run cargo hakari generate

Effectively, this means we use the tokio runtime's spawn_blocking-thread-pool
to execute the layer reads, instead of doing the reads on the
tokio runtime's main executor threads.

The use of the thread pool adds some overhead, but, not blocking
the main executor threads is more important, because they can now
execute other async tasks while we do the IO.

With a sufficiently large spawn_blocking-thread-pool, we also get more
IO parallelism between timelines than with blocking the main executor
threads. So, we might push the pageserver's NVMe closer to its limits.
But right now, there's lots of headroom.

.cargo/config.toml

@@ -14,4 +14,3 @@ opt-level = 1
 
 [alias]
 build_testing = ["build", "--features", "testing"]
-neon = ["run", "--bin", "neon_local"]

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

@@ -1,186 +0,0 @@
-name: 'Create Allure report'
-description: 'Generate Allure report from uploaded by actions/allure-report-store tests results'
-
-outputs:
-  report-url:
-    description: 'Allure report URL'
-    value: ${{ steps.generate-report.outputs.report-url }}
-  report-json-url:
-    description: 'Allure report JSON URL'
-    value: ${{ steps.generate-report.outputs.report-json-url }}
-
-runs:
-  using: "composite"
-
-  steps:
-    # We're using some of env variables quite offen, so let's set them once.
-    #
-    # It would be nice to have them set in common runs.env[0] section, but it doesn't work[1]
-    #
-    # - [0] https://docs.github.com/en/actions/creating-actions/metadata-syntax-for-github-actions#runsenv
-    # - [1] https://github.com/neondatabase/neon/pull/3907#discussion_r1154703456
-    #
-    - name: Set variables
-      shell: bash -euxo pipefail {0}
-      run: |
-        PR_NUMBER=$(jq --raw-output .pull_request.number "$GITHUB_EVENT_PATH" || true)
-        if [ "${PR_NUMBER}" != "null" ]; then
-          BRANCH_OR_PR=pr-${PR_NUMBER}
-        elif [ "${GITHUB_REF_NAME}" = "main" ] || [ "${GITHUB_REF_NAME}" = "release" ]; then
-          # Shortcut for special branches
-          BRANCH_OR_PR=${GITHUB_REF_NAME}
-        else
-          BRANCH_OR_PR=branch-$(printf "${GITHUB_REF_NAME}" | tr -c "[:alnum:]._-" "-")
-        fi
-
-        LOCK_FILE=reports/${BRANCH_OR_PR}/lock.txt
-
-        WORKDIR=/tmp/${BRANCH_OR_PR}-$(date +%s)
-        mkdir -p ${WORKDIR}
-
-        echo "BRANCH_OR_PR=${BRANCH_OR_PR}" >> $GITHUB_ENV
-        echo "LOCK_FILE=${LOCK_FILE}"       >> $GITHUB_ENV
-        echo "WORKDIR=${WORKDIR}"           >> $GITHUB_ENV
-        echo "BUCKET=${BUCKET}"             >> $GITHUB_ENV
-      env:
-        BUCKET: neon-github-public-dev
-
-    # TODO: We can replace with a special docker image with Java and Allure pre-installed
-    - uses: actions/setup-java@v3
-      with:
-        distribution: 'temurin'
-        java-version: '17'
-
-    - name: Install Allure
-      shell: bash -euxo pipefail {0}
-      run: |
-        if ! which allure; then
-          ALLURE_ZIP=allure-${ALLURE_VERSION}.zip
-          wget -q https://github.com/allure-framework/allure2/releases/download/${ALLURE_VERSION}/${ALLURE_ZIP}
-          echo "${ALLURE_ZIP_SHA256} ${ALLURE_ZIP}" | sha256sum --check
-          unzip -q ${ALLURE_ZIP}
-          echo "$(pwd)/allure-${ALLURE_VERSION}/bin" >> $GITHUB_PATH
-          rm -f ${ALLURE_ZIP}
-        fi
-      env:
-        ALLURE_VERSION: 2.22.1
-        ALLURE_ZIP_SHA256: fdc7a62d94b14c5e0bf25198ae1feded6b005fdbed864b4d3cb4e5e901720b0b
-
-    # Potentially we could have several running build for the same key (for example, for the main branch), so we use improvised lock for this
-    - name: Acquire lock
-      shell: bash -euxo pipefail {0}
-      run: |
-        LOCK_TIMEOUT=300 # seconds
-
-        LOCK_CONTENT="${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT}"
-        echo ${LOCK_CONTENT} > ${WORKDIR}/lock.txt
-
-        # Do it up to 5 times to avoid race condition
-        for _ in $(seq 1 5); do
-          for i in $(seq 1 ${LOCK_TIMEOUT}); do
-            LOCK_ACQUIRED=$(aws s3api head-object --bucket neon-github-public-dev --key ${LOCK_FILE} | jq --raw-output '.LastModified' || true)
-            # `date --date="..."` is supported only by gnu date (i.e. it doesn't work on BSD/macOS)
-            if [ -z "${LOCK_ACQUIRED}" ] || [ "$(( $(date +%s) - $(date --date="${LOCK_ACQUIRED}" +%s) ))" -gt "${LOCK_TIMEOUT}" ]; then
-              break
-            fi
-            sleep 1
-          done
-
-          aws s3 mv --only-show-errors ${WORKDIR}/lock.txt "s3://${BUCKET}/${LOCK_FILE}"
-
-          # Double-check that exactly THIS run has acquired the lock
-          aws s3 cp --only-show-errors "s3://${BUCKET}/${LOCK_FILE}" ./lock.txt
-          if [ "$(cat lock.txt)" = "${LOCK_CONTENT}" ]; then
-            break
-          fi
-        done
-
-    - name: Generate and publish final Allure report
-      id: generate-report
-      shell: bash -euxo pipefail {0}
-      run: |
-        REPORT_PREFIX=reports/${BRANCH_OR_PR}
-        RAW_PREFIX=reports-raw/${BRANCH_OR_PR}/${GITHUB_RUN_ID}
-
-        # Get previously uploaded data for this run
-        ZSTD_NBTHREADS=0
-
-        S3_FILEPATHS=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${RAW_PREFIX}/ | jq --raw-output  '.Contents[].Key')
-        if [ -z "$S3_FILEPATHS" ]; then
-          # There's no previously uploaded data for this $GITHUB_RUN_ID
-          exit 0
-        fi
-        for S3_FILEPATH in ${S3_FILEPATHS}; do
-          time aws s3 cp --only-show-errors "s3://${BUCKET}/${S3_FILEPATH}" "${WORKDIR}"
-
-          archive=${WORKDIR}/$(basename $S3_FILEPATH)
-          mkdir -p ${archive%.tar.zst}
-          time tar -xf ${archive} -C ${archive%.tar.zst}
-          rm -f ${archive}
-        done
-
-        # Get history trend
-        time aws s3 cp --recursive --only-show-errors "s3://${BUCKET}/${REPORT_PREFIX}/latest/history" "${WORKDIR}/latest/history" || true
-
-        # Generate report
-        time allure generate --clean --output ${WORKDIR}/report ${WORKDIR}/*
-
-        # Replace a logo link with a redirect to the latest version of the report
-        sed -i 's|<a href="." class=|<a href="https://'${BUCKET}'.s3.amazonaws.com/'${REPORT_PREFIX}'/latest/index.html?nocache='"'+Date.now()+'"'" class=|g' ${WORKDIR}/report/app.js
-
-        # Upload a history and the final report (in this particular order to not to have duplicated history in 2 places)
-        time aws s3 mv --recursive --only-show-errors "${WORKDIR}/report/history" "s3://${BUCKET}/${REPORT_PREFIX}/latest/history"
-        time aws s3 mv --recursive --only-show-errors "${WORKDIR}/report" "s3://${BUCKET}/${REPORT_PREFIX}/${GITHUB_RUN_ID}"
-
-        REPORT_URL=https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/${GITHUB_RUN_ID}/index.html
-
-        # Generate redirect
-        cat <<EOF > ${WORKDIR}/index.html
-          <!DOCTYPE html>
-
-          <meta charset="utf-8">
-          <title>Redirecting to ${REPORT_URL}</title>
-          <meta http-equiv="refresh" content="0; URL=${REPORT_URL}">
-        EOF
-        time aws s3 cp --only-show-errors ${WORKDIR}/index.html "s3://${BUCKET}/${REPORT_PREFIX}/latest/index.html"
-
-        echo "report-url=${REPORT_URL}"                                   >> $GITHUB_OUTPUT
-        echo "report-json-url=${REPORT_URL%/index.html}/data/suites.json" >> $GITHUB_OUTPUT
-
-        echo "[Allure Report](${REPORT_URL})" >> ${GITHUB_STEP_SUMMARY}
-
-    - name: Release lock
-      if: always()
-      shell: bash -euxo pipefail {0}
-      run: |
-        aws s3 cp --only-show-errors "s3://${BUCKET}/${LOCK_FILE}" ./lock.txt || exit 0
-
-        if [ "$(cat lock.txt)" = "${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT}" ]; then
-          aws s3 rm "s3://${BUCKET}/${LOCK_FILE}"
-        fi
-
-    - name: Cleanup
-      if: always()
-      shell: bash -euxo pipefail {0}
-      run: |
-        if [ -d "${WORKDIR}" ]; then
-          rm -rf ${WORKDIR}
-        fi
-
-    - uses: actions/github-script@v6
-      if: always()
-      env:
-        REPORT_URL: ${{ steps.generate-report.outputs.report-url }}
-        COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
-      with:
-        script: |
-          const { REPORT_URL, COMMIT_SHA } = process.env
-
-          await github.rest.repos.createCommitStatus({
-            owner: context.repo.owner,
-            repo: context.repo.repo,
-            sha: `${COMMIT_SHA}`,
-            state: 'success',
-            target_url: `${REPORT_URL}`,
-            context: 'Allure report',
-          })

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

@@ -1,72 +0,0 @@
-name: 'Store Allure results'
-description: 'Upload test results to be used by actions/allure-report-generate'
-
-inputs:
-  report-dir:
-    description: 'directory with test results generated by tests'
-    required: true
-  unique-key:
-    description: 'string to distinguish different results in the same run'
-    required: true
-
-runs:
-  using: "composite"
-
-  steps:
-    - name: Set variables
-      shell: bash -euxo pipefail {0}
-      run: |
-        PR_NUMBER=$(jq --raw-output .pull_request.number "$GITHUB_EVENT_PATH" || true)
-        if [ "${PR_NUMBER}" != "null" ]; then
-          BRANCH_OR_PR=pr-${PR_NUMBER}
-        elif [ "${GITHUB_REF_NAME}" = "main" ] || [ "${GITHUB_REF_NAME}" = "release" ]; then
-          # Shortcut for special branches
-          BRANCH_OR_PR=${GITHUB_REF_NAME}
-        else
-          BRANCH_OR_PR=branch-$(printf "${GITHUB_REF_NAME}" | tr -c "[:alnum:]._-" "-")
-        fi
-
-        echo "BRANCH_OR_PR=${BRANCH_OR_PR}" >> $GITHUB_ENV
-        echo "REPORT_DIR=${REPORT_DIR}"     >> $GITHUB_ENV
-      env:
-        REPORT_DIR: ${{ inputs.report-dir }}
-
-    - name: Upload test results
-      shell: bash -euxo pipefail {0}
-      run: |
-        REPORT_PREFIX=reports/${BRANCH_OR_PR}
-        RAW_PREFIX=reports-raw/${BRANCH_OR_PR}/${GITHUB_RUN_ID}
-
-        # Add metadata
-        cat <<EOF > ${REPORT_DIR}/executor.json
-          {
-            "name": "GitHub Actions",
-            "type": "github",
-            "url": "https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/latest/index.html",
-            "buildOrder": ${GITHUB_RUN_ID},
-            "buildName": "GitHub Actions Run #${GITHUB_RUN_NUMBER}/${GITHUB_RUN_ATTEMPT}",
-            "buildUrl": "${GITHUB_SERVER_URL}/${GITHUB_REPOSITORY}/actions/runs/${GITHUB_RUN_ID}/attempts/${GITHUB_RUN_ATTEMPT}",
-            "reportUrl": "https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/${GITHUB_RUN_ID}/index.html",
-            "reportName": "Allure Report"
-          }
-        EOF
-
-        cat <<EOF > ${REPORT_DIR}/environment.properties
-          COMMIT_SHA=${COMMIT_SHA}
-        EOF
-
-        ARCHIVE="${UNIQUE_KEY}-${GITHUB_RUN_ATTEMPT}-$(date +%s).tar.zst"
-        ZSTD_NBTHREADS=0
-
-        time tar -C ${REPORT_DIR} -cf ${ARCHIVE} --zstd .
-        time aws s3 mv --only-show-errors ${ARCHIVE} "s3://${BUCKET}/${RAW_PREFIX}/${ARCHIVE}"
-      env:
-        UNIQUE_KEY: ${{ inputs.unique-key }}
-        COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
-        BUCKET: neon-github-public-dev
-
-    - name: Cleanup
-      if: always()
-      shell: bash -euxo pipefail {0}
-      run: |
-        rm -rf ${REPORT_DIR}

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

@@ -0,0 +1,254 @@
+name: 'Create Allure report'
+description: 'Create and publish Allure report'
+
+inputs:
+  action:
+    desctiption: 'generate or store'
+    required: true
+  build_type:
+    description: '`build_type` from run-python-test-set action'
+    required: true
+  test_selection:
+    description: '`test_selector` from run-python-test-set action'
+    required: false
+outputs:
+  report-url:
+    description: 'Allure report URL'
+    value: ${{ steps.generate-report.outputs.report-url }}
+  report-json-url:
+    description: 'Allure report JSON URL'
+    value: ${{ steps.generate-report.outputs.report-json-url }}
+
+runs:
+  using: "composite"
+
+  steps:
+    # We're using some of env variables quite offen, so let's set them once.
+    #
+    # It would be nice to have them set in common runs.env[0] section, but it doesn't work[1]
+    #
+    # - [0] https://docs.github.com/en/actions/creating-actions/metadata-syntax-for-github-actions#runsenv
+    # - [1] https://github.com/neondatabase/neon/pull/3907#discussion_r1154703456
+    #
+    - name: Set common environment variables
+      shell: bash -euxo pipefail {0}
+      run: |
+        echo "BUILD_TYPE=${BUILD_TYPE}"   >> $GITHUB_ENV
+        echo "BUCKET=${BUCKET}"           >> $GITHUB_ENV
+        echo "TEST_OUTPUT=${TEST_OUTPUT}" >> $GITHUB_ENV
+      env:
+        BUILD_TYPE: ${{ inputs.build_type }}
+        BUCKET: neon-github-public-dev
+        TEST_OUTPUT: /tmp/test_output
+
+    - name: Validate input parameters
+      shell: bash -euxo pipefail {0}
+      run: |
+        if [ "${{ inputs.action }}" != "store" ] && [ "${{ inputs.action }}" != "generate" ]; then
+          echo >&2 "Unknown inputs.action type '${{ inputs.action }}'; allowed 'generate' or 'store' only"
+          exit 1
+        fi
+
+        if [ -z "${{ inputs.test_selection }}" ] && [ "${{ inputs.action }}" == "store" ]; then
+          echo >&2 "inputs.test_selection must be set for 'store' action"
+          exit 2
+        fi
+
+    - name: Calculate variables
+      id: calculate-vars
+      shell: bash -euxo pipefail {0}
+      run: |
+        # TODO: for manually triggered workflows (via workflow_dispatch) we need to have a separate key
+
+        pr_number=$(jq --raw-output .pull_request.number "$GITHUB_EVENT_PATH" || true)
+        if [ "${pr_number}" != "null" ]; then
+          key=pr-${pr_number}
+        elif [ "${GITHUB_REF_NAME}" = "main" ]; then
+          # Shortcut for a special branch
+          key=main
+        elif [ "${GITHUB_REF_NAME}" = "release" ]; then
+          # Shortcut for a special branch
+          key=release
+        else
+          key=branch-$(printf "${GITHUB_REF_NAME}" | tr -c "[:alnum:]._-" "-")
+        fi
+        echo "KEY=${key}" >> $GITHUB_OUTPUT
+
+        # Sanitize test selection to remove `/` and any other special characters
+        # Use printf instead of echo to avoid having `\n` at the end of the string
+        test_selection=$(printf "${{ inputs.test_selection }}" | tr -c "[:alnum:]._-" "-" )
+        echo "TEST_SELECTION=${test_selection}" >> $GITHUB_OUTPUT
+
+    - uses: actions/setup-java@v3
+      if: ${{ inputs.action == 'generate' }}
+      with:
+        distribution: 'temurin'
+        java-version: '17'
+
+    - name: Install Allure
+      if: ${{ inputs.action == 'generate' }}
+      shell: bash -euxo pipefail {0}
+      run: |
+        if ! which allure; then
+          ALLURE_ZIP=allure-${ALLURE_VERSION}.zip
+          wget -q https://github.com/allure-framework/allure2/releases/download/${ALLURE_VERSION}/${ALLURE_ZIP}
+          echo "${ALLURE_ZIP_MD5}  ${ALLURE_ZIP}" | md5sum -c
+          unzip -q ${ALLURE_ZIP}
+          echo "$(pwd)/allure-${ALLURE_VERSION}/bin" >> $GITHUB_PATH
+          rm -f ${ALLURE_ZIP}
+        fi
+      env:
+        ALLURE_VERSION: 2.21.0
+        ALLURE_ZIP_MD5: c8db4dd8e2a7882583d569ed2c82879c
+
+    - name: Upload Allure results
+      if: ${{ inputs.action == 'store' }}
+      env:
+        REPORT_PREFIX: reports/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
+        RAW_PREFIX: reports-raw/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
+        TEST_SELECTION: ${{ steps.calculate-vars.outputs.TEST_SELECTION }}
+      shell: bash -euxo pipefail {0}
+      run: |
+        # Add metadata
+        cat <<EOF > $TEST_OUTPUT/allure/results/executor.json
+          {
+            "name": "GitHub Actions",
+            "type": "github",
+            "url": "https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/latest/index.html",
+            "buildOrder": ${GITHUB_RUN_ID},
+            "buildName": "GitHub Actions Run #${{ github.run_number }}/${GITHUB_RUN_ATTEMPT}",
+            "buildUrl": "${GITHUB_SERVER_URL}/${GITHUB_REPOSITORY}/actions/runs/${GITHUB_RUN_ID}/attempts/${GITHUB_RUN_ATTEMPT}",
+            "reportUrl": "https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/${GITHUB_RUN_ID}/index.html",
+            "reportName": "Allure Report"
+          }
+        EOF
+        cat <<EOF > $TEST_OUTPUT/allure/results/environment.properties
+          TEST_SELECTION=${{ inputs.test_selection }}
+          BUILD_TYPE=${BUILD_TYPE}
+        EOF
+
+        ARCHIVE="${GITHUB_RUN_ID}-${TEST_SELECTION}-${GITHUB_RUN_ATTEMPT}-$(date +%s).tar.zst"
+        ZSTD_NBTHREADS=0
+
+        tar -C ${TEST_OUTPUT}/allure/results -cf ${ARCHIVE} --zstd .
+        aws s3 mv --only-show-errors ${ARCHIVE} "s3://${BUCKET}/${RAW_PREFIX}/${ARCHIVE}"
+
+    # Potentially we could have several running build for the same key (for example for the main branch), so we use improvised lock for this
+    - name: Acquire Allure lock
+      if: ${{ inputs.action == 'generate' }}
+      shell: bash -euxo pipefail {0}
+      env:
+        LOCK_FILE: reports/${{ steps.calculate-vars.outputs.KEY }}/lock.txt
+        TEST_SELECTION: ${{ steps.calculate-vars.outputs.TEST_SELECTION }}
+      run: |
+        LOCK_TIMEOUT=300 # seconds
+
+        for _ in $(seq 1 5); do
+          for i in $(seq 1 ${LOCK_TIMEOUT}); do
+            LOCK_ADDED=$(aws s3api head-object --bucket neon-github-public-dev --key ${LOCK_FILE} | jq --raw-output '.LastModified' || true)
+            # `date --date="..."` is supported only by gnu date (i.e. it doesn't work on BSD/macOS)
+            if [ -z "${LOCK_ADDED}" ] || [ "$(( $(date +%s) - $(date --date="${LOCK_ADDED}" +%s) ))" -gt "${LOCK_TIMEOUT}" ]; then
+              break
+            fi
+            sleep 1
+          done
+          echo "${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT}-${TEST_SELECTION}" > lock.txt
+          aws s3 mv --only-show-errors lock.txt "s3://${BUCKET}/${LOCK_FILE}"
+
+          # A double-check that exactly WE have acquired the lock
+          aws s3 cp --only-show-errors "s3://${BUCKET}/${LOCK_FILE}" ./lock.txt
+          if [ "$(cat lock.txt)" = "${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT}-${TEST_SELECTION}" ]; then
+            break
+          fi
+        done
+
+    - name: Generate and publish final Allure report
+      if: ${{ inputs.action == 'generate' }}
+      id: generate-report
+      env:
+        REPORT_PREFIX: reports/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
+        RAW_PREFIX: reports-raw/${{ steps.calculate-vars.outputs.KEY }}/${{ inputs.build_type }}
+      shell: bash -euxo pipefail {0}
+      run: |
+        # Get previously uploaded data for this run
+        ZSTD_NBTHREADS=0
+
+        s3_filepaths=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${RAW_PREFIX}/${GITHUB_RUN_ID}- | jq --raw-output  '.Contents[].Key')
+        if [ -z "$s3_filepaths" ]; then
+          # There's no previously uploaded data for this run
+          exit 0
+        fi
+        for s3_filepath in ${s3_filepaths}; do
+          aws s3 cp --only-show-errors "s3://${BUCKET}/${s3_filepath}" "${TEST_OUTPUT}/allure/"
+
+          archive=${TEST_OUTPUT}/allure/$(basename $s3_filepath)
+          mkdir -p ${archive%.tar.zst}
+          tar -xf ${archive} -C ${archive%.tar.zst}
+          rm -f ${archive}
+        done
+
+        # Get history trend
+        aws s3 cp --recursive --only-show-errors "s3://${BUCKET}/${REPORT_PREFIX}/latest/history" "${TEST_OUTPUT}/allure/latest/history" || true
+
+        # Generate report
+        allure generate --clean --output $TEST_OUTPUT/allure/report $TEST_OUTPUT/allure/*
+
+        # Replace a logo link with a redirect to the latest version of the report
+        sed -i 's|<a href="." class=|<a href="https://'${BUCKET}'.s3.amazonaws.com/'${REPORT_PREFIX}'/latest/index.html" class=|g' $TEST_OUTPUT/allure/report/app.js
+
+        # Upload a history and the final report (in this particular order to not to have duplicated history in 2 places)
+        aws s3 mv --recursive --only-show-errors "${TEST_OUTPUT}/allure/report/history" "s3://${BUCKET}/${REPORT_PREFIX}/latest/history"
+        aws s3 mv --recursive --only-show-errors "${TEST_OUTPUT}/allure/report" "s3://${BUCKET}/${REPORT_PREFIX}/${GITHUB_RUN_ID}"
+
+        REPORT_URL=https://${BUCKET}.s3.amazonaws.com/${REPORT_PREFIX}/${GITHUB_RUN_ID}/index.html
+
+        # Generate redirect
+        cat <<EOF > ${TEST_OUTPUT}/allure/index.html
+          <!DOCTYPE html>
+
+          <meta charset="utf-8">
+          <title>Redirecting to ${REPORT_URL}</title>
+          <meta http-equiv="refresh" content="0; URL=${REPORT_URL}">
+        EOF
+        aws s3 cp --only-show-errors ${TEST_OUTPUT}/allure/index.html "s3://${BUCKET}/${REPORT_PREFIX}/latest/index.html"
+
+        echo "[Allure Report](${REPORT_URL})" >> ${GITHUB_STEP_SUMMARY}
+        echo "report-url=${REPORT_URL}" >> $GITHUB_OUTPUT
+        echo "report-json-url=${REPORT_URL%/index.html}/data/suites.json" >> $GITHUB_OUTPUT
+
+    - name: Release Allure lock
+      if: ${{ inputs.action == 'generate' && always() }}
+      shell: bash -euxo pipefail {0}
+      env:
+        LOCK_FILE: reports/${{ steps.calculate-vars.outputs.KEY }}/lock.txt
+        TEST_SELECTION: ${{ steps.calculate-vars.outputs.TEST_SELECTION }}
+      run: |
+        aws s3 cp --only-show-errors "s3://${BUCKET}/${LOCK_FILE}" ./lock.txt || exit 0
+
+        if [ "$(cat lock.txt)" = "${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT}-${TEST_SELECTION}" ]; then
+          aws s3 rm "s3://${BUCKET}/${LOCK_FILE}"
+        fi
+
+    - name: Cleanup
+      if: always()
+      shell: bash -euxo pipefail {0}
+      run: |
+        rm -rf ${TEST_OUTPUT}/allure
+
+    - uses: actions/github-script@v6
+      if: ${{ inputs.action == 'generate' && always() }}
+      env:
+        REPORT_URL: ${{ steps.generate-report.outputs.report-url }}
+        SHA: ${{ github.event.pull_request.head.sha || github.sha }}
+      with:
+        script: |
+          const { REPORT_URL, BUILD_TYPE, SHA } = process.env
+
+          await github.rest.repos.createCommitStatus({
+            owner: context.repo.owner,
+            repo: context.repo.repo,
+            sha: `${SHA}`,
+            state: 'success',
+            target_url: `${REPORT_URL}`,
+            context: `Allure report / ${BUILD_TYPE}`,
+          })

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

@@ -36,14 +36,18 @@ inputs:
     description: 'Region name for real s3 tests'
     required: false
     default: ''
+  real_s3_access_key_id:
+    description: 'Access key id'
+    required: false
+    default: ''
+  real_s3_secret_access_key:
+    description: 'Secret access key'
+    required: false
+    default: ''
   rerun_flaky:
     description: 'Whether to rerun flaky tests'
     required: false
     default: 'false'
-  pg_version:
-    description: 'Postgres version to use for tests'
-    required: false
-    default: 'v14'
 
 runs:
   using: "composite"
@@ -63,12 +67,12 @@ runs:
         path: /tmp/neon-previous
         prefix: latest
 
-    - name: Download compatibility snapshot
+    - name: Download compatibility snapshot for Postgres 14
       if: inputs.build_type != 'remote'
       uses: ./.github/actions/download
       with:
-        name: compatibility-snapshot-${{ inputs.build_type }}-pg${{ inputs.pg_version }}
-        path: /tmp/compatibility_snapshot_pg${{ inputs.pg_version }}
+        name: compatibility-snapshot-${{ inputs.build_type }}-pg14
+        path: /tmp/compatibility_snapshot_pg14
         prefix: latest
 
     - name: Checkout
@@ -96,18 +100,19 @@ runs:
         COMPATIBILITY_POSTGRES_DISTRIB_DIR: /tmp/neon-previous/pg_install
         TEST_OUTPUT: /tmp/test_output
         BUILD_TYPE: ${{ inputs.build_type }}
-        COMPATIBILITY_SNAPSHOT_DIR: /tmp/compatibility_snapshot_pg${{ inputs.pg_version }}
+        AWS_ACCESS_KEY_ID: ${{ inputs.real_s3_access_key_id }}
+        AWS_SECRET_ACCESS_KEY: ${{ inputs.real_s3_secret_access_key }}
+        COMPATIBILITY_SNAPSHOT_DIR: /tmp/compatibility_snapshot_pg14
         ALLOW_BACKWARD_COMPATIBILITY_BREAKAGE: contains(github.event.pull_request.labels.*.name, 'backward compatibility breakage')
         ALLOW_FORWARD_COMPATIBILITY_BREAKAGE: contains(github.event.pull_request.labels.*.name, 'forward compatibility breakage')
         RERUN_FLAKY: ${{ inputs.rerun_flaky }}
-        PG_VERSION: ${{ inputs.pg_version }}
       shell: bash -euxo pipefail {0}
       run: |
         # PLATFORM will be embedded in the perf test report
         # and it is needed to distinguish different environments
         export PLATFORM=${PLATFORM:-github-actions-selfhosted}
         export POSTGRES_DISTRIB_DIR=${POSTGRES_DISTRIB_DIR:-/tmp/neon/pg_install}
-        export DEFAULT_PG_VERSION=${PG_VERSION#v}
+        export DEFAULT_PG_VERSION=${DEFAULT_PG_VERSION:-14}
 
         if [ "${BUILD_TYPE}" = "remote" ]; then
           export REMOTE_ENV=1
@@ -187,18 +192,19 @@ runs:
           scripts/generate_and_push_perf_report.sh
         fi
 
-    - name: Upload compatibility snapshot
+    - name: Upload compatibility snapshot for Postgres 14
       if: github.ref_name == 'release'
       uses: ./.github/actions/upload
       with:
-        name: compatibility-snapshot-${{ inputs.build_type }}-pg${{ inputs.pg_version }}-${{ github.run_id }}
-        # Directory is created by test_compatibility.py::test_create_snapshot, keep the path in sync with the test
-        path: /tmp/test_output/compatibility_snapshot_pg${{ inputs.pg_version }}/
+        name: compatibility-snapshot-${{ inputs.build_type }}-pg14-${{ github.run_id }}
+        # The path includes a test name (test_create_snapshot) and directory that the test creates (compatibility_snapshot_pg14), keep the path in sync with the test
+        path: /tmp/test_output/test_create_snapshot/compatibility_snapshot_pg14/
         prefix: latest
 
-    - name: Upload test results
+    - name: Create Allure report
       if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-store
+      uses: ./.github/actions/allure-report
       with:
-        report-dir: /tmp/test_output/allure/results
-        unique-key: ${{ inputs.build_type }}
+        action: store
+        build_type: ${{ inputs.build_type }}
+        test_selection: ${{ inputs.test_selection }}

.github/pull_request_template.md

@@ -1,6 +1,6 @@
-## Problem
+## Describe your changes
 
-## Summary of changes
+## Issue ticket number and link
 
 ## Checklist before requesting a review
 

.github/workflows/benchmarking.yml

@@ -16,12 +16,12 @@ on:
   workflow_dispatch: # adds ability to run this manually
     inputs:
       region_id:
-        description: 'Project region id. If not set, the default region will be used'
+        description: 'Use a particular region. If not set the default region will be used'
         required: false
         default: 'aws-us-east-2'
       save_perf_report:
         type: boolean
-        description: 'Publish perf report. If not set, the report will be published only for the main branch'
+        description: 'Publish perf report or not. If not set, the report is published only for the main branch'
         required: false
 
 defaults:
@@ -93,7 +93,10 @@ jobs:
 
     - name: Create Allure report
       if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
+      uses: ./.github/actions/allure-report
+      with:
+        action: generate
+        build_type: ${{ env.BUILD_TYPE }}
 
     - name: Post to a Slack channel
       if: ${{ github.event.schedule && failure() }}
@@ -125,14 +128,13 @@ jobs:
         matrix='{
           "platform": [
             "neon-captest-new",
-            "neon-captest-reuse",
-            "neonvm-captest-new"
+            "neon-captest-reuse"
           ],
           "db_size": [ "10gb" ],
-          "include": [{ "platform": "neon-captest-freetier",   "db_size": "3gb"  },
-                      { "platform": "neon-captest-new",        "db_size": "50gb" },
-                      { "platform": "neonvm-captest-freetier", "db_size": "3gb"  },
-                      { "platform": "neonvm-captest-new",      "db_size": "50gb" }]
+          "include": [
+            { "platform": "neon-captest-freetier", "db_size": "3gb"  },
+            { "platform": "neon-captest-new",      "db_size": "50gb" }
+          ]
         }'
 
         if [ "$(date +%A)" = "Saturday" ]; then
@@ -198,7 +200,7 @@ jobs:
         echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
 
     - name: Create Neon Project
-      if: contains(fromJson('["neon-captest-new", "neon-captest-freetier", "neonvm-captest-new", "neonvm-captest-freetier"]'), matrix.platform)
+      if: contains(fromJson('["neon-captest-new", "neon-captest-freetier"]'), matrix.platform)
       id: create-neon-project
       uses: ./.github/actions/neon-project-create
       with:
@@ -206,7 +208,6 @@ jobs:
         postgres_version: ${{ env.DEFAULT_PG_VERSION }}
         api_key: ${{ secrets.NEON_STAGING_API_KEY }}
         compute_units: ${{ (matrix.platform == 'neon-captest-freetier' && '[0.25, 0.25]') || '[1, 1]' }}
-        provisioner: ${{ (contains(matrix.platform, 'neonvm-') && 'k8s-neonvm') || 'k8s-pod' }}
 
     - name: Set up Connection String
       id: set-up-connstr
@@ -215,7 +216,7 @@ jobs:
           neon-captest-reuse)
             CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR }}
             ;;
-          neon-captest-new | neon-captest-freetier | neonvm-captest-new | neonvm-captest-freetier)
+          neon-captest-new | neon-captest-freetier)
             CONNSTR=${{ steps.create-neon-project.outputs.dsn }}
             ;;
           rds-aurora)
@@ -225,7 +226,7 @@ jobs:
             CONNSTR=${{ secrets.BENCHMARK_RDS_POSTGRES_CONNSTR }}
             ;;
           *)
-            echo >&2 "Unknown PLATFORM=${PLATFORM}"
+            echo >&2 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'neon-captest-new', 'neon-captest-freetier', 'rds-aurora', or 'rds-postgres'"
             exit 1
             ;;
         esac
@@ -282,7 +283,10 @@ jobs:
 
     - name: Create Allure report
       if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
+      uses: ./.github/actions/allure-report
+      with:
+        action: generate
+        build_type: ${{ env.BUILD_TYPE }}
 
     - name: Post to a Slack channel
       if: ${{ github.event.schedule && failure() }}
@@ -376,7 +380,10 @@ jobs:
 
     - name: Create Allure report
       if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
+      uses: ./.github/actions/allure-report
+      with:
+        action: generate
+        build_type: ${{ env.BUILD_TYPE }}
 
     - name: Post to a Slack channel
       if: ${{ github.event.schedule && failure() }}
@@ -469,7 +476,10 @@ jobs:
 
     - name: Create Allure report
       if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
+      uses: ./.github/actions/allure-report
+      with:
+        action: generate
+        build_type: ${{ env.BUILD_TYPE }}
 
     - name: Post to a Slack channel
       if: ${{ github.event.schedule && failure() }}
@@ -556,13 +566,16 @@ jobs:
 
     - name: Create Allure report
       if: ${{ !cancelled() }}
-      uses: ./.github/actions/allure-report-generate
+      uses: ./.github/actions/allure-report
+      with:
+        action: generate
+        build_type: ${{ env.BUILD_TYPE }}
 
     - name: Post to a Slack channel
       if: ${{ github.event.schedule && failure() }}
       uses: slackapi/slack-github-action@v1
       with:
         channel-id: "C033QLM5P7D" # dev-staging-stream
-        slack-message: "Periodic User example perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
+        slack-message: "Periodic TPC-H perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
       env:
         SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

.github/workflows/build_and_test.yml

@@ -324,14 +324,12 @@ jobs:
     runs-on: [ self-hosted, gen3, large ]
     container:
       image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
-      # Default shared memory is 64mb
-      options: --init --shm-size=512mb
+      options: --init
     needs: [ build-neon ]
     strategy:
       fail-fast: false
       matrix:
         build_type: [ debug, release ]
-        pg_version: [ v14, v15 ]
     steps:
       - name: Checkout
         uses: actions/checkout@v3
@@ -346,24 +344,24 @@ jobs:
           test_selection: regress
           needs_postgres_source: true
           run_with_real_s3: true
-          real_s3_bucket: neon-github-ci-tests
-          real_s3_region: eu-central-1
+          real_s3_bucket: ci-tests-s3
+          real_s3_region: us-west-2
+          real_s3_access_key_id: "${{ secrets.AWS_ACCESS_KEY_ID_CI_TESTS_S3 }}"
+          real_s3_secret_access_key: "${{ secrets.AWS_SECRET_ACCESS_KEY_CI_TESTS_S3 }}"
           rerun_flaky: true
-          pg_version: ${{ matrix.pg_version }}
         env:
           TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR }}
           CHECK_ONDISK_DATA_COMPATIBILITY: nonempty
 
       - name: Merge and upload coverage data
-        if: matrix.build_type == 'debug' && matrix.pg_version == 'v14'
+        if: matrix.build_type == 'debug'
         uses: ./.github/actions/save-coverage-data
 
   benchmarks:
     runs-on: [ self-hosted, gen3, small ]
     container:
       image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
-      # Default shared memory is 64mb
-      options: --init --shm-size=512mb
+      options: --init
     needs: [ build-neon ]
     if: github.ref_name == 'main' || contains(github.event.pull_request.labels.*.name, 'run-benchmarks')
     strategy:
@@ -401,10 +399,21 @@ jobs:
     steps:
       - uses: actions/checkout@v3
 
-      - name: Create Allure report
+      - name: Create Allure report (debug)
         if: ${{ !cancelled() }}
-        id: create-allure-report
-        uses: ./.github/actions/allure-report-generate
+        id: create-allure-report-debug
+        uses: ./.github/actions/allure-report
+        with:
+          action: generate
+          build_type: debug
+
+      - name: Create Allure report (release)
+        if: ${{ !cancelled() }}
+        id: create-allure-report-release
+        uses: ./.github/actions/allure-report
+        with:
+          action: generate
+          build_type: release
 
       - uses: actions/github-script@v6
         if: >
@@ -414,37 +423,52 @@ jobs:
           # Retry script for 5XX server errors: https://github.com/actions/github-script#retries
           retries: 5
           script: |
-            const report = {
-              reportUrl:     "${{ steps.create-allure-report.outputs.report-url }}",
-              reportJsonUrl: "${{ steps.create-allure-report.outputs.report-json-url }}",
-            }
+            const reports = [{
+              buildType: "debug",
+              reportUrl: "${{ steps.create-allure-report-debug.outputs.report-url }}",
+              jsonUrl:   "${{ steps.create-allure-report-debug.outputs.report-json-url }}",
+            }, {
+              buildType: "release",
+              reportUrl: "${{ steps.create-allure-report-release.outputs.report-url }}",
+              jsonUrl:   "${{ steps.create-allure-report-release.outputs.report-json-url }}",
+            }]
 
             const script = require("./scripts/pr-comment-test-report.js")
             await script({
               github,
               context,
               fetch,
-              report,
+              reports,
             })
 
       - name: Store Allure test stat in the DB
-        if: ${{ !cancelled() && steps.create-allure-report.outputs.report-json-url }}
+        if: >
+          !cancelled() && (
+            steps.create-allure-report-debug.outputs.report-url ||
+            steps.create-allure-report-release.outputs.report-url
+          )
         env:
-          COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
-          REPORT_JSON_URL: ${{ steps.create-allure-report.outputs.report-json-url }}
+          SHA: ${{ github.event.pull_request.head.sha || github.sha }}
+          REPORT_JSON_URL_DEBUG: ${{ steps.create-allure-report-debug.outputs.report-json-url }}
+          REPORT_JSON_URL_RELEASE: ${{ steps.create-allure-report-release.outputs.report-json-url }}
           TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR }}
         run: |
           ./scripts/pysync
 
-          curl --fail --output suites.json "${REPORT_JSON_URL}"
-          export BUILD_TYPE=unified
-          export DATABASE_URL="$TEST_RESULT_CONNSTR"
+          for report_url in $REPORT_JSON_URL_DEBUG $REPORT_JSON_URL_RELEASE; do
+            if [ -z "$report_url" ]; then
+              continue
+            fi
 
-          poetry run python3 scripts/ingest_regress_test_result.py \
-            --revision ${COMMIT_SHA} \
-            --reference ${GITHUB_REF} \
-            --build-type ${BUILD_TYPE} \
-            --ingest suites.json
+            if [[ "$report_url" == "$REPORT_JSON_URL_DEBUG" ]]; then
+              BUILD_TYPE=debug
+            else
+              BUILD_TYPE=release
+            fi
+
+            curl --fail --output suites.json "${report_url}"
+            DATABASE_URL="$TEST_RESULT_CONNSTR" poetry run python3 scripts/ingest_regress_test_result.py --revision ${SHA} --reference ${GITHUB_REF} --build-type ${BUILD_TYPE} --ingest suites.json
+          done
 
   coverage-report:
     runs-on: [ self-hosted, gen3, small ]
@@ -490,43 +514,37 @@ jobs:
       - name: Merge coverage data
         run: scripts/coverage "--profraw-prefix=$GITHUB_JOB" --dir=/tmp/coverage merge
 
-      - name: Build coverage report
-        env:
-          COMMIT_URL: ${{ github.server_url }}/${{ github.repository }}/commit/${{ github.event.pull_request.head.sha || github.sha }}
+      - name: Build and upload coverage report
         run: |
+          COMMIT_SHA=${{ github.event.pull_request.head.sha }}
+          COMMIT_SHA=${COMMIT_SHA:-${{ github.sha }}}
+          COMMIT_URL=https://github.com/${{ github.repository }}/commit/$COMMIT_SHA
+
           scripts/coverage \
             --dir=/tmp/coverage report \
             --input-objects=/tmp/coverage/binaries.list \
-            --commit-url=${COMMIT_URL} \
+            --commit-url=$COMMIT_URL \
             --format=github
 
-      - name: Upload coverage report
-        id: upload-coverage-report
-        env:
-          BUCKET: neon-github-public-dev
-          COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
-        run: |
-          aws s3 cp --only-show-errors --recursive /tmp/coverage/report s3://neon-github-public-dev/code-coverage/${COMMIT_SHA}
+          REPORT_URL=https://${{ github.repository_owner }}.github.io/zenith-coverage-data/$COMMIT_SHA
 
-          REPORT_URL=https://${BUCKET}.s3.amazonaws.com/code-coverage/${COMMIT_SHA}/index.html
-          echo "report-url=${REPORT_URL}" >> $GITHUB_OUTPUT
+          scripts/git-upload \
+            --repo=https://${{ secrets.VIP_VAP_ACCESS_TOKEN }}@github.com/${{ github.repository_owner }}/zenith-coverage-data.git \
+            --message="Add code coverage for $COMMIT_URL" \
+            copy /tmp/coverage/report $COMMIT_SHA # COPY FROM TO_RELATIVE
 
-      - uses: actions/github-script@v6
-        env:
-          REPORT_URL: ${{ steps.upload-coverage-report.outputs.report-url }}
-          COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }}
-        with:
-          script: |
-            const { REPORT_URL, COMMIT_SHA } = process.env
-
-            await github.rest.repos.createCommitStatus({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              sha: `${COMMIT_SHA}`,
-              state: 'success',
-              target_url: `${REPORT_URL}`,
-              context: 'Code coverage report',
-            })
+          # Add link to the coverage report to the commit
+          curl -f -X POST \
+          https://api.github.com/repos/${{ github.repository }}/statuses/$COMMIT_SHA \
+          -H "Accept: application/vnd.github.v3+json" \
+          --user "${{ secrets.CI_ACCESS_TOKEN }}" \
+          --data \
+            "{
+              \"state\": \"success\",
+              \"context\": \"neon-coverage\",
+              \"description\": \"Coverage report is ready\",
+              \"target_url\": \"$REPORT_URL\"
+            }"
 
   trigger-e2e-tests:
     runs-on: [ self-hosted, gen3, small ]
@@ -661,9 +679,6 @@ jobs:
           project: nrdv0s4kcs
           push: true
           tags: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:depot-${{needs.tag.outputs.build-tag}}
-          build-args: |
-            GIT_VERSION=${{ github.sha }}
-            REPOSITORY=369495373322.dkr.ecr.eu-central-1.amazonaws.com
 
   compute-tools-image:
     runs-on: [ self-hosted, gen3, large ]
@@ -712,11 +727,7 @@ jobs:
 
   compute-node-image:
     runs-on: [ self-hosted, gen3, large ]
-    container:
-      image: gcr.io/kaniko-project/executor:v1.9.2-debug
-      # Workaround for "Resolving download.osgeo.org (download.osgeo.org)... failed: Temporary failure in name resolution.""
-      # Should be prevented by https://github.com/neondatabase/neon/issues/4281
-      options: --add-host=download.osgeo.org:140.211.15.30
+    container: gcr.io/kaniko-project/executor:v1.9.2-debug
     needs: [ tag ]
     strategy:
       fail-fast: false
@@ -778,7 +789,7 @@ jobs:
       run:
         shell: sh -eu {0}
     env:
-      VM_BUILDER_VERSION: v0.8.0
+      VM_BUILDER_VERSION: v0.4.6
 
     steps:
       - name: Checkout
@@ -788,18 +799,21 @@ jobs:
 
       - name: Downloading vm-builder
         run: |
-          curl -fL https://github.com/neondatabase/autoscaling/releases/download/$VM_BUILDER_VERSION/vm-builder -o vm-builder
+          curl -L https://github.com/neondatabase/neonvm/releases/download/$VM_BUILDER_VERSION/vm-builder -o vm-builder
           chmod +x vm-builder
 
-      # Note: we need a separate pull step here because otherwise vm-builder will try to pull, and
-      # it won't have the proper authentication (written at v0.6.0)
       - name: Pulling compute-node image
         run: |
           docker pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
 
+      - name: Building VM compute-node rootfs
+        run: |
+          docker build -t temp-vm-compute-node --build-arg SRC_IMAGE=369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}} -f Dockerfile.vm-compute-node .
+
       - name: Build vm image
         run: |
-          ./vm-builder -src=369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}} -dst=369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
+          # note: as of 2023-01-12, vm-builder requires a trailing ":latest" for local images
+          ./vm-builder -use-inittab -src=temp-vm-compute-node:latest -dst=369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
 
       - name: Pushing vm-compute-node image
         run: |
@@ -959,7 +973,7 @@ jobs:
   promote-compatibility-data:
     runs-on: [ self-hosted, gen3, small ]
     container:
-      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/base:pinned
+      image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
       options: --init
     needs: [ promote-images, tag, regress-tests ]
     if: github.ref_name == 'release' && github.event_name != 'workflow_dispatch'
@@ -970,13 +984,11 @@ jobs:
           PREFIX: artifacts/latest
         run: |
           # Update compatibility snapshot for the release
-          for pg_version in v14 v15; do
-            for build_type in debug release; do
-              OLD_FILENAME=compatibility-snapshot-${build_type}-pg${pg_version}-${GITHUB_RUN_ID}.tar.zst
-              NEW_FILENAME=compatibility-snapshot-${build_type}-pg${pg_version}.tar.zst
+          for build_type in debug release; do
+            OLD_FILENAME=compatibility-snapshot-${build_type}-pg14-${GITHUB_RUN_ID}.tar.zst
+            NEW_FILENAME=compatibility-snapshot-${build_type}-pg14.tar.zst
 
-              time aws s3 mv --only-show-errors s3://${BUCKET}/${PREFIX}/${OLD_FILENAME} s3://${BUCKET}/${PREFIX}/${NEW_FILENAME}
-            done
+            time aws s3 mv --only-show-errors s3://${BUCKET}/${PREFIX}/${OLD_FILENAME} s3://${BUCKET}/${PREFIX}/${NEW_FILENAME}
           done
 
           # Update Neon artifact for the release (reuse already uploaded artifact)

CONTRIBUTING.md

@@ -2,7 +2,7 @@
 
 Howdy! Usual good software engineering practices apply. Write
 tests. Write comments. Follow standard Rust coding practices where
-possible. Use `cargo fmt` and `cargo clippy` to tidy up formatting.
+possible. Use 'cargo fmt' and 'clippy' to tidy up formatting.
 
