub struct Response {
+    // Donor safekeeper host
+    pub safekeeper_host: String,
+    // TODO: add more fields?
+}
+
 /// Response for debug dump request.
 #[derive(Debug, Deserialize)]
 pub struct DebugDumpResponse {
@@ -393,10 +405,10 @@ pub struct DebugDumpResponse {
 
 /// Find the most advanced safekeeper and pull timeline from it.
 pub async fn handle_request(
-    request: PullTimelineRequest,
+    request: Request,
     sk_auth_token: Option<SecretString>,
     global_timelines: Arc<GlobalTimelines>,
-) -> Result<PullTimelineResponse> {
+) -> Result<Response> {
     let existing_tli = global_timelines.get(TenantTimelineId::new(
         request.tenant_id,
         request.timeline_id,
@@ -448,7 +460,7 @@ async fn pull_timeline(
     host: String,
     sk_auth_token: Option<SecretString>,
     global_timelines: Arc<GlobalTimelines>,
-) -> Result<PullTimelineResponse> {
+) -> Result<Response> {
     let ttid = TenantTimelineId::new(status.tenant_id, status.timeline_id);
     info!(
         "pulling timeline {} from safekeeper {}, commit_lsn={}, flush_lsn={}, term={}, epoch={}",
@@ -523,7 +535,7 @@ async fn pull_timeline(
         .load_temp_timeline(ttid, &tli_dir_path, false)
         .await?;
 
-    Ok(PullTimelineResponse {
+    Ok(Response {
         safekeeper_host: host,
     })
 }

safekeeper/src/safekeeper.rs

@@ -1004,7 +1004,7 @@ mod tests {
 
     use postgres_ffi::{XLogSegNo, WAL_SEGMENT_SIZE};
     use safekeeper_api::{
-        membership::{Configuration, MemberSet, SafekeeperGeneration, SafekeeperId},
+        membership::{Configuration, MemberSet, SafekeeperId},
         ServerInfo,
     };
 
@@ -1303,7 +1303,7 @@ mod tests {
             tenant_id,
             timeline_id,
             mconf: Configuration {
-                generation: SafekeeperGeneration::new(42),
+                generation: 42,
                 members: MemberSet::new(vec![SafekeeperId {
                     id: NodeId(1),
                     host: "hehe.org".to_owned(),

safekeeper/src/send_interpreted_wal.rs

@@ -295,10 +295,6 @@ impl InterpretedWalReader {
 
         let mut wal_decoder = WalStreamDecoder::new(start_pos, self.pg_version);
 
-        // Tracks the start of the PG WAL LSN from which the current batch of
-        // interpreted records originated.
-        let mut current_batch_wal_start_lsn: Option<Lsn> = None;
-
         loop {
             tokio::select! {
                 // Main branch for reading WAL and forwarding it
@@ -306,7 +302,7 @@ impl InterpretedWalReader {
                     let wal = wal_or_reset.map(|wor| wor.get_wal().expect("reset handled in select branch below"));
                     let WalBytes {
                         wal,
-                        wal_start_lsn,
+                        wal_start_lsn: _,
                         wal_end_lsn,
                         available_wal_end_lsn,
                     } = match wal {
@@ -319,12 +315,6 @@ impl InterpretedWalReader {
                         }
                     };
 
-                    // We will already have a value if the previous chunks of WAL
-                    // did not decode into anything useful.
-                    if current_batch_wal_start_lsn.is_none() {
-                        current_batch_wal_start_lsn = Some(wal_start_lsn);
-                    }
-
                     wal_decoder.feed_bytes(&wal);
 
                     // Deserialize and interpret WAL records from this batch of WAL.
@@ -373,9 +363,7 @@ impl InterpretedWalReader {
 
                     let max_next_record_lsn = match max_next_record_lsn {
                         Some(lsn) => lsn,
-                        None => {
-                            continue;
-                        }
+                        None => { continue; }
                     };
 
                     // Update the current position such that new receivers can decide
@@ -389,38 +377,21 @@ impl InterpretedWalReader {
                         }
                     }
 
-                    let batch_wal_start_lsn = current_batch_wal_start_lsn.take().unwrap();
-
                     // Send interpreted records downstream. Anything that has already been seen
                     // by a shard is filtered out.
                     let mut shard_senders_to_remove = Vec::new();
                     for (shard, states) in &mut self.shard_senders {
                         for state in states {
-                            let shard_sender_id = ShardSenderId::new(*shard, state.sender_id);
-
-                            let batch = if max_next_record_lsn > state.next_record_lsn {
-                                // This batch contains at least one record that this shard has not
-                                // seen yet.
-                                let records = records_by_sender.remove(&shard_sender_id).unwrap_or_default();
-
-                                InterpretedWalRecords {
-                                    records,
-                                    next_record_lsn: max_next_record_lsn,
-                                    raw_wal_start_lsn: Some(batch_wal_start_lsn),
-                                }
-                            } else if wal_end_lsn > state.next_record_lsn {
-                                // All the records in this batch were seen by the shard
-                                // However, the batch maps to a chunk of WAL that the
-                                // shard has not yet seen. Notify it of the start LSN
-                                // of the PG WAL chunk such that it doesn't look like a gap.
-                                InterpretedWalRecords {
-                                    records: Vec::default(),
-                                    next_record_lsn: state.next_record_lsn,
-                                    raw_wal_start_lsn: Some(batch_wal_start_lsn),
-                                }
-                            } else {
-                                // The shard has seen this chunk of WAL before. Skip it.
+                            if max_next_record_lsn <= state.next_record_lsn {
                                 continue;
+                            }
+
+                            let shard_sender_id = ShardSenderId::new(*shard, state.sender_id);
+                            let records = records_by_sender.remove(&shard_sender_id).unwrap_or_default();
+
+                            let batch = InterpretedWalRecords {
+                                records,
+                                next_record_lsn: Some(max_next_record_lsn),
                             };
 
                             let res = state.tx.send(Batch {
@@ -432,7 +403,7 @@ impl InterpretedWalReader {
                             if res.is_err() {
                                 shard_senders_to_remove.push(shard_sender_id);
                             } else {
-                                state.next_record_lsn = std::cmp::max(state.next_record_lsn, max_next_record_lsn);
+                                state.next_record_lsn = max_next_record_lsn;
                             }
                         }
                     }

scripts/generate_image_maps.py

@@ -6,9 +6,6 @@ build_tag = os.environ["BUILD_TAG"]
 branch = os.environ["BRANCH"]
 dev_acr = os.environ["DEV_ACR"]
 prod_acr = os.environ["PROD_ACR"]
-dev_aws = os.environ["DEV_AWS"]
-prod_aws = os.environ["PROD_AWS"]
-aws_region = os.environ["AWS_REGION"]
 
 components = {
     "neon": ["neon"],
@@ -27,11 +24,11 @@ components = {
 registries = {
     "dev": [
         "docker.io/neondatabase",
-        f"{dev_aws}.dkr.ecr.{aws_region}.amazonaws.com",
+        "369495373322.dkr.ecr.eu-central-1.amazonaws.com",
         f"{dev_acr}.azurecr.io/neondatabase",
     ],
     "prod": [
-        f"{prod_aws}.dkr.ecr.{aws_region}.amazonaws.com",
+        "093970136003.dkr.ecr.eu-central-1.amazonaws.com",
         f"{prod_acr}.azurecr.io/neondatabase",
     ],
 }

storage_controller/Cargo.toml

@@ -24,7 +24,6 @@ hex.workspace = true
 hyper0.workspace = true
 humantime.workspace = true
 itertools.workspace = true
-json-structural-diff.workspace = true
 lasso.workspace = true
 once_cell.workspace = true
 pageserver_api.workspace = true
@@ -35,7 +34,6 @@ reqwest = { workspace = true, features = ["stream"] }
 routerify.workspace = true
 safekeeper_api.workspace = true
 safekeeper_client.workspace = true
-tikv-jemallocator.workspace = true
 regex.workspace = true
 rustls-native-certs.workspace = true
 serde.workspace = true

storage_controller/migrations/2025-02-11-144848_pageserver_use_https/down.sql

@@ -1 +0,0 @@
-ALTER TABLE nodes DROP listen_https_port;

storage_controller/migrations/2025-02-11-144848_pageserver_use_https/up.sql

@@ -1 +0,0 @@
-ALTER TABLE nodes ADD listen_https_port INTEGER;

storage_controller/src/heartbeater.rs

@@ -10,10 +10,7 @@ use std::{
 };
 use tokio_util::sync::CancellationToken;
 