 There are soft spots in the code, which could use cleanup,
 refactoring, additional comments, and so forth. Let's try to raise the

Cargo.toml

@@ -3,7 +3,6 @@ members = [
     "compute_tools",
     "control_plane",
     "pageserver",
-    "pageserver/ctl",
     "proxy",
     "safekeeper",
     "storage_broker",
@@ -22,10 +21,9 @@ anyhow = { version = "1.0", features = ["backtrace"] }
 async-stream = "0.3"
 async-trait = "0.1"
 atty = "0.2.14"
-aws-config = { version = "0.55", default-features = false, features=["rustls"] }
-aws-sdk-s3 = "0.27"
-aws-smithy-http = "0.55"
-aws-credential-types = "0.55"
+aws-config = { version = "0.51.0", default-features = false, features=["rustls"] }
+aws-sdk-s3 = "0.21.0"
+aws-smithy-http = "0.51.0"
 aws-types = "0.55"
 base64 = "0.13.0"
 bincode = "1.3"
@@ -127,11 +125,11 @@ env_logger = "0.10"
 log = "0.4"
 
 ## Libraries from neondatabase/ git forks, ideally with changes to be upstreamed
-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="2e9b5f1ddc481d1a98fa79f6b9378ac4f170b7c9" }
-postgres-native-tls = { git = "https://github.com/neondatabase/rust-postgres.git", rev="2e9b5f1ddc481d1a98fa79f6b9378ac4f170b7c9" }
-postgres-protocol = { git = "https://github.com/neondatabase/rust-postgres.git", rev="2e9b5f1ddc481d1a98fa79f6b9378ac4f170b7c9" }
-postgres-types = { git = "https://github.com/neondatabase/rust-postgres.git", rev="2e9b5f1ddc481d1a98fa79f6b9378ac4f170b7c9" }
-tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="2e9b5f1ddc481d1a98fa79f6b9378ac4f170b7c9" }
+postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="0bc41d8503c092b040142214aac3cf7d11d0c19f" }
+postgres-native-tls = { git = "https://github.com/neondatabase/rust-postgres.git", rev="0bc41d8503c092b040142214aac3cf7d11d0c19f" }
+postgres-protocol = { git = "https://github.com/neondatabase/rust-postgres.git", rev="0bc41d8503c092b040142214aac3cf7d11d0c19f" }
+postgres-types = { git = "https://github.com/neondatabase/rust-postgres.git", rev="0bc41d8503c092b040142214aac3cf7d11d0c19f" }
+tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="0bc41d8503c092b040142214aac3cf7d11d0c19f" }
 tokio-tar = { git = "https://github.com/neondatabase/tokio-tar.git", rev="404df61437de0feef49ba2ccdbdd94eb8ad6e142" }
 
 ## Other git libraries
@@ -167,7 +165,7 @@ tonic-build = "0.9"
 
 # This is only needed for proxy's tests.
 # TODO: we should probably fork `tokio-postgres-rustls` instead.
-tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="2e9b5f1ddc481d1a98fa79f6b9378ac4f170b7c9" }
+tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="0bc41d8503c092b040142214aac3cf7d11d0c19f" }
 
 # Changes the MAX_THREADS limit from 4096 to 32768.
 # This is a temporary workaround for using tracing from many threads in safekeepers code,

Dockerfile

@@ -47,7 +47,8 @@ RUN set -e \
     && mold -run cargo build  \
       --bin pg_sni_router  \
       --bin pageserver  \
-      --bin pagectl  \
+      --bin pageserver_binutils  \
+      --bin draw_timeline_dir \
       --bin safekeeper  \
       --bin storage_broker  \
       --bin proxy  \
@@ -72,7 +73,8 @@ RUN set -e \
 
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/pg_sni_router       /usr/local/bin
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/pageserver          /usr/local/bin
-COPY --from=build --chown=neon:neon /home/nonroot/target/release/pagectl             /usr/local/bin
+COPY --from=build --chown=neon:neon /home/nonroot/target/release/pageserver_binutils /usr/local/bin
+COPY --from=build --chown=neon:neon /home/nonroot/target/release/draw_timeline_dir   /usr/local/bin
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/safekeeper          /usr/local/bin
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/storage_broker         /usr/local/bin
 COPY --from=build --chown=neon:neon /home/nonroot/target/release/proxy               /usr/local/bin

Dockerfile.compute-node

@@ -415,23 +415,6 @@ RUN apt-get update && \
     make -j $(getconf _NPROCESSORS_ONLN) install && \
     echo 'trusted = true' >> /usr/local/pgsql/share/extension/kq_imcx.control
 
-#########################################################################################
-#
-# Layer "pg-cron-pg-build"
-# compile pg_cron extension
-#
-#########################################################################################
-FROM build-deps AS pg-cron-pg-build
-COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
-
-ENV PATH "/usr/local/pgsql/bin/:$PATH"
-RUN wget https://github.com/citusdata/pg_cron/archive/refs/tags/v1.5.2.tar.gz -O pg_cron.tar.gz && \
-    echo "6f7f0980c03f1e2a6a747060e67bf4a303ca2a50e941e2c19daeed2b44dec744 pg_cron.tar.gz" | sha256sum --check && \
-    mkdir pg_cron-src && cd pg_cron-src && tar xvzf ../pg_cron.tar.gz --strip-components=1 -C . && \
-    make -j $(getconf _NPROCESSORS_ONLN) && \
-    make -j $(getconf _NPROCESSORS_ONLN) install && \
-    echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_cron.control
-
 #########################################################################################
 #
 # Layer "rust extensions"
@@ -546,7 +529,6 @@ COPY --from=plpgsql-check-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=timescaledb-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=pg-hint-plan-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY --from=kq-imcx-pg-build /usr/local/pgsql/ /usr/local/pgsql/
-COPY --from=pg-cron-pg-build /usr/local/pgsql/ /usr/local/pgsql/
 COPY pgxn/ pgxn/
 
 RUN make -j $(getconf _NPROCESSORS_ONLN) \
@@ -632,7 +614,6 @@ RUN apt update &&  \
         libxml2 \
         libxslt1.1 \
         libzstd1 \
-        libcurl4-openssl-dev \
         procps && \
     rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
     localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8

Dockerfile.vm-compute-node

@@ -0,0 +1,70 @@
+# Note: this file *mostly* just builds on Dockerfile.compute-node
+
+ARG SRC_IMAGE
+ARG VM_INFORMANT_VERSION=v0.1.14
+# on libcgroup update, make sure to check bootstrap.sh for changes
+ARG LIBCGROUP_VERSION=v2.0.3
+
+# Pull VM informant, to copy from later
+FROM neondatabase/vm-informant:$VM_INFORMANT_VERSION as informant
+
+# Build cgroup-tools
+#
+# At time of writing (2023-03-14), debian bullseye has a version of cgroup-tools (technically
+# libcgroup) that doesn't support cgroup v2 (version 0.41-11). Unfortunately, the vm-informant
+# requires cgroup v2, so we'll build cgroup-tools ourselves.
+FROM debian:bullseye-slim as libcgroup-builder
+ARG LIBCGROUP_VERSION
+
+RUN set -exu \
+	&& apt update \
+	&& apt install --no-install-recommends -y \
+		git \
+		ca-certificates \
+		automake \
+		cmake \
+		make \
+		gcc \
+		byacc \
+		flex \
+		libtool \
+		libpam0g-dev \
+	&& git clone --depth 1 -b $LIBCGROUP_VERSION https://github.com/libcgroup/libcgroup \
+	&& INSTALL_DIR="/libcgroup-install" \
+	&& mkdir -p "$INSTALL_DIR/bin" "$INSTALL_DIR/include" \
+	&& cd libcgroup \
+	# extracted from bootstrap.sh, with modified flags:
+	&& (test -d m4 || mkdir m4) \
+	&& autoreconf -fi \
+	&& rm -rf autom4te.cache \
+	&& CFLAGS="-O3" ./configure --prefix="$INSTALL_DIR" --sysconfdir=/etc --localstatedir=/var --enable-opaque-hierarchy="name=systemd" \
+	# actually build the thing...
+	&& make install
+
+# Combine, starting from non-VM compute node image.
+FROM $SRC_IMAGE as base
+
+# Temporarily set user back to root so we can run adduser, set inittab
+USER root
+RUN adduser vm-informant --disabled-password --no-create-home
+
+RUN set -e \
+	&& rm -f /etc/inittab \
+	&& touch /etc/inittab
+
+RUN set -e \
+	&& echo "::sysinit:cgconfigparser -l /etc/cgconfig.conf -s 1664" >> /etc/inittab \
+	&& CONNSTR="dbname=postgres user=cloud_admin sslmode=disable" \
+	&& ARGS="--auto-restart --cgroup=neon-postgres --pgconnstr=\"$CONNSTR\"" \
+	&& echo "::respawn:su vm-informant -c '/usr/local/bin/vm-informant $ARGS'" >> /etc/inittab
+
+USER postgres
+
+ADD vm-cgconfig.conf /etc/cgconfig.conf
+COPY --from=informant /usr/bin/vm-informant /usr/local/bin/vm-informant
+
+COPY --from=libcgroup-builder /libcgroup-install/bin/* /usr/bin/
+COPY --from=libcgroup-builder /libcgroup-install/lib/* /usr/lib/
+COPY --from=libcgroup-builder /libcgroup-install/sbin/* /usr/sbin/
+
+ENTRYPOINT ["/usr/sbin/cgexec", "-g", "*:neon-postgres", "/usr/local/bin/compute_ctl"]

README.md

@@ -17,7 +17,7 @@ The Neon storage engine consists of two major components:
 - Pageserver. Scalable storage backend for the compute nodes.
 - Safekeepers. The safekeepers form a redundant WAL service that received WAL from the compute node, and stores it durably until it has been processed by the pageserver and uploaded to cloud storage.
 
-See developer documentation in [SUMMARY.md](/docs/SUMMARY.md) for more information.
+See developer documentation in [/docs/SUMMARY.md](/docs/SUMMARY.md) for more information.
 