-use pageserver_api::{
-    controller_api::{NodeAvailability, SkSchedulingPolicy},
-    models::PageserverUtilization,
-};
+use pageserver_api::{controller_api::NodeAvailability, models::PageserverUtilization};
 
 use thiserror::Error;
 use utils::{id::NodeId, logging::SecretString};
@@ -140,13 +137,8 @@ where
                 request = self.receiver.recv() => {
                     match request {
                         Some(req) => {
-                            if req.reply.is_closed() {
-                                // Prevent a possibly infinite buildup of the receiver channel, if requests arrive faster than we can handle them
-                                continue;
-                            }
                             let res = self.heartbeat(req.servers).await;
-                            // Ignore the return value in order to not panic if the heartbeat function's future was cancelled
-                            _ = req.reply.send(res);
+                            req.reply.send(res).unwrap();
                         },
                         None => { return; }
                     }
@@ -319,9 +311,6 @@ impl HeartBeat<Safekeeper, SafekeeperState> for HeartbeaterTask<Safekeeper, Safe
 
         let mut heartbeat_futs = FuturesUnordered::new();
         for (node_id, sk) in &*safekeepers {
-            if sk.scheduling_policy() == SkSchedulingPolicy::Decomissioned {
-                continue;
-            }
             heartbeat_futs.push({
                 let jwt_token = self
                     .jwt_token
@@ -351,13 +340,7 @@ impl HeartBeat<Safekeeper, SafekeeperState> for HeartbeaterTask<Safekeeper, Safe
                             // We ignore the node in this case.
                             return None;
                         }
-                        Err(e) => {
-                            tracing::info!(
-                                "Marking safekeeper {} at as offline: {e}",
-                                sk.base_url()
-                            );
-                            SafekeeperState::Offline
-                        }
+                        Err(_) => SafekeeperState::Offline,
                     };
 
                     Some((*node_id, status))

storage_controller/src/http.rs

@@ -9,10 +9,7 @@ use crate::service::{LeadershipStatus, Service, RECONCILE_TIMEOUT, STARTUP_RECON
 use anyhow::Context;
 use futures::Future;
 use http_utils::{
-    endpoint::{
-        self, auth_middleware, check_permission_with, profile_cpu_handler, profile_heap_handler,
-        request_span,
-    },
+    endpoint::{self, auth_middleware, check_permission_with, request_span},
     error::ApiError,
     failpoints::failpoints_handler,
     json::{json_request, json_response},
@@ -57,7 +54,7 @@ pub struct HttpState {
     service: Arc<crate::service::Service>,
     auth: Option<Arc<SwappableJwtAuth>>,
     neon_metrics: NeonMetrics,
-    allowlist_routes: &'static [&'static str],
+    allowlist_routes: Vec<Uri>,
 }
 
 impl HttpState {
@@ -66,17 +63,15 @@ impl HttpState {
         auth: Option<Arc<SwappableJwtAuth>>,
         build_info: BuildInfo,
     ) -> Self {
+        let allowlist_routes = ["/status", "/ready", "/metrics"]
+            .iter()
+            .map(|v| v.parse().unwrap())
+            .collect::<Vec<_>>();
         Self {
             service,
             auth,
             neon_metrics: NeonMetrics::new(build_info),
-            allowlist_routes: &[
-                "/status",
-                "/ready",
-                "/metrics",
-                "/profile/cpu",
-                "/profile/heap",
-            ],
+            allowlist_routes,
         }
     }
 }
@@ -521,24 +516,6 @@ async fn handle_tenant_timeline_block_unblock_gc(
     json_response(StatusCode::OK, ())
 }
 