 ## Running local installation
 
@@ -130,11 +130,11 @@ Python (3.9 or higher), and install python3 packages using `./scripts/pysync` (r
 ```sh
 # Create repository in .neon with proper paths to binaries and data
 # Later that would be responsibility of a package install script
-> cargo neon init
+> ./target/debug/neon_local init
 Starting pageserver at '127.0.0.1:64000' in '.neon'.
 
 # start pageserver, safekeeper, and broker for their intercommunication
-> cargo neon start
+> ./target/debug/neon_local start
 Starting neon broker at 127.0.0.1:50051
 storage_broker started, pid: 2918372
 Starting pageserver at '127.0.0.1:64000' in '.neon'.
@@ -143,19 +143,19 @@ Starting safekeeper at '127.0.0.1:5454' in '.neon/safekeepers/sk1'.
 safekeeper 1 started, pid: 2918437
 
 # create initial tenant and use it as a default for every future neon_local invocation
-> cargo neon tenant create --set-default
+> ./target/debug/neon_local tenant create --set-default
 tenant 9ef87a5bf0d92544f6fafeeb3239695c successfully created on the pageserver
 Created an initial timeline 'de200bd42b49cc1814412c7e592dd6e9' at Lsn 0/16B5A50 for tenant: 9ef87a5bf0d92544f6fafeeb3239695c
 Setting tenant 9ef87a5bf0d92544f6fafeeb3239695c as a default one
 
 # start postgres compute node
-> cargo neon endpoint start main
+> ./target/debug/neon_local endpoint start main
 Starting new endpoint main (PostgreSQL v14) on timeline de200bd42b49cc1814412c7e592dd6e9 ...
 Extracting base backup to create postgres instance: path=.neon/pgdatadirs/tenants/9ef87a5bf0d92544f6fafeeb3239695c/main port=55432
 Starting postgres at 'host=127.0.0.1 port=55432 user=cloud_admin dbname=postgres'
 
 # check list of running postgres instances
-> cargo neon endpoint list
+> ./target/debug/neon_local endpoint list
  ENDPOINT  ADDRESS          TIMELINE                          BRANCH NAME  LSN        STATUS
  main      127.0.0.1:55432  de200bd42b49cc1814412c7e592dd6e9  main         0/16B5BA8  running
 ```
@@ -177,22 +177,22 @@ postgres=# select * from t;
 3. And create branches and run postgres on them:
 ```sh
 # create branch named migration_check
-> cargo neon timeline branch --branch-name migration_check
+> ./target/debug/neon_local timeline branch --branch-name migration_check
 Created timeline 'b3b863fa45fa9e57e615f9f2d944e601' at Lsn 0/16F9A00 for tenant: 9ef87a5bf0d92544f6fafeeb3239695c. Ancestor timeline: 'main'
 
 # check branches tree
-> cargo neon timeline list
+> ./target/debug/neon_local timeline list
 (L) main [de200bd42b49cc1814412c7e592dd6e9]
 (L) ┗━ @0/16F9A00: migration_check [b3b863fa45fa9e57e615f9f2d944e601]
 
 # start postgres on that branch
-> cargo neon endpoint start migration_check --branch-name migration_check
+> ./target/debug/neon_local endpoint start migration_check --branch-name migration_check
 Starting new endpoint migration_check (PostgreSQL v14) on timeline b3b863fa45fa9e57e615f9f2d944e601 ...
 Extracting base backup to create postgres instance: path=.neon/pgdatadirs/tenants/9ef87a5bf0d92544f6fafeeb3239695c/migration_check port=55433
 Starting postgres at 'host=127.0.0.1 port=55433 user=cloud_admin dbname=postgres'
 
 # check the new list of running postgres instances
-> cargo neon endpoint list
+> ./target/debug/neon_local endpoint list
  ENDPOINT         ADDRESS          TIMELINE                          BRANCH NAME      LSN        STATUS
  main             127.0.0.1:55432  de200bd42b49cc1814412c7e592dd6e9  main             0/16F9A38  running
  migration_check  127.0.0.1:55433  b3b863fa45fa9e57e615f9f2d944e601  migration_check  0/16F9A70  running
@@ -221,7 +221,7 @@ postgres=# select * from t;
 4. If you want to run tests afterward (see below), you must stop all the running of the pageserver, safekeeper, and postgres instances
    you have just started. You can terminate them all with one command:
 ```sh
-> cargo neon stop
+> ./target/debug/neon_local stop
 ```
 
 ## Running tests
@@ -238,9 +238,9 @@ CARGO_BUILD_FLAGS="--features=testing" make
 
 ## Documentation
 
-[docs](/docs) Contains a top-level overview of all available markdown documentation.
+[/docs/](/docs/) Contains a top-level overview of all available markdown documentation.
 
-- [sourcetree.md](/docs/sourcetree.md) contains overview of source tree layout.
+- [/docs/sourcetree.md](/docs/sourcetree.md) contains overview of source tree layout.
 
 To view your `rustdoc` documentation in a browser, try running `cargo doc --no-deps --open`
 
@@ -265,6 +265,6 @@ To get more familiar with this aspect, refer to:
 
 ## Join the development
 
-- Read [CONTRIBUTING.md](/CONTRIBUTING.md) to learn about project code style and practices.
-- To get familiar with a source tree layout, use [sourcetree.md](/docs/sourcetree.md).
+- Read `CONTRIBUTING.md` to learn about project code style and practices.
+- To get familiar with a source tree layout, use [/docs/sourcetree.md](/docs/sourcetree.md).
 - To learn more about PostgreSQL internals, check http://www.interdb.jp/pg/index.html

compute_tools/src/compute.rs

@@ -362,8 +362,6 @@ impl ComputeNode {
         };
 
         // Proceed with post-startup configuration. Note, that order of operations is important.
-        // Disable DDL forwarding because control plane already knows about these roles/databases.
-        client.simple_query("SET neon.forward_ddl = false")?;
         let spec = &compute_state.pspec.as_ref().expect("spec must be set").spec;
         handle_roles(spec, &mut client)?;
         handle_databases(spec, &mut client)?;
@@ -405,9 +403,7 @@ impl ComputeNode {
         self.pg_reload_conf(&mut client)?;
 
         // Proceed with post-startup configuration. Note, that order of operations is important.
-        // Disable DDL forwarding because control plane already knows about these roles/databases.
         if spec.mode == ComputeMode::Primary {
-            client.simple_query("SET neon.forward_ddl = false")?;
             handle_roles(&spec, &mut client)?;
             handle_databases(&spec, &mut client)?;
             handle_role_deletions(&spec, self.connstr.as_str(), &mut client)?;

compute_tools/src/pg_helpers.rs

@@ -121,8 +121,9 @@ impl RoleExt for Role {
     /// string of arguments.
     fn to_pg_options(&self) -> String {
         // XXX: consider putting LOGIN as a default option somewhere higher, e.g. in control-plane.
-        let mut params: String = self.options.as_pg_options();
-        params.push_str(" LOGIN");
+        // For now, we do not use generic `options` for roles. Once used, add
+        // `self.options.as_pg_options()` somewhere here.
+        let mut params: String = "LOGIN".to_string();
 
         if let Some(pass) = &self.encrypted_password {
             // Some time ago we supported only md5 and treated all encrypted_password as md5.

compute_tools/src/spec.rs

@@ -62,7 +62,7 @@ fn do_control_plane_request(
     }
 }
 
-/// Request spec from the control-plane by compute_id. If `NEON_CONTROL_PLANE_TOKEN`
+/// Request spec from the control-plane by compute_id. If `NEON_CONSOLE_JWT`
 /// env variable is set, it will be used for authorization.
 pub fn get_spec_from_control_plane(
     base_uri: &str,

compute_tools/tests/pg_helpers_tests.rs

@@ -16,7 +16,7 @@ mod pg_helpers_tests {
         );
         assert_eq!(
             spec.cluster.roles.first().unwrap().to_pg_options(),
-            " LOGIN PASSWORD 'md56b1d16b78004bbd51fa06af9eda75972'"
+            "LOGIN PASSWORD 'md56b1d16b78004bbd51fa06af9eda75972'"
         );
     }
 

control_plane/src/bin/neon_local.rs

@@ -41,7 +41,7 @@ const DEFAULT_PAGESERVER_ID: NodeId = NodeId(1);
 const DEFAULT_BRANCH_NAME: &str = "main";
 project_git_version!(GIT_VERSION);
 
-const DEFAULT_PG_VERSION: &str = "15";
+const DEFAULT_PG_VERSION: &str = "14";
 
 fn default_conf() -> String {
     format!(

control_plane/src/endpoint.rs

@@ -134,7 +134,6 @@ pub struct Endpoint {
 
     // port and address of the Postgres server
     pub address: SocketAddr,
-    // postgres major version in the format: 14, 15, etc.
     pg_version: u32,
 
     // These are not part of the endpoint as such, but the environment
@@ -382,11 +381,6 @@ impl Endpoint {
                 conf.append("primary_conninfo", connstr.as_str());
                 conf.append("primary_slot_name", slot_name.as_str());
                 conf.append("hot_standby", "on");
-                // prefetching of blocks referenced in WAL doesn't make sense for us
-                // Neon hot standby ignores pages that are not in the shared_buffers
-                if self.pg_version >= 15 {
-                    conf.append("recovery_prefetch", "off");
-                }
             }
         }
 

control_plane/src/local_env.rs

@@ -24,7 +24,7 @@ use utils::{
 
 use crate::safekeeper::SafekeeperNode;
 
-pub const DEFAULT_PG_VERSION: u32 = 15;
+pub const DEFAULT_PG_VERSION: u32 = 14;
 
 //
 // This data structures represents neon_local CLI config

control_plane/src/pageserver.rs

@@ -8,7 +8,9 @@ use std::process::{Child, Command};
 use std::{io, result};
 
 use anyhow::{bail, Context};
-use pageserver_api::models::{self, TenantInfo, TimelineInfo};
+use pageserver_api::models::{
+    TenantConfigRequest, TenantCreateRequest, TenantInfo, TimelineCreateRequest, TimelineInfo,
+};
 use postgres_backend::AuthType;
 use postgres_connection::{parse_host_port, PgConnectionConfig};
 use reqwest::blocking::{Client, RequestBuilder, Response};
@@ -314,8 +316,8 @@ impl PageServerNode {
         settings: HashMap<&str, &str>,
     ) -> anyhow::Result<TenantId> {
         let mut settings = settings.clone();
-
-        let config = models::TenantConfig {
+        let request = TenantCreateRequest {
+            new_tenant_id,
             checkpoint_distance: settings
                 .remove("checkpoint_distance")
                 .map(|x| x.parse::<u64>())
@@ -370,10 +372,6 @@ impl PageServerNode {
                 .remove("evictions_low_residence_duration_metric_threshold")
                 .map(|x| x.to_string()),
         };
-        let request = models::TenantCreateRequest {
-            new_tenant_id,
-            config,
-        };
         if !settings.is_empty() {
             bail!("Unrecognized tenant settings: {settings:?}")
         }
@@ -393,81 +391,67 @@ impl PageServerNode {
             })
     }
 
-    pub fn tenant_config(
-        &self,
-        tenant_id: TenantId,
-        mut settings: HashMap<&str, &str>,
-    ) -> anyhow::Result<()> {
-        let config = {
-            // Braces to make the diff easier to read
-            models::TenantConfig {
+    pub fn tenant_config(&self, tenant_id: TenantId, settings: HashMap<&str, &str>) -> Result<()> {
+        self.http_request(Method::PUT, format!("{}/tenant/config", self.http_base_url))?
+            .json(&TenantConfigRequest {
+                tenant_id,
                 checkpoint_distance: settings
-                    .remove("checkpoint_distance")
+                    .get("checkpoint_distance")
                     .map(|x| x.parse::<u64>())
                     .transpose()
                     .context("Failed to parse 'checkpoint_distance' as an integer")?,
-                checkpoint_timeout: settings.remove("checkpoint_timeout").map(|x| x.to_string()),
+                checkpoint_timeout: settings.get("checkpoint_timeout").map(|x| x.to_string()),
                 compaction_target_size: settings
-                    .remove("compaction_target_size")
+                    .get("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(|x| x.to_string()),
+                compaction_period: settings.get("compaction_period").map(|x| x.to_string()),
                 compaction_threshold: settings
-                    .remove("compaction_threshold")
+                    .get("compaction_threshold")
                     .map(|x| x.parse::<usize>())
                     .transpose()
                     .context("Failed to parse 'compaction_threshold' as an integer")?,
                 gc_horizon: settings
-                    .remove("gc_horizon")
+                    .get("gc_horizon")
                     .map(|x| x.parse::<u64>())
                     .transpose()
                     .context("Failed to parse 'gc_horizon' as an integer")?,
-                gc_period: settings.remove("gc_period").map(|x| x.to_string()),
+                gc_period: settings.get("gc_period").map(|x| x.to_string()),
                 image_creation_threshold: settings
-                    .remove("image_creation_threshold")
+                    .get("image_creation_threshold")
                     .map(|x| x.parse::<usize>())
                     .transpose()
                     .context("Failed to parse 'image_creation_threshold' as non zero integer")?,
-                pitr_interval: settings.remove("pitr_interval").map(|x| x.to_string()),
+                pitr_interval: settings.get("pitr_interval").map(|x| x.to_string()),
                 walreceiver_connect_timeout: settings
-                    .remove("walreceiver_connect_timeout")
-                    .map(|x| x.to_string()),
-                lagging_wal_timeout: settings
-                    .remove("lagging_wal_timeout")
+                    .get("walreceiver_connect_timeout")
                     .map(|x| x.to_string()),
+                lagging_wal_timeout: settings.get("lagging_wal_timeout").map(|x| x.to_string()),
                 max_lsn_wal_lag: settings
-                    .remove("max_lsn_wal_lag")
+                    .get("max_lsn_wal_lag")
                     .map(|x| x.parse::<NonZeroU64>())
                     .transpose()
                     .context("Failed to parse 'max_lsn_wal_lag' as non zero integer")?,
                 trace_read_requests: settings
-                    .remove("trace_read_requests")
+                    .get("trace_read_requests")
                     .map(|x| x.parse::<bool>())
                     .transpose()
                     .context("Failed to parse 'trace_read_requests' as bool")?,
                 eviction_policy: settings
-                    .remove("eviction_policy")
-                    .map(serde_json::from_str)
+                    .get("eviction_policy")
+                    .map(|x| serde_json::from_str(x))
                     .transpose()
                     .context("Failed to parse 'eviction_policy' json")?,
                 min_resident_size_override: settings
-                    .remove("min_resident_size_override")
+                    .get("min_resident_size_override")
                     .map(|x| x.parse::<u64>())
                     .transpose()
                     .context("Failed to parse 'min_resident_size_override' as an integer")?,
                 evictions_low_residence_duration_metric_threshold: settings
-                    .remove("evictions_low_residence_duration_metric_threshold")
+                    .get("evictions_low_residence_duration_metric_threshold")
                     .map(|x| x.to_string()),
-            }
-        };
-
-        if !settings.is_empty() {
-            bail!("Unrecognized tenant settings: {settings:?}")
-        }
-
-        self.http_request(Method::PUT, format!("{}/tenant/config", self.http_base_url))?
-            .json(&models::TenantConfigRequest { tenant_id, config })
+            })
             .send()?
             .error_from_body()?;
 
@@ -499,7 +483,7 @@ impl PageServerNode {
             Method::POST,
             format!("{}/tenant/{}/timeline", self.http_base_url, tenant_id),
         )?
-        .json(&models::TimelineCreateRequest {
+        .json(&TimelineCreateRequest {
             new_timeline_id,
             ancestor_start_lsn,
             ancestor_timeline_id,

docs/pageserver-thread-mgmt.md

@@ -4,11 +4,6 @@ The pageserver uses Tokio for handling concurrency. Everything runs in
 Tokio tasks, although some parts are written in blocking style and use
 spawn_blocking().
 
-We currently use std blocking functions for disk I/O, however.  The
-current model is that we consider disk I/Os to be short enough that we
-perform them while running in a Tokio task. Changing all the disk I/O
-calls to async is a TODO.
-
 Each Tokio task is tracked by the `task_mgr` module. It maintains a
 registry of tasks, and which tenant or timeline they are operating
 on.
@@ -26,86 +21,19 @@ also a `shudown_watcher()` Future that can be used with `tokio::select!`
 or similar, to wake up on shutdown.
 
 
-### Async cancellation safety
+### Sync vs async
 
-In async Rust, futures can be "cancelled" at any await point, by
-dropping the Future. For example, `tokio::select!` returns as soon as
-one of the Futures returns, and drops the others. `tokio::timeout!` is
-another example. In the Rust ecosystem, some functions are
-cancellation-safe, meaning they can be safely dropped without
-side-effects, while others are not. See documentation of
-`tokio::select!` for examples.
+We use async to wait for incoming data on network connections, and to
+perform other long-running operations. For example, each WAL receiver
+connection is handled by a tokio Task. Once a piece of WAL has been
+received from the network, the task calls the blocking functions in
+the Repository to process the WAL.
 
-In the pageserver and safekeeper, async code is *not*
-cancellation-safe by default. Unless otherwise marked, any async
-function that you call cannot be assumed to be async
-cancellation-safe, and must be polled to completion.
+The core storage code in `layered_repository/` is synchronous, with
+blocking locks and I/O calls. The current model is that we consider
+disk I/Os to be short enough that we perform them while running in a
+Tokio task. If that becomes a problem, we should use `spawn_blocking`
+before entering the synchronous parts of the code, or switch to using
+tokio I/O functions.
 
-The downside of non-cancellation safe code is that you have to be very
-careful when using `tokio::select!`, `tokio::timeout!`, and other such
-functions that can cause a Future to be dropped. They can only be used
-with functions that are explicitly documented to be cancellation-safe,
-or you need to spawn a separate task to shield from the cancellation.
-
-At the entry points to the code, we also take care to poll futures to
-completion, or shield the rest of the code from surprise cancellations
-by spawning a separate task. The code that handles incoming HTTP
-requests, for example, spawns a separate task for each request,
-because Hyper will drop the request-handling Future if the HTTP
-connection is lost.  (FIXME: our HTTP handlers do not do that
-currently, but we should fix that. See [issue
-3478](https://github.com/neondatabase/neon/issues/3478)).
-
-
-#### How to cancel, then?
-
-If our code is not cancellation-safe, how do you cancel long-running
-tasks? Use CancellationTokens.
-
-TODO: More details on that. And we have an ongoing discussion on what
-to do if cancellations might come from multiple sources.
-
-#### Exceptions
-Some library functions are cancellation-safe, and are explicitly marked
-as such. For example, `utils::seqwait`.
-
-#### Rationale
-
-The alternative would be to make all async code cancellation-safe,
-unless otherwise marked. That way, you could use `tokio::select!` more
-liberally. The reasons we didn't choose that are explained in this
-section.
-
-Writing code in a cancellation-safe manner is tedious, as you need to
-scrutinize every `.await` and ensure that if the `.await` call never
-returns, the system is in a safe, consistent state. In some ways, you
-need to do that with `?` and early `returns`, too, but `.await`s are
-easier to miss. It is also easier to perform cleanup tasks when a
-function returns an `Err` than when an `.await` simply never
-returns. You can use `scopeguard` and Drop guards to perform cleanup
-tasks, but it is more tedious. An `.await` that never returns is more
-similar to a panic.
-
-Note that even if you only use building blocks that themselves are
-cancellation-safe, it doesn't mean that the code as whole is
-cancellation-safe. For example, consider the following code:
-
-```
-while let Some(i) = work_inbox.recv().await {
-	if let Err(_) = results_outbox.send(i).await {
-		println!("receiver dropped");
-		return;
-		}
-	}
-}
-```
-
-It reads messages from one channel, sends them to another channel. If
-this code is cancelled at the `results_outbox.send(i).await`, the
-message read from the receiver is lost. That may or may not be OK,
-depending on the context.
-
-Another reason to not require cancellation-safety is historical: we
-already had a lot of async code that was not scrutinized for
-cancellation-safety when this issue was raised. Scrutinizing all
-existing code is no fun.
+Be very careful when mixing sync and async code!

docs/rfcs/023-the-state-of-pageserver-tenant-relocation.md

@@ -1,232 +0,0 @@
-# The state of pageserver tenant relocation
-
-Created on 17.03.23
-
-## Motivation
-
-There were previous write ups on the subject. The design of tenant relocation was planned at the time when we had quite different landscape. I e there was no on-demand download/eviction. They were on the horizon but we still planned for cases when they were not available. Some other things have changed. Now safekeepers offload wal to s3 so we're not risking overflowing their disks. Having all of the above, it makes sense to recap and take a look at the options we have now, which adjustments we'd like to make to original process, etc.
-
-Related (in chronological order):
-
-- Tracking issue with initial discussion: [#886](https://github.com/neondatabase/neon/issues/886)
-- [015. Storage Messaging](015-storage-messaging.md)
-- [020. Pageserver S3 Coordination](020-pageserver-s3-coordination.md)
-
-## Summary
-
-The RFC consists of a walkthrough of prior art on tenant relocation and corresponding problems. It describes 3 approaches.
-
-1. Simplistic approach that uses ignore and is the fastest to implement. The main downside is a requirement of short downtime.
-2. More complicated approach that avoids even short downtime.
-3. Even more complicated approach that will allow multiple pageservers to operate concurrently on the same tenant possibly allowing for HA cluster topologies and horizontal scaling of reads (i e compute talks to multiple pageservers).
-
-The order in which solutions are described is a bit different. We start from 2, then move to possible compromises (aka simplistic approach) and then move to discussing directions for solving HA/Pageserver replica case with 3.
-
-## Components
-
-pageserver, control-plane, safekeepers (a bit)
-
-## Requirements
-
-Relocation procedure should move tenant from one pageserver to another without downtime introduced by storage side. For now restarting compute for applying new configuration is fine.
-
-- component restarts
-- component outage
-- pageserver loss
-
-## The original proposed implementation
-
-The starting point is this sequence:
-
-```mermaid
-sequenceDiagram
-    autonumber
-    participant CP as Control Plane
-    participant PS1 as Pageserver 1
-    participant PS2 as Pageserver 2
-    participant S3
-
-    CP->>PS2: Attach tenant X
-    PS2->>S3: Fetch timelines, indexes for them
-    PS2->>CP: Accepted
-    CP->>CP: Change pageserver id in project
-    CP->>PS1: Detach
-```
-
-Which problems do we have with naive approach?
-
-### Concurrent GC and Compaction
-
-The problem is that they can run on both, PS1 and PS2. Consider this example from [Pageserver S3 Coordination RFC](020-pageserver-s3-coordination.md)
-
-```mermaid
-sequenceDiagram
-    autonumber
-    participant PS1
-    participant S3
-    participant PS2
-
-    PS1->>S3: Uploads L1, L2 <br/> Index contains L1 L2
-    PS2->>S3: Attach called, sees L1, L2
-    PS1->>S3: Compaction comes <br/> Removes L1, adds L3
-    note over S3: Index now L2, L3
-    PS2->>S3: Uploads new layer L4 <br/> (added to previous view of the index)
-    note over S3: Index now L1, L2, L4
-```
-
-At this point it is not possible to restore the state from index, it contains L2 which
-is no longer available in s3 and doesnt contain L3 added by compaction by the
-first pageserver. So if any of the pageservers restart, initial sync will fail
-(or in on-demand world it will fail a bit later during page request from
-missing layer)
-
-The problem lies in shared index_part.json. Having intersecting layers from append only edits is expected to work, though this is an uncharted territory without tests.
-
-#### Options
-
-There are several options on how to restrict concurrent access to index file.
-
-First and the simplest one is external orchestration. Control plane which runs migration can use special api call on pageserver to stop background processes (gc, compaction), and even possibly all uploads.
-
-So the sequence becomes:
-
-```mermaid
-sequenceDiagram
-    autonumber
-    participant CP as Control Plane
-    participant PS1 as Pageserver 1
-    participant PS2 as Pageserver 2
-    participant S3
-
-    CP->>PS1: Pause background jobs, pause uploading new layers.
-    CP->>PS2: Attach tenant X.
-    PS2->>S3: Fetch timelines, index, start background operations
-    PS2->>CP: Accepted
-    CP->>CP: Monitor PS2 last record lsn, ensure OK lag
-    CP->>CP: Change pageserver id in project
-    CP->>PS1: Detach
-```
-
-The downside of this sequence is the potential rollback process. What if something goes wrong on new pageserver? Can we safely roll back to source pageserver?
-
-There are two questions:
-
-#### How can we detect that something went wrong?
-
-We can run usual availability check (consists of compute startup and an update of one row).
-Note that we cant run separate compute for that before touching compute that client runs actual workload on, because we cant have two simultaneous computes running in read-write mode on the same timeline (enforced by safekeepers consensus algorithm). So we can either run some readonly check first (basebackup) and then change pageserver id and run availability check. If it failed we can roll it back to the old one.
-
-#### What can go wrong? And how we can safely roll-back?
-
-In the sequence above during attach we start background processes/uploads. They change state in remote storage so it is possible that after rollback remote state will be different from one that was observed by source pageserver. So if target pageserver goes wild then source pageserver may fail to start with changed remote state.
-
-Proposed option would be to implement a barrier (read-only) mode when pageserver does not update remote state.
-
-So the sequence for happy path becomes this one:
-
-```mermaid
-sequenceDiagram
-    autonumber
-    participant CP as Control Plane
-    participant PS1 as Pageserver 1
-    participant PS2 as Pageserver 2
-    participant S3
-
-    CP->>PS1: Pause background jobs, pause uploading new layers.
-    CP->>PS2: Attach tenant X in remote readonly mode.
-    PS2->>S3: Fetch timelines, index
-    PS2->>CP: Accepted
-    CP->>CP: Monitor PS2 last record lsn, ensure OK lag
-    CP->>CP: Change pageserver id in project
-    CP->>CP: Run successful availability check
-    CP->>PS2: Start uploads, background tasks
-    CP->>PS1: Detach
-```
-
-With this sequence we restrict any changes to remote storage to one pageserver. So there is no concurrent access at all, not only for index_part.json, but for everything else too. This approach makes it possible to roll back after failure on new pageserver.
-
-The sequence with roll back process:
-
-```mermaid
-sequenceDiagram
-    autonumber
-    participant CP as Control Plane
-    participant PS1 as Pageserver 1
-    participant PS2 as Pageserver 2
-    participant S3
-
-    CP->>PS1: Pause background jobs, pause uploading new layers.
-    CP->>PS2: Attach tenant X in remote readonly mode.
-    PS2->>S3: Fetch timelines, index
-    PS2->>CP: Accepted
-    CP->>CP: Monitor PS2 last record lsn, ensure OK lag
-    CP->>CP: Change pageserver id in project
-    CP->>CP: Availability check Failed
-    CP->>CP: Change pageserver id back
-    CP->>PS1: Resume remote operations
-    CP->>PS2: Ignore (instead of detach for investigation purposes)
-```
-
-## Concurrent branch creation
-
-Another problem is a possibility of concurrent branch creation calls.
-
-I e during migration create_branch can be called on old pageserver and newly created branch wont be seen on new pageserver. Prior art includes prototyping an approach of trying to mirror such branches, but currently it lost its importance, because now attach is fast because we dont need to download all data, and additionally to the best of my knowledge of control plane internals (cc @ololobus to confirm) operations on one project are executed sequentially, so it is not possible to have such case. So branch create operation will be executed only when relocation is completed. As a safety measure we can forbid branch creation for tenants that are in readonly remote state.
-
-## Simplistic approach
-
-The difference of simplistic approach from one described above is that it calls ignore on source tenant first and then calls attach on target pageserver. Approach above does it in opposite order thus opening a possibility for race conditions we strive to avoid.
-
-The approach largely follows this guide: <https://github.com/neondatabase/cloud/wiki/Cloud:-Ad-hoc-tenant-relocation>
-
-The happy path sequence:
-
-```mermaid
-sequenceDiagram
-    autonumber
-    participant CP as Control Plane
-    participant PS1 as Pageserver 1
-    participant PS2 as Pageserver 2
-    participant SK as Safekeeper
-    participant S3
-
-    CP->>CP: Enable maintenance mode
-    CP->>PS1: Ignore
-    CP->>PS2: Attach
-    PS2->>CP: Accepted
-    loop Delete layers for each timeline
-        CP->>PS2: Get last record lsn
-        CP->>SK: Get commit lsn
-        CP->>CP: OK? Timed out?
-    end
-    CP->>CP: Change pageserver id in project
-    CP->>CP: Run successful availability check
-    CP->>CP: Disable maintenance mode
-    CP->>PS1: Detach ignored
-```
-
-The sequence contains exactly the same rollback problems as in previous approach described above. They can be resolved the same way.
-
-Most probably we'd like to move forward without this safety measure and implement it on top of this approach to make progress towards the downtime-less one.
-
-## Lease based approach
-
-In order to allow for concurrent operation on the same data on remote storage for multiple pageservers we need to go further than external orchestration.
-
-NOTE: [020. Pageserver S3 Coordination](020-pageserver-s3-coordination.md) discusses one more approach that relies on duplication of index_part.json for each pageserver operating on the timeline. This approach still requires external coordination which makes certain things easier but requires additional bookkeeping to account for multiple index_part.json files. Discussion/comparison with proposed lease based approach
-
-The problems are outlined in [020. Pageserver S3 Coordination](020-pageserver-s3-coordination.md) and suggested solution includes [Coordination based approach](020-pageserver-s3-coordination.md#coordination-based-approach). This way it will allow to do basic leader election for pageservers so they can decide which node will be responsible for running GC and compaction. The process is based on extensive communication via storage broker and consists of a lease that is taken by one of the pageservers that extends it to continue serving a leader role.
-
-There are two options for ingesting new data into pageserver in follower role. One option is to avoid WAL ingestion at all and rely on notifications from leader to discover new layers on s3. Main downside of this approach is that follower will always lag behind the primary node because it wont have the last layer until it is uploaded to remote storage. In case of a primary failure follower will be required to reingest last segment (up to 256Mb of WAL currently) which slows down recovery. Additionally if compute is connected to follower pageserver it will observe latest data with a delay. Queries from compute will likely experience bigger delays when recent lsn is required.
-
-The second option is to consume WAL stream on both pageservers. In this case the only problem is non deterministic layer generation. Additional bookkeeping will be required to deduplicate layers from primary with local ones. Some process needs to somehow merge them to remove duplicated data. Additionally we need to have good testing coverage to ensure that our implementation of `get_page@lsn` properly handles intersecting layers.
-
-There is another tradeoff. Approaches may be different in amount of traffic between system components. With first approach there can be increased traffic between follower and remote storage. But only in case follower has some activity that actually requests pages (!). With other approach traffic increase will be permanent and will be caused by two WAL streams instead of one.
-
-## Summary
-
-Proposed implementation strategy:
-
-Go with the simplest approach for now. Then work on tech debt, increase test coverage. Then gradually move forward to second approach by implementing safety measures first, finishing with switch of order between ignore and attach operation.
-
-And only then go to lease based approach to solve HA/Pageserver replica use cases.

docs/rfcs/024-ingress-egress-metrics.md

@@ -1,64 +0,0 @@
-# Ingress-Egress Metrics
-
-# Goals
-* Track ingress/egress traffic directly from a compute node, which may be sending stuff due to 
-  a custom extension.
-* Lay the groundwork for further traffic management on the compute node. 
-
-# Background
-As we move towards allowing custom extensions on the compute node, we need to cope with the fact
-that these extensions are untrusted. They could send or receive things from the open internet or
-even try to send bogus requests around on the internal network. Although we currently use the proxy
-to measure incoming/outgoing traffic, custom extensions will potentially be able to interact with
-the outside without involving the proxy. This has a high potential for abuse, so we need some ability
-to measure and lock down usage. This RFC concerns itself with the first part (measuring).
-
-# Possible Implementations
-* Service Mesh: There are lots of different implementations, but typically they have a 
-  proxy (proxy-per-node) that all traffic within the cluster goes through, and this proxy(-ies)
-  forwards metrics to the control plane. This seems more complicated than we need because: 
-  - We only really need this for our compute nodes (not safekeepers and pageservers)
-  - We control our compute nodes and have full power to customize it (which afaict is not an 
-    assumption that most service meshes can make)
-  - We already have a control plane, having two is more trouble than it's worth
-  - Having all traffic go through yet another proxy will add some amount of latency
-  - "You wanted a banana but what you got was a gorilla holding the banana and the entire jungle."
-
-* eBPF Program: This option essentially implements what we actually need from the service mesh
-  in its minimal form. The program will have very minimal overhead and will be run on each
-  compute node. It has the advantages of being small, fully-customizable, and performant. One
-  thing to consider is that it's not well-supported on MacOS, but given that we always run
-  everything in Linux docker containers this doesn't seem like an issue. 
-  
-# In Detail
-The eBPF program itself will be written in C and compiled with the eBPF backend as part of the 
-Compute Image. The program will be loaded into the kernel by `compute_ctl` just before
-it starts Postgres. The program will construct an eBPF ArrayMap that contains two elements:
-ingress bytes and egress bytes.
-```C
-struct 
-{
-        __uint(type, BPF_MAP_TYPE_ARRAY);
-        __type(key, u32);
-        __type(value, u64);
-        __uint(max_entries, 2);
-} ingress_egress_counter_map SEC(".maps");
-```
-
-On the kernel side, the eBPF program will intercept all packets,
-check the IP address to see if it is going in/outside of the network, and if so atomically increment
-the corresponding counter. On the userspace side, `compute_ctl` will `mmap` this array of counters
-(it isn't explicitly documented anywhere that you can do this, but I found
-[this patch](https://lwn.net/Articles/804180) that lets you mmap. Once per minute, `compute_ctl`
-will `atomic_exchange` these counters with `0`, wrap them up in a JSON, and send them to the
-control plane, in much the same format as in 021-metering.md.
-
-```json
-{
-    "metric" : "compute_ingress_bytes",
-    "endpoint_id" : "super-gigachad-117",
-    "event_start_time" : ...,
-    "event_stop_time" : ...,
-    "value" : ...
-}
-```

libs/pageserver_api/src/models.rs

@@ -48,33 +48,13 @@ pub enum TenantState {
 }
 
 impl TenantState {
-    pub fn attachment_status(&self) -> TenantAttachmentStatus {
-        use TenantAttachmentStatus::*;
+    pub fn has_in_progress_downloads(&self) -> bool {
         match self {
-            // The attach procedure writes the marker file before adding the Attaching tenant to the tenants map.
-            // So, technically, we can return Attached here.
-            // However, as soon as Console observes Attached, it will proceed with the Postgres-level health check.
-            // But, our attach task might still be fetching the remote timelines, etc.
-            // So, return `Maybe` while Attaching, making Console wait for the attach task to finish.
-            Self::Attaching => Maybe,
-            // tenant mgr startup distinguishes attaching from loading via marker file.
-            // If it's loading, there is no attach marker file, i.e., attach had finished in the past.
-            Self::Loading => Attached,
-            // We only reach Active after successful load / attach.
-            // So, call atttachment status Attached.
-            Self::Active => Attached,
-            // If the (initial or resumed) attach procedure fails, the tenant becomes Broken.
-            // However, it also becomes Broken if the regular load fails.
-            // We would need a separate TenantState variant to distinguish these cases.
-            // However, there's no practical difference from Console's perspective.
-            // It will run a Postgres-level health check as soon as it observes Attached.
-            // That will fail on Broken tenants.
-            // Console can then rollback the attach, or, wait for operator to fix the Broken tenant.
-            Self::Broken { .. } => Attached,
-            // Why is Stopping a Maybe case? Because, during pageserver shutdown,
-            // we set the Stopping state irrespective of whether the tenant
-            // has finished attaching or not.
-            Self::Stopping => Maybe,
+            Self::Loading => true,
+            Self::Attaching => true,
+            Self::Active => false,
+            Self::Stopping => false,
+            Self::Broken { .. } => false,
         }
     }
 
@@ -131,26 +111,11 @@ pub struct TimelineCreateRequest {
 }
 
 #[serde_as]
-#[derive(Serialize, Deserialize, Debug, Default)]
-#[serde(deny_unknown_fields)]
+#[derive(Serialize, Deserialize, Default)]
 pub struct TenantCreateRequest {
     #[serde(default)]
     #[serde_as(as = "Option<DisplayFromStr>")]
     pub new_tenant_id: Option<TenantId>,
-    #[serde(flatten)]
-    pub config: TenantConfig, // as we have a flattened field, we should reject all unknown fields in it
-}
-
-impl std::ops::Deref for TenantCreateRequest {
-    type Target = TenantConfig;
-
-    fn deref(&self) -> &Self::Target {
-        &self.config
-    }
-}
-
-#[derive(Serialize, Deserialize, Debug, Default)]
-pub struct TenantConfig {
     pub checkpoint_distance: Option<u64>,
     pub checkpoint_timeout: Option<String>,
     pub compaction_target_size: Option<u64>,
@@ -193,26 +158,37 @@ impl TenantCreateRequest {
 }
 
 #[serde_as]
-#[derive(Serialize, Deserialize, Debug)]
-#[serde(deny_unknown_fields)]
+#[derive(Serialize, Deserialize)]
 pub struct TenantConfigRequest {
     #[serde_as(as = "DisplayFromStr")]
     pub tenant_id: TenantId,
-    #[serde(flatten)]
-    pub config: TenantConfig, // as we have a flattened field, we should reject all unknown fields in it
-}
-
-impl std::ops::Deref for TenantConfigRequest {
-    type Target = TenantConfig;
-
-    fn deref(&self) -> &Self::Target {
-        &self.config
-    }
+    #[serde(default)]
+    pub checkpoint_distance: Option<u64>,
+    pub checkpoint_timeout: Option<String>,
+    pub compaction_target_size: Option<u64>,
+    pub compaction_period: Option<String>,
+    pub compaction_threshold: Option<usize>,
+    pub gc_horizon: Option<u64>,
+    pub gc_period: Option<String>,
+    pub image_creation_threshold: Option<usize>,
+    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 trace_read_requests: Option<bool>,
+    // We defer the parsing of the eviction_policy field to the request handler.
+    // Otherwise we'd have to move the types for eviction policy into this package.
+    // We might do that once the eviction feature has stabilizied.
+    // For now, this field is not even documented in the openapi_spec.yml.
+    pub eviction_policy: Option<serde_json::Value>,
+    pub min_resident_size_override: Option<u64>,
+    pub evictions_low_residence_duration_metric_threshold: Option<String>,
 }
 
 impl TenantConfigRequest {
     pub fn new(tenant_id: TenantId) -> TenantConfigRequest {
-        let config = TenantConfig {
+        TenantConfigRequest {
+            tenant_id,
             checkpoint_distance: None,
             checkpoint_timeout: None,
             compaction_target_size: None,
@@ -229,52 +205,20 @@ impl TenantConfigRequest {
             eviction_policy: None,
             min_resident_size_override: None,
             evictions_low_residence_duration_metric_threshold: None,
-        };
-        TenantConfigRequest { tenant_id, config }
+        }
     }
 }
 
-#[derive(Debug, Serialize, Deserialize)]
-pub struct TenantAttachRequest {
-    pub config: TenantAttachConfig,
-}
-
-/// Newtype to enforce deny_unknown_fields on TenantConfig for
-/// its usage inside `TenantAttachRequest`.
-#[derive(Debug, Serialize, Deserialize)]
-#[serde(deny_unknown_fields)]
-pub struct TenantAttachConfig {
-    #[serde(flatten)]
-    allowing_unknown_fields: TenantConfig,
-}
-
-impl std::ops::Deref for TenantAttachConfig {
-    type Target = TenantConfig;
-
-    fn deref(&self) -> &Self::Target {
-        &self.allowing_unknown_fields
-    }
-}
-
-/// See [`TenantState::attachment_status`] and the OpenAPI docs for context.
-#[derive(Serialize, Deserialize, Clone)]
-#[serde(rename_all = "snake_case")]
-pub enum TenantAttachmentStatus {
-    Maybe,
-    Attached,
-}
-
 #[serde_as]
 #[derive(Serialize, Deserialize, Clone)]
 pub struct TenantInfo {
     #[serde_as(as = "DisplayFromStr")]
     pub id: TenantId,
-    // NB: intentionally not part of OpenAPI, we don't want to commit to a specific set of TenantState's
     pub state: TenantState,
     /// Sum of the size of all layer files.
     /// If a layer is present in both local FS and S3, it counts only once.
     pub current_physical_size: Option<u64>, // physical size is only included in `tenant_status` endpoint
-    pub attachment_status: TenantAttachmentStatus,
+    pub has_in_progress_downloads: Option<bool>,
 }
 
 /// This represents the output of the "timeline_detail" and "timeline_list" API calls.
@@ -747,7 +691,7 @@ mod tests {
             id: TenantId::generate(),
             state: TenantState::Active,
             current_physical_size: Some(42),
-            attachment_status: TenantAttachmentStatus::Attached,
+            has_in_progress_downloads: Some(false),
         };
         let expected_active = json!({
             "id": original_active.id.to_string(),
@@ -755,7 +699,7 @@ mod tests {
                 "slug": "Active",
             },
             "current_physical_size": 42,
-            "attachment_status": "attached",
+            "has_in_progress_downloads": false,
         });
 
         let original_broken = TenantInfo {
@@ -765,7 +709,7 @@ mod tests {
                 backtrace: "backtrace info".into(),
             },
             current_physical_size: Some(42),
-            attachment_status: TenantAttachmentStatus::Attached,
+            has_in_progress_downloads: Some(false),
         };
         let expected_broken = json!({
             "id": original_broken.id.to_string(),
@@ -777,7 +721,7 @@ mod tests {
                 }
             },
             "current_physical_size": 42,
-            "attachment_status": "attached",
+            "has_in_progress_downloads": false,
         });
 
         assert_eq!(
@@ -792,43 +736,4 @@ mod tests {
         assert!(format!("{:?}", &original_broken.state).contains("reason"));
         assert!(format!("{:?}", &original_broken.state).contains("backtrace info"));
     }
-
-    #[test]
-    fn test_reject_unknown_field() {
-        let id = TenantId::generate();
-        let create_request = json!({
-            "new_tenant_id": id.to_string(),
-            "unknown_field": "unknown_value".to_string(),
-        });
-        let err = serde_json::from_value::<TenantCreateRequest>(create_request).unwrap_err();
-        assert!(
-            err.to_string().contains("unknown field `unknown_field`"),
-            "expect unknown field `unknown_field` error, got: {}",
-            err
-        );
-
-        let id = TenantId::generate();
-        let config_request = json!({
-            "tenant_id": id.to_string(),
-            "unknown_field": "unknown_value".to_string(),
-        });
-        let err = serde_json::from_value::<TenantConfigRequest>(config_request).unwrap_err();
-        assert!(
-            err.to_string().contains("unknown field `unknown_field`"),
-            "expect unknown field `unknown_field` error, got: {}",
-            err
-        );
-
-        let attach_request = json!({
-            "config": {
-                "unknown_field": "unknown_value".to_string(),
-            },
-        });
-        let err = serde_json::from_value::<TenantAttachRequest>(attach_request).unwrap_err();
-        assert!(
-            err.to_string().contains("unknown field `unknown_field`"),
-            "expect unknown field `unknown_field` error, got: {}",
-            err
-        );
-    }
 }

libs/postgres_ffi/src/pg_constants.rs

@@ -146,10 +146,6 @@ pub const XLH_UPDATE_OLD_ALL_VISIBLE_CLEARED: u8 = (1 << 0) as u8;
 pub const XLH_UPDATE_NEW_ALL_VISIBLE_CLEARED: u8 = (1 << 1) as u8;
 pub const XLH_DELETE_ALL_VISIBLE_CLEARED: u8 = (1 << 0) as u8;
 
-// From replication/message.h
-pub const XLOG_LOGICAL_MESSAGE: u8 = 0x00;
-
-// From rmgrlist.h
 pub const RM_XLOG_ID: u8 = 0;
 pub const RM_XACT_ID: u8 = 1;
 pub const RM_SMGR_ID: u8 = 2;
@@ -161,7 +157,6 @@ pub const RM_RELMAP_ID: u8 = 7;
 pub const RM_STANDBY_ID: u8 = 8;
 pub const RM_HEAP2_ID: u8 = 9;
 pub const RM_HEAP_ID: u8 = 10;
-pub const RM_LOGICALMSG_ID: u8 = 21;
 
 // from xlogreader.h
 pub const XLR_INFO_MASK: u8 = 0x0F;

libs/remote_storage/Cargo.toml

@@ -12,7 +12,6 @@ aws-smithy-http.workspace = true
 aws-types.workspace = true
 aws-config.workspace = true
 aws-sdk-s3.workspace = true
-aws-credential-types.workspace = true
 hyper = { workspace = true, features = ["stream"] }
 serde.workspace = true
 serde_json.workspace = true

libs/remote_storage/src/s3_bucket.rs

@@ -9,15 +9,14 @@ use std::sync::Arc;
 use anyhow::Context;
 use aws_config::{
     environment::credentials::EnvironmentVariableCredentialsProvider,
-    imds::credentials::ImdsCredentialsProvider, meta::credentials::CredentialsProviderChain,
+    imds::credentials::ImdsCredentialsProvider,
+    meta::credentials::{CredentialsProviderChain, LazyCachingCredentialsProvider},
 };
-use aws_credential_types::cache::CredentialsCache;
 use aws_sdk_s3::{
-    config::{Config, Region},
-    error::SdkError,
-    operation::get_object::GetObjectError,
-    primitives::ByteStream,
-    Client,
+    config::Config,
+    error::{GetObjectError, GetObjectErrorKind},
+    types::{ByteStream, SdkError},
+    Client, Endpoint, Region,
 };
 use aws_smithy_http::body::SdkBody;
 use hyper::Body;
@@ -126,23 +125,28 @@ impl S3Bucket {
 
         let credentials_provider = {
             // uses "AWS_ACCESS_KEY_ID", "AWS_SECRET_ACCESS_KEY"
-            CredentialsProviderChain::first_try(
-                "env",
-                EnvironmentVariableCredentialsProvider::new(),
-            )
+            let env_creds = EnvironmentVariableCredentialsProvider::new();
             // uses imds v2
-            .or_else("imds", ImdsCredentialsProvider::builder().build())
+            let imds = ImdsCredentialsProvider::builder().build();
+
+            // finally add caching.
+            // this might change in future, see https://github.com/awslabs/aws-sdk-rust/issues/629
+            LazyCachingCredentialsProvider::builder()
+                .load(CredentialsProviderChain::first_try("env", env_creds).or_else("imds", imds))
+                .build()
         };
 
         let mut config_builder = Config::builder()
             .region(Region::new(aws_config.bucket_region.clone()))
-            .credentials_cache(CredentialsCache::lazy())
             .credentials_provider(credentials_provider);
 
         if let Some(custom_endpoint) = aws_config.endpoint.clone() {
-            config_builder = config_builder
-                .endpoint_url(custom_endpoint)
-                .force_path_style(true);
+            let endpoint = Endpoint::immutable(
+                custom_endpoint
+                    .parse()
+                    .expect("Failed to parse S3 custom endpoint"),
+            );
+            config_builder.set_endpoint_resolver(Some(Arc::new(endpoint)));
         }
         let client = Client::from_conf(config_builder.build());
 
@@ -225,9 +229,14 @@ impl S3Bucket {
                     ))),
                 })
             }
-            Err(SdkError::ServiceError(e)) if matches!(e.err(), GetObjectError::NoSuchKey(_)) => {
-                Err(DownloadError::NotFound)
-            }
+            Err(SdkError::ServiceError {
+                err:
+                    GetObjectError {
+                        kind: GetObjectErrorKind::NoSuchKey(..),
+                        ..
+                    },
+                ..
+            }) => Err(DownloadError::NotFound),
             Err(e) => {
                 metrics::inc_get_object_fail();
                 Err(DownloadError::Other(anyhow::anyhow!(

libs/utils/scripts/restore_from_wal.sh

@@ -1,21 +1,21 @@
 #!/bin/bash
-
-set -euxo pipefail
-
 PG_BIN=$1
 WAL_PATH=$2
 DATA_DIR=$3
 PORT=$4
-SYSID=$(od -A n -j 24 -N 8 -t d8 "$WAL_PATH"/000000010000000000000002* | cut -c 3-)
-rm -fr "$DATA_DIR"
-env -i LD_LIBRARY_PATH="$PG_BIN"/../lib "$PG_BIN"/initdb -E utf8 -U cloud_admin -D "$DATA_DIR" --sysid="$SYSID"
-echo port="$PORT" >> "$DATA_DIR"/postgresql.conf
-REDO_POS=0x$("$PG_BIN"/pg_controldata -D "$DATA_DIR" | grep -F "REDO location"| cut -c 42-)
+SYSID=`od -A n -j 24 -N 8 -t d8 $WAL_PATH/000000010000000000000002* | cut -c 3-`
+rm -fr $DATA_DIR
+env -i LD_LIBRARY_PATH=$PG_BIN/../lib $PG_BIN/initdb -E utf8 -U cloud_admin -D $DATA_DIR --sysid=$SYSID
+echo port=$PORT >> $DATA_DIR/postgresql.conf
+REDO_POS=0x`$PG_BIN/pg_controldata -D $DATA_DIR | fgrep "REDO location"| cut -c 42-`
 declare -i WAL_SIZE=$REDO_POS+114
-"$PG_BIN"/pg_ctl -D "$DATA_DIR" -l logfile start
-"$PG_BIN"/pg_ctl -D "$DATA_DIR" -l logfile stop -m immediate
-cp "$DATA_DIR"/pg_wal/000000010000000000000001 .
-cp "$WAL_PATH"/* "$DATA_DIR"/pg_wal/
-for partial in "$DATA_DIR"/pg_wal/*.partial ; do mv "$partial" "${partial%.partial}" ; done
-dd if=000000010000000000000001 of="$DATA_DIR"/pg_wal/000000010000000000000001 bs=$WAL_SIZE count=1 conv=notrunc
+$PG_BIN/pg_ctl -D $DATA_DIR -l logfile start
+$PG_BIN/pg_ctl -D $DATA_DIR -l logfile stop -m immediate
+cp $DATA_DIR/pg_wal/000000010000000000000001 .
+cp $WAL_PATH/* $DATA_DIR/pg_wal/
+if [ -f $DATA_DIR/pg_wal/*.partial ]
+then
+	(cd $DATA_DIR/pg_wal ; for partial in \*.partial ; do mv $partial `basename $partial .partial` ; done)
+fi
+dd if=000000010000000000000001 of=$DATA_DIR/pg_wal/000000010000000000000001 bs=$WAL_SIZE count=1 conv=notrunc
 rm -f 000000010000000000000001

libs/utils/scripts/restore_from_wal_archive.sh

@@ -0,0 +1,20 @@
+PG_BIN=$1
+WAL_PATH=$2
+DATA_DIR=$3
+PORT=$4
+SYSID=`od -A n -j 24 -N 8 -t d8 $WAL_PATH/000000010000000000000002* | cut -c 3-`
+rm -fr $DATA_DIR /tmp/pg_wals
+mkdir /tmp/pg_wals
+env -i LD_LIBRARY_PATH=$PG_BIN/../lib $PG_BIN/initdb -E utf8 -U cloud_admin -D $DATA_DIR --sysid=$SYSID
+echo port=$PORT >> $DATA_DIR/postgresql.conf
+REDO_POS=0x`$PG_BIN/pg_controldata -D $DATA_DIR | fgrep "REDO location"| cut -c 42-`
+declare -i WAL_SIZE=$REDO_POS+114
+cp $WAL_PATH/* /tmp/pg_wals
+if [ -f $DATA_DIR/pg_wal/*.partial ]
+then
+	(cd /tmp/pg_wals ; for partial in \*.partial ; do  mv $partial `basename $partial .partial` ; done)
+fi
+dd if=$DATA_DIR/pg_wal/000000010000000000000001 of=/tmp/pg_wals/000000010000000000000001 bs=$WAL_SIZE count=1 conv=notrunc
+echo > $DATA_DIR/recovery.signal
+rm -f $DATA_DIR/pg_wal/*
+echo "restore_command = 'cp /tmp/pg_wals/%f %p'" >> $DATA_DIR/postgresql.conf

libs/utils/src/http/json.rs

@@ -8,26 +8,12 @@ use super::error::ApiError;
 pub async fn json_request<T: for<'de> Deserialize<'de>>(
     request: &mut Request<Body>,
 ) -> Result<T, ApiError> {
-    json_request_or_empty_body(request)
-        .await?
-        .context("missing request body")
-        .map_err(ApiError::BadRequest)
-}
-
-/// Will be removed as part of https://github.com/neondatabase/neon/issues/4282
-pub async fn json_request_or_empty_body<T: for<'de> Deserialize<'de>>(
-    request: &mut Request<Body>,
-) -> Result<Option<T>, ApiError> {
-    let body = hyper::body::aggregate(request.body_mut())
+    let whole_body = hyper::body::aggregate(request.body_mut())
         .await
         .context("Failed to read request body")
         .map_err(ApiError::BadRequest)?;
-    if body.remaining() == 0 {
-        return Ok(None);
-    }
-    serde_json::from_reader(body.reader())
+    serde_json::from_reader(whole_body.reader())
         .context("Failed to parse json request")
-        .map(Some)
         .map_err(ApiError::BadRequest)
 }
 

libs/utils/src/lib.rs

@@ -60,43 +60,24 @@ pub mod tracing_span_assert;
 
 pub mod rate_limit;
 
-mod failpoint_macro_helpers {
-
-    /// use with fail::cfg("$name", "return(2000)")
-    ///
-    /// The effect is similar to a "sleep(2000)" action, i.e. we sleep for the
-    /// specified time (in milliseconds). The main difference is that we use async
-    /// tokio sleep function. Another difference is that we print lines to the log,
-    /// which can be useful in tests to check that the failpoint was hit.
-    #[macro_export]
-    macro_rules! failpoint_sleep_millis_async {
-        ($name:literal) => {{
-            // If the failpoint is used with a "return" action, set should_sleep to the
-            // returned value (as string). Otherwise it's set to None.
-            let should_sleep = (|| {
-                ::fail::fail_point!($name, |x| x);
-                ::std::option::Option::None
-            })();
-
-            // Sleep if the action was a returned value
-            if let ::std::option::Option::Some(duration_str) = should_sleep {
-                $crate::failpoint_sleep_helper($name, duration_str).await
-            }
-        }};
-    }
-
-    // Helper function used by the macro. (A function has nicer scoping so we
-    // don't need to decorate everything with "::")
-    pub async fn failpoint_sleep_helper(name: &'static str, duration_str: String) {
-        let millis = duration_str.parse::<u64>().unwrap();
-        let d = std::time::Duration::from_millis(millis);
-
-        tracing::info!("failpoint {:?}: sleeping for {:?}", name, d);
-        tokio::time::sleep(d).await;
-        tracing::info!("failpoint {:?}: sleep done", name);
-    }
+/// use with fail::cfg("$name", "return(2000)")
+#[macro_export]
+macro_rules! failpoint_sleep_millis_async {
+    ($name:literal) => {{
+        let should_sleep: Option<std::time::Duration> = (|| {
+            fail::fail_point!($name, |v: Option<_>| {
+                let millis = v.unwrap().parse::<u64>().unwrap();
+                Some(Duration::from_millis(millis))
+            });
+            None
+        })();
+        if let Some(d) = should_sleep {
+            tracing::info!("failpoint {:?}: sleeping for {:?}", $name, d);
+            tokio::time::sleep(d).await;
+            tracing::info!("failpoint {:?}: sleep done", $name);
+        }
+    }};
 }
-pub use failpoint_macro_helpers::failpoint_sleep_helper;
 
 /// This is a shortcut to embed git sha into binaries and avoid copying the same build script to all packages
 ///

libs/utils/src/seqwait.rs

@@ -144,8 +144,6 @@ where
     ///
     /// This call won't complete until someone has called `advance`
     /// with a number greater than or equal to the one we're waiting for.
-    ///
-    /// This function is async cancellation-safe.
     pub async fn wait_for(&self, num: V) -> Result<(), SeqWaitError> {
         match self.queue_for_wait(num) {
             Ok(None) => Ok(()),
@@ -161,8 +159,6 @@ where
     ///
     /// If that hasn't happened after the specified timeout duration,
     /// [`SeqWaitError::Timeout`] will be returned.
-    ///
-    /// This function is async cancellation-safe.
     pub async fn wait_for_timeout(
         &self,
         num: V,

pageserver/Cargo.toml

@@ -75,6 +75,8 @@ enum-map.workspace = true
 enumset.workspace = true
 strum.workspace = true
 strum_macros.workspace = true
+# feature "send_guard" markes it so that lock guards are Send
+parking_lot = { workspace = true, default-features = false, features = [ "send_guard" ] }
 
 [dev-dependencies]
 criterion.workspace = true

pageserver/ctl/Cargo.toml

@@ -1,18 +0,0 @@
-[package]
-name = "pagectl"
-version = "0.1.0"
-edition.workspace = true
-license.workspace = true
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-anyhow.workspace = true
-bytes.workspace = true
-clap = { workspace = true, features = ["string"] }
-git-version.workspace = true
-pageserver = { path = ".." }
-postgres_ffi.workspace = true
-utils.workspace = true
-svg_fmt.workspace = true
-workspace_hack.workspace = true

pageserver/ctl/src/layers.rs

@@ -1,169 +0,0 @@
-use std::path::{Path, PathBuf};
-
-use anyhow::Result;
-use clap::Subcommand;
-use pageserver::tenant::block_io::BlockCursor;
-use pageserver::tenant::disk_btree::DiskBtreeReader;
-use pageserver::tenant::storage_layer::delta_layer::{BlobRef, Summary};
-use pageserver::{page_cache, virtual_file};
-use pageserver::{
-    repository::{Key, KEY_SIZE},
-    tenant::{
-        block_io::FileBlockReader, disk_btree::VisitDirection,
-        storage_layer::delta_layer::DELTA_KEY_SIZE,
-    },
-    virtual_file::VirtualFile,
-};
-use std::fs;
-use utils::bin_ser::BeSer;
-
-use crate::layer_map_analyzer::parse_filename;
-
-#[derive(Subcommand)]
-pub(crate) enum LayerCmd {
-    /// List all tenants and timelines under the pageserver path
-    ///
-    /// Example: `cargo run --bin pagectl layer list .neon/`
-    List { path: PathBuf },
-    /// List all layers of a given tenant and timeline
-    ///
-    /// Example: `cargo run --bin pagectl layer list .neon/`
-    ListLayer {
-        path: PathBuf,
-        tenant: String,
-        timeline: String,
-    },
-    /// Dump all information of a layer file
-    DumpLayer {
-        path: PathBuf,
-        tenant: String,
-        timeline: String,
-        /// The id from list-layer command
-        id: usize,
-    },
-}
-
-fn read_delta_file(path: impl AsRef<Path>) -> Result<()> {
-    use pageserver::tenant::blob_io::BlobCursor;
-    use pageserver::tenant::block_io::BlockReader;
-
-    let path = path.as_ref();
-    virtual_file::init(10);
-    page_cache::init(100);
-    let file = FileBlockReader::new(VirtualFile::open(path)?);
-    let summary_blk = file.read_blk(0)?;
-    let actual_summary = Summary::des_prefix(summary_blk.as_ref())?;
-    let tree_reader = DiskBtreeReader::<_, DELTA_KEY_SIZE>::new(
-        actual_summary.index_start_blk,
-        actual_summary.index_root_blk,
-        &file,
-    );
-    // TODO(chi): dedup w/ `delta_layer.rs` by exposing the API.
-    let mut all = vec![];
-    tree_reader.visit(
-        &[0u8; DELTA_KEY_SIZE],
-        VisitDirection::Forwards,
-        |key, value_offset| {
-            let curr = Key::from_slice(&key[..KEY_SIZE]);
-            all.push((curr, BlobRef(value_offset)));
-            true
-        },
-    )?;
-    let mut cursor = BlockCursor::new(&file);
-    for (k, v) in all {
-        let value = cursor.read_blob(v.pos())?;
-        println!("key:{} value_len:{}", k, value.len());
-    }
-    // TODO(chi): special handling for last key?
-    Ok(())
-}
-
-pub(crate) fn main(cmd: &LayerCmd) -> Result<()> {
-    match cmd {
-        LayerCmd::List { path } => {
-            for tenant in fs::read_dir(path.join("tenants"))? {
-                let tenant = tenant?;
-                if !tenant.file_type()?.is_dir() {
-                    continue;
-                }
-                println!("tenant {}", tenant.file_name().to_string_lossy());
-                for timeline in fs::read_dir(tenant.path().join("timelines"))? {
-                    let timeline = timeline?;
-                    if !timeline.file_type()?.is_dir() {
-                        continue;
-                    }
-                    println!("- timeline {}", timeline.file_name().to_string_lossy());
-                }
-            }
-        }
-        LayerCmd::ListLayer {
-            path,
-            tenant,
-            timeline,
-        } => {
-            let timeline_path = path
-                .join("tenants")
-                .join(tenant)
-                .join("timelines")
-                .join(timeline);
-            let mut idx = 0;
-            for layer in fs::read_dir(timeline_path)? {
-                let layer = layer?;
-                if let Some(layer_file) = parse_filename(&layer.file_name().into_string().unwrap())
-                {
-                    println!(
-                        "[{:3}]  key:{}-{}\n       lsn:{}-{}\n       delta:{}",
-                        idx,
-                        layer_file.key_range.start,
-                        layer_file.key_range.end,
-                        layer_file.lsn_range.start,
-                        layer_file.lsn_range.end,
-                        layer_file.is_delta,
-                    );
-                    idx += 1;
-                }
-            }
-        }
-        LayerCmd::DumpLayer {
-            path,
-            tenant,
-            timeline,
-            id,
-        } => {
-            let timeline_path = path
-                .join("tenants")
-                .join(tenant)
-                .join("timelines")
-                .join(timeline);
-            let mut idx = 0;
-            for layer in fs::read_dir(timeline_path)? {
-                let layer = layer?;
-                if let Some(layer_file) = parse_filename(&layer.file_name().into_string().unwrap())
-                {
-                    if *id == idx {
-                        // TODO(chi): dedup code
-                        println!(
-                            "[{:3}]  key:{}-{}\n       lsn:{}-{}\n       delta:{}",
-                            idx,
-                            layer_file.key_range.start,
-                            layer_file.key_range.end,
-                            layer_file.lsn_range.start,
-                            layer_file.lsn_range.end,
-                            layer_file.is_delta,
-                        );
-
-                        if layer_file.is_delta {
-                            read_delta_file(layer.path())?;
-                        } else {
-                            anyhow::bail!("not supported yet :(");
-                        }
-
-                        break;
-                    }
-                    idx += 1;
-                }
-            }
-        }
-    }
-    Ok(())
-}

pageserver/ctl/src/main.rs

@@ -1,179 +0,0 @@
-//! A helper tool to manage pageserver binary files.
-//! Accepts a file as an argument, attempts to parse it with all ways possible
-//! and prints its interpreted context.
-//!
-//! Separate, `metadata` subcommand allows to print and update pageserver's metadata file.
-
-mod draw_timeline_dir;
-mod layer_map_analyzer;
-mod layers;
-
-use clap::{Parser, Subcommand};
-use layers::LayerCmd;
-use pageserver::{
-    context::{DownloadBehavior, RequestContext},
-    page_cache,
-    task_mgr::TaskKind,
-    tenant::{dump_layerfile_from_path, metadata::TimelineMetadata},
-    virtual_file,
-};
-use postgres_ffi::ControlFileData;
-use std::path::{Path, PathBuf};
-use utils::{lsn::Lsn, project_git_version};
-
-project_git_version!(GIT_VERSION);
-
-#[derive(Parser)]
-#[command(
-    version = GIT_VERSION,
-    about = "Neon Pageserver binutils",
-    long_about = "Reads pageserver (and related) binary files management utility"
-)]
-#[command(propagate_version = true)]
-struct CliOpts {
-    #[command(subcommand)]
-    command: Commands,
-}
-
-#[derive(Subcommand)]
-enum Commands {
-    Metadata(MetadataCmd),
-    PrintLayerFile(PrintLayerFileCmd),
-    DrawTimeline {},
-    AnalyzeLayerMap(AnalyzeLayerMapCmd),
-    #[command(subcommand)]
-    Layer(LayerCmd),
-}
-
-/// Read and update pageserver metadata file
-#[derive(Parser)]
-struct MetadataCmd {
-    /// Input metadata file path
-    metadata_path: PathBuf,
-    /// Replace disk consistent Lsn
-    disk_consistent_lsn: Option<Lsn>,
-    /// Replace previous record Lsn
-    prev_record_lsn: Option<Lsn>,
-    /// Replace latest gc cuttoff
-    latest_gc_cuttoff: Option<Lsn>,
-}
-
-#[derive(Parser)]
-struct PrintLayerFileCmd {
-    /// Pageserver data path
-    path: PathBuf,
-}
-
-#[derive(Parser)]
-struct AnalyzeLayerMapCmd {
-    /// Pageserver data path
-    path: PathBuf,
-    /// Max holes
-    max_holes: Option<usize>,
-}
-
-fn main() -> anyhow::Result<()> {
-    let cli = CliOpts::parse();
-
-    match cli.command {
-        Commands::Layer(cmd) => {
-            layers::main(&cmd)?;
-        }
-        Commands::Metadata(cmd) => {
-            handle_metadata(&cmd)?;
-        }
-        Commands::DrawTimeline {} => {
-            draw_timeline_dir::main()?;
-        }
-        Commands::AnalyzeLayerMap(cmd) => {
-            layer_map_analyzer::main(&cmd)?;
-        }
-        Commands::PrintLayerFile(cmd) => {
-            if let Err(e) = read_pg_control_file(&cmd.path) {
-                println!(
-                    "Failed to read input file as a pg control one: {e:#}\n\
-                    Attempting to read it as layer file"
-                );
-                print_layerfile(&cmd.path)?;
-            }
-        }
-    };
-    Ok(())
-}
-
-fn read_pg_control_file(control_file_path: &Path) -> anyhow::Result<()> {
-    let control_file = ControlFileData::decode(&std::fs::read(control_file_path)?)?;
-    println!("{control_file:?}");
-    let control_file_initdb = Lsn(control_file.checkPoint);
-    println!(
-        "pg_initdb_lsn: {}, aligned: {}",
-        control_file_initdb,
-        control_file_initdb.align()
-    );
-    Ok(())
-}
-
-fn print_layerfile(path: &Path) -> anyhow::Result<()> {
-    // Basic initialization of things that don't change after startup
-    virtual_file::init(10);
-    page_cache::init(100);
-    let ctx = RequestContext::new(TaskKind::DebugTool, DownloadBehavior::Error);
-    dump_layerfile_from_path(path, true, &ctx)
-}
-
-fn handle_metadata(
-    MetadataCmd {
-        metadata_path: path,
-        disk_consistent_lsn,
-        prev_record_lsn,
-        latest_gc_cuttoff,
-    }: &MetadataCmd,
-) -> Result<(), anyhow::Error> {
-    let metadata_bytes = std::fs::read(path)?;
-    let mut meta = TimelineMetadata::from_bytes(&metadata_bytes)?;
-    println!("Current metadata:\n{meta:?}");
-    let mut update_meta = false;
-    if let Some(disk_consistent_lsn) = disk_consistent_lsn {
-        meta = TimelineMetadata::new(
-            *disk_consistent_lsn,
-            meta.prev_record_lsn(),
-            meta.ancestor_timeline(),
-            meta.ancestor_lsn(),
-            meta.latest_gc_cutoff_lsn(),
-            meta.initdb_lsn(),
-            meta.pg_version(),
-        );
-        update_meta = true;
-    }
-    if let Some(prev_record_lsn) = prev_record_lsn {
-        meta = TimelineMetadata::new(
-            meta.disk_consistent_lsn(),
-            Some(*prev_record_lsn),
-            meta.ancestor_timeline(),
-            meta.ancestor_lsn(),
-            meta.latest_gc_cutoff_lsn(),
-            meta.initdb_lsn(),
-            meta.pg_version(),
-        );
-        update_meta = true;
-    }
-    if let Some(latest_gc_cuttoff) = latest_gc_cuttoff {
-        meta = TimelineMetadata::new(
-            meta.disk_consistent_lsn(),
-            meta.prev_record_lsn(),
-            meta.ancestor_timeline(),
-            meta.ancestor_lsn(),
-            *latest_gc_cuttoff,
-            meta.initdb_lsn(),
-            meta.pg_version(),
-        );
-        update_meta = true;
-    }
-
-    if update_meta {
-        let metadata_bytes = meta.to_bytes()?;
-        std::fs::write(path, metadata_bytes)?;
-    }
-
-    Ok(())
-}

pageserver/src/bin/draw_timeline_dir.rs

@@ -12,7 +12,7 @@
 //! Example use:
 //! ```
 //! $ ls test_output/test_pgbench\[neon-45-684\]/repo/tenants/$TENANT/timelines/$TIMELINE | \
-//! $   grep "__" | cargo run --release --bin pagectl draw-timeline-dir > out.svg
+//! $   grep "__" | cargo run --release --bin draw_timeline_dir > out.svg
 //! $ firefox out.svg
 //! ```
 //!
@@ -62,7 +62,7 @@ fn parse_filename(name: &str) -> (Range<Key>, Range<Lsn>) {
     (keys, lsns)
 }
 
-pub fn main() -> Result<()> {
+fn main() -> Result<()> {
     // Parse layer filenames from stdin
     let mut ranges: Vec<(Range<Key>, Range<Lsn>)> = vec![];
     let stdin = io::stdin();

pageserver/src/bin/layer_map_analyzer.rs

@@ -6,7 +6,7 @@ use anyhow::Result;
 use std::cmp::Ordering;
 use std::collections::BinaryHeap;
 use std::ops::Range;
-use std::{fs, path::Path, str};
+use std::{env, fs, path::Path, path::PathBuf, str, str::FromStr};
 
 use pageserver::page_cache::PAGE_SZ;
 use pageserver::repository::{Key, KEY_SIZE};
@@ -18,14 +18,12 @@ use pageserver::virtual_file::VirtualFile;
 
 use utils::{bin_ser::BeSer, lsn::Lsn};
 
-use crate::AnalyzeLayerMapCmd;
-
 const MIN_HOLE_LENGTH: i128 = (128 * 1024 * 1024 / PAGE_SZ) as i128;
 const DEFAULT_MAX_HOLES: usize = 10;
 
 /// Wrapper for key range to provide reverse ordering by range length for BinaryHeap
 #[derive(PartialEq, Eq)]
-pub struct Hole(Range<Key>);
+struct Hole(Range<Key>);
 
 impl Ord for Hole {
     fn cmp(&self, other: &Self) -> Ordering {
@@ -41,11 +39,11 @@ impl PartialOrd for Hole {
     }
 }
 
-pub(crate) struct LayerFile {
-    pub key_range: Range<Key>,
-    pub lsn_range: Range<Lsn>,
-    pub is_delta: bool,
-    pub holes: Vec<Hole>,
+struct LayerFile {
+    key_range: Range<Key>,
+    lsn_range: Range<Lsn>,
+    is_delta: bool,
+    holes: Vec<Hole>,
 }
 
 impl LayerFile {
@@ -69,7 +67,7 @@ impl LayerFile {
     }
 }
 
-pub(crate) fn parse_filename(name: &str) -> Option<LayerFile> {
+fn parse_filename(name: &str) -> Option<LayerFile> {
     let split: Vec<&str> = name.split("__").collect();
     if split.len() != 2 {
         return None;
@@ -129,9 +127,18 @@ fn get_holes(path: &Path, max_holes: usize) -> Result<Vec<Hole>> {
     Ok(holes)
 }
 
-pub(crate) fn main(cmd: &AnalyzeLayerMapCmd) -> Result<()> {
-    let storage_path = &cmd.path;
-    let max_holes = cmd.max_holes.unwrap_or(DEFAULT_MAX_HOLES);
+fn main() -> Result<()> {
+    let args: Vec<String> = env::args().collect();
+    if args.len() < 2 {
+        println!("Usage: layer_map_analyzer PAGESERVER_DATA_DIR [MAX_HOLES]");
+        return Ok(());
+    }
+    let storage_path = PathBuf::from_str(&args[1])?;
+    let max_holes = if args.len() > 2 {
+        args[2].parse::<usize>().unwrap()
+    } else {
+        DEFAULT_MAX_HOLES
+    };
 
     // Initialize virtual_file (file desriptor cache) and page cache which are needed to access layer persistent B-Tree.
     pageserver::virtual_file::init(10);

pageserver/src/bin/pageserver.rs

@@ -9,7 +9,6 @@ use clap::{Arg, ArgAction, Command};
 use fail::FailScenario;
 use metrics::launch_timestamp::{set_launch_timestamp_metric, LaunchTimestamp};
 use pageserver::disk_usage_eviction_task::{self, launch_disk_usage_global_eviction_task};
-use pageserver::task_mgr::WALRECEIVER_RUNTIME;
 use remote_storage::GenericRemoteStorage;
 use tracing::*;
 
@@ -19,7 +18,9 @@ use pageserver::{
     context::{DownloadBehavior, RequestContext},
     http, page_cache, page_service, task_mgr,
     task_mgr::TaskKind,
-    task_mgr::{BACKGROUND_RUNTIME, COMPUTE_REQUEST_RUNTIME, MGMT_REQUEST_RUNTIME},
+    task_mgr::{
+        BACKGROUND_RUNTIME, COMPUTE_REQUEST_RUNTIME, MGMT_REQUEST_RUNTIME, WALRECEIVER_RUNTIME,
+    },
     tenant::mgr,
     virtual_file,
 };
@@ -275,18 +276,7 @@ fn start_pageserver(
     let pageserver_listener = tcp_listener::bind(pg_addr)?;
 
     // Launch broker client
-    // The storage_broker::connect call needs to happen inside a tokio runtime thread.
-    let broker_client = WALRECEIVER_RUNTIME
-        .block_on(async {
-            // Note: we do not attempt connecting here (but validate endpoints sanity).
-            storage_broker::connect(conf.broker_endpoint.clone(), conf.broker_keepalive_interval)
-        })
-        .with_context(|| {
-            format!(
-                "create broker client for uri={:?} keepalive_interval={:?}",
-                &conf.broker_endpoint, conf.broker_keepalive_interval,
-            )
-        })?;
+    WALRECEIVER_RUNTIME.block_on(pageserver::broker_client::init_broker_client(conf))?;
 
     // Initialize authentication for incoming connections
     let http_auth;
@@ -336,11 +326,7 @@ fn start_pageserver(
     let remote_storage = create_remote_storage_client(conf)?;
 
     // Scan the local 'tenants/' directory and start loading the tenants
-    BACKGROUND_RUNTIME.block_on(mgr::init_tenant_mgr(
-        conf,
-        broker_client.clone(),
-        remote_storage.clone(),
-    ))?;
+    BACKGROUND_RUNTIME.block_on(mgr::init_tenant_mgr(conf, remote_storage.clone()))?;
 
     // shared state between the disk-usage backed eviction background task and the http endpoint
     // that allows triggering disk-usage based eviction manually. note that the http endpoint
@@ -365,7 +351,6 @@ fn start_pageserver(
             conf,
             launch_ts,
             http_auth,
-            broker_client.clone(),
             remote_storage,
             disk_usage_eviction_state,
         )?
@@ -442,7 +427,6 @@ fn start_pageserver(
             async move {
                 page_service::libpq_listener_main(
                     conf,
-                    broker_client,
                     pg_auth,
                     pageserver_listener,
                     conf.pg_auth_type,

pageserver/src/bin/pageserver_binutils.rs

@@ -0,0 +1,157 @@
+//! A helper tool to manage pageserver binary files.
+//! Accepts a file as an argument, attempts to parse it with all ways possible
+//! and prints its interpreted context.
+//!
+//! Separate, `metadata` subcommand allows to print and update pageserver's metadata file.
+use std::{
+    path::{Path, PathBuf},
+    str::FromStr,
+};
+
+use anyhow::Context;
+use clap::{value_parser, Arg, Command};
+
+use pageserver::{
+    context::{DownloadBehavior, RequestContext},
+    page_cache,
+    task_mgr::TaskKind,
+    tenant::{dump_layerfile_from_path, metadata::TimelineMetadata},
+    virtual_file,
+};
+use postgres_ffi::ControlFileData;
+use utils::{lsn::Lsn, project_git_version};
+
+project_git_version!(GIT_VERSION);
+
+const METADATA_SUBCOMMAND: &str = "metadata";
+
+fn main() -> anyhow::Result<()> {
+    let arg_matches = cli().get_matches();
+
+    match arg_matches.subcommand() {
+        Some((subcommand_name, subcommand_matches)) => {
+            let path = subcommand_matches
+                .get_one::<PathBuf>("metadata_path")
+                .context("'metadata_path' argument is missing")?
+                .to_path_buf();
+            anyhow::ensure!(
+                subcommand_name == METADATA_SUBCOMMAND,
+                "Unknown subcommand {subcommand_name}"
+            );
+            handle_metadata(&path, subcommand_matches)?;
+        }
+        None => {
+            let path = arg_matches
+                .get_one::<PathBuf>("path")
+                .context("'path' argument is missing")?
+                .to_path_buf();
+            println!(
+                "No subcommand specified, attempting to guess the format for file {}",
+                path.display()
+            );
+            if let Err(e) = read_pg_control_file(&path) {
+                println!(
+                    "Failed to read input file as a pg control one: {e:#}\n\
+                    Attempting to read it as layer file"
+                );
+                print_layerfile(&path)?;
+            }
+        }
+    };
+    Ok(())
+}
+
+fn read_pg_control_file(control_file_path: &Path) -> anyhow::Result<()> {
+    let control_file = ControlFileData::decode(&std::fs::read(control_file_path)?)?;
+    println!("{control_file:?}");
+    let control_file_initdb = Lsn(control_file.checkPoint);
+    println!(
+        "pg_initdb_lsn: {}, aligned: {}",
+        control_file_initdb,
+        control_file_initdb.align()
+    );
+    Ok(())
+}
+
+fn print_layerfile(path: &Path) -> anyhow::Result<()> {
+    // Basic initialization of things that don't change after startup
+    virtual_file::init(10);
+    page_cache::init(100);
+    let ctx = RequestContext::new(TaskKind::DebugTool, DownloadBehavior::Error);
+    dump_layerfile_from_path(path, true, &ctx)
+}
+
+fn handle_metadata(path: &Path, arg_matches: &clap::ArgMatches) -> Result<(), anyhow::Error> {
+    let metadata_bytes = std::fs::read(path)?;
+    let mut meta = TimelineMetadata::from_bytes(&metadata_bytes)?;
+    println!("Current metadata:\n{meta:?}");
+    let mut update_meta = false;
+    if let Some(disk_consistent_lsn) = arg_matches.get_one::<String>("disk_consistent_lsn") {
+        meta = TimelineMetadata::new(
+            Lsn::from_str(disk_consistent_lsn)?,
+            meta.prev_record_lsn(),
+            meta.ancestor_timeline(),
+            meta.ancestor_lsn(),
+            meta.latest_gc_cutoff_lsn(),
+            meta.initdb_lsn(),
+            meta.pg_version(),
+        );
+        update_meta = true;
+    }
+    if let Some(prev_record_lsn) = arg_matches.get_one::<String>("prev_record_lsn") {
+        meta = TimelineMetadata::new(
+            meta.disk_consistent_lsn(),
+            Some(Lsn::from_str(prev_record_lsn)?),
+            meta.ancestor_timeline(),
+            meta.ancestor_lsn(),
+            meta.latest_gc_cutoff_lsn(),
+            meta.initdb_lsn(),
+            meta.pg_version(),
+        );
+        update_meta = true;
+    }
+
+    if update_meta {
+        let metadata_bytes = meta.to_bytes()?;
+        std::fs::write(path, metadata_bytes)?;
+    }
+
+    Ok(())
+}
+
+fn cli() -> Command {
+    Command::new("Neon Pageserver binutils")
+        .about("Reads pageserver (and related) binary files management utility")
+        .version(GIT_VERSION)
+        .arg(
+            Arg::new("path")
+                .help("Input file path")
+                .value_parser(value_parser!(PathBuf))
+                .required(false),
+        )
+        .subcommand(
+            Command::new(METADATA_SUBCOMMAND)
+                .about("Read and update pageserver metadata file")
+                .arg(
+                    Arg::new("metadata_path")
+                        .help("Input metadata file path")
+                        .value_parser(value_parser!(PathBuf))
+                        .required(false),
+                )
+                .arg(
+                    Arg::new("disk_consistent_lsn")
+                        .long("disk_consistent_lsn")
+                        .help("Replace disk consistent Lsn"),
+                )
+                .arg(
+                    Arg::new("prev_record_lsn")
+                        .long("prev_record_lsn")
+                        .help("Replace previous record Lsn"),
+                ),
+        )
+}
+
+#[test]
+fn verify_cli() {
+    cli().debug_assert();
+}

pageserver/src/broker_client.rs

@@ -0,0 +1,48 @@
+//! The broker client instance of the pageserver, created during pageserver startup.
+//! Used by each timelines' [`walreceiver`].
+
+use crate::config::PageServerConf;
+
+use anyhow::Context;
+use once_cell::sync::OnceCell;
+use storage_broker::BrokerClientChannel;
+use tracing::*;
+
+static BROKER_CLIENT: OnceCell<BrokerClientChannel> = OnceCell::new();
+
+///
+/// Initialize the broker client. This must be called once at page server startup.
+///
+pub async fn init_broker_client(conf: &'static PageServerConf) -> anyhow::Result<()> {
+    let broker_endpoint = conf.broker_endpoint.clone();
+
+    // Note: we do not attempt connecting here (but validate endpoints sanity).
+    let broker_client =
+        storage_broker::connect(broker_endpoint.clone(), conf.broker_keepalive_interval).context(
+            format!(
+                "Failed to create broker client to {}",
+                &conf.broker_endpoint
+            ),
+        )?;
+
+    if BROKER_CLIENT.set(broker_client).is_err() {
+        panic!("broker already initialized");
+    }
+
+    info!(
+        "Initialized broker client with endpoints: {}",
+        broker_endpoint
+    );
+    Ok(())
+}
+
+///
+/// Get a handle to the broker client
+///
+pub fn get_broker_client() -> &'static BrokerClientChannel {
+    BROKER_CLIENT.get().expect("broker client not initialized")
+}
+
+pub fn is_broker_client_initialized() -> bool {
+    BROKER_CLIENT.get().is_some()
+}

pageserver/src/config.rs

@@ -797,8 +797,7 @@ impl PageServerConf {
             )?);
         }
         if let Some(max_lsn_wal_lag) = item.get("max_lsn_wal_lag") {
-            t_conf.max_lsn_wal_lag =
-                Some(deserialize_from_item("max_lsn_wal_lag", max_lsn_wal_lag)?);
+            t_conf.max_lsn_wal_lag = Some(parse_toml_from_str("max_lsn_wal_lag", max_lsn_wal_lag)?);
         }
         if let Some(trace_read_requests) = item.get("trace_read_requests") {
             t_conf.trace_read_requests =

pageserver/src/consumption_metrics.rs

@@ -5,7 +5,7 @@
 //!
 use crate::context::{DownloadBehavior, RequestContext};
 use crate::task_mgr::{self, TaskKind, BACKGROUND_RUNTIME};
-use crate::tenant::{mgr, LogicalSizeCalculationCause};
+use crate::tenant::mgr;
 use anyhow;
 use chrono::Utc;
 use consumption_metrics::{idempotency_key, Event, EventChunk, EventType, CHUNK_SIZE};
@@ -164,8 +164,7 @@ pub async fn collect_metrics_iteration(
                     timeline_written_size,
                 ));
 
-                let span = info_span!("collect_metrics_iteration", tenant_id = %timeline.tenant_id, timeline_id = %timeline.timeline_id);
-                match span.in_scope(|| timeline.get_current_logical_size(ctx)) {
+                match timeline.get_current_logical_size(ctx) {
                     // Only send timeline logical size when it is fully calculated.
                     Ok((size, is_exact)) if is_exact => {
                         current_metrics.push((
@@ -335,9 +334,7 @@ pub async fn calculate_synthetic_size_worker(
 
                     if let Ok(tenant) = mgr::get_tenant(tenant_id, true).await
                     {
-                        if let Err(e) = tenant.calculate_synthetic_size(
-                            LogicalSizeCalculationCause::ConsumptionMetricsSyntheticSize,
-                            ctx).await {
+                        if let Err(e) = tenant.calculate_synthetic_size(ctx).await {
                             error!("failed to calculate synthetic size for tenant {}: {}", tenant_id, e);
                         }
                     }

pageserver/src/http/openapi_spec.yml

@@ -346,46 +346,28 @@ paths:
         starts writing to the tenant's S3 state unless it receives one of the
         distinguished errors below that state otherwise.
 
-        If a client receives a not-distinguished response, e.g., a network timeout,
-        it MUST retry the /attach request and poll again for the tenant's
-        attachment status.
-
-        After the client has received a 202, it MUST poll the tenant's
-        attachment status (field `attachment_status`) to reach state `attached`.
-        If the `attachment_status` is missing, the client MUST retry the `/attach`
-        request (goto previous paragraph). This is a robustness measure in case the tenant
-        status endpoint is buggy, but the attach operation is ongoing.
+        The method to identify whether a request has arrived at the pageserver, and
+        whether it has succeeded, is to poll for the tenant status to reach "Active"
+        or "Broken" state. These values are currently not explicitly documented in
+        the API spec.
+        Polling for `has_in_progress_downloads == false` is INCORRECT because that
+        value can turn `false` during shutdown while the Attach operation is still
+        unfinished.
 
         There is no way to cancel an in-flight request.
 
-        In any case, the client
-        * MUST NOT ASSUME that the /attach request has been lost in the network,
-        * MUST NOT ASSUME that the request has been lost, based on the observation
-          that a subsequent tenant status request returns 404. The request may
-          still be in flight. It must be retried.
+        If a client receives a not-distinguished response, e.g., a network timeout,
+        it MUST retry the /attach request and poll again for tenant status.
 
-        The client SHOULD supply a `TenantConfig` for the tenant in the request body.
-        Settings specified in the config override the pageserver's defaults.
-        It is guaranteed that the config settings are applied before the pageserver
-        starts operating on the tenant. E.g., if the config specifies a specific
-        PITR interval for a tenant, then that setting will be in effect before the
-        pageserver starts the garbage collection loop. This enables a client to
-        guarantee a specific PITR setting across detach/attach cycles.
-        The pageserver will reject the request if it cannot parse the config, or
-        if there are any unknown fields in it.
-
-        If the client does not supply a config, the pageserver will use its defaults.
-        This behavior is deprecated: https://github.com/neondatabase/neon/issues/4282
-      requestBody:
-        required: false
-        content:
-          application/json:
-            schema:
-              $ref: "#/components/schemas/TenantAttachRequest"
+        In any case, it must
+        * NOT ASSUME that the /attach request has been lost in the network,
+        * NOT ASSUME that the request has been lost based on a subsequent
+          tenant status request returning 404. (The request may still be in flight!)
       responses:
         "202":
           description: Tenant attaching scheduled
         "400":
+          description: Error when no tenant id found in path parameters
           content:
             application/json:
               schema:
@@ -759,16 +741,13 @@ paths:
                 $ref: "#/components/schemas/Error"
     post:
       description: |
-        Create a tenant. Returns new tenant id on success.
-
+        Create a tenant. Returns new tenant id on success.\
         If no new tenant id is specified in parameters, it would be generated. It's an error to recreate the same tenant.
-
-        Invalid fields in the tenant config will cause the request to be rejected with status 400.
       requestBody:
         content:
           application/json:
             schema:
-              $ref: "#/components/schemas/TenantCreateRequest"
+              $ref: "#/components/schemas/TenantCreateInfo"
       responses:
         "201":
           description: New tenant created successfully
@@ -811,13 +790,11 @@ paths:
     put:
       description: |
         Update tenant's config.
-
-        Invalid fields in the tenant config will cause the request to be rejected with status 400.
       requestBody:
         content:
           application/json:
             schema:
-              $ref: "#/components/schemas/TenantConfigRequest"
+              $ref: "#/components/schemas/TenantConfigInfo"
       responses:
         "200":
           description: OK
@@ -869,7 +846,7 @@ paths:
           content:
             application/json:
               schema:
-                $ref: "#/components/schemas/TenantConfigResponse"
+                $ref: "#/components/schemas/TenantConfig"
         "400":
           description: Malformed get tenanant config request
           content:
@@ -911,55 +888,42 @@ components:
       type: object
       required:
         - id
-        - attachment_status
       properties:
         id:
           type: string
         current_physical_size:
           type: integer
-        attachment_status:
-          description: |
-            Status of this tenant's attachment to this pageserver.
-
-            - `maybe` means almost nothing, don't read anything into it
-              except for the fact that the pageserver _might_ be already
-              writing to the tenant's S3 state, so, DO NOT ATTACH the
-              tenant to any other pageserver, or we risk split-brain.
-            - `attached` means that the attach operation has completed,
-              maybe successfully, maybe not. Perform a health check at
-              the Postgres level to determine healthiness of the tenant.
-
-            See the tenant `/attach` endpoint for more information.
+        has_in_progress_downloads:
+          type: boolean
+    TenantCreateInfo:
+      type: object
+      properties:
+        new_tenant_id:
           type: string
-          enum: [ "maybe", "attached" ]
-    TenantCreateRequest:
-      allOf:
-        - $ref: '#/components/schemas/TenantConfig'
-        - type: object
-          properties:
-            new_tenant_id:
-              type: string
-              format: hex
-    TenantAttachRequest:
-      type: object
-      required:
-        - config
-      properties:
-        config:
-          $ref: '#/components/schemas/TenantConfig'
-    TenantConfigRequest:
-      allOf:
-        - $ref: '#/components/schemas/TenantConfig'
-        - type: object
-          required:
-            - tenant_id
-          properties:
-            tenant_id:
-              type: string
-              format: hex
-    TenantConfig:
+          format: hex
+        tenant_id:
+          type: string
+          format: hex
+        gc_period:
+          type: string
+        gc_horizon:
+          type: integer
+        pitr_interval:
+          type: string
+        checkpoint_distance:
+          type: integer
+        checkpoint_timeout:
+          type: string
+        compaction_period:
+          type: string
+        compaction_threshold:
+          type: string
+    TenantConfigInfo:
       type: object
       properties:
+        tenant_id:
+          type: string
+          format: hex
         gc_period:
           type: string
         gc_horizon:
@@ -986,13 +950,13 @@ components:
           type: integer
         trace_read_requests:
           type: boolean
-    TenantConfigResponse:
+    TenantConfig:
       type: object
       properties:
         tenant_specific_overrides:
-          $ref: "#/components/schemas/TenantConfig"
+          $ref: "#/components/schemas/TenantConfigInfo"
         effective_config:
-          $ref: "#/components/schemas/TenantConfig"
+          $ref: "#/components/schemas/TenantConfigInfo"
     TimelineInfo:
       type: object
       required:

pageserver/src/http/routes.rs

@@ -5,14 +5,12 @@ use anyhow::{anyhow, Context, Result};
 use hyper::StatusCode;
 use hyper::{Body, Request, Response, Uri};
 use metrics::launch_timestamp::LaunchTimestamp;
-use pageserver_api::models::{DownloadRemoteLayersTaskSpawnRequest, TenantAttachRequest};
+use pageserver_api::models::DownloadRemoteLayersTaskSpawnRequest;
 use remote_storage::GenericRemoteStorage;
-use storage_broker::BrokerClientChannel;
 use tenant_size_model::{SizeResult, StorageModel};
 use tokio_util::sync::CancellationToken;
 use tracing::*;
 use utils::http::endpoint::RequestSpan;
-use utils::http::json::json_request_or_empty_body;
 use utils::http::request::{get_request_param, must_get_query_param, parse_query_param};
 
 use super::models::{
@@ -21,16 +19,13 @@ use super::models::{
 };
 use crate::context::{DownloadBehavior, RequestContext};
 use crate::disk_usage_eviction_task;
-use crate::metrics::{StorageTimeOperation, STORAGE_TIME_GLOBAL};
 use crate::pgdatadir_mapping::LsnForTimestamp;
 use crate::task_mgr::TaskKind;
 use crate::tenant::config::TenantConfOpt;
-use crate::tenant::mgr::{
-    GetTenantError, SetNewTenantConfigError, TenantMapInsertError, TenantStateError,
-};
+use crate::tenant::mgr::{TenantMapInsertError, TenantStateError};
 use crate::tenant::size::ModelInputs;
 use crate::tenant::storage_layer::LayerAccessStatsReset;
-use crate::tenant::{LogicalSizeCalculationCause, PageReconstructError, Timeline};
+use crate::tenant::{PageReconstructError, Timeline};
 use crate::{config::PageServerConf, tenant::mgr};
 use utils::{
     auth::JwtAuth,
@@ -54,7 +49,6 @@ struct State {
     auth: Option<Arc<JwtAuth>>,
     allowlist_routes: Vec<Uri>,
     remote_storage: Option<GenericRemoteStorage>,
-    broker_client: storage_broker::BrokerClientChannel,
     disk_usage_eviction_state: Arc<disk_usage_eviction_task::State>,
 }
 
@@ -63,7 +57,6 @@ impl State {
         conf: &'static PageServerConf,
         auth: Option<Arc<JwtAuth>>,
         remote_storage: Option<GenericRemoteStorage>,
-        broker_client: storage_broker::BrokerClientChannel,
         disk_usage_eviction_state: Arc<disk_usage_eviction_task::State>,
     ) -> anyhow::Result<Self> {
         let allowlist_routes = ["/v1/status", "/v1/doc", "/swagger.yml"]
@@ -75,7 +68,6 @@ impl State {
             auth,
             allowlist_routes,
             remote_storage,