-async fn handle_tenant_timeline_download_heatmap_layers(
-    service: Arc<Service>,
-    req: Request<Body>,
-) -> Result<Response<Body>, ApiError> {
-    let tenant_shard_id: TenantShardId = parse_request_param(&req, "tenant_shard_id")?;
-
-    check_permissions(&req, Scope::PageServerApi)?;
-
-    let timeline_id: TimelineId = parse_request_param(&req, "timeline_id")?;
-    let concurrency: Option<usize> = parse_query_param(&req, "concurrency")?;
-
-    service
-        .tenant_timeline_download_heatmap_layers(tenant_shard_id, timeline_id, concurrency)
-        .await?;
-
-    json_response(StatusCode::OK, ())
-}
-
 // For metric labels where we would like to include the approximate path, but exclude high-cardinality fields like query parameters
 // and tenant/timeline IDs.  Since we are proxying to arbitrary paths, we don't have routing templates to
 // compare to, so we can just filter out our well known ID format with regexes.
@@ -598,10 +575,7 @@ async fn handle_tenant_timeline_passthrough(
 
     let _timer = latency.start_timer(labels.clone());
 
-    let client = mgmt_api::Client::new(
-        node.base_url(),
-        service.get_config().pageserver_jwt_token.as_deref(),
-    );
+    let client = mgmt_api::Client::new(node.base_url(), service.get_config().jwt_token.as_deref());
     let resp = client.get_raw(path).await.map_err(|e|
         // We return 503 here because if we can't successfully send a request to the pageserver,
         // either we aren't available or the pageserver is unavailable.
@@ -1357,7 +1331,10 @@ async fn handle_safekeeper_scheduling_policy(
         .set_safekeeper_scheduling_policy(id, body.scheduling_policy)
         .await?;
 
-    json_response(StatusCode::OK, ())
+    Ok(Response::builder()
+        .status(StatusCode::NO_CONTENT)
+        .body(Body::empty())
+        .unwrap())
 }
 
 /// Common wrapper for request handlers that call into Service and will operate on tenants: they must only
@@ -1421,26 +1398,23 @@ pub fn prologue_leadership_status_check_middleware<
         let state = get_state(&req);
         let leadership_status = state.service.get_leadership_status();
 
-        enum AllowedRoutes {
+        enum AllowedRoutes<'a> {
             All,
-            Some(&'static [&'static str]),
+            Some(Vec<&'a str>),
         }
 
         let allowed_routes = match leadership_status {
             LeadershipStatus::Leader => AllowedRoutes::All,
             LeadershipStatus::SteppedDown => AllowedRoutes::All,
-            LeadershipStatus::Candidate => AllowedRoutes::Some(&[
-                "/ready",
-                "/status",
-                "/metrics",
-                "/profile/cpu",
-                "/profile/heap",
-            ]),
+            LeadershipStatus::Candidate => {
+                AllowedRoutes::Some(["/ready", "/status", "/metrics"].to_vec())
+            }
         };
 
+        let uri = req.uri().to_string();
         match allowed_routes {
             AllowedRoutes::All => Ok(req),
-            AllowedRoutes::Some(allowed) if allowed.contains(&req.uri().path()) => Ok(req),
+            AllowedRoutes::Some(allowed) if allowed.contains(&uri.as_str()) => Ok(req),
             _ => {
                 tracing::info!(
                     "Request {} not allowed due to current leadership state",
@@ -1549,8 +1523,7 @@ enum ForwardOutcome {
 
 /// Potentially forward the request to the current storage controler leader.
 /// More specifically we forward when:
-/// 1. Request is not one of:
-///    ["/control/v1/step_down", "/status", "/ready", "/metrics", "/profile/cpu", "/profile/heap"]
+/// 1. Request is not one of ["/control/v1/step_down", "/status", "/ready", "/metrics"]
 /// 2. Current instance is in [`LeadershipStatus::SteppedDown`] state
 /// 3. There is a leader in the database to forward to
 /// 4. Leader from step (3) is not the current instance
@@ -1571,17 +1544,10 @@ enum ForwardOutcome {
 /// Hence, if we are in the edge case scenario the leader persisted in the database is the
 /// stepped down instance that received the request. Condition (4) above covers this scenario.
 async fn maybe_forward(req: Request<Body>) -> ForwardOutcome {
-    const NOT_FOR_FORWARD: &[&str] = &[
-        "/control/v1/step_down",
-        "/status",
-        "/ready",
-        "/metrics",
-        "/profile/cpu",
-        "/profile/heap",
-    ];
+    const NOT_FOR_FORWARD: [&str; 4] = ["/control/v1/step_down", "/status", "/ready", "/metrics"];
 
-    let uri = req.uri();
-    let uri_for_forward = !NOT_FOR_FORWARD.contains(&uri.path());
+    let uri = req.uri().to_string();
+    let uri_for_forward = !NOT_FOR_FORWARD.contains(&uri.as_str());
 
     // Fast return before trying to take any Service locks, if we will never forward anyway
     if !uri_for_forward {
@@ -1781,7 +1747,7 @@ pub fn make_router(
     if auth.is_some() {
         router = router.middleware(auth_middleware(|request| {
             let state = get_state(request);
-            if state.allowlist_routes.contains(&request.uri().path()) {
+            if state.allowlist_routes.contains(request.uri()) {
                 None
             } else {
                 state.auth.as_deref()
@@ -1794,19 +1760,13 @@ pub fn make_router(
         .get("/metrics", |r| {
             named_request_span(r, measured_metrics_handler, RequestName("metrics"))
         })
-        // Non-prefixed generic endpoints (status, metrics, profiling)
+        // Non-prefixed generic endpoints (status, metrics)
         .get("/status", |r| {
             named_request_span(r, handle_status, RequestName("status"))
         })
         .get("/ready", |r| {
             named_request_span(r, handle_ready, RequestName("ready"))
         })
-        .get("/profile/cpu", |r| {
-            named_request_span(r, profile_cpu_handler, RequestName("profile_cpu"))
-        })
-        .get("/profile/heap", |r| {
-            named_request_span(r, profile_heap_handler, RequestName("profile_heap"))
-        })
         // Upcalls for the pageserver: point the pageserver's `control_plane_api` config to this prefix
         .post("/upcall/v1/re-attach", |r| {
             named_request_span(r, handle_re_attach, RequestName("upcall_v1_reattach"))
@@ -2118,16 +2078,6 @@ pub fn make_router(
                 )
             },
         )
-        .post(
-            "/v1/tenant/:tenant_shard_id/timeline/:timeline_id/download_heatmap_layers",
-            |r| {
-                tenant_service_handler(
-                    r,
-                    handle_tenant_timeline_download_heatmap_layers,
-                    RequestName("v1_tenant_timeline_download_heatmap_layers"),
-                )
-            },
-        )
         // Tenant detail GET passthrough to shard zero:
         .get("/v1/tenant/:tenant_id", |r| {
             tenant_service_handler(

storage_controller/src/main.rs

@@ -27,16 +27,6 @@ use utils::{project_build_tag, project_git_version, tcp_listener};
 project_git_version!(GIT_VERSION);
 project_build_tag!(BUILD_TAG);
 
-#[global_allocator]
-static GLOBAL: tikv_jemallocator::Jemalloc = tikv_jemallocator::Jemalloc;
-
-/// Configure jemalloc to profile heap allocations by sampling stack traces every 2 MB (1 << 21).
-/// This adds roughly 3% overhead for allocations on average, which is acceptable considering
-/// performance-sensitive code will avoid allocations as far as possible anyway.
-#[allow(non_upper_case_globals)]
-#[export_name = "malloc_conf"]
-pub static malloc_conf: &[u8] = b"prof:true,prof_active:true,lg_prof_sample:21\0";
-
 #[derive(Parser)]
 #[command(author, version, about, long_about = None)]
 #[command(arg_required_else_help(true))]
@@ -53,10 +43,6 @@ struct Cli {
     #[arg(long)]
     jwt_token: Option<String>,
 
-    /// Token for authenticating this service with the safekeepers it controls
-    #[arg(long)]
-    safekeeper_jwt_token: Option<String>,
-
     /// Token for authenticating this service with the control plane, when calling
     /// the compute notification endpoint
     #[arg(long)]
@@ -130,10 +116,6 @@ struct Cli {
 
     #[arg(long)]
     long_reconcile_threshold: Option<humantime::Duration>,
-
-    // Flag to use https for requests to pageserver API.
-    #[arg(long, default_value = "false")]
-    use_https_pageserver_api: bool,
 }
 