-            broker_client,
             disk_usage_eviction_state,
         })
     }
@@ -113,9 +105,6 @@ impl From<PageReconstructError> for ApiError {
             PageReconstructError::Cancelled => {
                 ApiError::InternalServerError(anyhow::anyhow!("request was cancelled"))
             }
-            PageReconstructError::AncestorStopping(_) => {
-                ApiError::InternalServerError(anyhow::Error::new(pre))
-            }
             PageReconstructError::WalRedo(pre) => {
                 ApiError::InternalServerError(anyhow::Error::new(pre))
             }
@@ -146,36 +135,6 @@ impl From<TenantStateError> for ApiError {
     }
 }
 
-impl From<GetTenantError> for ApiError {
-    fn from(tse: GetTenantError) -> ApiError {
-        match tse {
-            GetTenantError::NotFound(tid) => ApiError::NotFound(anyhow!("tenant {}", tid)),
-            e @ GetTenantError::NotActive(_) => {
-                // Why is this not `ApiError::NotFound`?
-                // Because we must be careful to never return 404 for a tenant if it does
-                // in fact exist locally. If we did, the caller could draw the conclusion
-                // that it can attach the tenant to another PS and we'd be in split-brain.
-                //
-                // (We can produce this variant only in `mgr::get_tenant(..., active=true)` calls).
-                ApiError::InternalServerError(anyhow::Error::new(e))
-            }
-        }
-    }
-}
-
-impl From<SetNewTenantConfigError> for ApiError {
-    fn from(e: SetNewTenantConfigError) -> ApiError {
-        match e {
-            SetNewTenantConfigError::GetTenant(tid) => {
-                ApiError::NotFound(anyhow!("tenant {}", tid))
-            }
-            e @ SetNewTenantConfigError::Persist(_) => {
-                ApiError::InternalServerError(anyhow::Error::new(e))
-            }
-        }
-    }
-}
-
 impl From<crate::tenant::DeleteTimelineError> for ApiError {
     fn from(value: crate::tenant::DeleteTimelineError) -> Self {
         use crate::tenant::DeleteTimelineError::*;
@@ -195,7 +154,7 @@ impl From<crate::tenant::mgr::DeleteTimelineError> for ApiError {
         match value {
             // Report Precondition failed so client can distinguish between
             // "tenant is missing" case from "timeline is missing"
-            Tenant(GetTenantError::NotFound(..)) => {
+            Tenant(TenantStateError::NotFound(..)) => {
                 ApiError::PreconditionFailed("Requested tenant is missing")
             }
             Tenant(t) => ApiError::from(t),
@@ -210,8 +169,6 @@ async fn build_timeline_info(
     include_non_incremental_logical_size: bool,
     ctx: &RequestContext,
 ) -> anyhow::Result<TimelineInfo> {
-    crate::tenant::debug_assert_current_span_has_tenant_and_timeline_id();
-
     let mut info = build_timeline_info_common(timeline, ctx)?;
     if include_non_incremental_logical_size {
         // XXX we should be using spawn_ondemand_logical_size_calculation here.
@@ -234,7 +191,6 @@ fn build_timeline_info_common(
     timeline: &Arc<Timeline>,
     ctx: &RequestContext,
 ) -> anyhow::Result<TimelineInfo> {
-    crate::tenant::debug_assert_current_span_has_tenant_and_timeline_id();
     let last_record_lsn = timeline.get_last_record_lsn();
     let (wal_source_connstr, last_received_msg_lsn, last_received_msg_ts) = {
         let guard = timeline.last_received_wal.lock().unwrap();
@@ -307,31 +263,25 @@ async fn timeline_create_handler(mut request: Request<Body>) -> Result<Response<
 
     let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Error);
 
-    let state = get_state(&request);
-
-    async {
-        let tenant = mgr::get_tenant(tenant_id, true).await?;
-        match tenant.create_timeline(
-            new_timeline_id,
-            request_data.ancestor_timeline_id.map(TimelineId::from),
-            request_data.ancestor_start_lsn,
-            request_data.pg_version.unwrap_or(crate::DEFAULT_PG_VERSION),
-            state.broker_client.clone(),
-            &ctx,
-        )
-        .await {
-            Ok(Some(new_timeline)) => {
-                // Created. Construct a TimelineInfo for it.
-                let timeline_info = build_timeline_info_common(&new_timeline, &ctx)
-                    .map_err(ApiError::InternalServerError)?;
-                json_response(StatusCode::CREATED, timeline_info)
-            }
-            Ok(None) => json_response(StatusCode::CONFLICT, ()), // timeline already exists
-            Err(err) => Err(ApiError::InternalServerError(err)),
-        }
-    }
+    let tenant = mgr::get_tenant(tenant_id, true).await?;
+    match tenant.create_timeline(
+        new_timeline_id,
+        request_data.ancestor_timeline_id.map(TimelineId::from),
+        request_data.ancestor_start_lsn,
+        request_data.pg_version.unwrap_or(crate::DEFAULT_PG_VERSION),
+        &ctx,
+    )
     .instrument(info_span!("timeline_create", tenant = %tenant_id, new_timeline = ?request_data.new_timeline_id, timeline_id = %new_timeline_id, lsn=?request_data.ancestor_start_lsn, pg_version=?request_data.pg_version))
-    .await
+    .await {
+        Ok(Some(new_timeline)) => {
+            // Created. Construct a TimelineInfo for it.
+            let timeline_info = build_timeline_info_common(&new_timeline, &ctx)
+                .map_err(ApiError::InternalServerError)?;
+            json_response(StatusCode::CREATED, timeline_info)
+        }
+        Ok(None) => json_response(StatusCode::CONFLICT, ()), // timeline already exists
+        Err(err) => Err(ApiError::InternalServerError(err)),
+    }
 }
 
 async fn timeline_list_handler(request: Request<Body>) -> Result<Response<Body>, ApiError> {
@@ -353,7 +303,6 @@ async fn timeline_list_handler(request: Request<Body>) -> Result<Response<Body>,
                 include_non_incremental_logical_size.unwrap_or(false),
                 &ctx,
             )
-            .instrument(info_span!("build_timeline_info", timeline_id = %timeline.timeline_id))
             .await
             .context("Failed to convert tenant timeline {timeline_id} into the local one: {e:?}")
             .map_err(ApiError::InternalServerError)?;
@@ -426,16 +375,11 @@ async fn get_lsn_by_timestamp_handler(request: Request<Body>) -> Result<Response
     json_response(StatusCode::OK, result)
 }
 
-async fn tenant_attach_handler(mut request: Request<Body>) -> Result<Response<Body>, ApiError> {
+// TODO makes sense to provide tenant config right away the same way as it handled in tenant_create
+async fn tenant_attach_handler(request: Request<Body>) -> Result<Response<Body>, ApiError> {
     let tenant_id: TenantId = parse_request_param(&request, "tenant_id")?;
     check_permission(&request, Some(tenant_id))?;
 
-    let maybe_body: Option<TenantAttachRequest> = json_request_or_empty_body(&mut request).await?;
-    let tenant_conf = match maybe_body {
-        Some(request) => TenantConfOpt::try_from(&*request.config).map_err(ApiError::BadRequest)?,
-        None => TenantConfOpt::default(),
-    };
-
     let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Warn);
 
     info!("Handling tenant attach {tenant_id}");
@@ -443,16 +387,9 @@ async fn tenant_attach_handler(mut request: Request<Body>) -> Result<Response<Bo
     let state = get_state(&request);
 
     if let Some(remote_storage) = &state.remote_storage {
-        mgr::attach_tenant(
-            state.conf,
-            tenant_id,
-            tenant_conf,
-            state.broker_client.clone(),
-            remote_storage.clone(),
-            &ctx,
-        )
-        .instrument(info_span!("tenant_attach", tenant = %tenant_id))
-        .await?;
+        mgr::attach_tenant(state.conf, tenant_id, remote_storage.clone(), &ctx)
+            .instrument(info_span!("tenant_attach", tenant = %tenant_id))
+            .await?;
     } else {
         return Err(ApiError::BadRequest(anyhow!(
             "attach_tenant is not possible because pageserver was configured without remote storage"
@@ -497,15 +434,9 @@ async fn tenant_load_handler(request: Request<Body>) -> Result<Response<Body>, A
     let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Warn);
 
     let state = get_state(&request);
-    mgr::load_tenant(
-        state.conf,
-        tenant_id,
-        state.broker_client.clone(),
-        state.remote_storage.clone(),
-        &ctx,
-    )
-    .instrument(info_span!("load", tenant = %tenant_id))
-    .await?;
+    mgr::load_tenant(state.conf, tenant_id, state.remote_storage.clone(), &ctx)
+        .instrument(info_span!("load", tenant = %tenant_id))
+        .await?;
 
     json_response(StatusCode::ACCEPTED, ())
 }
@@ -536,7 +467,7 @@ async fn tenant_list_handler(request: Request<Body>) -> Result<Response<Body>, A
             id: *id,
             state: state.clone(),
             current_physical_size: None,
-            attachment_status: state.attachment_status(),
+            has_in_progress_downloads: Some(state.has_in_progress_downloads()),
         })
         .collect::<Vec<TenantInfo>>();
 
@@ -557,15 +488,16 @@ async fn tenant_status(request: Request<Body>) -> Result<Response<Body>, ApiErro
         }
 
         let state = tenant.current_state();
-        Result::<_, ApiError>::Ok(TenantInfo {
+        Ok(TenantInfo {
             id: tenant_id,
             state: state.clone(),
             current_physical_size: Some(current_physical_size),
-            attachment_status: state.attachment_status(),
+            has_in_progress_downloads: Some(state.has_in_progress_downloads()),
         })
     }
     .instrument(info_span!("tenant_status_handler", tenant = %tenant_id))
-    .await?;
+    .await
+    .map_err(ApiError::InternalServerError)?;
 
     json_response(StatusCode::OK, tenant_info)
 }
@@ -595,11 +527,7 @@ async fn tenant_size_handler(request: Request<Body>) -> Result<Response<Body>, A
 
     // this can be long operation
     let inputs = tenant
-        .gather_size_inputs(
-            retention_period,
-            LogicalSizeCalculationCause::TenantSizeHandler,
-            &ctx,
-        )
+        .gather_size_inputs(retention_period, &ctx)
         .await
         .map_err(ApiError::InternalServerError)?;
 
@@ -766,17 +694,11 @@ pub fn html_response(status: StatusCode, data: String) -> Result<Response<Body>,
 async fn tenant_create_handler(mut request: Request<Body>) -> Result<Response<Body>, ApiError> {
     check_permission(&request, None)?;
 
-    let _timer = STORAGE_TIME_GLOBAL
-        .get_metric_with_label_values(&[StorageTimeOperation::CreateTenant.into()])
-        .expect("bug")
-        .start_timer();
-
     let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Warn);
 
     let request_data: TenantCreateRequest = json_request(&mut request).await?;
 
-    let tenant_conf =
-        TenantConfOpt::try_from(&request_data.config).map_err(ApiError::BadRequest)?;
+    let tenant_conf = TenantConfOpt::try_from(&request_data).map_err(ApiError::BadRequest)?;
 
     let target_tenant_id = request_data
         .new_tenant_id
@@ -789,7 +711,6 @@ async fn tenant_create_handler(mut request: Request<Body>) -> Result<Response<Bo
         state.conf,
         tenant_conf,
         target_tenant_id,
-        state.broker_client.clone(),
         state.remote_storage.clone(),
         &ctx,
     )
@@ -808,7 +729,6 @@ async fn tenant_create_handler(mut request: Request<Body>) -> Result<Response<Bo
         res.context("created tenant failed to become active")
             .map_err(ApiError::InternalServerError)?;
     }
-
     json_response(
         StatusCode::CREATED,
         TenantCreateResponse(new_tenant.tenant_id()),
@@ -846,8 +766,7 @@ async fn update_tenant_config_handler(
     let tenant_id = request_data.tenant_id;
     check_permission(&request, Some(tenant_id))?;
 
-    let tenant_conf =
-        TenantConfOpt::try_from(&request_data.config).map_err(ApiError::BadRequest)?;
+    let tenant_conf = TenantConfOpt::try_from(&request_data).map_err(ApiError::BadRequest)?;
 
     let state = get_state(&request);
     mgr::set_new_tenant_config(state.conf, tenant_conf, tenant_id)
@@ -1145,7 +1064,6 @@ pub fn make_router(
     conf: &'static PageServerConf,
     launch_ts: &'static LaunchTimestamp,
     auth: Option<Arc<JwtAuth>>,
-    broker_client: BrokerClientChannel,
     remote_storage: Option<GenericRemoteStorage>,
     disk_usage_eviction_state: Arc<disk_usage_eviction_task::State>,
 ) -> anyhow::Result<RouterBuilder<hyper::Body, ApiError>> {
@@ -1192,14 +1110,8 @@ pub fn make_router(
 
     Ok(router
         .data(Arc::new(
-            State::new(
-                conf,
-                auth,
-                remote_storage,
-                broker_client,
-                disk_usage_eviction_state,
-            )
-            .context("Failed to initialize router state")?,
+            State::new(conf, auth, remote_storage, disk_usage_eviction_state)
+                .context("Failed to initialize router state")?,
         ))
         .get("/v1/status", |r| RequestSpan(status_handler).handle(r))
         .put(

pageserver/src/keyspace.rs

@@ -5,7 +5,7 @@ use std::ops::Range;
 ///
 /// Represents a set of Keys, in a compact form.
 ///
-#[derive(Clone, Debug, Default)]
+#[derive(Clone, Debug)]
 pub struct KeySpace {
     /// Contiguous ranges of keys that belong to the key space. In key order,
     /// and with no overlap.
@@ -61,18 +61,6 @@ impl KeySpace {
 
         KeyPartitioning { parts }
     }
-
-    ///
-    /// Check if key space contains overlapping range
-    ///
-    pub fn overlaps(&self, range: &Range<Key>) -> bool {
-        match self.ranges.binary_search_by_key(&range.end, |r| r.start) {
-            Ok(0) => false,
-            Err(0) => false,
-            Ok(index) => self.ranges[index - 1].end > range.start,
-            Err(index) => self.ranges[index - 1].end > range.start,
-        }
-    }
 }
 
 ///
@@ -141,226 +129,3 @@ impl KeySpaceAccum {
         }
     }
 }
-
-///
-/// A helper object, to collect a set of keys and key ranges into a KeySpace
-/// object. Key ranges may be inserted in any order and can overlap.
-///
-#[derive(Clone, Debug, Default)]
-pub struct KeySpaceRandomAccum {
-    ranges: Vec<Range<Key>>,
-}
-
-impl KeySpaceRandomAccum {
-    pub fn new() -> Self {
-        Self { ranges: Vec::new() }
-    }
-
-    pub fn add_key(&mut self, key: Key) {
-        self.add_range(singleton_range(key))
-    }
-
-    pub fn add_range(&mut self, range: Range<Key>) {
-        self.ranges.push(range);
-    }
-
-    pub fn to_keyspace(mut self) -> KeySpace {
-        let mut ranges = Vec::new();
-        if !self.ranges.is_empty() {
-            self.ranges.sort_by_key(|r| r.start);
-            let mut start = self.ranges.first().unwrap().start;
-            let mut end = self.ranges.first().unwrap().end;
-            for r in self.ranges {
-                assert!(r.start >= start);
-                if r.start > end {
-                    ranges.push(start..end);
-                    start = r.start;
-                    end = r.end;
-                } else if r.end > end {
-                    end = r.end;
-                }
-            }
-            ranges.push(start..end);
-        }
-        KeySpace { ranges }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use std::fmt::Write;
-
-    // Helper function to create a key range.
-    //
-    // Make the tests below less verbose.
-    fn kr(irange: Range<i128>) -> Range<Key> {
-        Key::from_i128(irange.start)..Key::from_i128(irange.end)
-    }
-
-    #[allow(dead_code)]
-    fn dump_keyspace(ks: &KeySpace) {
-        for r in ks.ranges.iter() {
-            println!("  {}..{}", r.start.to_i128(), r.end.to_i128());
-        }
-    }
-
-    fn assert_ks_eq(actual: &KeySpace, expected: Vec<Range<Key>>) {
-        if actual.ranges != expected {
-            let mut msg = String::new();
-
-            writeln!(msg, "expected:").unwrap();
-            for r in &expected {
-                writeln!(msg, "  {}..{}", r.start.to_i128(), r.end.to_i128()).unwrap();
-            }
-            writeln!(msg, "got:").unwrap();
-            for r in &actual.ranges {
-                writeln!(msg, "  {}..{}", r.start.to_i128(), r.end.to_i128()).unwrap();
-            }
-            panic!("{}", msg);
-        }
-    }
-
-    #[test]
-    fn keyspace_add_range() {
-        // two separate ranges
-        //
-        // #####
-        //         #####
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(0..10));
-        ks.add_range(kr(20..30));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..10), kr(20..30)]);
-
-        // two separate ranges, added in reverse order
-        //
-        //         #####
-        // #####
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(20..30));
-        ks.add_range(kr(0..10));
-
-        // add range that is adjacent to the end of an existing range
-        //
-        // #####
-        //      #####
-        ks.add_range(kr(0..10));
-        ks.add_range(kr(10..30));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-
-        // add range that is adjacent to the start of an existing range
-        //
-        //      #####
-        // #####
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(10..30));
-        ks.add_range(kr(0..10));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-
-        // add range that overlaps with the end of an existing range
-        //
-        // #####
-        //    #####
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(0..10));
-        ks.add_range(kr(5..30));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-
-        // add range that overlaps with the start of an existing range
-        //
-        //    #####
-        // #####
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(5..30));
-        ks.add_range(kr(0..10));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-
-        // add range that is fully covered by an existing range
-        //
-        // #########
-        //   #####
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(0..30));
-        ks.add_range(kr(10..20));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-
-        // add range that extends an existing range from both ends
-        //
-        //   #####
-        // #########
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(10..20));
-        ks.add_range(kr(0..30));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-
-        // add a range that overlaps with two existing ranges, joining them
-        //
-        // #####   #####
-        //    #######
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(0..10));
-        ks.add_range(kr(20..30));
-        ks.add_range(kr(5..25));
-        assert_ks_eq(&ks.to_keyspace(), vec![kr(0..30)]);
-    }
-
-    #[test]
-    fn keyspace_overlaps() {
-        let mut ks = KeySpaceRandomAccum::default();
-        ks.add_range(kr(10..20));
-        ks.add_range(kr(30..40));
-        let ks = ks.to_keyspace();
-
-        //        #####      #####
-        // xxxx
-        assert!(!ks.overlaps(&kr(0..5)));
-
-        //        #####      #####
-        //   xxxx
-        assert!(!ks.overlaps(&kr(5..9)));
-
-        //        #####      #####
-        //    xxxx
-        assert!(!ks.overlaps(&kr(5..10)));
-
-        //        #####      #####
-        //     xxxx
-        assert!(ks.overlaps(&kr(5..11)));
-
-        //        #####      #####
-        //        xxxx
-        assert!(ks.overlaps(&kr(10..15)));
-
-        //        #####      #####
-        //         xxxx
-        assert!(ks.overlaps(&kr(15..20)));
-
-        //        #####      #####
-        //           xxxx
-        assert!(ks.overlaps(&kr(15..25)));
-
-        //        #####      #####
-        //              xxxx
-        assert!(!ks.overlaps(&kr(22..28)));
-
-        //        #####      #####
-        //               xxxx
-        assert!(!ks.overlaps(&kr(25..30)));
-
-        //        #####      #####
-        //                      xxxx
-        assert!(ks.overlaps(&kr(35..35)));
-
-        //        #####      #####
-        //                        xxxx
-        assert!(!ks.overlaps(&kr(40..45)));
-
-        //        #####      #####
-        //                        xxxx
-        assert!(!ks.overlaps(&kr(45..50)));
-
-        //        #####      #####
-        //        xxxxxxxxxxx
-        assert!(ks.overlaps(&kr(0..30))); // XXXXX This fails currently!
-    }
-}

pageserver/src/lib.rs

@@ -1,5 +1,6 @@
 mod auth;
 pub mod basebackup;
+pub mod broker_client;
 pub mod config;
 pub mod consumption_metrics;
 pub mod context;
@@ -35,7 +36,7 @@ use tracing::info;
 /// backwards-compatible changes to the metadata format.
 pub const STORAGE_FORMAT_VERSION: u16 = 3;
 
-pub const DEFAULT_PG_VERSION: u32 = 15;
+pub const DEFAULT_PG_VERSION: u32 = 14;
 
 // Magic constants used to identify different kinds of files
 pub const IMAGE_FILE_MAGIC: u16 = 0x5A60;

pageserver/src/metrics.rs

@@ -8,7 +8,6 @@ use metrics::{
 use once_cell::sync::Lazy;
 use pageserver_api::models::TenantState;
 use strum::VariantNames;
-use strum_macros::{EnumVariantNames, IntoStaticStr};
 use utils::id::{TenantId, TimelineId};
 
 /// Prometheus histogram buckets (in seconds) for operations in the critical
@@ -25,33 +24,15 @@ const CRITICAL_OP_BUCKETS: &[f64] = &[
 ];
 
 // Metrics collected on operations on the storage repository.
-#[derive(Debug, EnumVariantNames, IntoStaticStr)]
-#[strum(serialize_all = "kebab_case")]
-pub enum StorageTimeOperation {
-    #[strum(serialize = "layer flush")]
-    LayerFlush,
-
-    #[strum(serialize = "compact")]
-    Compact,
-
-    #[strum(serialize = "create images")]
-    CreateImages,
-
-    #[strum(serialize = "logical size")]
-    LogicalSize,
-
-    #[strum(serialize = "imitate logical size")]
-    ImitateLogicalSize,
-
-    #[strum(serialize = "load layer map")]
-    LoadLayerMap,
-
-    #[strum(serialize = "gc")]
-    Gc,
-
-    #[strum(serialize = "create tenant")]
-    CreateTenant,
-}
+const STORAGE_TIME_OPERATIONS: &[&str] = &[
+    "layer flush",
+    "compact",
+    "create images",
+    "init logical size",
+    "logical size",
+    "load layer map",
+    "gc",
+];
 
 pub static STORAGE_TIME_SUM_PER_TIMELINE: Lazy<CounterVec> = Lazy::new(|| {
     register_counter_vec!(
@@ -205,16 +186,6 @@ static PERSISTENT_BYTES_WRITTEN: Lazy<IntCounterVec> = Lazy::new(|| {
     .expect("failed to define a metric")
 });
 
-pub(crate) static EVICTION_ITERATION_DURATION: Lazy<HistogramVec> = Lazy::new(|| {
-    register_histogram_vec!(
-        "pageserver_eviction_iteration_duration_seconds_global",
-        "Time spent on a single eviction iteration",
-        &["period_secs", "threshold_secs"],
-        STORAGE_OP_BUCKETS.into(),
-    )
-    .expect("failed to define a metric")
-});
-
 static EVICTIONS: Lazy<IntCounterVec> = Lazy::new(|| {
     register_int_counter_vec!(
         "pageserver_evictions",
@@ -507,15 +478,6 @@ pub static TENANT_TASK_EVENTS: Lazy<IntCounterVec> = Lazy::new(|| {
     .expect("Failed to register tenant_task_events metric")
 });
 
-pub static BACKGROUND_LOOP_PERIOD_OVERRUN_COUNT: Lazy<IntCounterVec> = Lazy::new(|| {
-    register_int_counter_vec!(
-        "pageserver_background_loop_period_overrun_count",
-        "Incremented whenever warn_when_period_overrun() logs a warning.",
-        &["task", "period"],
-    )
-    .expect("failed to define a metric")
-});
-
 // walreceiver metrics
 
 pub static WALRECEIVER_STARTED_CONNECTIONS: Lazy<IntCounter> = Lazy::new(|| {
@@ -690,9 +652,7 @@ pub struct StorageTimeMetrics {
 }
 
 impl StorageTimeMetrics {
-    pub fn new(operation: StorageTimeOperation, tenant_id: &str, timeline_id: &str) -> Self {
-        let operation: &'static str = operation.into();
-
+    pub fn new(operation: &str, tenant_id: &str, timeline_id: &str) -> Self {
         let timeline_sum = STORAGE_TIME_SUM_PER_TIMELINE
             .get_metric_with_label_values(&[operation, tenant_id, timeline_id])
             .unwrap();
@@ -728,7 +688,6 @@ pub struct TimelineMetrics {
     pub compact_time_histo: StorageTimeMetrics,
     pub create_images_time_histo: StorageTimeMetrics,
     pub logical_size_histo: StorageTimeMetrics,
-    pub imitate_logical_size_histo: StorageTimeMetrics,
     pub load_layer_map_histo: StorageTimeMetrics,
     pub garbage_collect_histo: StorageTimeMetrics,
     pub last_record_gauge: IntGauge,
@@ -756,23 +715,14 @@ impl TimelineMetrics {
         let materialized_page_cache_hit_counter = MATERIALIZED_PAGE_CACHE_HIT
             .get_metric_with_label_values(&[&tenant_id, &timeline_id])
             .unwrap();
-        let flush_time_histo =
-            StorageTimeMetrics::new(StorageTimeOperation::LayerFlush, &tenant_id, &timeline_id);
-        let compact_time_histo =
-            StorageTimeMetrics::new(StorageTimeOperation::Compact, &tenant_id, &timeline_id);
+        let flush_time_histo = StorageTimeMetrics::new("layer flush", &tenant_id, &timeline_id);
+        let compact_time_histo = StorageTimeMetrics::new("compact", &tenant_id, &timeline_id);
         let create_images_time_histo =
-            StorageTimeMetrics::new(StorageTimeOperation::CreateImages, &tenant_id, &timeline_id);
-        let logical_size_histo =
-            StorageTimeMetrics::new(StorageTimeOperation::LogicalSize, &tenant_id, &timeline_id);
-        let imitate_logical_size_histo = StorageTimeMetrics::new(
-            StorageTimeOperation::ImitateLogicalSize,
-            &tenant_id,
-            &timeline_id,
-        );
+            StorageTimeMetrics::new("create images", &tenant_id, &timeline_id);
+        let logical_size_histo = StorageTimeMetrics::new("logical size", &tenant_id, &timeline_id);
         let load_layer_map_histo =
-            StorageTimeMetrics::new(StorageTimeOperation::LoadLayerMap, &tenant_id, &timeline_id);
-        let garbage_collect_histo =
-            StorageTimeMetrics::new(StorageTimeOperation::Gc, &tenant_id, &timeline_id);
+            StorageTimeMetrics::new("load layer map", &tenant_id, &timeline_id);
+        let garbage_collect_histo = StorageTimeMetrics::new("gc", &tenant_id, &timeline_id);
         let last_record_gauge = LAST_RECORD_LSN
             .get_metric_with_label_values(&[&tenant_id, &timeline_id])
             .unwrap();
@@ -806,7 +756,6 @@ impl TimelineMetrics {
             compact_time_histo,
             create_images_time_histo,
             logical_size_histo,
-            imitate_logical_size_histo,
             garbage_collect_histo,
             load_layer_map_histo,
             last_record_gauge,
@@ -840,7 +789,7 @@ impl Drop for TimelineMetrics {
             .write()
             .unwrap()
             .remove(tenant_id, timeline_id);
-        for op in StorageTimeOperation::VARIANTS {
+        for op in STORAGE_TIME_OPERATIONS {
             let _ =
                 STORAGE_TIME_SUM_PER_TIMELINE.remove_label_values(&[op, tenant_id, timeline_id]);
             let _ =
@@ -1267,7 +1216,4 @@ pub fn preinitialize_metrics() {
     // Initialize it eagerly, so that our alert rule can distinguish absence of the metric from metric value 0.
     assert_eq!(UNEXPECTED_ONDEMAND_DOWNLOADS.get(), 0);
     UNEXPECTED_ONDEMAND_DOWNLOADS.reset();
-
-    // Same as above for this metric, but, it's a Vec-type metric for which we don't know all the labels.
-    BACKGROUND_LOOP_PERIOD_OVERRUN_COUNT.reset();
 }

pageserver/src/page_service.rs

@@ -50,9 +50,7 @@ use crate::import_datadir::import_wal_from_tar;
 use crate::metrics::{LIVE_CONNECTIONS_COUNT, SMGR_QUERY_TIME};
 use crate::task_mgr;
 use crate::task_mgr::TaskKind;
-use crate::tenant;
 use crate::tenant::mgr;
-use crate::tenant::mgr::GetTenantError;
 use crate::tenant::{Tenant, Timeline};
 use crate::trace::Tracer;
 
@@ -174,7 +172,6 @@ async fn read_tar_eof(mut reader: (impl AsyncRead + Unpin)) -> anyhow::Result<()
 ///
 pub async fn libpq_listener_main(
     conf: &'static PageServerConf,
-    broker_client: storage_broker::BrokerClientChannel,
     auth: Option<Arc<JwtAuth>>,
     listener: TcpListener,
     auth_type: AuthType,
@@ -216,14 +213,7 @@ pub async fn libpq_listener_main(
                     None,
                     "serving compute connection task",
                     false,
-                    page_service_conn_main(
-                        conf,
-                        broker_client.clone(),
-                        local_auth,
-                        socket,
-                        auth_type,
-                        connection_ctx,
-                    ),
+                    page_service_conn_main(conf, local_auth, socket, auth_type, connection_ctx),
                 );
             }
             Err(err) => {
@@ -240,7 +230,6 @@ pub async fn libpq_listener_main(
 
 async fn page_service_conn_main(
     conf: &'static PageServerConf,
-    broker_client: storage_broker::BrokerClientChannel,
     auth: Option<Arc<JwtAuth>>,
     socket: tokio::net::TcpStream,
     auth_type: AuthType,
@@ -267,17 +256,14 @@ async fn page_service_conn_main(
     //
     // no write timeout is used, because the kernel is assumed to error writes after some time.
     let mut socket = tokio_io_timeout::TimeoutReader::new(socket);
-
-    // timeout should be lower, but trying out multiple days for
-    // <https://github.com/neondatabase/neon/issues/4205>
-    socket.set_timeout(Some(std::time::Duration::from_secs(60 * 60 * 24 * 3)));
+    socket.set_timeout(Some(std::time::Duration::from_secs(60 * 10)));
     let socket = std::pin::pin!(socket);
 
     // XXX: pgbackend.run() should take the connection_ctx,
     // and create a child per-query context when it invokes process_query.
     // But it's in a shared crate, so, we store connection_ctx inside PageServerHandler
     // and create the per-query context in process_query ourselves.
-    let mut conn_handler = PageServerHandler::new(conf, broker_client, auth, connection_ctx);
+    let mut conn_handler = PageServerHandler::new(conf, auth, connection_ctx);
     let pgbackend = PostgresBackend::new_from_io(socket, peer_addr, auth_type, None)?;
 
     match pgbackend
@@ -335,7 +321,6 @@ impl PageRequestMetrics {
 
 struct PageServerHandler {
     _conf: &'static PageServerConf,
-    broker_client: storage_broker::BrokerClientChannel,
     auth: Option<Arc<JwtAuth>>,
     claims: Option<Claims>,
 
@@ -349,13 +334,11 @@ struct PageServerHandler {
 impl PageServerHandler {
     pub fn new(
         conf: &'static PageServerConf,
-        broker_client: storage_broker::BrokerClientChannel,
         auth: Option<Arc<JwtAuth>>,
         connection_ctx: RequestContext,
     ) -> Self {
         PageServerHandler {
             _conf: conf,
-            broker_client,
             auth,
             claims: None,
             connection_ctx,
@@ -508,12 +491,7 @@ impl PageServerHandler {
 
         let mut copyin_reader = pin!(StreamReader::new(copyin_stream(pgb)));
         timeline
-            .import_basebackup_from_tar(
-                &mut copyin_reader,
-                base_lsn,
-                self.broker_client.clone(),
-                &ctx,
-            )
+            .import_basebackup_from_tar(&mut copyin_reader, base_lsn, &ctx)
             .await?;
 
         // Read the end of the tar archive.
@@ -1150,9 +1128,7 @@ enum GetActiveTenantError {
         wait_time: Duration,
     },
     #[error(transparent)]
-    NotFound(GetTenantError),
-    #[error(transparent)]
-    WaitTenantActive(tenant::WaitToBecomeActiveError),
+    Other(#[from] anyhow::Error),
 }
 
 impl From<GetActiveTenantError> for QueryError {
@@ -1161,8 +1137,7 @@ impl From<GetActiveTenantError> for QueryError {
             GetActiveTenantError::WaitForActiveTimeout { .. } => QueryError::Disconnected(
                 ConnectionError::Io(io::Error::new(io::ErrorKind::TimedOut, e.to_string())),
             ),
-            GetActiveTenantError::WaitTenantActive(e) => QueryError::Other(anyhow::Error::new(e)),
-            GetActiveTenantError::NotFound(e) => QueryError::Other(anyhow::Error::new(e)),
+            GetActiveTenantError::Other(e) => QueryError::Other(e),
         }
     }
 }
@@ -1178,16 +1153,13 @@ async fn get_active_tenant_with_timeout(
 ) -> Result<Arc<Tenant>, GetActiveTenantError> {
     let tenant = match mgr::get_tenant(tenant_id, false).await {
         Ok(tenant) => tenant,
-        Err(e @ GetTenantError::NotFound(_)) => return Err(GetActiveTenantError::NotFound(e)),
-        Err(GetTenantError::NotActive(_)) => {
-            unreachable!("we're calling get_tenant with active=false")
-        }
+        Err(e) => return Err(GetActiveTenantError::Other(e.into())),
     };
     let wait_time = Duration::from_secs(30);
     match tokio::time::timeout(wait_time, tenant.wait_to_become_active()).await {
         Ok(Ok(())) => Ok(tenant),
         // no .context(), the error message is good enough and some tests depend on it
-        Ok(Err(e)) => Err(GetActiveTenantError::WaitTenantActive(e)),
+        Ok(Err(wait_error)) => Err(GetActiveTenantError::Other(wait_error)),
         Err(_) => {
             let latest_state = tenant.current_state();
             if latest_state == TenantState::Active {
@@ -1202,34 +1174,13 @@ async fn get_active_tenant_with_timeout(
     }
 }
 
-#[derive(Debug, thiserror::Error)]
-enum GetActiveTimelineError {
-    #[error(transparent)]
-    Tenant(GetActiveTenantError),
-    #[error(transparent)]
-    Timeline(anyhow::Error),
-}
-
-impl From<GetActiveTimelineError> for QueryError {
-    fn from(e: GetActiveTimelineError) -> Self {
-        match e {
-            GetActiveTimelineError::Tenant(e) => e.into(),
-            GetActiveTimelineError::Timeline(e) => QueryError::Other(e),
-        }
-    }
-}
-
 /// Shorthand for getting a reference to a Timeline of an Active tenant.
 async fn get_active_tenant_timeline(
     tenant_id: TenantId,
     timeline_id: TimelineId,
     ctx: &RequestContext,
-) -> Result<Arc<Timeline>, GetActiveTimelineError> {
-    let tenant = get_active_tenant_with_timeout(tenant_id, ctx)
-        .await
-        .map_err(GetActiveTimelineError::Tenant)?;
-    let timeline = tenant
-        .get_timeline(timeline_id, true)
-        .map_err(GetActiveTimelineError::Timeline)?;
+) -> Result<Arc<Timeline>, GetActiveTenantError> {
+    let tenant = get_active_tenant_with_timeout(tenant_id, ctx).await?;
+    let timeline = tenant.get_timeline(timeline_id, true)?;
     Ok(timeline)
 }

pageserver/src/pgdatadir_mapping.rs

@@ -500,8 +500,6 @@ impl Timeline {
         cancel: CancellationToken,
         ctx: &RequestContext,
     ) -> Result<u64, CalculateLogicalSizeError> {
-        crate::tenant::debug_assert_current_span_has_tenant_and_timeline_id();
-
         // Fetch list of database dirs and iterate them
         let buf = self.get(DBDIR_KEY, lsn, ctx).await.context("read dbdir")?;
         let dbdir = DbDirectory::des(&buf).context("deserialize db directory")?;
@@ -1600,7 +1598,9 @@ pub fn create_test_timeline(
     pg_version: u32,
     ctx: &RequestContext,
 ) -> anyhow::Result<std::sync::Arc<Timeline>> {
-    let tline = tenant.create_test_timeline(timeline_id, Lsn(8), pg_version, ctx)?;
+    let tline = tenant
+        .create_empty_timeline(timeline_id, Lsn(8), pg_version, ctx)?
+        .initialize(ctx)?;
     let mut m = tline.begin_modification(Lsn(8));
     m.init_empty()?;
     m.commit()?;

pageserver/src/tenant/config.rs

@@ -9,7 +9,7 @@
 //! may lead to a data loss.
 //!
 use anyhow::Context;
-use pageserver_api::models;
+use pageserver_api::models::{TenantConfigRequest, TenantCreateRequest};
 use serde::{Deserialize, Serialize};
 use std::num::NonZeroU64;
 use std::time::Duration;
@@ -292,10 +292,10 @@ fn bad_duration<'a>(field_name: &'static str, value: &'a str) -> impl 'a + Fn()
     move || format!("Cannot parse `{field_name}` duration {value:?}")
 }
 
-impl TryFrom<&'_ models::TenantConfig> for TenantConfOpt {
+impl TryFrom<&'_ TenantCreateRequest> for TenantConfOpt {
     type Error = anyhow::Error;
 
-    fn try_from(request_data: &'_ models::TenantConfig) -> Result<Self, Self::Error> {
+    fn try_from(request_data: &TenantCreateRequest) -> Result<Self, Self::Error> {
         let mut tenant_conf = TenantConfOpt::default();
 
         if let Some(gc_period) = &request_data.gc_period {
@@ -377,6 +377,84 @@ impl TryFrom<&'_ models::TenantConfig> for TenantConfOpt {
     }
 }
 
+impl TryFrom<&'_ TenantConfigRequest> for TenantConfOpt {
+    type Error = anyhow::Error;
+
+    fn try_from(request_data: &TenantConfigRequest) -> Result<Self, Self::Error> {
+        let mut tenant_conf = TenantConfOpt::default();
+        if let Some(gc_period) = &request_data.gc_period {
+            tenant_conf.gc_period = Some(
+                humantime::parse_duration(gc_period)
+                    .with_context(bad_duration("gc_period", gc_period))?,
+            );
+        }
+        tenant_conf.gc_horizon = request_data.gc_horizon;
+        tenant_conf.image_creation_threshold = request_data.image_creation_threshold;
+
+        if let Some(pitr_interval) = &request_data.pitr_interval {
+            tenant_conf.pitr_interval = Some(
+                humantime::parse_duration(pitr_interval)
+                    .with_context(bad_duration("pitr_interval", pitr_interval))?,
+            );
+        }
+        if let Some(walreceiver_connect_timeout) = &request_data.walreceiver_connect_timeout {
+            tenant_conf.walreceiver_connect_timeout = Some(
+                humantime::parse_duration(walreceiver_connect_timeout).with_context(
+                    bad_duration("walreceiver_connect_timeout", walreceiver_connect_timeout),
+                )?,
+            );
+        }
+        if let Some(lagging_wal_timeout) = &request_data.lagging_wal_timeout {
+            tenant_conf.lagging_wal_timeout = Some(
+                humantime::parse_duration(lagging_wal_timeout)
+                    .with_context(bad_duration("lagging_wal_timeout", lagging_wal_timeout))?,
+            );
+        }
+        tenant_conf.max_lsn_wal_lag = request_data.max_lsn_wal_lag;
+        tenant_conf.trace_read_requests = request_data.trace_read_requests;
+
+        tenant_conf.checkpoint_distance = request_data.checkpoint_distance;
+        if let Some(checkpoint_timeout) = &request_data.checkpoint_timeout {
+            tenant_conf.checkpoint_timeout = Some(
+                humantime::parse_duration(checkpoint_timeout)
+                    .with_context(bad_duration("checkpoint_timeout", checkpoint_timeout))?,
+            );
+        }
+        tenant_conf.compaction_target_size = request_data.compaction_target_size;
+        tenant_conf.compaction_threshold = request_data.compaction_threshold;
+
+        if let Some(compaction_period) = &request_data.compaction_period {
+            tenant_conf.compaction_period = Some(
+                humantime::parse_duration(compaction_period)
+                    .with_context(bad_duration("compaction_period", compaction_period))?,
+            );
+        }
+
+        if let Some(eviction_policy) = &request_data.eviction_policy {
+            tenant_conf.eviction_policy = Some(
+                serde::Deserialize::deserialize(eviction_policy)
+                    .context("parse field `eviction_policy`")?,
+            );
+        }
+
+        tenant_conf.min_resident_size_override = request_data.min_resident_size_override;
+
+        if let Some(evictions_low_residence_duration_metric_threshold) =
+            &request_data.evictions_low_residence_duration_metric_threshold
+        {
+            tenant_conf.evictions_low_residence_duration_metric_threshold = Some(
+                humantime::parse_duration(evictions_low_residence_duration_metric_threshold)
+                    .with_context(bad_duration(
+                        "evictions_low_residence_duration_metric_threshold",
+                        evictions_low_residence_duration_metric_threshold,
+                    ))?,
+            );
+        }
+
+        Ok(tenant_conf)
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;

pageserver/src/tenant/mgr.rs

@@ -19,7 +19,7 @@ use crate::config::PageServerConf;
 use crate::context::{DownloadBehavior, RequestContext};
 use crate::task_mgr::{self, TaskKind};
 use crate::tenant::config::TenantConfOpt;
-use crate::tenant::{create_tenant_files, CreateTenantFilesMode, Tenant, TenantState};
+use crate::tenant::{Tenant, TenantState};
 use crate::IGNORED_TENANT_FILE_NAME;
 
 use utils::fs_ext::PathExt;
@@ -58,10 +58,9 @@ static TENANTS: Lazy<RwLock<TenantsMap>> = Lazy::new(|| RwLock::new(TenantsMap::
 /// Initialize repositories with locally available timelines.
 /// Timelines that are only partially available locally (remote storage has more data than this pageserver)
 /// are scheduled for download and added to the tenant once download is completed.
-#[instrument(skip_all)]
+#[instrument(skip(conf, remote_storage))]
 pub async fn init_tenant_mgr(
     conf: &'static PageServerConf,
-    broker_client: storage_broker::BrokerClientChannel,
     remote_storage: Option<GenericRemoteStorage>,
 ) -> anyhow::Result<()> {
     // Scan local filesystem for attached tenants
@@ -117,7 +116,6 @@ pub async fn init_tenant_mgr(
                     match schedule_local_tenant_processing(
                         conf,
                         &tenant_dir_path,
-                        broker_client.clone(),
                         remote_storage.clone(),
                         &ctx,
                     ) {
@@ -152,7 +150,6 @@ pub async fn init_tenant_mgr(
 pub fn schedule_local_tenant_processing(
     conf: &'static PageServerConf,
     tenant_path: &Path,
-    broker_client: storage_broker::BrokerClientChannel,
     remote_storage: Option<GenericRemoteStorage>,
     ctx: &RequestContext,
 ) -> anyhow::Result<Arc<Tenant>> {
@@ -189,25 +186,15 @@ pub fn schedule_local_tenant_processing(
     let tenant = if conf.tenant_attaching_mark_file_path(&tenant_id).exists() {
         info!("tenant {tenant_id} has attaching mark file, resuming its attach operation");
         if let Some(remote_storage) = remote_storage {
-            match Tenant::spawn_attach(conf, tenant_id, broker_client, remote_storage, ctx) {
-                Ok(tenant) => tenant,
-                Err(e) => {
-                    error!("Failed to spawn_attach tenant {tenant_id}, reason: {e:#}");
-                    Tenant::create_broken_tenant(conf, tenant_id, format!("{e:#}"))
-                }
-            }
+            Tenant::spawn_attach(conf, tenant_id, remote_storage, ctx)
         } else {
             warn!("tenant {tenant_id} has attaching mark file, but pageserver has no remote storage configured");
-            Tenant::create_broken_tenant(
-                conf,
-                tenant_id,
-                "attaching mark file present but no remote storage configured".to_string(),
-            )
+            Tenant::create_broken_tenant(conf, tenant_id)
         }
     } else {
         info!("tenant {tenant_id} is assumed to be loadable, starting load operation");
         // Start loading the tenant into memory. It will initially be in Loading state.
-        Tenant::spawn_load(conf, tenant_id, broker_client, remote_storage, ctx)
+        Tenant::spawn_load(conf, tenant_id, remote_storage, ctx)
     };
     Ok(tenant)
 }
@@ -278,45 +265,31 @@ pub async fn create_tenant(
     conf: &'static PageServerConf,
     tenant_conf: TenantConfOpt,
     tenant_id: TenantId,
-    broker_client: storage_broker::BrokerClientChannel,
     remote_storage: Option<GenericRemoteStorage>,
     ctx: &RequestContext,
 ) -> Result<Arc<Tenant>, TenantMapInsertError> {
-    tenant_map_insert(tenant_id, || {
+    tenant_map_insert(tenant_id, |vacant_entry| {
         // We're holding the tenants lock in write mode while doing local IO.
         // If this section ever becomes contentious, introduce a new `TenantState::Creating`
         // and do the work in that state.
-        let tenant_directory = super::create_tenant_files(conf, tenant_conf, tenant_id, CreateTenantFilesMode::Create)?;
-        // TODO: tenant directory remains on disk if we bail out from here on.
-        //       See https://github.com/neondatabase/neon/issues/4233
-
+        let tenant_directory = super::create_tenant_files(conf, tenant_conf, tenant_id)?;
         let created_tenant =
-            schedule_local_tenant_processing(conf, &tenant_directory, broker_client, remote_storage, ctx)?;
-        // TODO: tenant object & its background loops remain, untracked in tenant map, if we fail here.
-        //      See https://github.com/neondatabase/neon/issues/4233
-
+            schedule_local_tenant_processing(conf, &tenant_directory, remote_storage, ctx)?;
         let crated_tenant_id = created_tenant.tenant_id();
         anyhow::ensure!(
                 tenant_id == crated_tenant_id,
                 "loaded created tenant has unexpected tenant id (expect {tenant_id} != actual {crated_tenant_id})",
             );
+        vacant_entry.insert(Arc::clone(&created_tenant));
         Ok(created_tenant)
     }).await
 }
 
-#[derive(Debug, thiserror::Error)]
-pub enum SetNewTenantConfigError {
-    #[error(transparent)]
-    GetTenant(#[from] GetTenantError),
-    #[error(transparent)]
-    Persist(anyhow::Error),
-}
-
 pub async fn set_new_tenant_config(
     conf: &'static PageServerConf,
     new_tenant_conf: TenantConfOpt,
     tenant_id: TenantId,
-) -> Result<(), SetNewTenantConfigError> {
+) -> Result<(), TenantStateError> {
     info!("configuring tenant {tenant_id}");
     let tenant = get_tenant(tenant_id, true).await?;
 
@@ -326,32 +299,23 @@ pub async fn set_new_tenant_config(
         &tenant_config_path,
         new_tenant_conf,
         false,
-    )
-    .map_err(SetNewTenantConfigError::Persist)?;
+    )?;
     tenant.set_new_tenant_config(new_tenant_conf);
     Ok(())
 }
 
-#[derive(Debug, thiserror::Error)]
-pub enum GetTenantError {
-    #[error("Tenant {0} not found")]
-    NotFound(TenantId),
-    #[error("Tenant {0} is not active")]
-    NotActive(TenantId),
-}
-
 /// Gets the tenant from the in-memory data, erroring if it's absent or is not fitting to the query.
 /// `active_only = true` allows to query only tenants that are ready for operations, erroring on other kinds of tenants.
 pub async fn get_tenant(
     tenant_id: TenantId,
     active_only: bool,
-) -> Result<Arc<Tenant>, GetTenantError> {
+) -> Result<Arc<Tenant>, TenantStateError> {
     let m = TENANTS.read().await;
     let tenant = m
         .get(&tenant_id)
-        .ok_or(GetTenantError::NotFound(tenant_id))?;
+        .ok_or(TenantStateError::NotFound(tenant_id))?;
     if active_only && !tenant.is_active() {
-        Err(GetTenantError::NotActive(tenant_id))
+        Err(TenantStateError::NotActive(tenant_id))
     } else {
         Ok(Arc::clone(tenant))
     }
@@ -360,7 +324,7 @@ pub async fn get_tenant(
 #[derive(Debug, thiserror::Error)]
 pub enum DeleteTimelineError {
     #[error("Tenant {0}")]
-    Tenant(#[from] GetTenantError),
+    Tenant(#[from] TenantStateError),
 
     #[error("Timeline {0}")]
     Timeline(#[from] crate::tenant::DeleteTimelineError),
@@ -425,11 +389,10 @@ pub async fn detach_tenant(
 pub async fn load_tenant(
     conf: &'static PageServerConf,