 enum StrictMode {
@@ -157,8 +139,7 @@ impl Default for StrictMode {
 struct Secrets {
     database_url: String,
     public_key: Option<JwtAuth>,
-    pageserver_jwt_token: Option<String>,
-    safekeeper_jwt_token: Option<String>,
+    jwt_token: Option<String>,
     control_plane_jwt_token: Option<String>,
     peer_jwt_token: Option<String>,
 }
@@ -166,7 +147,6 @@ struct Secrets {
 impl Secrets {
     const DATABASE_URL_ENV: &'static str = "DATABASE_URL";
     const PAGESERVER_JWT_TOKEN_ENV: &'static str = "PAGESERVER_JWT_TOKEN";
-    const SAFEKEEPER_JWT_TOKEN_ENV: &'static str = "SAFEKEEPER_JWT_TOKEN";
     const CONTROL_PLANE_JWT_TOKEN_ENV: &'static str = "CONTROL_PLANE_JWT_TOKEN";
     const PEER_JWT_TOKEN_ENV: &'static str = "PEER_JWT_TOKEN";
     const PUBLIC_KEY_ENV: &'static str = "PUBLIC_KEY";
@@ -190,14 +170,7 @@ impl Secrets {
         let this = Self {
             database_url,
             public_key,
-            pageserver_jwt_token: Self::load_secret(
-                &args.jwt_token,
-                Self::PAGESERVER_JWT_TOKEN_ENV,
-            ),
-            safekeeper_jwt_token: Self::load_secret(
-                &args.safekeeper_jwt_token,
-                Self::SAFEKEEPER_JWT_TOKEN_ENV,
-            ),
+            jwt_token: Self::load_secret(&args.jwt_token, Self::PAGESERVER_JWT_TOKEN_ENV),
             control_plane_jwt_token: Self::load_secret(
                 &args.control_plane_jwt_token,
                 Self::CONTROL_PLANE_JWT_TOKEN_ENV,
@@ -277,17 +250,11 @@ async fn async_main() -> anyhow::Result<()> {
 
     let secrets = Secrets::load(&args).await?;
 
-    // TODO: once we've rolled out the safekeeper JWT token everywhere, put it into the validation code below
-    tracing::info!(
-        "safekeeper_jwt_token set: {:?}",
-        secrets.safekeeper_jwt_token.is_some()
-    );
-
     // Validate required secrets and arguments are provided in strict mode
     match strict_mode {
         StrictMode::Strict
             if (secrets.public_key.is_none()
-                || secrets.pageserver_jwt_token.is_none()
+                || secrets.jwt_token.is_none()
                 || secrets.control_plane_jwt_token.is_none()) =>
         {
             // Production systems should always have secrets configured: if public_key was not set
@@ -312,8 +279,7 @@ async fn async_main() -> anyhow::Result<()> {
     }
 
     let config = Config {
-        pageserver_jwt_token: secrets.pageserver_jwt_token,
-        safekeeper_jwt_token: secrets.safekeeper_jwt_token,
+        jwt_token: secrets.jwt_token,
         control_plane_jwt_token: secrets.control_plane_jwt_token,
         peer_jwt_token: secrets.peer_jwt_token,
         compute_hook_url: args.compute_hook_url,
@@ -345,7 +311,6 @@ async fn async_main() -> anyhow::Result<()> {
         address_for_peers: args.address_for_peers,
         start_as_candidate: args.start_as_candidate,
         http_service_port: args.listen.port() as i32,
-        use_https_pageserver_api: args.use_https_pageserver_api,
     };
 
     // Validate that we can connect to the database

storage_controller/src/node.rs

@@ -1,6 +1,5 @@
 use std::{str::FromStr, time::Duration};
 
-use anyhow::anyhow;
 use pageserver_api::{
     controller_api::{
         AvailabilityZone, NodeAvailability, NodeDescribeResponse, NodeRegisterRequest,
@@ -33,16 +32,12 @@ pub(crate) struct Node {
 
     listen_http_addr: String,
     listen_http_port: u16,
-    listen_https_port: Option<u16>,
 
     listen_pg_addr: String,
     listen_pg_port: u16,
 
     availability_zone_id: AvailabilityZone,
 
-    // Flag from storcon's config to use https for pageserver admin API.
-    // Invariant: if |true|, listen_https_port should contain a value.
-    use_https: bool,
     // This cancellation token means "stop any RPCs in flight to this node, and don't start
     // any more". It is not related to process shutdown.
     #[serde(skip)]
@@ -61,16 +56,7 @@ pub(crate) enum AvailabilityTransition {
 