     tenant_id: TenantId,
-    broker_client: storage_broker::BrokerClientChannel,
     remote_storage: Option<GenericRemoteStorage>,
     ctx: &RequestContext,
 ) -> Result<(), TenantMapInsertError> {
-    tenant_map_insert(tenant_id, || {
+    tenant_map_insert(tenant_id, |vacant_entry| {
         let tenant_path = conf.tenant_path(&tenant_id);
         let tenant_ignore_mark = conf.tenant_ignore_mark_file_path(tenant_id);
         if tenant_ignore_mark.exists() {
@@ -437,14 +400,14 @@ pub async fn load_tenant(
                 .with_context(|| format!("Failed to remove tenant ignore mark {tenant_ignore_mark:?} during tenant loading"))?;
         }
 
-        let new_tenant = schedule_local_tenant_processing(conf, &tenant_path, broker_client, remote_storage, ctx)
+        let new_tenant = schedule_local_tenant_processing(conf, &tenant_path, remote_storage, ctx)
             .with_context(|| {
                 format!("Failed to schedule tenant processing in path {tenant_path:?}")
             })?;
 
-        Ok(new_tenant)
-    }).await?;
-    Ok(())
+        vacant_entry.insert(new_tenant);
+        Ok(())
+    }).await
 }
 
 pub async fn ignore_tenant(
@@ -493,36 +456,21 @@ pub async fn list_tenants() -> Result<Vec<(TenantId, TenantState)>, TenantMapLis
 pub async fn attach_tenant(
     conf: &'static PageServerConf,
     tenant_id: TenantId,
-    tenant_conf: TenantConfOpt,
-    broker_client: storage_broker::BrokerClientChannel,
     remote_storage: GenericRemoteStorage,
     ctx: &RequestContext,
 ) -> Result<(), TenantMapInsertError> {
-    tenant_map_insert(tenant_id, || {
-        let tenant_dir = create_tenant_files(conf, tenant_conf, tenant_id, CreateTenantFilesMode::Attach)?;
-        // TODO: tenant directory remains on disk if we bail out from here on.
-        //       See https://github.com/neondatabase/neon/issues/4233
-
-        // Without the attach marker, schedule_local_tenant_processing will treat the attached tenant as fully attached
-        let marker_file_exists = conf
-            .tenant_attaching_mark_file_path(&tenant_id)
-            .try_exists()
-            .context("check for attach marker file existence")?;
-        anyhow::ensure!(marker_file_exists, "create_tenant_files should have created the attach marker file");
-
-        let attached_tenant = schedule_local_tenant_processing(conf, &tenant_dir, broker_client, Some(remote_storage), ctx)?;
-        // TODO: tenant object & its background loops remain, untracked in tenant map, if we fail here.
-        //      See https://github.com/neondatabase/neon/issues/4233
-
-        let attached_tenant_id = attached_tenant.tenant_id();
+    tenant_map_insert(tenant_id, |vacant_entry| {
+        let tenant_path = conf.tenant_path(&tenant_id);
         anyhow::ensure!(
-            tenant_id == attached_tenant_id,
-            "loaded created tenant has unexpected tenant id (expect {tenant_id} != actual {attached_tenant_id})",
+            !tenant_path.exists(),
+            "Cannot attach tenant {tenant_id}, local tenant directory already exists"
         );
-        Ok(attached_tenant)
+
+        let tenant = Tenant::spawn_attach(conf, tenant_id, remote_storage, ctx);
+        vacant_entry.insert(tenant);
+        Ok(())
     })
-    .await?;
-    Ok(())
+    .await
 }
 
 #[derive(Debug, thiserror::Error)]
@@ -543,12 +491,12 @@ pub enum TenantMapInsertError {
 ///
 /// NB: the closure should return quickly because the current implementation of tenants map
 /// serializes access through an `RwLock`.
-async fn tenant_map_insert<F>(
+async fn tenant_map_insert<F, V>(
     tenant_id: TenantId,
     insert_fn: F,
-) -> Result<Arc<Tenant>, TenantMapInsertError>
+) -> Result<V, TenantMapInsertError>
 where
-    F: FnOnce() -> anyhow::Result<Arc<Tenant>>,
+    F: FnOnce(hash_map::VacantEntry<TenantId, Arc<Tenant>>) -> anyhow::Result<V>,
 {
     let mut guard = TENANTS.write().await;
     let m = match &mut *guard {
@@ -561,11 +509,8 @@ where
             tenant_id,
             e.get().current_state(),
         )),
-        hash_map::Entry::Vacant(v) => match insert_fn() {
-            Ok(tenant) => {
-                v.insert(tenant.clone());
-                Ok(tenant)
-            }
+        hash_map::Entry::Vacant(v) => match insert_fn(v) {
+            Ok(v) => Ok(v),
             Err(e) => Err(TenantMapInsertError::Closure(e)),
         },
     }

pageserver/src/tenant/remote_timeline_client.rs

@@ -1264,7 +1264,9 @@ mod tests {
             let harness = TenantHarness::create(test_name)?;
             let (tenant, ctx) = runtime.block_on(harness.load());
             // create an empty timeline directory
-            let _ = tenant.create_test_timeline(TIMELINE_ID, Lsn(0), DEFAULT_PG_VERSION, &ctx)?;
+            let timeline =
+                tenant.create_empty_timeline(TIMELINE_ID, Lsn(0), DEFAULT_PG_VERSION, &ctx)?;
+            let _ = timeline.initialize(&ctx).unwrap();
 
             let remote_fs_dir = harness.conf.workdir.join("remote_fs");
             std::fs::create_dir_all(remote_fs_dir)?;

pageserver/src/tenant/size.rs

@@ -11,7 +11,7 @@ use tokio_util::sync::CancellationToken;
 use crate::context::RequestContext;
 use crate::pgdatadir_mapping::CalculateLogicalSizeError;
 
-use super::{LogicalSizeCalculationCause, Tenant};
+use super::Tenant;
 use crate::tenant::Timeline;
 use utils::id::TimelineId;
 use utils::lsn::Lsn;
@@ -126,7 +126,6 @@ pub(super) async fn gather_inputs(
     limit: &Arc<Semaphore>,
     max_retention_period: Option<u64>,
     logical_size_cache: &mut HashMap<(TimelineId, Lsn), u64>,
-    cause: LogicalSizeCalculationCause,
     ctx: &RequestContext,
 ) -> anyhow::Result<ModelInputs> {
     // refresh is needed to update gc related pitr_cutoff and horizon_cutoff
@@ -319,15 +318,7 @@ pub(super) async fn gather_inputs(
 
     // We left the 'size' field empty in all of the Segments so far.
     // Now find logical sizes for all of the points that might need or benefit from them.
-    fill_logical_sizes(
-        &timelines,
-        &mut segments,
-        limit,
-        logical_size_cache,
-        cause,
-        ctx,
-    )
-    .await?;
+    fill_logical_sizes(&timelines, &mut segments, limit, logical_size_cache, ctx).await?;
 
     Ok(ModelInputs {
         segments,
@@ -345,7 +336,6 @@ async fn fill_logical_sizes(
     segments: &mut [SegmentMeta],
     limit: &Arc<Semaphore>,
     logical_size_cache: &mut HashMap<(TimelineId, Lsn), u64>,
-    cause: LogicalSizeCalculationCause,
     ctx: &RequestContext,
 ) -> anyhow::Result<()> {
     let timeline_hash: HashMap<TimelineId, Arc<Timeline>> = HashMap::from_iter(
@@ -383,17 +373,13 @@ async fn fill_logical_sizes(
                 let timeline = Arc::clone(timeline_hash.get(&timeline_id).unwrap());
                 let parallel_size_calcs = Arc::clone(limit);
                 let ctx = ctx.attached_child();
-                joinset.spawn(
-                    calculate_logical_size(
-                        parallel_size_calcs,
-                        timeline,
-                        lsn,
-                        cause,
-                        ctx,
-                        cancel.child_token(),
-                    )
-                    .in_current_span(),
-                );
+                joinset.spawn(calculate_logical_size(
+                    parallel_size_calcs,
+                    timeline,
+                    lsn,
+                    ctx,
+                    cancel.child_token(),
+                ));
             }
             e.insert(cached_size);
         }
@@ -496,7 +482,6 @@ async fn calculate_logical_size(
     limit: Arc<tokio::sync::Semaphore>,
     timeline: Arc<crate::tenant::Timeline>,
     lsn: utils::lsn::Lsn,
-    cause: LogicalSizeCalculationCause,
     ctx: RequestContext,
     cancel: CancellationToken,
 ) -> Result<TimelineAtLsnSizeResult, RecvError> {
@@ -505,7 +490,7 @@ async fn calculate_logical_size(
         .expect("global semaphore should not had been closed");
 
     let size_res = timeline
-        .spawn_ondemand_logical_size_calculation(lsn, cause, ctx, cancel)
+        .spawn_ondemand_logical_size_calculation(lsn, ctx, cancel)
         .instrument(info_span!("spawn_ondemand_logical_size_calculation"))
         .await?;
     Ok(TimelineAtLsnSizeResult(timeline, lsn, size_res))

pageserver/src/tenant/storage_layer.rs

@@ -22,6 +22,7 @@ use pageserver_api::models::{
 };
 use std::ops::Range;
 use std::path::PathBuf;
+use std::pin::Pin;
 use std::sync::{Arc, Mutex};
 use std::time::{Duration, SystemTime, UNIX_EPOCH};
 use tracing::warn;
@@ -335,6 +336,13 @@ impl LayerAccessStats {
     }
 }
 
+pub(crate) type GetValueReconstructFuture = Pin<
+    Box<
+        dyn Send
+            + std::future::Future<Output = Result<(ValueReconstructState, ValueReconstructResult)>>,
+    >,
+>;
+
 /// Supertrait of the [`Layer`] trait that captures the bare minimum interface
 /// required by [`LayerMap`].
 ///
@@ -372,12 +380,12 @@ pub trait Layer: std::fmt::Debug + Send + Sync {
     /// the predecessor layer and call again with the same 'reconstruct_data' to
     /// collect more data.
     fn get_value_reconstruct_data(
-        &self,
+        self: Arc<Self>,
         key: Key,
         lsn_range: Range<Lsn>,
-        reconstruct_data: &mut ValueReconstructState,
-        ctx: &RequestContext,
-    ) -> Result<ValueReconstructResult>;
+        reconstruct_data: ValueReconstructState,
+        ctx: RequestContext,
+    ) -> GetValueReconstructFuture;
 
     /// A short ID string that uniquely identifies the given layer within a [`LayerMap`].
     fn short_id(&self) -> String;
@@ -486,12 +494,12 @@ impl Layer for LayerDescriptor {
     }
 
     fn get_value_reconstruct_data(
-        &self,
+        self: Arc<Self>,
         _key: Key,
         _lsn_range: Range<Lsn>,
-        _reconstruct_data: &mut ValueReconstructState,
-        _ctx: &RequestContext,
-    ) -> Result<ValueReconstructResult> {
+        _reconstruct_data: ValueReconstructState,
+        _ctx: RequestContext,
+    ) -> GetValueReconstructFuture {
         todo!("This method shouldn't be part of the Layer trait")
     }
 
@@ -542,7 +550,7 @@ impl From<LayerFileName> for LayerDescriptor {
 ///
 /// This is used by DeltaLayer and ImageLayer. Normally, this holds a reference to the
 /// global config, and paths to layer files are constructed using the tenant/timeline
-/// path from the config. But in the 'pagectl' binary, we need to construct a Layer
+/// path from the config. But in the 'pageserver_binutils' binary, we need to construct a Layer
 /// struct for a file on disk, without having a page server running, so that we have no
 /// config. In that case, we use the Path variant to hold the full path to the file on
 /// disk.

pageserver/src/tenant/storage_layer/delta_layer.rs

@@ -46,7 +46,7 @@ use std::io::{Seek, SeekFrom};
 use std::ops::Range;
 use std::os::unix::fs::FileExt;
 use std::path::{Path, PathBuf};
-use std::sync::{RwLock, RwLockReadGuard, RwLockWriteGuard};
+use std::sync::{Arc, RwLock, RwLockReadGuard, RwLockWriteGuard};
 use tracing::*;
 
 use utils::{
@@ -56,8 +56,8 @@ use utils::{
 };
 
 use super::{
-    DeltaFileName, Layer, LayerAccessStats, LayerAccessStatsReset, LayerFileName, LayerIter,
-    LayerKeyIter, PathOrConf,
+    DeltaFileName, GetValueReconstructFuture, Layer, LayerAccessStats, LayerAccessStatsReset,
+    LayerFileName, LayerIter, LayerKeyIter, PathOrConf,
 };
 
 ///
@@ -110,7 +110,7 @@ const WILL_INIT: u64 = 1;
 /// reading/deserializing records themselves.
 ///
 #[derive(Debug, Serialize, Deserialize, Copy, Clone)]
-pub struct BlobRef(pub u64);
+struct BlobRef(u64);
 
 impl BlobRef {
     pub fn will_init(&self) -> bool {
@@ -318,89 +318,94 @@ impl Layer for DeltaLayer {
     }
 
     fn get_value_reconstruct_data(
-        &self,
+        self: Arc<Self>,
         key: Key,
         lsn_range: Range<Lsn>,
-        reconstruct_state: &mut ValueReconstructState,
-        ctx: &RequestContext,
-    ) -> anyhow::Result<ValueReconstructResult> {
-        ensure!(lsn_range.start >= self.lsn_range.start);
-        let mut need_image = true;
+        mut reconstruct_state: ValueReconstructState,
+        ctx: RequestContext,
+    ) -> GetValueReconstructFuture {
+        Box::pin(async move {
+            tokio::task::spawn_blocking(move || {
+                ensure!(lsn_range.start >= self.lsn_range.start);
+                let mut need_image = true;
 
-        ensure!(self.key_range.contains(&key));
+                ensure!(self.key_range.contains(&key));
 
-        {
-            // Open the file and lock the metadata in memory
-            let inner = self.load(LayerAccessKind::GetValueReconstructData, ctx)?;
+                {
+                    // Open the file and lock the metadata in memory
+                    let inner = self.load(LayerAccessKind::GetValueReconstructData, &ctx)?;
 
-            // Scan the page versions backwards, starting from `lsn`.
-            let file = inner.file.as_ref().unwrap();
-            let tree_reader = DiskBtreeReader::<_, DELTA_KEY_SIZE>::new(
-                inner.index_start_blk,
-                inner.index_root_blk,
-                file,
-            );
-            let search_key = DeltaKey::from_key_lsn(&key, Lsn(lsn_range.end.0 - 1));
+                    // Scan the page versions backwards, starting from `lsn`.
+                    let file = inner.file.as_ref().unwrap();
+                    let tree_reader = DiskBtreeReader::<_, DELTA_KEY_SIZE>::new(
+                        inner.index_start_blk,
+                        inner.index_root_blk,
+                        file,
+                    );
+                    let search_key = DeltaKey::from_key_lsn(&key, Lsn(lsn_range.end.0 - 1));
 
-            let mut offsets: Vec<(Lsn, u64)> = Vec::new();
+                    let mut offsets: Vec<(Lsn, u64)> = Vec::new();
 
-            tree_reader.visit(&search_key.0, VisitDirection::Backwards, |key, value| {
-                let blob_ref = BlobRef(value);
-                if key[..KEY_SIZE] != search_key.0[..KEY_SIZE] {
-                    return false;
-                }
-                let entry_lsn = DeltaKey::extract_lsn_from_buf(key);
-                if entry_lsn < lsn_range.start {
-                    return false;
-                }
-                offsets.push((entry_lsn, blob_ref.pos()));
+                    tree_reader.visit(&search_key.0, VisitDirection::Backwards, |key, value| {
+                        let blob_ref = BlobRef(value);
+                        if key[..KEY_SIZE] != search_key.0[..KEY_SIZE] {
+                            return false;
+                        }
+                        let entry_lsn = DeltaKey::extract_lsn_from_buf(key);
+                        if entry_lsn < lsn_range.start {
+                            return false;
+                        }
+                        offsets.push((entry_lsn, blob_ref.pos()));
 
-                !blob_ref.will_init()
-            })?;
+                        !blob_ref.will_init()
+                    })?;
 
-            // Ok, 'offsets' now contains the offsets of all the entries we need to read
-            let mut cursor = file.block_cursor();
-            let mut buf = Vec::new();
-            for (entry_lsn, pos) in offsets {
-                cursor.read_blob_into_buf(pos, &mut buf).with_context(|| {
-                    format!(
-                        "Failed to read blob from virtual file {}",
-                        file.file.path.display()
-                    )
-                })?;
-                let val = Value::des(&buf).with_context(|| {
-                    format!(
-                        "Failed to deserialize file blob from virtual file {}",
-                        file.file.path.display()
-                    )
-                })?;
-                match val {
-                    Value::Image(img) => {
-                        reconstruct_state.img = Some((entry_lsn, img));
-                        need_image = false;
-                        break;
-                    }
-                    Value::WalRecord(rec) => {
-                        let will_init = rec.will_init();
-                        reconstruct_state.records.push((entry_lsn, rec));
-                        if will_init {
-                            // This WAL record initializes the page, so no need to go further back
-                            need_image = false;
-                            break;
+                    // Ok, 'offsets' now contains the offsets of all the entries we need to read
+                    let mut cursor = file.block_cursor();
+                    let mut buf = Vec::new();
+                    for (entry_lsn, pos) in offsets {
+                        cursor.read_blob_into_buf(pos, &mut buf).with_context(|| {
+                            format!(
+                                "Failed to read blob from virtual file {}",
+                                file.file.path.display()
+                            )
+                        })?;
+                        let val = Value::des(&buf).with_context(|| {
+                            format!(
+                                "Failed to deserialize file blob from virtual file {}",
+                                file.file.path.display()
+                            )
+                        })?;
+                        match val {
+                            Value::Image(img) => {
+                                reconstruct_state.img = Some((entry_lsn, img));
+                                need_image = false;
+                                break;
+                            }
+                            Value::WalRecord(rec) => {
+                                let will_init = rec.will_init();
+                                reconstruct_state.records.push((entry_lsn, rec));
+                                if will_init {
+                                    // This WAL record initializes the page, so no need to go further back
+                                    need_image = false;
+                                    break;
+                                }
+                            }
                         }
                     }
+                    // release metadata lock and close the file
                 }
-            }
-            // release metadata lock and close the file
-        }
-
-        // If an older page image is needed to reconstruct the page, let the
-        // caller know.
-        if need_image {
-            Ok(ValueReconstructResult::Continue)
-        } else {
-            Ok(ValueReconstructResult::Complete)
-        }
+                // If an older page image is needed to reconstruct the page, let the
+                // caller know.
+                if need_image {
+                    Ok((reconstruct_state, ValueReconstructResult::Continue))
+                } else {
+                    Ok((reconstruct_state, ValueReconstructResult::Complete))
+                }
+            })
+            .await
+            .context("spawn_blocking")?
+        })
     }
 }
 
@@ -619,7 +624,7 @@ impl DeltaLayer {
 
     /// Create a DeltaLayer struct representing an existing file on disk.
     ///
-    /// This variant is only used for debugging purposes, by the 'pagectl' binary.
+    /// This variant is only used for debugging purposes, by the 'pageserver_binutils' binary.
     pub fn new_for_path(path: &Path, file: File) -> Result<Self> {
         let mut summary_buf = Vec::new();
         summary_buf.resize(PAGE_SZ, 0);

pageserver/src/tenant/storage_layer/image_layer.rs

@@ -43,7 +43,7 @@ use std::io::{Seek, SeekFrom};
 use std::ops::Range;
 use std::os::unix::prelude::FileExt;
 use std::path::{Path, PathBuf};
-use std::sync::{RwLock, RwLockReadGuard};
+use std::sync::{Arc, RwLock, RwLockReadGuard};
 use tracing::*;
 
 use utils::{
@@ -53,7 +53,7 @@ use utils::{
 };
 
 use super::filename::{ImageFileName, LayerFileName};
-use super::{Layer, LayerAccessStatsReset, LayerIter, PathOrConf};
+use super::{GetValueReconstructFuture, Layer, LayerAccessStatsReset, LayerIter, PathOrConf};
 
 ///
 /// Header stored in the beginning of the file
@@ -197,38 +197,45 @@ impl Layer for ImageLayer {
 
     /// Look up given page in the file
     fn get_value_reconstruct_data(
-        &self,
+        self: Arc<Self>,
         key: Key,
         lsn_range: Range<Lsn>,
-        reconstruct_state: &mut ValueReconstructState,
-        ctx: &RequestContext,
-    ) -> anyhow::Result<ValueReconstructResult> {
-        assert!(self.key_range.contains(&key));
-        assert!(lsn_range.start >= self.lsn);
-        assert!(lsn_range.end >= self.lsn);
+        mut reconstruct_state: ValueReconstructState,
+        ctx: RequestContext,
+    ) -> GetValueReconstructFuture {
+        Box::pin(async move {
+            tokio::task::spawn_blocking(move || {
+                assert!(self.key_range.contains(&key));
+                assert!(lsn_range.start >= self.lsn);
+                assert!(lsn_range.end >= self.lsn);
 
-        let inner = self.load(LayerAccessKind::GetValueReconstructData, ctx)?;
+                let inner = self.load(LayerAccessKind::GetValueReconstructData, &ctx)?;
 
-        let file = inner.file.as_ref().unwrap();
-        let tree_reader = DiskBtreeReader::new(inner.index_start_blk, inner.index_root_blk, file);
+                let file = inner.file.as_ref().unwrap();
+                let tree_reader =
+                    DiskBtreeReader::new(inner.index_start_blk, inner.index_root_blk, file);
 
-        let mut keybuf: [u8; KEY_SIZE] = [0u8; KEY_SIZE];
-        key.write_to_byte_slice(&mut keybuf);
-        if let Some(offset) = tree_reader.get(&keybuf)? {
-            let blob = file.block_cursor().read_blob(offset).with_context(|| {
-                format!(
-                    "failed to read value from data file {} at offset {}",
-                    self.path().display(),
-                    offset
-                )
-            })?;
-            let value = Bytes::from(blob);
+                let mut keybuf: [u8; KEY_SIZE] = [0u8; KEY_SIZE];
+                key.write_to_byte_slice(&mut keybuf);
+                if let Some(offset) = tree_reader.get(&keybuf)? {
+                    let blob = file.block_cursor().read_blob(offset).with_context(|| {
+                        format!(
+                            "failed to read value from data file {} at offset {}",
+                            self.path().display(),
+                            offset
+                        )
+                    })?;
+                    let value = Bytes::from(blob);
 
-            reconstruct_state.img = Some((self.lsn, value));
-            Ok(ValueReconstructResult::Complete)
-        } else {
-            Ok(ValueReconstructResult::Missing)
-        }
+                    reconstruct_state.img = Some((self.lsn, value));
+                    Ok((reconstruct_state, ValueReconstructResult::Complete))
+                } else {
+                    Ok((reconstruct_state, ValueReconstructResult::Missing))
+                }
+            })
+            .await
+            .context("spawn_blocking")?
+        })
     }
 }
 
@@ -422,7 +429,7 @@ impl ImageLayer {
 
     /// Create an ImageLayer struct representing an existing file on disk.
     ///
-    /// This variant is only used for debugging purposes, by the 'pagectl' binary.
+    /// This variant is only used for debugging purposes, by the 'pageserver_binutils' binary.
     pub fn new_for_path(path: &Path, file: File) -> Result<ImageLayer> {
         let mut summary_buf = Vec::new();
         summary_buf.resize(PAGE_SZ, 0);

pageserver/src/tenant/storage_layer/inmemory_layer.rs

@@ -12,7 +12,7 @@ use crate::tenant::block_io::BlockReader;
 use crate::tenant::ephemeral_file::EphemeralFile;
 use crate::tenant::storage_layer::{ValueReconstructResult, ValueReconstructState};
 use crate::walrecord;
-use anyhow::{ensure, Result};
+use anyhow::{ensure, Context, Result};
 use pageserver_api::models::InMemoryLayerInfo;
 use std::cell::RefCell;
 use std::collections::HashMap;
@@ -27,9 +27,9 @@ use utils::{
 // while being able to use std::fmt::Write's methods
 use std::fmt::Write as _;
 use std::ops::Range;
-use std::sync::RwLock;
+use std::sync::{Arc, RwLock};
 
-use super::{DeltaLayer, DeltaLayerWriter, Layer};
+use super::{DeltaLayer, DeltaLayerWriter, GetValueReconstructFuture, Layer};
 
 thread_local! {
     /// A buffer for serializing object during [`InMemoryLayer::put_value`].
@@ -191,52 +191,60 @@ impl Layer for InMemoryLayer {
 
     /// Look up given value in the layer.
     fn get_value_reconstruct_data(
-        &self,
+        self: Arc<Self>,
         key: Key,
         lsn_range: Range<Lsn>,
-        reconstruct_state: &mut ValueReconstructState,
-        _ctx: &RequestContext,
-    ) -> anyhow::Result<ValueReconstructResult> {
-        ensure!(lsn_range.start >= self.start_lsn);
-        let mut need_image = true;
+        mut reconstruct_state: ValueReconstructState,
+        _ctx: RequestContext,
+    ) -> GetValueReconstructFuture {
+        Box::pin(async move {
+            // The in-memory layer isn't actually in-memory. It uses EphemeralFile.
+            // So, this does do IO.
+            tokio::task::spawn_blocking(move || {
+                ensure!(lsn_range.start >= self.start_lsn);
+                let mut need_image = true;
 
-        let inner = self.inner.read().unwrap();
+                let inner = self.inner.read().unwrap();
 
-        let mut reader = inner.file.block_cursor();
+                let mut reader = inner.file.block_cursor();
 
-        // Scan the page versions backwards, starting from `lsn`.
-        if let Some(vec_map) = inner.index.get(&key) {
-            let slice = vec_map.slice_range(lsn_range);
-            for (entry_lsn, pos) in slice.iter().rev() {
-                let buf = reader.read_blob(*pos)?;
-                let value = Value::des(&buf)?;
-                match value {
-                    Value::Image(img) => {
-                        reconstruct_state.img = Some((*entry_lsn, img));
-                        return Ok(ValueReconstructResult::Complete);
-                    }
-                    Value::WalRecord(rec) => {
-                        let will_init = rec.will_init();
-                        reconstruct_state.records.push((*entry_lsn, rec));
-                        if will_init {
-                            // This WAL record initializes the page, so no need to go further back
-                            need_image = false;
-                            break;
+                // Scan the page versions backwards, starting from `lsn`.
+                if let Some(vec_map) = inner.index.get(&key) {
+                    let slice = vec_map.slice_range(lsn_range);
+                    for (entry_lsn, pos) in slice.iter().rev() {
+                        let buf = reader.read_blob(*pos)?;
+                        let value = Value::des(&buf)?;
+                        match value {
+                            Value::Image(img) => {
+                                reconstruct_state.img = Some((*entry_lsn, img));
+                                return Ok((reconstruct_state, ValueReconstructResult::Complete));
+                            }
+                            Value::WalRecord(rec) => {
+                                let will_init = rec.will_init();
+                                reconstruct_state.records.push((*entry_lsn, rec));
+                                if will_init {
+                                    // This WAL record initializes the page, so no need to go further back
+                                    need_image = false;
+                                    break;
+                                }
+                            }
                         }
                     }
                 }
-            }
-        }
 
-        // release lock on 'inner'
+                // release lock on 'inner'
 
-        // If an older page image is needed to reconstruct the page, let the
-        // caller know.
-        if need_image {
-            Ok(ValueReconstructResult::Continue)
-        } else {
-            Ok(ValueReconstructResult::Complete)
-        }
+                // If an older page image is needed to reconstruct the page, let the
+                // caller know.
+                if need_image {
+                    Ok((reconstruct_state, ValueReconstructResult::Continue))
+                } else {
+                    Ok((reconstruct_state, ValueReconstructResult::Complete))
+                }
+            })
+            .await
+            .context("spawn_blocking")?
+        })
     }
 }
 

pageserver/src/tenant/storage_layer/remote_layer.rs

@@ -6,7 +6,7 @@ use crate::context::RequestContext;
 use crate::repository::Key;
 use crate::tenant::layer_map::BatchedUpdates;
 use crate::tenant::remote_timeline_client::index::LayerFileMetadata;
-use crate::tenant::storage_layer::{Layer, ValueReconstructResult, ValueReconstructState};
+use crate::tenant::storage_layer::{Layer, ValueReconstructState};
 use anyhow::{bail, Result};
 use pageserver_api::models::HistoricLayerInfo;
 use std::ops::Range;
@@ -21,8 +21,8 @@ use utils::{
 use super::filename::{DeltaFileName, ImageFileName, LayerFileName};
 use super::image_layer::ImageLayer;
 use super::{
-    DeltaLayer, LayerAccessStats, LayerAccessStatsReset, LayerIter, LayerKeyIter,
-    LayerResidenceStatus, PersistentLayer,
+    DeltaLayer, GetValueReconstructFuture, LayerAccessStats, LayerAccessStatsReset, LayerIter,
+    LayerKeyIter, LayerResidenceStatus, PersistentLayer,
 };
 
 /// RemoteLayer is a not yet downloaded [`ImageLayer`] or
@@ -83,16 +83,18 @@ impl Layer for RemoteLayer {
     }
 
     fn get_value_reconstruct_data(
-        &self,
+        self: Arc<Self>,
         _key: Key,
         _lsn_range: Range<Lsn>,
-        _reconstruct_state: &mut ValueReconstructState,
-        _ctx: &RequestContext,
-    ) -> Result<ValueReconstructResult> {
-        bail!(
-            "layer {} needs to be downloaded",
-            self.filename().file_name()
-        );
+        _reconstruct_state: ValueReconstructState,
+        _ctx: RequestContext,
+    ) -> GetValueReconstructFuture {
+        Box::pin(async move {
+            bail!(
+                "layer {} needs to be downloaded",
+                self.filename().file_name()
+            );
+        })
     }
 
     fn is_incremental(&self) -> bool {

pageserver/src/tenant/tasks.rs

@@ -9,12 +9,13 @@ use crate::context::{DownloadBehavior, RequestContext};
 use crate::metrics::TENANT_TASK_EVENTS;
 use crate::task_mgr;
 use crate::task_mgr::{TaskKind, BACKGROUND_RUNTIME};
+use crate::tenant::mgr;
 use crate::tenant::{Tenant, TenantState};
 use tokio_util::sync::CancellationToken;
 use tracing::*;
+use utils::id::TenantId;
 
-pub fn start_background_loops(tenant: &Arc<Tenant>) {
-    let tenant_id = tenant.tenant_id;
+pub fn start_background_loops(tenant_id: TenantId) {
     task_mgr::spawn(
         BACKGROUND_RUNTIME.handle(),
         TaskKind::Compaction,
@@ -22,14 +23,11 @@ pub fn start_background_loops(tenant: &Arc<Tenant>) {
         None,
         &format!("compactor for tenant {tenant_id}"),
         false,
-        {
-            let tenant = Arc::clone(tenant);
-            async move {
-                compaction_loop(tenant)
-                    .instrument(info_span!("compaction_loop", tenant_id = %tenant_id))
-                    .await;
-                Ok(())
-            }
+        async move {
+            compaction_loop(tenant_id)
+                .instrument(info_span!("compaction_loop", tenant_id = %tenant_id))
+                .await;
+            Ok(())
         },
     );
     task_mgr::spawn(
@@ -39,14 +37,11 @@ pub fn start_background_loops(tenant: &Arc<Tenant>) {
         None,
         &format!("garbage collector for tenant {tenant_id}"),
         false,
-        {
-            let tenant = Arc::clone(tenant);
-            async move {
-                gc_loop(tenant)
-                    .instrument(info_span!("gc_loop", tenant_id = %tenant_id))
-                    .await;
-                Ok(())
-            }
+        async move {
+            gc_loop(tenant_id)
+                .instrument(info_span!("gc_loop", tenant_id = %tenant_id))
+                .await;
+            Ok(())
         },
     );
 }
@@ -54,7 +49,7 @@ pub fn start_background_loops(tenant: &Arc<Tenant>) {
 ///
 /// Compaction task's main loop
 ///
-async fn compaction_loop(tenant: Arc<Tenant>) {
+async fn compaction_loop(tenant_id: TenantId) {
     let wait_duration = Duration::from_secs(2);
     info!("starting");
     TENANT_TASK_EVENTS.with_label_values(&["start"]).inc();
@@ -65,16 +60,16 @@ async fn compaction_loop(tenant: Arc<Tenant>) {
         loop {
             trace!("waking up");
 
-            tokio::select! {
+            let tenant = tokio::select! {
                 _ = cancel.cancelled() => {
                     info!("received cancellation request");
                     return;
                 },
-                tenant_wait_result = wait_for_active_tenant(&tenant) => match tenant_wait_result {
+                tenant_wait_result = wait_for_active_tenant(tenant_id, wait_duration) => match tenant_wait_result {
                     ControlFlow::Break(()) => return,
-                    ControlFlow::Continue(()) => (),
+                    ControlFlow::Continue(tenant) => tenant,
                 },
-            }
+            };
 
             let period = tenant.get_compaction_period();
 
@@ -124,7 +119,7 @@ async fn compaction_loop(tenant: Arc<Tenant>) {
 ///
 /// GC task's main loop
 ///
-async fn gc_loop(tenant: Arc<Tenant>) {
+async fn gc_loop(tenant_id: TenantId) {
     let wait_duration = Duration::from_secs(2);
     info!("starting");
     TENANT_TASK_EVENTS.with_label_values(&["start"]).inc();
@@ -132,22 +127,21 @@ async fn gc_loop(tenant: Arc<Tenant>) {
         let cancel = task_mgr::shutdown_token();
         // GC might require downloading, to find the cutoff LSN that corresponds to the
         // cutoff specified as time.
-        let ctx =
-            RequestContext::todo_child(TaskKind::GarbageCollector, DownloadBehavior::Download);
+        let ctx = RequestContext::todo_child(TaskKind::GarbageCollector, DownloadBehavior::Download);
         let mut first = true;
         loop {
             trace!("waking up");
 
-            tokio::select! {
+            let tenant = tokio::select! {
                 _ = cancel.cancelled() => {
                     info!("received cancellation request");
                     return;
                 },
-                tenant_wait_result = wait_for_active_tenant(&tenant) => match tenant_wait_result {
+                tenant_wait_result = wait_for_active_tenant(tenant_id, wait_duration) => match tenant_wait_result {
                     ControlFlow::Break(()) => return,
-                    ControlFlow::Continue(()) => (),
+                    ControlFlow::Continue(tenant) => tenant,
                 },
-            }
+            };
 
             let period = tenant.get_gc_period();
 
@@ -167,9 +161,7 @@ async fn gc_loop(tenant: Arc<Tenant>) {
                 Duration::from_secs(10)
             } else {
                 // Run gc
-                let res = tenant
-                    .gc_iteration(None, gc_horizon, tenant.get_pitr_interval(), &ctx)
-                    .await;
+                let res = tenant.gc_iteration(None, gc_horizon, tenant.get_pitr_interval(), &ctx).await;
                 if let Err(e) = res {
                     error!("Gc failed, retrying in {:?}: {e:?}", wait_duration);
                     wait_duration
@@ -195,10 +187,23 @@ async fn gc_loop(tenant: Arc<Tenant>) {
     trace!("GC loop stopped.");
 }
 
-async fn wait_for_active_tenant(tenant: &Arc<Tenant>) -> ControlFlow<()> {
+async fn wait_for_active_tenant(
+    tenant_id: TenantId,
+    wait: Duration,
+) -> ControlFlow<(), Arc<Tenant>> {
+    let tenant = loop {
+        match mgr::get_tenant(tenant_id, false).await {
+            Ok(tenant) => break tenant,
+            Err(e) => {
+                error!("Failed to get a tenant {tenant_id}: {e:#}");
+                tokio::time::sleep(wait).await;
+            }
+        }
+    };
+
     // if the tenant has a proper status already, no need to wait for anything
     if tenant.current_state() == TenantState::Active {
-        ControlFlow::Continue(())
+        ControlFlow::Continue(tenant)
     } else {
         let mut tenant_state_updates = tenant.subscribe_for_state_updates();
         loop {
@@ -208,7 +213,7 @@ async fn wait_for_active_tenant(tenant: &Arc<Tenant>) -> ControlFlow<()> {
                     match new_state {
                         TenantState::Active => {
                             debug!("Tenant state changed to active, continuing the task loop");
-                            return ControlFlow::Continue(());
+                            return ControlFlow::Continue(tenant);
                         }
                         state => {
                             debug!("Not running the task loop, tenant is not active: {state:?}");
@@ -254,7 +259,6 @@ pub(crate) async fn random_init_delay(
     }
 }
 
-/// Attention: the `task` and `period` beocme labels of a pageserver-wide prometheus metric.
 pub(crate) fn warn_when_period_overrun(elapsed: Duration, period: Duration, task: &str) {
     // Duration::ZERO will happen because it's the "disable [bgtask]" value.
     if elapsed >= period && period != Duration::ZERO {
@@ -267,8 +271,5 @@ pub(crate) fn warn_when_period_overrun(elapsed: Duration, period: Duration, task
             task,
             "task iteration took longer than the configured period"
         );
-        crate::metrics::BACKGROUND_LOOP_PERIOD_OVERRUN_COUNT
-            .with_label_values(&[task, &format!("{}", period.as_secs())])
-            .inc();
     }
 }

pageserver/src/tenant/timeline.rs

@@ -22,7 +22,8 @@ use tracing::*;
 use utils::id::TenantTimelineId;
 
 use std::cmp::{max, min, Ordering};
-use std::collections::{BinaryHeap, HashMap};
+use std::collections::BinaryHeap;
+use std::collections::HashMap;
 use std::fs;
 use std::ops::{Deref, Range};
 use std::path::{Path, PathBuf};
@@ -31,6 +32,7 @@ use std::sync::atomic::{AtomicI64, Ordering as AtomicOrdering};
 use std::sync::{Arc, Mutex, MutexGuard, RwLock, Weak};
 use std::time::{Duration, Instant, SystemTime};
 
+use crate::broker_client::{get_broker_client, is_broker_client_initialized};
 use crate::context::{DownloadBehavior, RequestContext};
 use crate::tenant::remote_timeline_client::{self, index::LayerFileMetadata};
 use crate::tenant::storage_layer::{
@@ -46,7 +48,7 @@ use crate::tenant::{
 };
 
 use crate::config::PageServerConf;
-use crate::keyspace::{KeyPartitioning, KeySpace, KeySpaceRandomAccum};
+use crate::keyspace::{KeyPartitioning, KeySpace};
 use crate::metrics::{TimelineMetrics, UNEXPECTED_ONDEMAND_DOWNLOADS};
 use crate::pgdatadir_mapping::LsnForTimestamp;
 use crate::pgdatadir_mapping::{is_rel_fsm_block_key, is_rel_vm_block_key};
@@ -119,18 +121,7 @@ pub struct Timeline {
 
     pub pg_version: u32,
 
-    pub(super) layers: RwLock<LayerMap<dyn PersistentLayer>>,
-
-    /// Set of key ranges which should be covered by image layers to
-    /// allow GC to remove old layers. This set is created by GC and its cutoff LSN is also stored.
-    /// It is used by compaction task when it checks if new image layer should be created.
-    /// Newly created image layer doesn't help to remove the delta layer, until the
-    /// newly created image layer falls off the PITR horizon. So on next GC cycle,
-    /// gc_timeline may still want the new image layer to be created. To avoid redundant
-    /// image layers creation we should check if image layer exists but beyond PITR horizon.
-    /// This is why we need remember GC cutoff LSN.
-    ///
-    wanted_image_layers: Mutex<Option<(Lsn, KeySpace)>>,
+    pub(super) layers: parking_lot::RwLock<LayerMap<dyn PersistentLayer>>,
 
     last_freeze_at: AtomicLsn,
     // Atomic would be more appropriate here.
@@ -226,7 +217,7 @@ pub struct Timeline {
     /// or None if WAL receiver has not received anything for this timeline
     /// yet.
     pub last_received_wal: Mutex<Option<WalReceiverInfo>>,
-    pub walreceiver: Mutex<Option<WalReceiver>>,
+    pub walreceiver: WalReceiver,
 
     /// Relation size cache
     pub rel_size_cache: RwLock<HashMap<RelTag, (Lsn, BlockNumber)>>,
@@ -405,9 +396,6 @@ pub enum PageReconstructError {
     /// The operation was cancelled
     Cancelled,
 
-    /// The ancestor of this is being stopped
-    AncestorStopping(TimelineId),
-
     /// An error happened replaying WAL records
     #[error(transparent)]
     WalRedo(#[from] crate::walredo::WalRedoError),
@@ -426,9 +414,6 @@ impl std::fmt::Debug for PageReconstructError {
                 )
             }
             Self::Cancelled => write!(f, "cancelled"),
-            Self::AncestorStopping(timeline_id) => {
-                write!(f, "ancestor timeline {timeline_id} is being stopped")
-            }
             Self::WalRedo(err) => err.fmt(f),
         }
     }
@@ -447,22 +432,11 @@ impl std::fmt::Display for PageReconstructError {
                 )
             }
             Self::Cancelled => write!(f, "cancelled"),
-            Self::AncestorStopping(timeline_id) => {
-                write!(f, "ancestor timeline {timeline_id} is being stopped")
-            }
             Self::WalRedo(err) => err.fmt(f),
         }
     }
 }
 
-#[derive(Clone, Copy)]
-pub enum LogicalSizeCalculationCause {
-    Initial,
-    ConsumptionMetricsSyntheticSize,
-    EvictionTaskImitation,
-    TenantSizeHandler,
-}
-
 /// Public interface functions
 impl Timeline {
     /// Get the LSN where this branch was created
@@ -530,12 +504,13 @@ impl Timeline {
             None => None,
         };
 
-        let mut reconstruct_state = ValueReconstructState {
+        let reconstruct_state = ValueReconstructState {
             records: Vec::new(),
             img: cached_page_img,
         };
 
-        self.get_reconstruct_data(key, lsn, &mut reconstruct_state, ctx)
+        let reconstruct_state = self
+            .get_reconstruct_data(key, lsn, reconstruct_state, ctx)
             .await?;
 
         self.metrics
@@ -573,7 +548,7 @@ impl Timeline {
     /// This method makes no distinction between local and remote layers.
     /// Hence, the result **does not represent local filesystem usage**.
     pub fn layer_size_sum(&self) -> u64 {
-        let layer_map = self.layers.read().unwrap();
+        let layer_map = self.layers.read();
         let mut size = 0;
         for l in layer_map.iter_historic_layers() {
             size += l.file_size();
@@ -621,27 +596,17 @@ impl Timeline {
             .await
         {
             Ok(()) => Ok(()),
-            Err(e) => {
-                // don't count the time spent waiting for lock below, and also in walreceiver.status(), towards the wait_lsn_time_histo
+            seqwait_error => {
                 drop(_timer);
-                let walreceiver_status = {
-                    match &*self.walreceiver.lock().unwrap() {
-                        None => "stopping or stopped".to_string(),
-                        Some(walreceiver) => match walreceiver.status() {
-                            Some(status) => status.to_human_readable_string(),
-                            None => "Not active".to_string(),
-                        },
-                    }
-                };
-                Err(anyhow::Error::new(e).context({
-                    format!(
-                        "Timed out while waiting for WAL record at LSN {} to arrive, last_record_lsn {} disk consistent LSN={}, WalReceiver status: {}",
-                        lsn,
-                        self.get_last_record_lsn(),
-                        self.get_disk_consistent_lsn(),
-                        walreceiver_status,
-                    )
-                }))
+                let walreceiver_status = self.walreceiver.status().await;
+                seqwait_error.with_context(|| format!(
+                    "Timed out while waiting for WAL record at LSN {} to arrive, last_record_lsn {} disk consistent LSN={}, {}",
+                    lsn,
+                    self.get_last_record_lsn(),
+                    self.get_disk_consistent_lsn(),
+                    walreceiver_status.map(|status| status.to_human_readable_string())
+                            .unwrap_or_else(|| "WalReceiver status: Not active".to_string()),
+                ))
             }
         }
     }
@@ -883,7 +848,7 @@ impl Timeline {
     /// safekeepers to regard pageserver as caught up and suspend activity.
     pub fn check_checkpoint_distance(self: &Arc<Timeline>) -> anyhow::Result<()> {
         let last_lsn = self.get_last_record_lsn();
-        let layers = self.layers.read().unwrap();
+        let layers = self.layers.read();
         if let Some(open_layer) = &layers.open_layer {
             let open_layer_size = open_layer.size()?;
             drop(layers);
@@ -916,10 +881,18 @@ impl Timeline {
         Ok(())
     }
 
-    pub fn activate(self: &Arc<Self>, broker_client: BrokerClientChannel, ctx: &RequestContext) {
-        self.launch_wal_receiver(ctx, broker_client);
+    pub fn activate(self: &Arc<Self>, ctx: &RequestContext) -> anyhow::Result<()> {
+        if is_broker_client_initialized() {
+            self.launch_wal_receiver(ctx, get_broker_client().clone())?;
+        } else if cfg!(test) {
+            info!("not launching WAL receiver because broker client hasn't been initialized");
+        } else {
+            anyhow::bail!("broker client not initialized");
+        }
+
         self.set_state(TimelineState::Active);
         self.launch_eviction_task();
+        Ok(())
     }
 
     pub fn set_state(&self, new_state: TimelineState) {
@@ -954,33 +927,8 @@ impl Timeline {
         self.state.subscribe()
     }
 
-    pub async fn wait_to_become_active(
-        &self,
-        _ctx: &RequestContext, // Prepare for use by cancellation
-    ) -> Result<(), TimelineState> {
-        let mut receiver = self.state.subscribe();
-        loop {
-            let current_state = *receiver.borrow_and_update();
-            match current_state {
-                TimelineState::Loading => {
-                    receiver
-                        .changed()
-                        .await
-                        .expect("holding a reference to self");
-                }
-                TimelineState::Active { .. } => {
-                    return Ok(());
-                }
-                TimelineState::Broken { .. } | TimelineState::Stopping => {
-                    // There's no chance the timeline can transition back into ::Active
-                    return Err(current_state);
-                }
-            }
-        }
-    }
-
     pub fn layer_map_info(&self, reset: LayerAccessStatsReset) -> LayerMapInfo {
-        let layer_map = self.layers.read().unwrap();
+        let layer_map = self.layers.read();
         let mut in_memory_layers = Vec::with_capacity(layer_map.frozen_layers.len() + 1);