 impl Node {
     pub(crate) fn base_url(&self) -> String {
-        if self.use_https {
-            format!(
-                "https://{}:{}",
-                self.listen_http_addr,
-                self.listen_https_port
-                    .expect("https port should be specified if use_https is on")
-            )
-        } else {
-            format!("http://{}:{}", self.listen_http_addr, self.listen_http_port)
-        }
+        format!("http://{}:{}", self.listen_http_addr, self.listen_http_port)
     }
 
     pub(crate) fn get_id(&self) -> NodeId {
@@ -96,20 +82,11 @@ impl Node {
         self.id == register_req.node_id
             && self.listen_http_addr == register_req.listen_http_addr
             && self.listen_http_port == register_req.listen_http_port
-            // Note: listen_https_port may change. See [`Self::need_update`] for mode details.
-            // && self.listen_https_port == register_req.listen_https_port
             && self.listen_pg_addr == register_req.listen_pg_addr
             && self.listen_pg_port == register_req.listen_pg_port
             && self.availability_zone_id == register_req.availability_zone_id
     }
 
-    // Do we need to update an existing record in DB on this registration request?
-    pub(crate) fn need_update(&self, register_req: &NodeRegisterRequest) -> bool {
-        // listen_https_port is checked here because it may change during migration to https.
-        // After migration, this check may be moved to registration_match.
-        self.listen_https_port != register_req.listen_https_port
-    }
-
     /// For a shard located on this node, populate a response object
     /// with this node's address information.
     pub(crate) fn shard_location(&self, shard_id: TenantShardId) -> TenantLocateResponseShard {
@@ -118,7 +95,6 @@ impl Node {
             node_id: self.id,
             listen_http_addr: self.listen_http_addr.clone(),
             listen_http_port: self.listen_http_port,
-            listen_https_port: self.listen_https_port,
             listen_pg_addr: self.listen_pg_addr.clone(),
             listen_pg_port: self.listen_pg_port,
         }
@@ -199,34 +175,25 @@ impl Node {
         }
     }
 
-    #[allow(clippy::too_many_arguments)]
     pub(crate) fn new(
         id: NodeId,
         listen_http_addr: String,
         listen_http_port: u16,
-        listen_https_port: Option<u16>,
         listen_pg_addr: String,
         listen_pg_port: u16,
         availability_zone_id: AvailabilityZone,
-        use_https: bool,
-    ) -> anyhow::Result<Self> {
-        if use_https && listen_https_port.is_none() {
-            return Err(anyhow!("https is enabled, but node has no https port"));
-        }
-
-        Ok(Self {
+    ) -> Self {
+        Self {
             id,
             listen_http_addr,
             listen_http_port,
-            listen_https_port,
             listen_pg_addr,
             listen_pg_port,
             scheduling: NodeSchedulingPolicy::Active,
             availability: NodeAvailability::Offline,
             availability_zone_id,
-            use_https,
             cancel: CancellationToken::new(),
-        })
+        }
     }
 
     pub(crate) fn to_persistent(&self) -> NodePersistence {
@@ -235,19 +202,14 @@ impl Node {
             scheduling_policy: self.scheduling.into(),
             listen_http_addr: self.listen_http_addr.clone(),
             listen_http_port: self.listen_http_port as i32,
-            listen_https_port: self.listen_https_port.map(|x| x as i32),
             listen_pg_addr: self.listen_pg_addr.clone(),
             listen_pg_port: self.listen_pg_port as i32,
             availability_zone_id: self.availability_zone_id.0.clone(),
         }
     }
 
-    pub(crate) fn from_persistent(np: NodePersistence, use_https: bool) -> anyhow::Result<Self> {
-        if use_https && np.listen_https_port.is_none() {
-            return Err(anyhow!("https is enabled, but node has no https port"));
-        }
-
-        Ok(Self {
+    pub(crate) fn from_persistent(np: NodePersistence) -> Self {
+        Self {
             id: NodeId(np.node_id as u64),
             // At startup we consider a node offline until proven otherwise.
             availability: NodeAvailability::Offline,
@@ -255,13 +217,11 @@ impl Node {
                 .expect("Bad scheduling policy in DB"),
             listen_http_addr: np.listen_http_addr,
             listen_http_port: np.listen_http_port as u16,
-            listen_https_port: np.listen_https_port.map(|x| x as u16),
             listen_pg_addr: np.listen_pg_addr,
             listen_pg_port: np.listen_pg_port as u16,
             availability_zone_id: AvailabilityZone(np.availability_zone_id),
-            use_https,
             cancel: CancellationToken::new(),
-        })
+        }
     }
 