         if let Some(open_layer) = &layer_map.open_layer {
             in_memory_layers.push(open_layer.info());
@@ -1100,7 +1048,7 @@ impl Timeline {
         }
 
         // start the batch update
-        let mut layer_map = self.layers.write().unwrap();
+        let mut layer_map = self.layers.write();
         let mut batch_updates = layer_map.batch_update();
 
         let mut results = Vec::with_capacity(layers_to_evict.len());
@@ -1328,7 +1276,15 @@ impl Timeline {
         let (layer_flush_done_tx, _) = tokio::sync::watch::channel((0, Ok(())));
 
         let tenant_conf_guard = tenant_conf.read().unwrap();
-
+        let wal_connect_timeout = tenant_conf_guard
+            .walreceiver_connect_timeout
+            .unwrap_or(conf.default_tenant_conf.walreceiver_connect_timeout);
+        let lagging_wal_timeout = tenant_conf_guard
+            .lagging_wal_timeout
+            .unwrap_or(conf.default_tenant_conf.lagging_wal_timeout);
+        let max_lsn_wal_lag = tenant_conf_guard
+            .max_lsn_wal_lag
+            .unwrap_or(conf.default_tenant_conf.max_lsn_wal_lag);
         let evictions_low_residence_duration_metric_threshold =
             Self::get_evictions_low_residence_duration_metric_threshold(
                 &tenant_conf_guard,
@@ -1337,6 +1293,18 @@ impl Timeline {
         drop(tenant_conf_guard);
 
         Arc::new_cyclic(|myself| {
+            let walreceiver = WalReceiver::new(
+                TenantTimelineId::new(tenant_id, timeline_id),
+                Weak::clone(myself),
+                WalReceiverConf {
+                    wal_connect_timeout,
+                    lagging_wal_timeout,
+                    max_lsn_wal_lag,
+                    auth_token: crate::config::SAFEKEEPER_AUTH_TOKEN.get().cloned(),
+                    availability_zone: conf.availability_zone.clone(),
+                },
+            );
+
             let mut result = Timeline {
                 conf,
                 tenant_conf,
@@ -1344,11 +1312,10 @@ impl Timeline {
                 timeline_id,
                 tenant_id,
                 pg_version,
-                layers: RwLock::new(LayerMap::default()),
-                wanted_image_layers: Mutex::new(None),
+                layers: parking_lot::RwLock::new(LayerMap::default()),
 
                 walredo_mgr,
-                walreceiver: Mutex::new(None),
+                walreceiver,
 
                 remote_client: remote_client.map(Arc::new),
 
@@ -1468,49 +1435,17 @@ impl Timeline {
         *flush_loop_state = FlushLoopState::Running;
     }
 
-    /// Creates and starts the wal receiver.
-    ///
-    /// This function is expected to be called at most once per Timeline's lifecycle
-    /// when the timeline is activated.
-    fn launch_wal_receiver(
-        self: &Arc<Self>,
+    pub(super) fn launch_wal_receiver(
+        &self,
         ctx: &RequestContext,
         broker_client: BrokerClientChannel,
-    ) {
+    ) -> anyhow::Result<()> {
         info!(
             "launching WAL receiver for timeline {} of tenant {}",
             self.timeline_id, self.tenant_id
         );
-
-        let tenant_conf_guard = self.tenant_conf.read().unwrap();
-        let wal_connect_timeout = tenant_conf_guard
-            .walreceiver_connect_timeout
-            .unwrap_or(self.conf.default_tenant_conf.walreceiver_connect_timeout);
-        let lagging_wal_timeout = tenant_conf_guard
-            .lagging_wal_timeout
-            .unwrap_or(self.conf.default_tenant_conf.lagging_wal_timeout);
-        let max_lsn_wal_lag = tenant_conf_guard
-            .max_lsn_wal_lag
-            .unwrap_or(self.conf.default_tenant_conf.max_lsn_wal_lag);
-        drop(tenant_conf_guard);
-
-        let mut guard = self.walreceiver.lock().unwrap();
-        assert!(
-            guard.is_none(),
-            "multiple launches / re-launches of WAL receiver are not supported"
-        );
-        *guard = Some(WalReceiver::start(
-            Arc::clone(self),
-            WalReceiverConf {
-                wal_connect_timeout,
-                lagging_wal_timeout,
-                max_lsn_wal_lag,
-                auth_token: crate::config::SAFEKEEPER_AUTH_TOKEN.get().cloned(),
-                availability_zone: self.conf.availability_zone.clone(),
-            },
-            broker_client,
-            ctx,
-        ));
+        self.walreceiver.start(ctx, broker_client)?;
+        Ok(())
     }
 
     ///
@@ -1518,7 +1453,7 @@ impl Timeline {
     /// Returns all timeline-related files that were found and loaded.
     ///
     pub(super) fn load_layer_map(&self, disk_consistent_lsn: Lsn) -> anyhow::Result<()> {
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
         let mut updates = layers.batch_update();
         let mut num_layers = 0;
 
@@ -1647,7 +1582,7 @@ impl Timeline {
 
         // We're holding a layer map lock for a while but this
         // method is only called during init so it's fine.
-        let mut layer_map = self.layers.write().unwrap();
+        let mut layer_map = self.layers.write();
         let mut updates = layer_map.batch_update();
         for remote_layer_name in &index_part.timeline_layers {
             let local_layer = local_only_layers.remove(remote_layer_name);
@@ -1800,7 +1735,6 @@ impl Timeline {
         let local_layers = self
             .layers
             .read()
-            .unwrap()
             .iter_historic_layers()
             .map(|l| (l.filename(), l))
             .collect::<HashMap<_, _>>();
@@ -1905,31 +1839,18 @@ impl Timeline {
                 // to spawn_ondemand_logical_size_calculation.
                 let cancel = CancellationToken::new();
                 let calculated_size = match self_clone
-                    .logical_size_calculation_task(lsn, LogicalSizeCalculationCause::Initial, &background_ctx, cancel)
+                    .logical_size_calculation_task(lsn, &background_ctx, cancel)
                     .await
                 {
                     Ok(s) => s,
                     Err(CalculateLogicalSizeError::Cancelled) => {
                         // Don't make noise, this is a common task.
-                        // In the unlikely case that there is another call to this function, we'll retry
+                        // In the unlikely case that there ihs another call to this function, we'll retry
                         // because initial_logical_size is still None.
                         info!("initial size calculation cancelled, likely timeline delete / tenant detach");
                         return Ok(());
                     }
-                    Err(CalculateLogicalSizeError::Other(err)) => {
-                        if let Some(e @ PageReconstructError::AncestorStopping(_)) =
-                            err.root_cause().downcast_ref()
-                        {
-                            // This can happen if the timeline parent timeline switches to
-                            // Stopping state while we're still calculating the initial
-                            // timeline size for the child, for example if the tenant is
-                            // being detached or the pageserver is shut down. Like with
-                            // CalculateLogicalSizeError::Cancelled, don't make noise.
-                            info!("initial size calculation failed because the timeline or its ancestor is Stopping, likely because the tenant is being detached: {e:#}");
-                            return Ok(());
-                        }
-                        return Err(err.context("Failed to calculate logical size"));
-                    }
+                    x @ Err(_) => x.context("Failed to calculate logical size")?,
                 };
 
                 // we cannot query current_logical_size.current_size() to know the current
@@ -1965,14 +1886,13 @@ impl Timeline {
                 // so that we prevent future callers from spawning this task
                 permit.forget();
                 Ok(())
-            }.in_current_span(),
+            },
         );
     }
 
     pub fn spawn_ondemand_logical_size_calculation(
         self: &Arc<Self>,
         lsn: Lsn,
-        cause: LogicalSizeCalculationCause,
         ctx: RequestContext,
         cancel: CancellationToken,
     ) -> oneshot::Receiver<Result<u64, CalculateLogicalSizeError>> {
@@ -1995,26 +1915,22 @@ impl Timeline {
             false,
             async move {
                 let res = self_clone
-                    .logical_size_calculation_task(lsn, cause, &ctx, cancel)
+                    .logical_size_calculation_task(lsn, &ctx, cancel)
                     .await;
                 let _ = sender.send(res).ok();
                 Ok(()) // Receiver is responsible for handling errors
-            }
-            .in_current_span(),
+            },
         );
         receiver
     }
 
-    #[instrument(skip_all)]
+    #[instrument(skip_all, fields(tenant = %self.tenant_id, timeline = %self.timeline_id))]
     async fn logical_size_calculation_task(
         self: &Arc<Self>,
         lsn: Lsn,
-        cause: LogicalSizeCalculationCause,
         ctx: &RequestContext,
         cancel: CancellationToken,
     ) -> Result<u64, CalculateLogicalSizeError> {
-        debug_assert_current_span_has_tenant_and_timeline_id();
-
         let mut timeline_state_updates = self.subscribe_for_state_updates();
         let self_calculation = Arc::clone(self);
 
@@ -2022,7 +1938,7 @@ impl Timeline {
             let cancel = cancel.child_token();
             let ctx = ctx.attached_child();
             self_calculation
-                .calculate_logical_size(lsn, cause, cancel, &ctx)
+                .calculate_logical_size(lsn, cancel, &ctx)
                 .await
         });
         let timeline_state_cancellation = async {
@@ -2077,7 +1993,6 @@ impl Timeline {
     pub async fn calculate_logical_size(
         &self,
         up_to_lsn: Lsn,
-        cause: LogicalSizeCalculationCause,
         cancel: CancellationToken,
         ctx: &RequestContext,
     ) -> Result<u64, CalculateLogicalSizeError> {
@@ -2111,15 +2026,7 @@ impl Timeline {
         if let Some(size) = self.current_logical_size.initialized_size(up_to_lsn) {
             return Ok(size);
         }
-        let storage_time_metrics = match cause {
-            LogicalSizeCalculationCause::Initial
-            | LogicalSizeCalculationCause::ConsumptionMetricsSyntheticSize
-            | LogicalSizeCalculationCause::TenantSizeHandler => &self.metrics.logical_size_histo,
-            LogicalSizeCalculationCause::EvictionTaskImitation => {
-                &self.metrics.imitate_logical_size_histo
-            }
-        };
-        let timer = storage_time_metrics.start_timer();
+        let timer = self.metrics.logical_size_histo.start_timer();
         let logical_size = self
             .get_current_logical_size_non_incremental(up_to_lsn, cancel, ctx)
             .await?;
@@ -2153,7 +2060,7 @@ impl Timeline {
     }
 
     fn find_layer(&self, layer_file_name: &str) -> Option<Arc<dyn PersistentLayer>> {
-        for historic_layer in self.layers.read().unwrap().iter_historic_layers() {
+        for historic_layer in self.layers.read().iter_historic_layers() {
             let historic_layer_name = historic_layer.filename().file_name();
             if layer_file_name == historic_layer_name {
                 return Some(historic_layer);
@@ -2236,13 +2143,31 @@ impl Timeline {
     ///
     /// This function takes the current timeline's locked LayerMap as an argument,
     /// so callers can avoid potential race conditions.
+    ///
+    // TODO: find a way to not hold the Timeline::layers lock during get_value_reconstruct_data calls.
+    //
+    // Since these calls do local disk IO, they'll be reasonably fast, until come disk IOPS bound.
+    // We have lots of headroom on current pageservers, so, it's going to be fine for now.
+    //
+    // We can't use tokio::sync::RwLock that easily because its guard is not Send, but,
+    // many tasks that access Timeline::layers run inside task_mgr tasks, which are required
+    // to be Send. It has been tried in origin/problame/asyncify-get-reconstruct-data--tokio-sync.
+    //
+    // The solution will probably be to have an immutable + multi-versioned layer map, allowing
+    // us to grab a snapshot of the layer map once and execute this function on the snapshot.
+    //
+    // Or, we could invest time to figure out whether we can drop the layer map lock after
+    // we grabbed the layer, do the IO, re-aquire, and continue the traversal.
+    //
+    // (Why is this allow() not inside the function? Because clippy doesn't respect it then).
+    #[allow(clippy::await_holding_lock)]
     async fn get_reconstruct_data(
         &self,
         key: Key,
         request_lsn: Lsn,
-        reconstruct_state: &mut ValueReconstructState,
+        mut reconstruct_state: ValueReconstructState,
         ctx: &RequestContext,
-    ) -> Result<(), PageReconstructError> {
+    ) -> Result<ValueReconstructState, PageReconstructError> {
         // Start from the current timeline.
         let mut timeline_owned;
         let mut timeline = self;
@@ -2269,12 +2194,12 @@ impl Timeline {
             // The function should have updated 'state'
             //info!("CALLED for {} at {}: {:?} with {} records, cached {}", key, cont_lsn, result, reconstruct_state.records.len(), cached_lsn);
             match result {
-                ValueReconstructResult::Complete => return Ok(()),
+                ValueReconstructResult::Complete => return Ok(reconstruct_state),
                 ValueReconstructResult::Continue => {
                     // If we reached an earlier cached page image, we're done.
                     if cont_lsn == cached_lsn + 1 {
                         self.metrics.materialized_page_cache_hit_counter.inc_by(1);
-                        return Ok(());
+                        return Ok(reconstruct_state);
                     }
                     if prev_lsn <= cont_lsn {
                         // Didn't make any progress in last iteration. Error out to avoid
@@ -2311,46 +2236,6 @@ impl Timeline {
                     Ok(timeline) => timeline,
                     Err(e) => return Err(PageReconstructError::from(e)),
                 };
-
-                // It's possible that the ancestor timeline isn't active yet, or
-                // is active but hasn't yet caught up to the branch point. Wait
-                // for it.
-                //
-                // This cannot happen while the pageserver is running normally,
-                // because you cannot create a branch from a point that isn't
-                // present in the pageserver yet. However, we don't wait for the
-                // branch point to be uploaded to cloud storage before creating
-                // a branch. I.e., the branch LSN need not be remote consistent
-                // for the branching operation to succeed.
-                //
-                // Hence, if we try to load a tenant in such a state where
-                // 1. the existence of the branch was persisted (in IndexPart and/or locally)
-                // 2. but the ancestor state is behind branch_lsn because it was not yet persisted
-                // then we will need to wait for the ancestor timeline to
-                // re-stream WAL up to branch_lsn before we access it.
-                //
-                // How can a tenant get in such a state?
-                // - ungraceful pageserver process exit
-                // - detach+attach => this is a bug, https://github.com/neondatabase/neon/issues/4219
-                //
-                // NB: this could be avoided by requiring
-                //   branch_lsn >= remote_consistent_lsn
-                // during branch creation.
-                match ancestor.wait_to_become_active(ctx).await {
-                    Ok(()) => {}
-                    Err(state) if state == TimelineState::Stopping => {
-                        return Err(PageReconstructError::AncestorStopping(ancestor.timeline_id));
-                    }
-                    Err(state) => {
-                        return Err(PageReconstructError::Other(anyhow::anyhow!(
-                            "Timeline {} will not become active. Current state: {:?}",
-                            ancestor.timeline_id,
-                            &state,
-                        )));
-                    }
-                }
-                ancestor.wait_lsn(timeline.ancestor_lsn, ctx).await?;
-
                 timeline_owned = ancestor;
                 timeline = &*timeline_owned;
                 prev_lsn = Lsn(u64::MAX);
@@ -2360,7 +2245,7 @@ impl Timeline {
             #[allow(clippy::never_loop)] // see comment at bottom of this loop
             'layer_map_search: loop {
                 let remote_layer = {
-                    let layers = timeline.layers.read().unwrap();
+                    let layers = timeline.layers.read();
 
                     // Check the open and frozen in-memory layers first, in order from newest
                     // to oldest.
@@ -2371,13 +2256,19 @@ impl Timeline {
                             // Get all the data needed to reconstruct the page version from this layer.
                             // But if we have an older cached page image, no need to go past that.
                             let lsn_floor = max(cached_lsn + 1, start_lsn);
-                            result = match open_layer.get_value_reconstruct_data(
-                                key,
-                                lsn_floor..cont_lsn,
-                                reconstruct_state,
-                                ctx,
-                            ) {
-                                Ok(result) => result,
+                            result = match Arc::clone(open_layer)
+                                .get_value_reconstruct_data(
+                                    key,
+                                    lsn_floor..cont_lsn,
+                                    reconstruct_state,
+                                    ctx.attached_child(),
+                                )
+                                .await
+                            {
+                                Ok((new_reconstruct_state, result)) => {
+                                    reconstruct_state = new_reconstruct_state;
+                                    result
+                                }
                                 Err(e) => return Err(PageReconstructError::from(e)),
                             };
                             cont_lsn = lsn_floor;
@@ -2397,13 +2288,19 @@ impl Timeline {
                         if cont_lsn > start_lsn {
                             //info!("CHECKING for {} at {} on frozen layer {}", key, cont_lsn, frozen_layer.filename().display());
                             let lsn_floor = max(cached_lsn + 1, start_lsn);
-                            result = match frozen_layer.get_value_reconstruct_data(
-                                key,
-                                lsn_floor..cont_lsn,
-                                reconstruct_state,
-                                ctx,
-                            ) {
-                                Ok(result) => result,
+                            result = match Arc::clone(frozen_layer)
+                                .get_value_reconstruct_data(
+                                    key,
+                                    lsn_floor..cont_lsn,
+                                    reconstruct_state,
+                                    ctx.attached_child(),
+                                )
+                                .await
+                            {
+                                Ok((new_reconstruct_state, result)) => {
+                                    reconstruct_state = new_reconstruct_state;
+                                    result
+                                }
                                 Err(e) => return Err(PageReconstructError::from(e)),
                             };
                             cont_lsn = lsn_floor;
@@ -2431,13 +2328,19 @@ impl Timeline {
                             // Get all the data needed to reconstruct the page version from this layer.
                             // But if we have an older cached page image, no need to go past that.
                             let lsn_floor = max(cached_lsn + 1, lsn_floor);
-                            result = match layer.get_value_reconstruct_data(
-                                key,
-                                lsn_floor..cont_lsn,
-                                reconstruct_state,
-                                ctx,
-                            ) {
-                                Ok(result) => result,
+                            result = match Arc::clone(&layer)
+                                .get_value_reconstruct_data(
+                                    key,
+                                    lsn_floor..cont_lsn,
+                                    reconstruct_state,
+                                    ctx.attached_child(),
+                                )
+                                .await
+                            {
+                                Ok((new_reconstruct_state, result)) => {
+                                    reconstruct_state = new_reconstruct_state;
+                                    result
+                                }
                                 Err(e) => return Err(PageReconstructError::from(e)),
                             };
                             cont_lsn = lsn_floor;
@@ -2540,7 +2443,7 @@ impl Timeline {
     /// Get a handle to the latest layer for appending.
     ///
     fn get_layer_for_write(&self, lsn: Lsn) -> anyhow::Result<Arc<InMemoryLayer>> {
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
 
         ensure!(lsn.is_aligned());
 
@@ -2613,7 +2516,7 @@ impl Timeline {
         } else {
             Some(self.write_lock.lock().unwrap())
         };
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
         if let Some(open_layer) = &layers.open_layer {
             let open_layer_rc = Arc::clone(open_layer);
             // Does this layer need freezing?
@@ -2651,7 +2554,7 @@ impl Timeline {
             let flush_counter = *layer_flush_start_rx.borrow();
             let result = loop {
                 let layer_to_flush = {
-                    let layers = self.layers.read().unwrap();
+                    let layers = self.layers.read();
                     layers.frozen_layers.front().cloned()
                     // drop 'layers' lock to allow concurrent reads and writes
                 };
@@ -2752,7 +2655,7 @@ impl Timeline {
         // The new on-disk layers are now in the layer map. We can remove the
         // in-memory layer from the map now.
         {
-            let mut layers = self.layers.write().unwrap();
+            let mut layers = self.layers.write();
             let l = layers.frozen_layers.pop_front();
 
             // Only one thread may call this function at a time (for this
@@ -2870,7 +2773,7 @@ impl Timeline {
 
         // Add it to the layer map
         let l = Arc::new(new_delta);
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
         let mut batch_updates = layers.batch_update();
         l.access_stats().record_residence_event(
             &batch_updates,
@@ -2925,33 +2828,9 @@ impl Timeline {
     fn time_for_new_image_layer(&self, partition: &KeySpace, lsn: Lsn) -> anyhow::Result<bool> {
         let threshold = self.get_image_creation_threshold();
 
-        let layers = self.layers.read().unwrap();
+        let layers = self.layers.read();
 
         let mut max_deltas = 0;
-        {
-            let wanted_image_layers = self.wanted_image_layers.lock().unwrap();
-            if let Some((cutoff_lsn, wanted)) = &*wanted_image_layers {
-                let img_range =
-                    partition.ranges.first().unwrap().start..partition.ranges.last().unwrap().end;
-                if wanted.overlaps(&img_range) {
-                    //
-                    // gc_timeline only pays attention to image layers that are older than the GC cutoff,
-                    // but create_image_layers creates image layers at last-record-lsn.
-                    // So it's possible that gc_timeline wants a new image layer to be created for a key range,
-                    // but the range is already covered by image layers at more recent LSNs. Before we
-                    // create a new image layer, check if the range is already covered at more recent LSNs.
-                    if !layers
-                        .image_layer_exists(&img_range, &(Lsn::min(lsn, *cutoff_lsn)..lsn + 1))?
-                    {
-                        debug!(
-                            "Force generation of layer {}-{} wanted by GC, cutoff={}, lsn={})",
-                            img_range.start, img_range.end, cutoff_lsn, lsn
-                        );
-                        return Ok(true);
-                    }
-                }
-            }
-        }
 
         for part_range in &partition.ranges {
             let image_coverage = layers.image_coverage(part_range, lsn)?;
@@ -3071,12 +2950,6 @@ impl Timeline {
                 image_layers.push(image_layer);
             }
         }
-        // All layers that the GC wanted us to create have now been created.
-        //
-        // It's possible that another GC cycle happened while we were compacting, and added
-        // something new to wanted_image_layers, and we now clear that before processing it.
-        // That's OK, because the next GC iteration will put it back in.
-        *self.wanted_image_layers.lock().unwrap() = None;
 
         // Sync the new layer to disk before adding it to the layer map, to make sure
         // we don't garbage collect something based on the new layer, before it has
@@ -3098,7 +2971,7 @@ impl Timeline {
 
         let mut layer_paths_to_upload = HashMap::with_capacity(image_layers.len());
 
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
         let mut updates = layers.batch_update();
         let timeline_path = self.conf.timeline_path(&self.timeline_id, &self.tenant_id);
         for l in image_layers {
@@ -3165,7 +3038,7 @@ impl Timeline {
         target_file_size: u64,
         ctx: &RequestContext,
     ) -> Result<CompactLevel0Phase1Result, CompactionError> {
-        let layers = self.layers.read().unwrap();
+        let layers = self.layers.read();
         let mut level0_deltas = layers.get_level0_deltas()?;
         drop(layers);
 
@@ -3288,7 +3161,7 @@ impl Timeline {
         // Determine N largest holes where N is number of compacted layers.
         let max_holes = deltas_to_compact.len();
         let last_record_lsn = self.get_last_record_lsn();
-        let layers = self.layers.read().unwrap(); // Is'n it better to hold original layers lock till here?
+        let layers = self.layers.read(); // Is'n it better to hold original layers lock till here?
         let min_hole_range = (target_file_size / page_cache::PAGE_SZ as u64) as i128;
         let min_hole_coverage_size = 3; // TODO: something more flexible?
 
@@ -3525,7 +3398,7 @@ impl Timeline {
                 .context("wait for layer upload ops to complete")?;
         }
 
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
         let mut updates = layers.batch_update();
         let mut new_layer_paths = HashMap::with_capacity(new_layers.len());
         for l in new_layers {
@@ -3774,7 +3647,6 @@ impl Timeline {
         }
 
         let mut layers_to_remove = Vec::new();
-        let mut wanted_image_layers = KeySpaceRandomAccum::default();
 
         // Scan all layers in the timeline (remote or on-disk).
         //
@@ -3785,7 +3657,7 @@ impl Timeline {
         // 4. newer on-disk image layers cover the layer's whole key range
         //
         // TODO holding a write lock is too agressive and avoidable
-        let mut layers = self.layers.write().unwrap();
+        let mut layers = self.layers.write();
         'outer: for l in layers.iter_historic_layers() {
             result.layers_total += 1;
 
@@ -3858,15 +3730,6 @@ impl Timeline {
                     "keeping {} because it is the latest layer",
                     l.filename().file_name()
                 );
-                // Collect delta key ranges that need image layers to allow garbage
-                // collecting the layers.
-                // It is not so obvious whether we need to propagate information only about
-                // delta layers. Image layers can form "stairs" preventing old image from been deleted.
-                // But image layers are in any case less sparse than delta layers. Also we need some
-                // protection from replacing recent image layers with new one after each GC iteration.
-                if l.is_incremental() && !LayerMap::is_l0(&*l) {
-                    wanted_image_layers.add_range(l.get_key_range());
-                }
                 result.layers_not_updated += 1;
                 continue 'outer;
             }
@@ -3879,10 +3742,6 @@ impl Timeline {
             );
             layers_to_remove.push(Arc::clone(&l));
         }
-        self.wanted_image_layers
-            .lock()
-            .unwrap()
-            .replace((new_gc_cutoff, wanted_image_layers.to_keyspace()));
 
         let mut updates = layers.batch_update();
         if !layers_to_remove.is_empty() {
@@ -4081,7 +3940,7 @@ impl Timeline {
 
                     // Download complete. Replace the RemoteLayer with the corresponding
                     // Delta- or ImageLayer in the layer map.
-                    let mut layers = self_clone.layers.write().unwrap();
+                    let mut layers = self_clone.layers.write();
                     let mut updates = layers.batch_update();
                     let new_layer = remote_layer.create_downloaded_layer(&updates, self_clone.conf, *size);
                     {
@@ -4239,7 +4098,7 @@ impl Timeline {
     ) {
         let mut downloads = Vec::new();
         {
-            let layers = self.layers.read().unwrap();
+            let layers = self.layers.read();
             layers
                 .iter_historic_layers()
                 .filter_map(|l| l.downcast_remote_layer())
@@ -4342,7 +4201,7 @@ impl LocalLayerInfoForDiskUsageEviction {
 
 impl Timeline {
     pub(crate) fn get_local_layers_for_disk_usage_eviction(&self) -> DiskUsageEvictionInfo {
-        let layers = self.layers.read().unwrap();
+        let layers = self.layers.read();
 
         let mut max_layer_size: Option<u64> = None;
         let mut resident_layers = Vec::new();
@@ -4484,6 +4343,12 @@ pub(crate) fn debug_assert_current_span_has_tenant_and_timeline_id() {}
 pub(crate) fn debug_assert_current_span_has_tenant_and_timeline_id() {
     use utils::tracing_span_assert;
 
+    pub static TENANT_ID_EXTRACTOR: once_cell::sync::Lazy<
+        tracing_span_assert::MultiNameExtractor<2>,
+    > = once_cell::sync::Lazy::new(|| {
+        tracing_span_assert::MultiNameExtractor::new("TenantId", ["tenant_id", "tenant"])
+    });
+
     pub static TIMELINE_ID_EXTRACTOR: once_cell::sync::Lazy<
         tracing_span_assert::MultiNameExtractor<2>,
     > = once_cell::sync::Lazy::new(|| {
@@ -4491,7 +4356,7 @@ pub(crate) fn debug_assert_current_span_has_tenant_and_timeline_id() {
     });
 
     match tracing_span_assert::check_fields_present([
-        &*super::TENANT_ID_EXTRACTOR,
+        &*TENANT_ID_EXTRACTOR,
         &*TIMELINE_ID_EXTRACTOR,
     ]) {
         Ok(()) => (),

pageserver/src/tenant/timeline/eviction_task.rs

@@ -30,7 +30,7 @@ use crate::{
     tenant::{
         config::{EvictionPolicy, EvictionPolicyLayerAccessThreshold},
         storage_layer::PersistentLayer,
-        LogicalSizeCalculationCause, Tenant,
+        Tenant,
     },
 };
 
@@ -120,13 +120,6 @@ impl Timeline {
                 }
                 let elapsed = start.elapsed();
                 crate::tenant::tasks::warn_when_period_overrun(elapsed, p.period, "eviction");
-                crate::metrics::EVICTION_ITERATION_DURATION
-                    .get_metric_with_label_values(&[
-                        &format!("{}", p.period.as_secs()),
-                        &format!("{}", p.threshold.as_secs()),
-                    ])
-                    .unwrap()
-                    .observe(elapsed.as_secs_f64());
                 ControlFlow::Continue(start + p.period)
             }
         }
@@ -185,7 +178,7 @@ impl Timeline {
         // We don't want to hold the layer map lock during eviction.
         // So, we just need to deal with this.
         let candidates: Vec<Arc<dyn PersistentLayer>> = {
-            let layers = self.layers.read().unwrap();
+            let layers = self.layers.read();
             let mut candidates = Vec::new();
             for hist_layer in layers.iter_historic_layers() {
                 if hist_layer.is_remote_layer() {
@@ -342,12 +335,7 @@ impl Timeline {
 
         // imitiate on-restart initial logical size
         let size = self
-            .calculate_logical_size(
-                lsn,
-                LogicalSizeCalculationCause::EvictionTaskImitation,
-                cancel.clone(),
-                ctx,
-            )
+            .calculate_logical_size(lsn, cancel.clone(), ctx)
             .instrument(info_span!("calculate_logical_size"))
             .await;
 
@@ -419,15 +407,9 @@ impl Timeline {
             .inner();
 
         let mut throwaway_cache = HashMap::new();
-        let gather = crate::tenant::size::gather_inputs(
-            tenant,
-            limit,
-            None,
-            &mut throwaway_cache,
-            LogicalSizeCalculationCause::EvictionTaskImitation,
-            ctx,
-        )
-        .instrument(info_span!("gather_inputs"));
+        let gather =
+            crate::tenant::size::gather_inputs(tenant, limit, None, &mut throwaway_cache, ctx)
+                .instrument(info_span!("gather_inputs"));
 
         tokio::select! {
             _ = cancel.cancelled() => {}

pageserver/src/tenant/timeline/walreceiver.rs

@@ -29,14 +29,16 @@ use crate::tenant::timeline::walreceiver::connection_manager::{
     connection_manager_loop_step, ConnectionManagerState,
 };
 
+use anyhow::Context;
 use std::future::Future;
 use std::num::NonZeroU64;
 use std::ops::ControlFlow;
-use std::sync::Arc;
+use std::sync::atomic::{self, AtomicBool};
+use std::sync::{Arc, Weak};
 use std::time::Duration;
 use storage_broker::BrokerClientChannel;
 use tokio::select;
-use tokio::sync::watch;
+use tokio::sync::{watch, RwLock};
 use tokio_util::sync::CancellationToken;
 use tracing::*;
 
@@ -60,23 +62,46 @@ pub struct WalReceiverConf {
 
 pub struct WalReceiver {
     timeline: TenantTimelineId,
-    manager_status: Arc<std::sync::RwLock<Option<ConnectionManagerStatus>>>,
+    timeline_ref: Weak<Timeline>,
+    conf: WalReceiverConf,
+    started: AtomicBool,
+    manager_status: Arc<RwLock<Option<ConnectionManagerStatus>>>,
 }
 
 impl WalReceiver {
-    pub fn start(
-        timeline: Arc<Timeline>,
+    pub fn new(
+        timeline: TenantTimelineId,
+        timeline_ref: Weak<Timeline>,
         conf: WalReceiverConf,
-        mut broker_client: BrokerClientChannel,
-        ctx: &RequestContext,
     ) -> Self {
+        Self {
+            timeline,
+            timeline_ref,
+            conf,
+            started: AtomicBool::new(false),
+            manager_status: Arc::new(RwLock::new(None)),
+        }
+    }
+
+    pub fn start(
+        &self,
+        ctx: &RequestContext,
+        mut broker_client: BrokerClientChannel,
+    ) -> anyhow::Result<()> {
+        if self.started.load(atomic::Ordering::Acquire) {
+            anyhow::bail!("Wal receiver is already started");
+        }
+
+        let timeline = self.timeline_ref.upgrade().with_context(|| {
+            format!("walreceiver start on a dropped timeline {}", self.timeline)
+        })?;
+
         let tenant_id = timeline.tenant_id;
         let timeline_id = timeline.timeline_id;
         let walreceiver_ctx =
             ctx.detached_child(TaskKind::WalReceiverManager, DownloadBehavior::Error);
-
-        let loop_status = Arc::new(std::sync::RwLock::new(None));
-        let manager_status = Arc::clone(&loop_status);
+        let wal_receiver_conf = self.conf.clone();
+        let loop_status = Arc::clone(&self.manager_status);
         task_mgr::spawn(
             WALRECEIVER_RUNTIME.handle(),
             TaskKind::WalReceiverManager,
@@ -88,7 +113,7 @@ impl WalReceiver {
                 info!("WAL receiver manager started, connecting to broker");
                 let mut connection_manager_state = ConnectionManagerState::new(
                     timeline,
-                    conf,
+                    wal_receiver_conf,
                 );
                 loop {
                     select! {
@@ -112,29 +137,29 @@ impl WalReceiver {
                 }
 
                 connection_manager_state.shutdown().await;
-                *loop_status.write().unwrap() = None;
+                *loop_status.write().await = None;
                 Ok(())
             }
             .instrument(info_span!(parent: None, "wal_connection_manager", tenant = %tenant_id, timeline = %timeline_id))
         );
 
-        Self {
-            timeline: TenantTimelineId::new(tenant_id, timeline_id),
-            manager_status,
-        }
+        self.started.store(true, atomic::Ordering::Release);
+
+        Ok(())
     }
 
-    pub async fn stop(self) {
+    pub async fn stop(&self) {
         task_mgr::shutdown_tasks(
             Some(TaskKind::WalReceiverManager),
             Some(self.timeline.tenant_id),
             Some(self.timeline.timeline_id),
         )
         .await;
+        self.started.store(false, atomic::Ordering::Release);
     }
 
-    pub(super) fn status(&self) -> Option<ConnectionManagerStatus> {
-        self.manager_status.read().unwrap().clone()
+    pub(super) async fn status(&self) -> Option<ConnectionManagerStatus> {
+        self.manager_status.read().await.clone()
     }
 }
 

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

@@ -28,7 +28,8 @@ use storage_broker::proto::SubscribeSafekeeperInfoRequest;
 use storage_broker::proto::TenantTimelineId as ProtoTenantTimelineId;
 use storage_broker::BrokerClientChannel;
 use storage_broker::Streaming;
-use tokio::select;
+use tokio::sync::RwLock;
+use tokio::{select, sync::watch};
 use tracing::*;
 
 use crate::{exponential_backoff, DEFAULT_BASE_BACKOFF_SECONDS, DEFAULT_MAX_BACKOFF_SECONDS};
@@ -47,15 +48,15 @@ pub(super) async fn connection_manager_loop_step(
     broker_client: &mut BrokerClientChannel,
     connection_manager_state: &mut ConnectionManagerState,
     ctx: &RequestContext,
-    manager_status: &std::sync::RwLock<Option<ConnectionManagerStatus>>,
+    manager_status: &RwLock<Option<ConnectionManagerStatus>>,
 ) -> ControlFlow<(), ()> {
-    match connection_manager_state
+    let mut timeline_state_updates = connection_manager_state
         .timeline
-        .wait_to_become_active(ctx)
-        .await
-    {
-        Ok(()) => {}
-        Err(_) => {
+        .subscribe_for_state_updates();
+
+    match wait_for_active_timeline(&mut timeline_state_updates).await {
+        ControlFlow::Continue(()) => {}
+        ControlFlow::Break(()) => {
             info!("Timeline dropped state updates sender before becoming active, stopping wal connection manager loop");
             return ControlFlow::Break(());
         }
@@ -71,10 +72,6 @@ pub(super) async fn connection_manager_loop_step(
         timeline_id: connection_manager_state.timeline.timeline_id,
     };
 
-    let mut timeline_state_updates = connection_manager_state
-        .timeline
-        .subscribe_for_state_updates();
-
     // Subscribe to the broker updates. Stream shares underlying TCP connection
     // with other streams on this client (other connection managers). When
     // object goes out of scope, stream finishes in drop() automatically.
@@ -194,7 +191,35 @@ pub(super) async fn connection_manager_loop_step(
                 .change_connection(new_candidate, ctx)
                 .await
         }
-        *manager_status.write().unwrap() = Some(connection_manager_state.manager_status());
+        *manager_status.write().await = Some(connection_manager_state.manager_status());
+    }
+}
+
+async fn wait_for_active_timeline(
+    timeline_state_updates: &mut watch::Receiver<TimelineState>,
+) -> ControlFlow<(), ()> {
+    let current_state = *timeline_state_updates.borrow();
+    if current_state == TimelineState::Active {
+        return ControlFlow::Continue(());
+    }
+
+    loop {
+        match timeline_state_updates.changed().await {
+            Ok(()) => {
+                let new_state = *timeline_state_updates.borrow();
+                match new_state {
+                    TimelineState::Active => {
+                        debug!("Timeline state changed to active, continuing the walreceiver connection manager");
+                        return ControlFlow::Continue(());
+                    }
+                    state => {
+                        debug!("Not running the walreceiver connection manager, timeline is not active: {state:?}");
+                        continue;
+                    }
+                }
+            }
+            Err(_sender_dropped_error) => return ControlFlow::Break(()),
+        }
     }
 }
 
@@ -1308,8 +1333,9 @@ mod tests {
     async fn dummy_state(harness: &TenantHarness<'_>) -> ConnectionManagerState {
         let (tenant, ctx) = harness.load().await;
         let timeline = tenant
-            .create_test_timeline(TIMELINE_ID, Lsn(0), crate::DEFAULT_PG_VERSION, &ctx)
+            .create_empty_timeline(TIMELINE_ID, Lsn(0), crate::DEFAULT_PG_VERSION, &ctx)
             .expect("Failed to create an empty timeline for dummy wal connection manager");
+        let timeline = timeline.initialize(&ctx).unwrap();
 
         ConnectionManagerState {
             id: TenantTimelineId {

pageserver/src/walingest.rs

@@ -305,15 +305,6 @@ impl<'a> WalIngest<'a> {
                     self.checkpoint_modified = true;
                 }
             }
-        } else if decoded.xl_rmid == pg_constants::RM_LOGICALMSG_ID {
-            let info = decoded.xl_info & pg_constants::XLR_RMGR_INFO_MASK;
-            if info == pg_constants::XLOG_LOGICAL_MESSAGE {
-                // This is a convenient way to make the WAL ingestion pause at
-                // particular point in the WAL. For more fine-grained control,
-                // we could peek into the message and only pause if it contains
-                // a particular string, for example, but this is enough for now.
-                utils::failpoint_sleep_millis_async!("wal-ingest-logical-message-sleep");
-            }
         }
 
         // Iterate through all the blocks that the record modifies, and

pageserver/src/walrecord.rs

@@ -379,6 +379,17 @@ impl XlXactParsedRecord {
                 });
             }
         }
+        if xinfo & pg_constants::XACT_XINFO_HAS_INVALS != 0 {
+            let nmsgs = buf.get_i32_le();
+            for _i in 0..nmsgs {
+                let sizeof_shared_invalidation_message = 0;
+                buf.advance(sizeof_shared_invalidation_message);
+            }
+        }
+        if xinfo & pg_constants::XACT_XINFO_HAS_TWOPHASE != 0 {
+            xid = buf.get_u32_le();
+            trace!("XLOG_XACT_COMMIT-XACT_XINFO_HAS_TWOPHASE");
+        }
 
         if xinfo & postgres_ffi::v15::bindings::XACT_XINFO_HAS_DROPPED_STATS != 0 {
             let nitems = buf.get_i32_le();
@@ -386,23 +397,7 @@ impl XlXactParsedRecord {
                 "XLOG_XACT_COMMIT-XACT_XINFO_HAS_DROPPED_STAT nitems {}",
                 nitems
             );
-            let sizeof_xl_xact_stats_item = 12;
-            buf.advance((nitems * sizeof_xl_xact_stats_item).try_into().unwrap());
-        }
-
-        if xinfo & pg_constants::XACT_XINFO_HAS_INVALS != 0 {
-            let nmsgs = buf.get_i32_le();
-            let sizeof_shared_invalidation_message = 16;
-            buf.advance(
-                (nmsgs * sizeof_shared_invalidation_message)
-                    .try_into()
-                    .unwrap(),
-            );
-        }
-
-        if xinfo & pg_constants::XACT_XINFO_HAS_TWOPHASE != 0 {
-            xid = buf.get_u32_le();
-            debug!("XLOG_XACT_COMMIT-XACT_XINFO_HAS_TWOPHASE xid {}", xid);
+            //FIXME: do we need to handle dropped stats here?
         }
 
         XlXactParsedRecord {

pgxn/neon/Makefile

@@ -11,12 +11,10 @@ OBJS = \
 	pagestore_smgr.o \
 	relsize_cache.o \
 	walproposer.o \
-	walproposer_utils.o \
-	control_plane_connector.o
+	walproposer_utils.o
 
 PG_CPPFLAGS = -I$(libpq_srcdir)
 SHLIB_LINK_INTERNAL = $(libpq)
-SHLIB_LINK = -lcurl
 
 EXTENSION = neon
 DATA = neon--1.0.sql

pgxn/neon/control_plane_connector.c

@@ -1,830 +0,0 @@
-/*-------------------------------------------------------------------------
- *
- * control_plane_connector.c
- *	  Captures updates to roles/databases using ProcessUtility_hook and
- *        sends them to the control ProcessUtility_hook. The changes are sent
- *        via HTTP to the URL specified by the GUC neon.console_url when the
- *        transaction commits. Forwarding may be disabled temporarily by
- *        setting neon.forward_ddl to false.
- *
- *        Currently, the transaction may abort AFTER
- *        changes have already been forwarded, and that case is not handled.
- *        Subtransactions are handled using a stack of hash tables, which
- *        accumulate changes. On subtransaction commit, the top of the stack
- *        is merged with the table below it.
- *
- * IDENTIFICATION
- *	 contrib/neon/control_plane_connector.c
- *
- *-------------------------------------------------------------------------
- */
-#include "postgres.h"
-#include "tcop/pquery.h"
-#include "tcop/utility.h"
-#include "access/xact.h"
-#include "utils/hsearch.h"
-#include "utils/memutils.h"
-#include "commands/defrem.h"
-#include "miscadmin.h"
-#include "utils/acl.h"
-#include "fmgr.h"
-#include "utils/guc.h"
-#include "port.h"
-#include <curl/curl.h>
-#include "utils/jsonb.h"
-
-static ProcessUtility_hook_type PreviousProcessUtilityHook = NULL;
-
-/* GUCs */
-static char *ConsoleURL = NULL;
-static bool ForwardDDL = true;
-
-/* Curl structures for sending the HTTP requests */
-static CURL * CurlHandle;
-static struct curl_slist *ContentHeader = NULL;
-
-/*
- * CURL docs say that this buffer must exist until we call curl_easy_cleanup
- * (which we never do), so we make this a static
- */
-static char CurlErrorBuf[CURL_ERROR_SIZE];
-
-typedef enum
-{
-	Op_Set,						/* An upsert: Either a creation or an alter */
-	Op_Delete,
-}			OpType;
-
-typedef struct
-{
-	char		name[NAMEDATALEN];
-	Oid			owner;
-	char		old_name[NAMEDATALEN];
-	OpType		type;
-}			DbEntry;
-
-typedef struct
-{
-	char		name[NAMEDATALEN];
-	char		old_name[NAMEDATALEN];
-	const char *password;
-	OpType		type;