     /// Wrapper for issuing requests to pageserver management API: takes care of generic
@@ -325,9 +285,8 @@ impl Node {
             warn_threshold,
             max_retries,
             &format!(
-                "Call to node {} ({}) management API",
-                self.id,
-                self.base_url(),
+                "Call to node {} ({}:{}) management API",
+                self.id, self.listen_http_addr, self.listen_http_port
             ),
             cancel,
         )
@@ -343,7 +302,6 @@ impl Node {
             availability_zone_id: self.availability_zone_id.0.clone(),
             listen_http_addr: self.listen_http_addr.clone(),
             listen_http_port: self.listen_http_port,
-            listen_https_port: self.listen_https_port,
             listen_pg_addr: self.listen_pg_addr.clone(),
             listen_pg_port: self.listen_pg_port,
         }

storage_controller/src/pageserver_client.rs

@@ -280,22 +280,6 @@ impl PageserverClient {
         )
     }
 
-    pub(crate) async fn timeline_download_heatmap_layers(
-        &self,
-        tenant_shard_id: TenantShardId,
-        timeline_id: TimelineId,
-        concurrency: Option<usize>,
-    ) -> Result<()> {
-        measured_request!(
-            "download_heatmap_layers",
-            crate::metrics::Method::Post,
-            &self.node_id_label,
-            self.inner
-                .timeline_download_heatmap_layers(tenant_shard_id, timeline_id, concurrency)
-                .await
-        )
-    }
-
     pub(crate) async fn get_utilization(&self) -> Result<PageserverUtilization> {
         measured_request!(
             "utilization",

storage_controller/src/persistence.rs

@@ -375,23 +375,18 @@ impl Persistence {
         Ok(nodes)
     }
 
-    pub(crate) async fn update_node<V>(
+    pub(crate) async fn update_node(
         &self,
         input_node_id: NodeId,
-        values: V,
-    ) -> DatabaseResult<()>
-    where
-        V: diesel::AsChangeset<Target = crate::schema::nodes::table> + Clone + Send + Sync,
-        V::Changeset: diesel::query_builder::QueryFragment<diesel::pg::Pg> + Send, // valid Postgres SQL
-    {
+        input_scheduling: NodeSchedulingPolicy,
+    ) -> DatabaseResult<()> {
         use crate::schema::nodes::dsl::*;
         let updated = self
             .with_measured_conn(DatabaseOperation::UpdateNode, move |conn| {
-                let values = values.clone();
                 Box::pin(async move {
                     let updated = diesel::update(nodes)
                         .filter(node_id.eq(input_node_id.0 as i64))
-                        .set(values)
+                        .set((scheduling_policy.eq(String::from(input_scheduling)),))
                         .execute(conn)
                         .await?;
                     Ok(updated)
@@ -408,32 +403,6 @@ impl Persistence {
         }
     }
 
-    pub(crate) async fn update_node_scheduling_policy(
-        &self,
-        input_node_id: NodeId,
-        input_scheduling: NodeSchedulingPolicy,
-    ) -> DatabaseResult<()> {
-        use crate::schema::nodes::dsl::*;
-        self.update_node(
-            input_node_id,
-            scheduling_policy.eq(String::from(input_scheduling)),
-        )
-        .await
-    }
-
-    pub(crate) async fn update_node_on_registration(
-        &self,
-        input_node_id: NodeId,
-        input_https_port: Option<u16>,
-    ) -> DatabaseResult<()> {
-        use crate::schema::nodes::dsl::*;
-        self.update_node(
-            input_node_id,
-            listen_https_port.eq(input_https_port.map(|x| x as i32)),
-        )
-        .await
-    }
-
     /// At startup, load the high level state for shards, such as their config + policy.  This will
     /// be enriched at runtime with state discovered on pageservers.
     ///
@@ -1483,7 +1452,6 @@ pub(crate) struct NodePersistence {
     pub(crate) listen_pg_addr: String,
     pub(crate) listen_pg_port: i32,
     pub(crate) availability_zone_id: String,
-    pub(crate) listen_https_port: Option<i32>,
 }
 
 /// Tenant metadata health status that are stored durably.