-}			RoleEntry;
-
-/*
- * We keep one of these for each subtransaction in a stack. When a subtransaction
- * commits, we merge the top of the stack into the table below it. It is allocated in the
- * subtransaction's context.
- */
-typedef struct DdlHashTable
-{
-	struct DdlHashTable *prev_table;
-	HTAB	   *db_table;
-	HTAB	   *role_table;
-}			DdlHashTable;
-
-static DdlHashTable RootTable;
-static DdlHashTable * CurrentDdlTable = &RootTable;
-
-static void
-PushKeyValue(JsonbParseState **state, char *key, char *value)
-{
-	JsonbValue	k,
-				v;
-
-	k.type = jbvString;
-	k.val.string.len = strlen(key);
-	k.val.string.val = key;
-	v.type = jbvString;
-	v.val.string.len = strlen(value);
-	v.val.string.val = value;
-	pushJsonbValue(state, WJB_KEY, &k);
-	pushJsonbValue(state, WJB_VALUE, &v);
-}
-
-static char *
-ConstructDeltaMessage()
-{
-	JsonbParseState *state = NULL;
-
-	pushJsonbValue(&state, WJB_BEGIN_OBJECT, NULL);
-	if (RootTable.db_table)
-	{
-		JsonbValue	dbs;
-
-		dbs.type = jbvString;
-		dbs.val.string.val = "dbs";
-		dbs.val.string.len = strlen(dbs.val.string.val);
-		pushJsonbValue(&state, WJB_KEY, &dbs);
-		pushJsonbValue(&state, WJB_BEGIN_ARRAY, NULL);
-
-		HASH_SEQ_STATUS status;
-		DbEntry    *entry;
-
-		hash_seq_init(&status, RootTable.db_table);
-		while ((entry = hash_seq_search(&status)) != NULL)
-		{
-			pushJsonbValue(&state, WJB_BEGIN_OBJECT, NULL);
-			PushKeyValue(&state, "op", entry->type == Op_Set ? "set" : "del");
-			PushKeyValue(&state, "name", entry->name);
-			if (entry->owner != InvalidOid)
-			{
-				PushKeyValue(&state, "owner", GetUserNameFromId(entry->owner, false));
-			}
-			if (entry->old_name[0] != '\0')
-			{
-				PushKeyValue(&state, "old_name", entry->old_name);
-			}
-			pushJsonbValue(&state, WJB_END_OBJECT, NULL);
-		}
-		pushJsonbValue(&state, WJB_END_ARRAY, NULL);
-	}
-
-	if (RootTable.role_table)
-	{
-		JsonbValue	roles;
-
-		roles.type = jbvString;
-		roles.val.string.val = "roles";
-		roles.val.string.len = strlen(roles.val.string.val);
-		pushJsonbValue(&state, WJB_KEY, &roles);
-		pushJsonbValue(&state, WJB_BEGIN_ARRAY, NULL);
-
-		HASH_SEQ_STATUS status;
-		RoleEntry  *entry;
-
-		hash_seq_init(&status, RootTable.role_table);
-		while ((entry = hash_seq_search(&status)) != NULL)
-		{
-			pushJsonbValue(&state, WJB_BEGIN_OBJECT, NULL);
-			PushKeyValue(&state, "op", entry->type == Op_Set ? "set" : "del");
-			PushKeyValue(&state, "name", entry->name);
-			if (entry->password)
-			{
-				PushKeyValue(&state, "password", (char *) entry->password);
-			}
-			if (entry->old_name[0] != '\0')
-			{
-				PushKeyValue(&state, "old_name", entry->old_name);
-			}
-			pushJsonbValue(&state, WJB_END_OBJECT, NULL);
-		}
-		pushJsonbValue(&state, WJB_END_ARRAY, NULL);
-	}
-	JsonbValue *result = pushJsonbValue(&state, WJB_END_OBJECT, NULL);
-	Jsonb	   *jsonb = JsonbValueToJsonb(result);
-
-	return JsonbToCString(NULL, &jsonb->root, 0 /* estimated_len */ );
-}
-
-#define ERROR_SIZE 1024
-
-typedef struct
-{
-	char		str[ERROR_SIZE];
-	size_t		size;
-}			ErrorString;
-
-static size_t
-ErrorWriteCallback(char *ptr, size_t size, size_t nmemb, void *userdata)
-{
-	/* Docs say size is always 1 */
-	ErrorString *str = userdata;
-
-	size_t		to_write = nmemb;
-
-	/* +1 for null terminator */
-	if (str->size + nmemb + 1 >= ERROR_SIZE)
-		to_write = ERROR_SIZE - str->size - 1;
-
-	/* Ignore everyrthing past the first ERROR_SIZE bytes */
-	if (to_write == 0)
-		return nmemb;
-	memcpy(str->str + str->size, ptr, to_write);
-	str->size += to_write;
-	str->str[str->size] = '\0';
-	return nmemb;
-}
-
-static void
-SendDeltasToControlPlane()
-{
-	if (!RootTable.db_table && !RootTable.role_table)
-		return;
-	if (!ConsoleURL)
-	{
-		elog(LOG, "ConsoleURL not set, skipping forwarding");
-		return;
-	}
-	if (!ForwardDDL)
-		return;
-
-	char	   *message = ConstructDeltaMessage();
-	ErrorString str = {};
-
-	curl_easy_setopt(CurlHandle, CURLOPT_CUSTOMREQUEST, "PATCH");
-	curl_easy_setopt(CurlHandle, CURLOPT_HTTPHEADER, ContentHeader);
-	curl_easy_setopt(CurlHandle, CURLOPT_POSTFIELDS, message);
-	curl_easy_setopt(CurlHandle, CURLOPT_URL, ConsoleURL);
-	curl_easy_setopt(CurlHandle, CURLOPT_ERRORBUFFER, CurlErrorBuf);
-	curl_easy_setopt(CurlHandle, CURLOPT_TIMEOUT, 3L /* seconds */ );
-	curl_easy_setopt(CurlHandle, CURLOPT_WRITEDATA, &str);
-	curl_easy_setopt(CurlHandle, CURLOPT_WRITEFUNCTION, ErrorWriteCallback);
-
-	const int	num_retries = 5;
-	int			curl_status;
-
-	for (int i = 0; i < num_retries; i++)
-	{
-		if ((curl_status = curl_easy_perform(CurlHandle)) == 0)
-			break;
-		elog(LOG, "Curl request failed on attempt %d: %s", i, CurlErrorBuf);
-		pg_usleep(1000 * 1000);
-	}
-	if (curl_status != 0)
-	{
-		elog(ERROR, "Failed to perform curl request: %s", CurlErrorBuf);
-	}
-	else
-	{
-		long		response_code;
-
-		if (curl_easy_getinfo(CurlHandle, CURLINFO_RESPONSE_CODE, &response_code) != CURLE_UNKNOWN_OPTION)
-		{
-			bool		error_exists = str.size != 0;
-
-			if (response_code != 200)
-			{
-				if (error_exists)
-				{
-					elog(ERROR,
-						 "Received HTTP code %ld from control plane: %s",
-						 response_code,
-						 str.str);
-				}
-				else
-				{
-					elog(ERROR,
-						 "Received HTTP code %ld from control plane",
-						 response_code);
-				}
-			}
-		}
-	}
-}
-
-static void
-InitDbTableIfNeeded()
-{
-	if (!CurrentDdlTable->db_table)
-	{
-		HASHCTL		db_ctl = {};
-
-		db_ctl.keysize = NAMEDATALEN;
-		db_ctl.entrysize = sizeof(DbEntry);
-		db_ctl.hcxt = CurTransactionContext;
-		CurrentDdlTable->db_table = hash_create(
-												"Dbs Created",
-												4,
-												&db_ctl,
-												HASH_ELEM | HASH_STRINGS | HASH_CONTEXT);
-	}
-}
-
-static void
-InitRoleTableIfNeeded()
-{
-	if (!CurrentDdlTable->role_table)
-	{
-		HASHCTL		role_ctl = {};
-
-		role_ctl.keysize = NAMEDATALEN;
-		role_ctl.entrysize = sizeof(RoleEntry);
-		role_ctl.hcxt = CurTransactionContext;
-		CurrentDdlTable->role_table = hash_create(
-												  "Roles Created",
-												  4,
-												  &role_ctl,
-												  HASH_ELEM | HASH_STRINGS | HASH_CONTEXT);
-	}
-}
-
-static void
-PushTable()
-{
-	DdlHashTable *new_table = MemoryContextAlloc(CurTransactionContext, sizeof(DdlHashTable));
-
-	new_table->prev_table = CurrentDdlTable;
-	new_table->role_table = NULL;
-	new_table->db_table = NULL;
-	CurrentDdlTable = new_table;
-}
-
-static void
-MergeTable()
-{
-	DdlHashTable *old_table = CurrentDdlTable;
-
-	CurrentDdlTable = old_table->prev_table;
-
-	if (old_table->db_table)
-	{
-		InitDbTableIfNeeded();
-		DbEntry    *entry;
-		HASH_SEQ_STATUS status;
-
-		hash_seq_init(&status, old_table->db_table);
-		while ((entry = hash_seq_search(&status)) != NULL)
-		{
-			DbEntry    *to_write = hash_search(
-											   CurrentDdlTable->db_table,
-											   entry->name,
-											   HASH_ENTER,
-											   NULL);
-
-			to_write->type = entry->type;
-			if (entry->owner != InvalidOid)
-				to_write->owner = entry->owner;
-			strlcpy(to_write->old_name, entry->old_name, NAMEDATALEN);
-			if (entry->old_name[0] != '\0')
-			{
-				bool		found_old = false;
-				DbEntry    *old = hash_search(
-											  CurrentDdlTable->db_table,
-											  entry->old_name,
-											  HASH_FIND,
-											  &found_old);
-
-				if (found_old)
-				{
-					if (old->old_name[0] != '\0')
-						strlcpy(to_write->old_name, old->old_name, NAMEDATALEN);
-					else
-						strlcpy(to_write->old_name, entry->old_name, NAMEDATALEN);
-					hash_search(
-								CurrentDdlTable->db_table,
-								entry->old_name,
-								HASH_REMOVE,
-								NULL);
-				}
-			}
-		}
-		hash_destroy(old_table->db_table);
-	}
-
-	if (old_table->role_table)
-	{
-		InitRoleTableIfNeeded();
-		RoleEntry  *entry;
-		HASH_SEQ_STATUS status;
-
-		hash_seq_init(&status, old_table->role_table);
-		while ((entry = hash_seq_search(&status)) != NULL)
-		{
-			RoleEntry  *to_write = hash_search(
-											   CurrentDdlTable->role_table,
-											   entry->name,
-											   HASH_ENTER,
-											   NULL);
-
-			to_write->type = entry->type;
-			if (entry->password)
-				to_write->password = entry->password;
-			strlcpy(to_write->old_name, entry->old_name, NAMEDATALEN);
-			if (entry->old_name[0] != '\0')
-			{
-				bool		found_old = false;
-				RoleEntry  *old = hash_search(
-											  CurrentDdlTable->role_table,
-											  entry->old_name,
-											  HASH_FIND,
-											  &found_old);
-
-				if (found_old)
-				{
-					if (old->old_name[0] != '\0')
-						strlcpy(to_write->old_name, old->old_name, NAMEDATALEN);
-					else
-						strlcpy(to_write->old_name, entry->old_name, NAMEDATALEN);
-					hash_search(CurrentDdlTable->role_table,
-								entry->old_name,
-								HASH_REMOVE,
-								NULL);
-				}
-			}
-		}
-		hash_destroy(old_table->role_table);
-	}
-}
-
-static void
-PopTable()
-{
-	/*
-	 * Current table gets freed because it is allocated in aborted
-	 * subtransaction's memory context.
-	 */
-	CurrentDdlTable = CurrentDdlTable->prev_table;
-}
-
-static void
-NeonSubXactCallback(
-					SubXactEvent event,
-					SubTransactionId mySubid,
-					SubTransactionId parentSubid,
-					void *arg)
-{
-	switch (event)
-	{
-		case SUBXACT_EVENT_START_SUB:
-			return PushTable();
-		case SUBXACT_EVENT_COMMIT_SUB:
-			return MergeTable();
-		case SUBXACT_EVENT_ABORT_SUB:
-			return PopTable();
-		default:
-			return;
-	}
-}
-
-static void
-NeonXactCallback(XactEvent event, void *arg)
-{
-	if (event == XACT_EVENT_PRE_COMMIT || event == XACT_EVENT_PARALLEL_PRE_COMMIT)
-	{
-		SendDeltasToControlPlane();
-	}
-	RootTable.role_table = NULL;
-	RootTable.db_table = NULL;
-	Assert(CurrentDdlTable == &RootTable);
-}
-
-static void
-HandleCreateDb(CreatedbStmt *stmt)
-{
-	InitDbTableIfNeeded();
-	DefElem    *downer = NULL;
-	ListCell   *option;
-
-	foreach(option, stmt->options)
-	{
-		DefElem    *defel = lfirst(option);
-
-		if (strcmp(defel->defname, "owner") == 0)
-			downer = defel;
-	}
-	bool		found = false;
-	DbEntry    *entry = hash_search(
-									CurrentDdlTable->db_table,
-									stmt->dbname,
-									HASH_ENTER,
-									&found);
-
-	if (!found)
-		memset(entry->old_name, 0, sizeof(entry->old_name));
-
-	entry->type = Op_Set;
-	if (downer && downer->arg)
-		entry->owner = get_role_oid(defGetString(downer), false);
-	else
-		entry->owner = GetUserId();
-}
-
-static void
-HandleAlterOwner(AlterOwnerStmt *stmt)
-{
-	if (stmt->objectType != OBJECT_DATABASE)
-		return;
-	InitDbTableIfNeeded();
-	const char *name = strVal(stmt->object);
-	bool		found = false;
-	DbEntry    *entry = hash_search(
-									CurrentDdlTable->db_table,
-									name,
-									HASH_ENTER,
-									&found);
-
-	if (!found)
-		memset(entry->old_name, 0, sizeof(entry->old_name));
-
-	entry->owner = get_role_oid(get_rolespec_name(stmt->newowner), false);
-	entry->type = Op_Set;
-}
-
-static void
-HandleDbRename(RenameStmt *stmt)
-{
-	Assert(stmt->renameType == OBJECT_DATABASE);
-	InitDbTableIfNeeded();
-	bool		found = false;
-	DbEntry    *entry = hash_search(
-									CurrentDdlTable->db_table,
-									stmt->subname,
-									HASH_FIND,
-									&found);
-	DbEntry    *entry_for_new_name = hash_search(
-												 CurrentDdlTable->db_table,
-												 stmt->newname,
-												 HASH_ENTER,
-												 NULL);
-
-	entry_for_new_name->type = Op_Set;
-	if (found)
-	{
-		if (entry->old_name[0] != '\0')
-			strlcpy(entry_for_new_name->old_name, entry->old_name, NAMEDATALEN);
-		else
-			strlcpy(entry_for_new_name->old_name, entry->name, NAMEDATALEN);
-		entry_for_new_name->owner = entry->owner;
-		hash_search(
-					CurrentDdlTable->db_table,
-					stmt->subname,
-					HASH_REMOVE,
-					NULL);
-	}
-	else
-	{
-		strlcpy(entry_for_new_name->old_name, stmt->subname, NAMEDATALEN);
-		entry_for_new_name->owner = InvalidOid;
-	}
-}
-
-static void
-HandleDropDb(DropdbStmt *stmt)
-{
-	InitDbTableIfNeeded();
-	bool		found = false;
-	DbEntry    *entry = hash_search(
-									CurrentDdlTable->db_table,
-									stmt->dbname,
-									HASH_ENTER,
-									&found);
-
-	entry->type = Op_Delete;
-	entry->owner = InvalidOid;
-	if (!found)
-		memset(entry->old_name, 0, sizeof(entry->old_name));
-}
-
-static void
-HandleCreateRole(CreateRoleStmt *stmt)
-{
-	InitRoleTableIfNeeded();
-	bool		found = false;
-	RoleEntry  *entry = hash_search(
-									CurrentDdlTable->role_table,
-									stmt->role,
-									HASH_ENTER,
-									&found);
-	DefElem    *dpass = NULL;
-	ListCell   *option;
-
-	foreach(option, stmt->options)
-	{
-		DefElem    *defel = lfirst(option);
-
-		if (strcmp(defel->defname, "password") == 0)
-			dpass = defel;
-	}
-	if (!found)
-		memset(entry->old_name, 0, sizeof(entry->old_name));
-	if (dpass && dpass->arg)
-		entry->password = MemoryContextStrdup(CurTransactionContext, strVal(dpass->arg));
-	else
-		entry->password = NULL;
-	entry->type = Op_Set;
-}
-
-static void
-HandleAlterRole(AlterRoleStmt *stmt)
-{
-	InitRoleTableIfNeeded();
-	DefElem    *dpass = NULL;
-	ListCell   *option;
-
-	foreach(option, stmt->options)
-	{
-		DefElem    *defel = lfirst(option);
-
-		if (strcmp(defel->defname, "password") == 0)
-			dpass = defel;
-	}
-	/* We only care about updates to the password */
-	if (!dpass)
-		return;
-	bool		found = false;
-	RoleEntry  *entry = hash_search(
-									CurrentDdlTable->role_table,
-									stmt->role->rolename,
-									HASH_ENTER,
-									&found);
-
-	if (!found)
-		memset(entry->old_name, 0, sizeof(entry->old_name));
-	if (dpass->arg)
-		entry->password = MemoryContextStrdup(CurTransactionContext, strVal(dpass->arg));
-	else
-		entry->password = NULL;
-	entry->type = Op_Set;
-}
-
-static void
-HandleRoleRename(RenameStmt *stmt)
-{
-	InitRoleTableIfNeeded();
-	Assert(stmt->renameType == OBJECT_ROLE);
-	bool		found = false;
-	RoleEntry  *entry = hash_search(
-									CurrentDdlTable->role_table,
-									stmt->subname,
-									HASH_FIND,
-									&found);
-
-	RoleEntry  *entry_for_new_name = hash_search(
-												 CurrentDdlTable->role_table,
-												 stmt->newname,
-												 HASH_ENTER,
-												 NULL);
-
-	entry_for_new_name->type = Op_Set;
-	if (found)
-	{
-		if (entry->old_name[0] != '\0')
-			strlcpy(entry_for_new_name->old_name, entry->old_name, NAMEDATALEN);
-		else
-			strlcpy(entry_for_new_name->old_name, entry->name, NAMEDATALEN);
-		entry_for_new_name->password = entry->password;
-		hash_search(
-					CurrentDdlTable->role_table,
-					entry->name,
-					HASH_REMOVE,
-					NULL);
-	}
-	else
-	{
-		strlcpy(entry_for_new_name->old_name, stmt->subname, NAMEDATALEN);
-		entry_for_new_name->password = NULL;
-	}
-}
-
-static void
-HandleDropRole(DropRoleStmt *stmt)
-{
-	InitRoleTableIfNeeded();
-	ListCell   *item;
-
-	foreach(item, stmt->roles)
-	{
-		RoleSpec   *spec = lfirst(item);
-		bool		found = false;
-		RoleEntry  *entry = hash_search(
-										CurrentDdlTable->role_table,
-										spec->rolename,
-										HASH_ENTER,
-										&found);
-
-		entry->type = Op_Delete;
-		entry->password = NULL;
-		if (!found)
-			memset(entry->old_name, 0, sizeof(entry));
-	}
-}
-
-static void
-HandleRename(RenameStmt *stmt)
-{
-	if (stmt->renameType == OBJECT_DATABASE)
-		return HandleDbRename(stmt);
-	else if (stmt->renameType == OBJECT_ROLE)
-		return HandleRoleRename(stmt);
-}
-
-static void
-NeonProcessUtility(
-				   PlannedStmt *pstmt,
-				   const char *queryString,
-				   bool readOnlyTree,
-				   ProcessUtilityContext context,
-				   ParamListInfo params,
-				   QueryEnvironment *queryEnv,
-				   DestReceiver *dest,
-				   QueryCompletion *qc)
-{
-	Node	   *parseTree = pstmt->utilityStmt;
-
-	switch (nodeTag(parseTree))
-	{
-		case T_CreatedbStmt:
-			HandleCreateDb(castNode(CreatedbStmt, parseTree));
-			break;
-		case T_AlterOwnerStmt:
-			HandleAlterOwner(castNode(AlterOwnerStmt, parseTree));
-			break;
-		case T_RenameStmt:
-			HandleRename(castNode(RenameStmt, parseTree));
-			break;
-		case T_DropdbStmt:
-			HandleDropDb(castNode(DropdbStmt, parseTree));
-			break;
-		case T_CreateRoleStmt:
-			HandleCreateRole(castNode(CreateRoleStmt, parseTree));
-			break;
-		case T_AlterRoleStmt:
-			HandleAlterRole(castNode(AlterRoleStmt, parseTree));
-			break;
-		case T_DropRoleStmt:
-			HandleDropRole(castNode(DropRoleStmt, parseTree));
-			break;
-		default:
-			break;
-	}
-
-	if (PreviousProcessUtilityHook)
-	{
-		PreviousProcessUtilityHook(
-								   pstmt,
-								   queryString,
-								   readOnlyTree,
-								   context,
-								   params,
-								   queryEnv,
-								   dest,
-								   qc);
-	}
-	else
-	{
-		standard_ProcessUtility(
-								pstmt,
-								queryString,
-								readOnlyTree,
-								context,
-								params,
-								queryEnv,
-								dest,
-								qc);
-	}
-}
-
-extern void
-InitControlPlaneConnector()
-{
-	PreviousProcessUtilityHook = ProcessUtility_hook;
-	ProcessUtility_hook = NeonProcessUtility;
-	RegisterXactCallback(NeonXactCallback, NULL);
-	RegisterSubXactCallback(NeonSubXactCallback, NULL);
-
-	DefineCustomStringVariable(
-							   "neon.console_url",
-							   "URL of the Neon Console, which will be forwarded changes to dbs and roles",
-							   NULL,
-							   &ConsoleURL,
-							   NULL,
-							   PGC_POSTMASTER,
-							   0,
-							   NULL,
-							   NULL,
-							   NULL);
-
-	DefineCustomBoolVariable(
-							 "neon.forward_ddl",
-							 "Controls whether to forward DDL to the control plane",
-							 NULL,
-							 &ForwardDDL,
-							 true,
-							 PGC_SUSET,
-							 0,
-							 NULL,
-							 NULL,
-							 NULL);
-
-	const char *jwt_token = getenv("NEON_CONTROL_PLANE_TOKEN");
-
-	if (!jwt_token)
-	{
-		elog(LOG, "Missing NEON_CONTROL_PLANE_TOKEN environment variable, forwarding will not be authenticated");
-	}
-
-	if (curl_global_init(CURL_GLOBAL_DEFAULT))
-	{
-		elog(ERROR, "Failed to initialize curl");
-	}
-	if ((CurlHandle = curl_easy_init()) == NULL)
-	{
-		elog(ERROR, "Failed to initialize curl handle");
-	}
-	if ((ContentHeader = curl_slist_append(ContentHeader, "Content-Type: application/json")) == NULL)
-	{
-		elog(ERROR, "Failed to initialize content header");
-	}
-
-	if (jwt_token)
-	{
-		char		auth_header[8192];
-
-		snprintf(auth_header, sizeof(auth_header), "Authorization: Bearer %s", jwt_token);
-		if ((ContentHeader = curl_slist_append(ContentHeader, auth_header)) == NULL)
-		{
-			elog(ERROR, "Failed to initialize authorization header");
-		}
-	}
-}

pgxn/neon/control_plane_connector.h

@@ -1,6 +0,0 @@
-#ifndef CONTROL_PLANE_CONNECTOR_H
-#define CONTROL_PLANE_CONNECTOR_H
-
-void		InitControlPlaneConnector();
-
-#endif

pgxn/neon/libpagestore.c

@@ -192,9 +192,8 @@ retry:
 		{
 			if (!PQconsumeInput(pageserver_conn))
 			{
-				char	   *msg = pchomp(PQerrorMessage(pageserver_conn));
-				neon_log(LOG, "could not get response from pageserver: %s", msg);
-				pfree(msg);
+				neon_log(LOG, "could not get response from pageserver: %s",
+						 PQerrorMessage(pageserver_conn));
 				return -1;
 			}
 		}
@@ -344,7 +343,7 @@ pageserver_receive(void)
 			resp = NULL;
 		}
 		else if (rc == -2)
-			neon_log(ERROR, "could not read COPY data: %s", pchomp(PQerrorMessage(pageserver_conn)));
+			neon_log(ERROR, "could not read COPY data: %s", PQerrorMessage(pageserver_conn));
 		else
 			neon_log(ERROR, "unexpected PQgetCopyData return value: %d", rc);
 	}
@@ -368,7 +367,7 @@ pageserver_flush(void)
 	}
 	else if (PQflush(pageserver_conn))
 	{
-		char	   *msg = pchomp(PQerrorMessage(pageserver_conn));
+		char	   *msg = PQerrorMessage(pageserver_conn);
 
 		pageserver_disconnect();
 		neon_log(ERROR, "failed to flush page requests: %s", msg);

pgxn/neon/neon.c

@@ -25,7 +25,6 @@
 #include "neon.h"
 #include "walproposer.h"
 #include "pagestore_client.h"
-#include "control_plane_connector.h"
 
 PG_MODULE_MAGIC;
 void		_PG_init(void);
@@ -35,11 +34,7 @@ _PG_init(void)
 {
 	pg_init_libpagestore();
 	pg_init_walproposer();
-	InitControlPlaneConnector();
 
-        // Important: This must happen after other parts of the extension
-        // are loaded, otherwise any settings to GUCs that were set before
-        // the extension was loaded will be removed.
 	EmitWarningsOnPlaceholders("neon");
 }
 

pgxn/neon/pagestore_client.h

@@ -52,7 +52,7 @@ typedef struct
 #define NEON_TAG "[NEON_SMGR] "
 #define neon_log(tag, fmt, ...) ereport(tag,                                  \
 										(errmsg(NEON_TAG fmt, ##__VA_ARGS__), \
-										 errhidestmt(true), errhidecontext(true), errposition(0), internalerrposition(0)))
+										 errhidestmt(true), errhidecontext(true), internalerrposition(0)))
 
 /*
  * supertype of all the Neon*Request structs below

poetry.lock

@@ -1,4 +1,4 @@
-# This file is automatically @generated by Poetry 1.4.2 and should not be changed by hand.
+# This file is automatically @generated by Poetry and should not be changed by hand.
 
 [[package]]
 name = "aiohttp"
@@ -79,30 +79,30 @@ sa = ["sqlalchemy[postgresql-psycopg2binary] (>=1.3,<1.5)"]
 
 [[package]]
 name = "allure-pytest"
-version = "2.13.2"
+version = "2.13.1"
 description = "Allure pytest integration"
 category = "main"
 optional = false
 python-versions = "*"
 files = [
-    {file = "allure-pytest-2.13.2.tar.gz", hash = "sha256:22243159e8ec81ce2b5254b4013802198821b1b42f118f69d4a289396607c7b3"},
-    {file = "allure_pytest-2.13.2-py3-none-any.whl", hash = "sha256:17de9dbee7f61c8e66a5b5e818b00e419dbcea44cb55c24319401ba813220690"},
+    {file = "allure-pytest-2.13.1.tar.gz", hash = "sha256:68d69456eeb65af4061ec06a80bc941163b0616e8216554d36b070a6bf070e08"},
+    {file = "allure_pytest-2.13.1-py3-none-any.whl", hash = "sha256:a8de2fc3b3effe2d8f98801646920de3f055b779710f4c806dbee7c613c24633"},
 ]
 
 [package.dependencies]
-allure-python-commons = "2.13.2"
+allure-python-commons = "2.13.1"
 pytest = ">=4.5.0"
 
 [[package]]
 name = "allure-python-commons"
-version = "2.13.2"
+version = "2.13.1"
 description = "Common module for integrate allure with python-based frameworks"
 category = "main"
 optional = false
 python-versions = ">=3.6"
 files = [
-    {file = "allure-python-commons-2.13.2.tar.gz", hash = "sha256:8a03681330231b1deadd86b97ff68841c6591320114ae638570f1ed60d7a2033"},
-    {file = "allure_python_commons-2.13.2-py3-none-any.whl", hash = "sha256:2bb3646ec3fbf5b36d178a5e735002bc130ae9f9ba80f080af97d368ba375051"},
+    {file = "allure-python-commons-2.13.1.tar.gz", hash = "sha256:3fc13e1da8ebb23f9ab5c9c72ad04595023cdd5078dbb8604939997faebed5cb"},
+    {file = "allure_python_commons-2.13.1-py3-none-any.whl", hash = "sha256:d08e04867bddf44fef55def3d67f4bc25af58a1bf9fcffcf4ec3331f7f2ef0d0"},
 ]
 
 [package.dependencies]
@@ -172,6 +172,17 @@ dev = ["Cython (>=0.29.24,<0.30.0)", "Sphinx (>=4.1.2,<4.2.0)", "flake8 (>=5.0.4
 docs = ["Sphinx (>=4.1.2,<4.2.0)", "sphinx-rtd-theme (>=0.5.2,<0.6.0)", "sphinxcontrib-asyncio (>=0.3.0,<0.4.0)"]
 test = ["flake8 (>=5.0.4,<5.1.0)", "uvloop (>=0.15.3)"]
 
+[[package]]
+name = "atomicwrites"
+version = "1.4.1"
+description = "Atomic file writes."
+category = "main"
+optional = false
+python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*"
+files = [
+    {file = "atomicwrites-1.4.1.tar.gz", hash = "sha256:81b2c9071a49367a7f770170e5eec8cb66567cfbbc8c73d20ce5ca4a8d71cf11"},
+]
+
 [[package]]
 name = "attrs"
 version = "21.4.0"
@@ -228,49 +239,49 @@ wrapt = "*"
 
 [[package]]
 name = "backoff"
-version = "2.2.1"
+version = "1.11.1"
 description = "Function decoration for backoff and retry"
 category = "main"
 optional = false
-python-versions = ">=3.7,<4.0"
+python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
 files = [
-    {file = "backoff-2.2.1-py3-none-any.whl", hash = "sha256:63579f9a0628e06278f7e47b7d7d5b6ce20dc65c5e96a6f3ca99a6adca0396e8"},
-    {file = "backoff-2.2.1.tar.gz", hash = "sha256:03f829f5bb1923180821643f8753b0502c3b682293992485b0eef2807afa5cba"},
+    {file = "backoff-1.11.1-py2.py3-none-any.whl", hash = "sha256:61928f8fa48d52e4faa81875eecf308eccfb1016b018bb6bd21e05b5d90a96c5"},
+    {file = "backoff-1.11.1.tar.gz", hash = "sha256:ccb962a2378418c667b3c979b504fdeb7d9e0d29c0579e3b13b86467177728cb"},
 ]
 
 [[package]]
 name = "black"
-version = "23.3.0"
+version = "23.1.0"
 description = "The uncompromising code formatter."
 category = "dev"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "black-23.3.0-cp310-cp310-macosx_10_16_arm64.whl", hash = "sha256:0945e13506be58bf7db93ee5853243eb368ace1c08a24c65ce108986eac65915"},
-    {file = "black-23.3.0-cp310-cp310-macosx_10_16_universal2.whl", hash = "sha256:67de8d0c209eb5b330cce2469503de11bca4085880d62f1628bd9972cc3366b9"},
-    {file = "black-23.3.0-cp310-cp310-macosx_10_16_x86_64.whl", hash = "sha256:7c3eb7cea23904399866c55826b31c1f55bbcd3890ce22ff70466b907b6775c2"},
-    {file = "black-23.3.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:32daa9783106c28815d05b724238e30718f34155653d4d6e125dc7daec8e260c"},
-    {file = "black-23.3.0-cp310-cp310-win_amd64.whl", hash = "sha256:35d1381d7a22cc5b2be2f72c7dfdae4072a3336060635718cc7e1ede24221d6c"},
-    {file = "black-23.3.0-cp311-cp311-macosx_10_16_arm64.whl", hash = "sha256:a8a968125d0a6a404842fa1bf0b349a568634f856aa08ffaff40ae0dfa52e7c6"},
-    {file = "black-23.3.0-cp311-cp311-macosx_10_16_universal2.whl", hash = "sha256:c7ab5790333c448903c4b721b59c0d80b11fe5e9803d8703e84dcb8da56fec1b"},
-    {file = "black-23.3.0-cp311-cp311-macosx_10_16_x86_64.whl", hash = "sha256:a6f6886c9869d4daae2d1715ce34a19bbc4b95006d20ed785ca00fa03cba312d"},
-    {file = "black-23.3.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6f3c333ea1dd6771b2d3777482429864f8e258899f6ff05826c3a4fcc5ce3f70"},
-    {file = "black-23.3.0-cp311-cp311-win_amd64.whl", hash = "sha256:11c410f71b876f961d1de77b9699ad19f939094c3a677323f43d7a29855fe326"},
-    {file = "black-23.3.0-cp37-cp37m-macosx_10_16_x86_64.whl", hash = "sha256:1d06691f1eb8de91cd1b322f21e3bfc9efe0c7ca1f0e1eb1db44ea367dff656b"},
-    {file = "black-23.3.0-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:50cb33cac881766a5cd9913e10ff75b1e8eb71babf4c7104f2e9c52da1fb7de2"},
-    {file = "black-23.3.0-cp37-cp37m-win_amd64.whl", hash = "sha256:e114420bf26b90d4b9daa597351337762b63039752bdf72bf361364c1aa05925"},
-    {file = "black-23.3.0-cp38-cp38-macosx_10_16_arm64.whl", hash = "sha256:48f9d345675bb7fbc3dd85821b12487e1b9a75242028adad0333ce36ed2a6d27"},
-    {file = "black-23.3.0-cp38-cp38-macosx_10_16_universal2.whl", hash = "sha256:714290490c18fb0126baa0fca0a54ee795f7502b44177e1ce7624ba1c00f2331"},
-    {file = "black-23.3.0-cp38-cp38-macosx_10_16_x86_64.whl", hash = "sha256:064101748afa12ad2291c2b91c960be28b817c0c7eaa35bec09cc63aa56493c5"},
-    {file = "black-23.3.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:562bd3a70495facf56814293149e51aa1be9931567474993c7942ff7d3533961"},
-    {file = "black-23.3.0-cp38-cp38-win_amd64.whl", hash = "sha256:e198cf27888ad6f4ff331ca1c48ffc038848ea9f031a3b40ba36aced7e22f2c8"},
-    {file = "black-23.3.0-cp39-cp39-macosx_10_16_arm64.whl", hash = "sha256:3238f2aacf827d18d26db07524e44741233ae09a584273aa059066d644ca7b30"},
-    {file = "black-23.3.0-cp39-cp39-macosx_10_16_universal2.whl", hash = "sha256:f0bd2f4a58d6666500542b26354978218a9babcdc972722f4bf90779524515f3"},
-    {file = "black-23.3.0-cp39-cp39-macosx_10_16_x86_64.whl", hash = "sha256:92c543f6854c28a3c7f39f4d9b7694f9a6eb9d3c5e2ece488c327b6e7ea9b266"},
-    {file = "black-23.3.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3a150542a204124ed00683f0db1f5cf1c2aaaa9cc3495b7a3b5976fb136090ab"},
-    {file = "black-23.3.0-cp39-cp39-win_amd64.whl", hash = "sha256:6b39abdfb402002b8a7d030ccc85cf5afff64ee90fa4c5aebc531e3ad0175ddb"},
-    {file = "black-23.3.0-py3-none-any.whl", hash = "sha256:ec751418022185b0c1bb7d7736e6933d40bbb14c14a0abcf9123d1b159f98dd4"},
-    {file = "black-23.3.0.tar.gz", hash = "sha256:1c7b8d606e728a41ea1ccbd7264677e494e87cf630e399262ced92d4a8dac940"},
+    {file = "black-23.1.0-cp310-cp310-macosx_10_16_arm64.whl", hash = "sha256:b6a92a41ee34b883b359998f0c8e6eb8e99803aa8bf3123bf2b2e6fec505a221"},
+    {file = "black-23.1.0-cp310-cp310-macosx_10_16_universal2.whl", hash = "sha256:57c18c5165c1dbe291d5306e53fb3988122890e57bd9b3dcb75f967f13411a26"},
+    {file = "black-23.1.0-cp310-cp310-macosx_10_16_x86_64.whl", hash = "sha256:9880d7d419bb7e709b37e28deb5e68a49227713b623c72b2b931028ea65f619b"},
+    {file = "black-23.1.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e6663f91b6feca5d06f2ccd49a10f254f9298cc1f7f49c46e498a0771b507104"},
+    {file = "black-23.1.0-cp310-cp310-win_amd64.whl", hash = "sha256:9afd3f493666a0cd8f8df9a0200c6359ac53940cbde049dcb1a7eb6ee2dd7074"},
+    {file = "black-23.1.0-cp311-cp311-macosx_10_16_arm64.whl", hash = "sha256:bfffba28dc52a58f04492181392ee380e95262af14ee01d4bc7bb1b1c6ca8d27"},
+    {file = "black-23.1.0-cp311-cp311-macosx_10_16_universal2.whl", hash = "sha256:c1c476bc7b7d021321e7d93dc2cbd78ce103b84d5a4cf97ed535fbc0d6660648"},
+    {file = "black-23.1.0-cp311-cp311-macosx_10_16_x86_64.whl", hash = "sha256:382998821f58e5c8238d3166c492139573325287820963d2f7de4d518bd76958"},
+    {file = "black-23.1.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2bf649fda611c8550ca9d7592b69f0637218c2369b7744694c5e4902873b2f3a"},
+    {file = "black-23.1.0-cp311-cp311-win_amd64.whl", hash = "sha256:121ca7f10b4a01fd99951234abdbd97728e1240be89fde18480ffac16503d481"},
+    {file = "black-23.1.0-cp37-cp37m-macosx_10_16_x86_64.whl", hash = "sha256:a8471939da5e824b891b25751955be52ee7f8a30a916d570a5ba8e0f2eb2ecad"},
+    {file = "black-23.1.0-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:8178318cb74f98bc571eef19068f6ab5613b3e59d4f47771582f04e175570ed8"},
+    {file = "black-23.1.0-cp37-cp37m-win_amd64.whl", hash = "sha256:a436e7881d33acaf2536c46a454bb964a50eff59b21b51c6ccf5a40601fbef24"},
+    {file = "black-23.1.0-cp38-cp38-macosx_10_16_arm64.whl", hash = "sha256:a59db0a2094d2259c554676403fa2fac3473ccf1354c1c63eccf7ae65aac8ab6"},
+    {file = "black-23.1.0-cp38-cp38-macosx_10_16_universal2.whl", hash = "sha256:0052dba51dec07ed029ed61b18183942043e00008ec65d5028814afaab9a22fd"},
+    {file = "black-23.1.0-cp38-cp38-macosx_10_16_x86_64.whl", hash = "sha256:49f7b39e30f326a34b5c9a4213213a6b221d7ae9d58ec70df1c4a307cf2a1580"},
+    {file = "black-23.1.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:162e37d49e93bd6eb6f1afc3e17a3d23a823042530c37c3c42eeeaf026f38468"},
+    {file = "black-23.1.0-cp38-cp38-win_amd64.whl", hash = "sha256:8b70eb40a78dfac24842458476135f9b99ab952dd3f2dab738c1881a9b38b753"},
+    {file = "black-23.1.0-cp39-cp39-macosx_10_16_arm64.whl", hash = "sha256:a29650759a6a0944e7cca036674655c2f0f63806ddecc45ed40b7b8aa314b651"},
+    {file = "black-23.1.0-cp39-cp39-macosx_10_16_universal2.whl", hash = "sha256:bb460c8561c8c1bec7824ecbc3ce085eb50005883a6203dcfb0122e95797ee06"},
+    {file = "black-23.1.0-cp39-cp39-macosx_10_16_x86_64.whl", hash = "sha256:c91dfc2c2a4e50df0026f88d2215e166616e0c80e86004d0003ece0488db2739"},
+    {file = "black-23.1.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2a951cc83ab535d248c89f300eccbd625e80ab880fbcfb5ac8afb5f01a258ac9"},
+    {file = "black-23.1.0-cp39-cp39-win_amd64.whl", hash = "sha256:0680d4380db3719ebcfb2613f34e86c8e6d15ffeabcf8ec59355c5e7b85bb555"},
+    {file = "black-23.1.0-py3-none-any.whl", hash = "sha256:7a0f701d314cfa0896b9001df70a530eb2472babb76086344e688829efd97d32"},
+    {file = "black-23.1.0.tar.gz", hash = "sha256:b0bd97bea8903f5a2ba7219257a44e3f1f9d00073d6cc1add68f0beec69692ac"},
 ]
 
 [package.dependencies]
@@ -940,21 +951,6 @@ six = ">=1.9.0"
 gmpy = ["gmpy"]
 gmpy2 = ["gmpy2"]
 
-[[package]]
-name = "exceptiongroup"
-version = "1.1.1"
-description = "Backport of PEP 654 (exception groups)"
-category = "main"
-optional = false
-python-versions = ">=3.7"
-files = [
-    {file = "exceptiongroup-1.1.1-py3-none-any.whl", hash = "sha256:232c37c63e4f682982c8b6459f33a8981039e5fb8756b2074364e5055c498c9e"},
-    {file = "exceptiongroup-1.1.1.tar.gz", hash = "sha256:d484c3090ba2889ae2928419117447a14daf3c1231d5e30d0aae34f354f01785"},
-]
-
-[package.extras]
-test = ["pytest (>=6)"]
-
 [[package]]
 name = "execnet"
 version = "1.9.0"
@@ -1414,38 +1410,38 @@ files = [
 
 [[package]]
 name = "mypy"
-version = "1.3.0"
+version = "1.1.1"
 description = "Optional static typing for Python"
 category = "dev"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "mypy-1.3.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:c1eb485cea53f4f5284e5baf92902cd0088b24984f4209e25981cc359d64448d"},
-    {file = "mypy-1.3.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:4c99c3ecf223cf2952638da9cd82793d8f3c0c5fa8b6ae2b2d9ed1e1ff51ba85"},
-    {file = "mypy-1.3.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:550a8b3a19bb6589679a7c3c31f64312e7ff482a816c96e0cecec9ad3a7564dd"},
-    {file = "mypy-1.3.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:cbc07246253b9e3d7d74c9ff948cd0fd7a71afcc2b77c7f0a59c26e9395cb152"},
-    {file = "mypy-1.3.0-cp310-cp310-win_amd64.whl", hash = "sha256:a22435632710a4fcf8acf86cbd0d69f68ac389a3892cb23fbad176d1cddaf228"},
-    {file = "mypy-1.3.0-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:6e33bb8b2613614a33dff70565f4c803f889ebd2f859466e42b46e1df76018dd"},
-    {file = "mypy-1.3.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:7d23370d2a6b7a71dc65d1266f9a34e4cde9e8e21511322415db4b26f46f6b8c"},
-    {file = "mypy-1.3.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:658fe7b674769a0770d4b26cb4d6f005e88a442fe82446f020be8e5f5efb2fae"},
-    {file = "mypy-1.3.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:6e42d29e324cdda61daaec2336c42512e59c7c375340bd202efa1fe0f7b8f8ca"},
-    {file = "mypy-1.3.0-cp311-cp311-win_amd64.whl", hash = "sha256:d0b6c62206e04061e27009481cb0ec966f7d6172b5b936f3ead3d74f29fe3dcf"},
-    {file = "mypy-1.3.0-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:76ec771e2342f1b558c36d49900dfe81d140361dd0d2df6cd71b3db1be155409"},
-    {file = "mypy-1.3.0-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ebc95f8386314272bbc817026f8ce8f4f0d2ef7ae44f947c4664efac9adec929"},
-    {file = "mypy-1.3.0-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:faff86aa10c1aa4a10e1a301de160f3d8fc8703b88c7e98de46b531ff1276a9a"},
-    {file = "mypy-1.3.0-cp37-cp37m-win_amd64.whl", hash = "sha256:8c5979d0deb27e0f4479bee18ea0f83732a893e81b78e62e2dda3e7e518c92ee"},
-    {file = "mypy-1.3.0-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:c5d2cc54175bab47011b09688b418db71403aefad07cbcd62d44010543fc143f"},
-    {file = "mypy-1.3.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:87df44954c31d86df96c8bd6e80dfcd773473e877ac6176a8e29898bfb3501cb"},
-    {file = "mypy-1.3.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:473117e310febe632ddf10e745a355714e771ffe534f06db40702775056614c4"},
-    {file = "mypy-1.3.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:74bc9b6e0e79808bf8678d7678b2ae3736ea72d56eede3820bd3849823e7f305"},
-    {file = "mypy-1.3.0-cp38-cp38-win_amd64.whl", hash = "sha256:44797d031a41516fcf5cbfa652265bb994e53e51994c1bd649ffcd0c3a7eccbf"},
-    {file = "mypy-1.3.0-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:ddae0f39ca146972ff6bb4399f3b2943884a774b8771ea0a8f50e971f5ea5ba8"},
-    {file = "mypy-1.3.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:1c4c42c60a8103ead4c1c060ac3cdd3ff01e18fddce6f1016e08939647a0e703"},
-    {file = "mypy-1.3.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e86c2c6852f62f8f2b24cb7a613ebe8e0c7dc1402c61d36a609174f63e0ff017"},
-    {file = "mypy-1.3.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:f9dca1e257d4cc129517779226753dbefb4f2266c4eaad610fc15c6a7e14283e"},
-    {file = "mypy-1.3.0-cp39-cp39-win_amd64.whl", hash = "sha256:95d8d31a7713510685b05fbb18d6ac287a56c8f6554d88c19e73f724a445448a"},
-    {file = "mypy-1.3.0-py3-none-any.whl", hash = "sha256:a8763e72d5d9574d45ce5881962bc8e9046bf7b375b0abf031f3e6811732a897"},
-    {file = "mypy-1.3.0.tar.gz", hash = "sha256:e1f4d16e296f5135624b34e8fb741eb0eadedca90862405b1f1fde2040b9bd11"},
+    {file = "mypy-1.1.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:39c7119335be05630611ee798cc982623b9e8f0cff04a0b48dfc26100e0b97af"},
+    {file = "mypy-1.1.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:61bf08362e93b6b12fad3eab68c4ea903a077b87c90ac06c11e3d7a09b56b9c1"},
+    {file = "mypy-1.1.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dbb19c9f662e41e474e0cff502b7064a7edc6764f5262b6cd91d698163196799"},
+    {file = "mypy-1.1.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:315ac73cc1cce4771c27d426b7ea558fb4e2836f89cb0296cbe056894e3a1f78"},
+    {file = "mypy-1.1.1-cp310-cp310-win_amd64.whl", hash = "sha256:5cb14ff9919b7df3538590fc4d4c49a0f84392237cbf5f7a816b4161c061829e"},
+    {file = "mypy-1.1.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:26cdd6a22b9b40b2fd71881a8a4f34b4d7914c679f154f43385ca878a8297389"},
+    {file = "mypy-1.1.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:5b5f81b40d94c785f288948c16e1f2da37203c6006546c5d947aab6f90aefef2"},
+    {file = "mypy-1.1.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:21b437be1c02712a605591e1ed1d858aba681757a1e55fe678a15c2244cd68a5"},
+    {file = "mypy-1.1.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:d809f88734f44a0d44959d795b1e6f64b2bbe0ea4d9cc4776aa588bb4229fc1c"},
+    {file = "mypy-1.1.1-cp311-cp311-win_amd64.whl", hash = "sha256:a380c041db500e1410bb5b16b3c1c35e61e773a5c3517926b81dfdab7582be54"},
+    {file = "mypy-1.1.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:b7c7b708fe9a871a96626d61912e3f4ddd365bf7f39128362bc50cbd74a634d5"},
+    {file = "mypy-1.1.1-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c1c10fa12df1232c936830839e2e935d090fc9ee315744ac33b8a32216b93707"},
+    {file = "mypy-1.1.1-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:0a28a76785bf57655a8ea5eb0540a15b0e781c807b5aa798bd463779988fa1d5"},
+    {file = "mypy-1.1.1-cp37-cp37m-win_amd64.whl", hash = "sha256:ef6a01e563ec6a4940784c574d33f6ac1943864634517984471642908b30b6f7"},
+    {file = "mypy-1.1.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:d64c28e03ce40d5303450f547e07418c64c241669ab20610f273c9e6290b4b0b"},
+    {file = "mypy-1.1.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:64cc3afb3e9e71a79d06e3ed24bb508a6d66f782aff7e56f628bf35ba2e0ba51"},
+    {file = "mypy-1.1.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ce61663faf7a8e5ec6f456857bfbcec2901fbdb3ad958b778403f63b9e606a1b"},
+    {file = "mypy-1.1.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:2b0c373d071593deefbcdd87ec8db91ea13bd8f1328d44947e88beae21e8d5e9"},
+    {file = "mypy-1.1.1-cp38-cp38-win_amd64.whl", hash = "sha256:2888ce4fe5aae5a673386fa232473014056967f3904f5abfcf6367b5af1f612a"},
+    {file = "mypy-1.1.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:19ba15f9627a5723e522d007fe708007bae52b93faab00f95d72f03e1afa9598"},
+    {file = "mypy-1.1.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:59bbd71e5c58eed2e992ce6523180e03c221dcd92b52f0e792f291d67b15a71c"},
+    {file = "mypy-1.1.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9401e33814cec6aec8c03a9548e9385e0e228fc1b8b0a37b9ea21038e64cdd8a"},
+    {file = "mypy-1.1.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:4b398d8b1f4fba0e3c6463e02f8ad3346f71956b92287af22c9b12c3ec965a9f"},
+    {file = "mypy-1.1.1-cp39-cp39-win_amd64.whl", hash = "sha256:69b35d1dcb5707382810765ed34da9db47e7f95b3528334a3c999b0c90fe523f"},
+    {file = "mypy-1.1.1-py3-none-any.whl", hash = "sha256:4e4e8b362cdf99ba00c2b218036002bdcdf1e0de085cdb296a49df03fb31dfc4"},
+    {file = "mypy-1.1.1.tar.gz", hash = "sha256:ae9ceae0f5b9059f33dbc62dea087e942c0ccab4b7a003719cb70f9b8abfa32f"},
 ]
 
 [package.dependencies]
@@ -1725,6 +1721,18 @@ files = [
     {file = "psycopg2_binary-2.9.3-cp39-cp39-win_amd64.whl", hash = "sha256:accfe7e982411da3178ec690baaceaad3c278652998b2c45828aaac66cd8285f"},
 ]
 
+[[package]]
+name = "py"
+version = "1.11.0"
+description = "library with cross-python path, ini-parsing, io, code, log facilities"
+category = "main"
+optional = false
+python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*"
+files = [
+    {file = "py-1.11.0-py2.py3-none-any.whl", hash = "sha256:607c53218732647dff4acdfcd50cb62615cedf612e72d1724fb1a0cc6405b378"},
+    {file = "py-1.11.0.tar.gz", hash = "sha256:51c75c4126074b472f746a24399ad32f6053d1b34b68d2fa41e558e6f4a98719"},
+]
+
 [[package]]
 name = "pyasn1"
 version = "0.4.8"
@@ -1833,56 +1841,57 @@ files = [
 
 [[package]]
 name = "pytest"
-version = "7.3.1"
+version = "6.2.5"
 description = "pytest: simple powerful testing with Python"
 category = "main"
 optional = false
-python-versions = ">=3.7"
+python-versions = ">=3.6"
 files = [
-    {file = "pytest-7.3.1-py3-none-any.whl", hash = "sha256:3799fa815351fea3a5e96ac7e503a96fa51cc9942c3753cda7651b93c1cfa362"},
-    {file = "pytest-7.3.1.tar.gz", hash = "sha256:434afafd78b1d78ed0addf160ad2b77a30d35d4bdf8af234fe621919d9ed15e3"},
+    {file = "pytest-6.2.5-py3-none-any.whl", hash = "sha256:7310f8d27bc79ced999e760ca304d69f6ba6c6649c0b60fb0e04a4a77cacc134"},
+    {file = "pytest-6.2.5.tar.gz", hash = "sha256:131b36680866a76e6781d13f101efb86cf674ebb9762eb70d3082b6f29889e89"},
 ]
 
 [package.dependencies]
+atomicwrites = {version = ">=1.0", markers = "sys_platform == \"win32\""}
+attrs = ">=19.2.0"
 colorama = {version = "*", markers = "sys_platform == \"win32\""}
-exceptiongroup = {version = ">=1.0.0rc8", markers = "python_version < \"3.11\""}
 iniconfig = "*"
 packaging = "*"
 pluggy = ">=0.12,<2.0"
-tomli = {version = ">=1.0.0", markers = "python_version < \"3.11\""}
+py = ">=1.8.2"
+toml = "*"
 
 [package.extras]
-testing = ["argcomplete", "attrs (>=19.2.0)", "hypothesis (>=3.56)", "mock", "nose", "pygments (>=2.7.2)", "requests", "xmlschema"]
+testing = ["argcomplete", "hypothesis (>=3.56)", "mock", "nose", "requests", "xmlschema"]
 
 [[package]]
 name = "pytest-asyncio"
-version = "0.21.0"
+version = "0.19.0"
 description = "Pytest support for asyncio"
 category = "main"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "pytest-asyncio-0.21.0.tar.gz", hash = "sha256:2b38a496aef56f56b0e87557ec313e11e1ab9276fc3863f6a7be0f1d0e415e1b"},
-    {file = "pytest_asyncio-0.21.0-py3-none-any.whl", hash = "sha256:f2b3366b7cd501a4056858bd39349d5af19742aed2d81660b7998b6341c7eb9c"},
+    {file = "pytest-asyncio-0.19.0.tar.gz", hash = "sha256:ac4ebf3b6207259750bc32f4c1d8fcd7e79739edbc67ad0c58dd150b1d072fed"},
+    {file = "pytest_asyncio-0.19.0-py3-none-any.whl", hash = "sha256:7a97e37cfe1ed296e2e84941384bdd37c376453912d397ed39293e0916f521fa"},
 ]
 
 [package.dependencies]
-pytest = ">=7.0.0"
+pytest = ">=6.1.0"
 
 [package.extras]
-docs = ["sphinx (>=5.3)", "sphinx-rtd-theme (>=1.0)"]
 testing = ["coverage (>=6.2)", "flaky (>=3.5.0)", "hypothesis (>=5.7.1)", "mypy (>=0.931)", "pytest-trio (>=0.7.0)"]
 
 [[package]]
 name = "pytest-httpserver"
-version = "1.0.8"
+version = "1.0.6"
 description = "pytest-httpserver is a httpserver for pytest"
 category = "main"
 optional = false
-python-versions = ">=3.8,<4.0"
+python-versions = ">=3.7,<4.0"
 files = [
-    {file = "pytest_httpserver-1.0.8-py3-none-any.whl", hash = "sha256:24cd3d9f6a0b927c7bfc400d0b3fda7442721b8267ce29942bf307b190f0bb09"},
-    {file = "pytest_httpserver-1.0.8.tar.gz", hash = "sha256:e052f69bc8a9073db02484681e8e47004dd1fb3763b0ae833bd899e5895c559a"},
+    {file = "pytest_httpserver-1.0.6-py3-none-any.whl", hash = "sha256:ac2379acc91fe8bdbe2911c93af8dd130e33b5899fb9934d15669480739c6d32"},
+    {file = "pytest_httpserver-1.0.6.tar.gz", hash = "sha256:9040d07bf59ac45d8de3db1d4468fd2d1d607975e4da4c872ecc0402cdbf7b3e"},
 ]
 
 [package.dependencies]
@@ -1905,14 +1914,14 @@ pytest = ">=3.2.5"
 
 [[package]]
 name = "pytest-order"
-version = "1.1.0"
+version = "1.0.1"
 description = "pytest plugin to run your tests in a specific order"
 category = "main"
 optional = false
 python-versions = ">=3.6"
 files = [
-    {file = "pytest-order-1.1.0.tar.gz", hash = "sha256:139d25b30826b78eebb42722f747eab14c44b88059d7a71d4f79d14a057269a5"},
-    {file = "pytest_order-1.1.0-py3-none-any.whl", hash = "sha256:3b3730969c97900fa5cd31ecff80847680ed56b2490954565c14949ba60d9371"},
+    {file = "pytest-order-1.0.1.tar.gz", hash = "sha256:5dd6b929fbd7eaa6d0ee07586f65c623babb0afe72b4843c5f15055d6b3b1b1f"},
+    {file = "pytest_order-1.0.1-py3-none-any.whl", hash = "sha256:bbe6e63a8e23741ab3e810d458d1ea7317e797b70f9550512d77d6e9e8fd1bbb"},
 ]
 
 [package.dependencies]
@@ -1954,14 +1963,14 @@ pytest = ">=5.0.0"
 
 [[package]]
 name = "pytest-xdist"
-version = "3.3.1"
-description = "pytest xdist plugin for distributed testing, most importantly across multiple CPUs"
+version = "3.0.2"
+description = "pytest xdist plugin for distributed testing and loop-on-failing modes"
 category = "main"
 optional = false
-python-versions = ">=3.7"
+python-versions = ">=3.6"
 files = [
-    {file = "pytest-xdist-3.3.1.tar.gz", hash = "sha256:d5ee0520eb1b7bcca50a60a518ab7a7707992812c578198f8b44fdfac78e8c93"},
-    {file = "pytest_xdist-3.3.1-py3-none-any.whl", hash = "sha256:ff9daa7793569e6a68544850fd3927cd257cc03a7ef76c95e86915355e82b5f2"},
+    {file = "pytest-xdist-3.0.2.tar.gz", hash = "sha256:688da9b814370e891ba5de650c9327d1a9d861721a524eb917e620eec3e90291"},
+    {file = "pytest_xdist-3.0.2-py3-none-any.whl", hash = "sha256:9feb9a18e1790696ea23e1434fa73b325ed4998b0e9fcb221f16fd1945e6df1b"},
 ]
 
 [package.dependencies]
@@ -2083,21 +2092,21 @@ files = [
 
 [[package]]
 name = "requests"
-version = "2.31.0"
+version = "2.28.1"
 description = "Python HTTP for Humans."
 category = "main"
 optional = false
-python-versions = ">=3.7"
+python-versions = ">=3.7, <4"
 files = [
-    {file = "requests-2.31.0-py3-none-any.whl", hash = "sha256:58cd2187c01e70e6e26505bca751777aa9f2ee0b7f4300988b709f44e013003f"},
-    {file = "requests-2.31.0.tar.gz", hash = "sha256:942c5a758f98d790eaed1a29cb6eefc7ffb0d1cf7af05c3d2791656dbd6ad1e1"},
+    {file = "requests-2.28.1-py3-none-any.whl", hash = "sha256:8fefa2a1a1365bf5520aac41836fbee479da67864514bdb821f31ce07ce65349"},
+    {file = "requests-2.28.1.tar.gz", hash = "sha256:7c5599b102feddaa661c826c56ab4fee28bfd17f5abca1ebbe3e7f19d7c97983"},
 ]
 
 [package.dependencies]
 certifi = ">=2017.4.17"
-charset-normalizer = ">=2,<4"
+charset-normalizer = ">=2,<3"
 idna = ">=2.5,<4"
-urllib3 = ">=1.21.1,<3"
+urllib3 = ">=1.21.1,<1.27"
 
 [package.extras]
 socks = ["PySocks (>=1.5.6,!=1.5.7)"]
@@ -2139,29 +2148,29 @@ pyasn1 = ">=0.1.3"
 
 [[package]]
 name = "ruff"
-version = "0.0.269"
+version = "0.0.255"
 description = "An extremely fast Python linter, written in Rust."
 category = "dev"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "ruff-0.0.269-py3-none-macosx_10_7_x86_64.whl", hash = "sha256:3569bcdee679045c09c0161fabc057599759c49219a08d9a4aad2cc3982ccba3"},
-    {file = "ruff-0.0.269-py3-none-macosx_10_9_x86_64.macosx_11_0_arm64.macosx_10_9_universal2.whl", hash = "sha256:56347da63757a56cbce7d4b3d6044ca4f1941cd1bbff3714f7554360c3361f83"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6da8ee25ef2f0cc6cc8e6e20942c1d44d25a36dce35070d7184655bc14f63f63"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:bd81b8e681b9eaa6cf15484f3985bd8bd97c3d114e95bff3e8ea283bf8865062"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:1f19f59ca3c28742955241fb452f3346241ddbd34e72ac5cb3d84fadebcf6bc8"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:f062059b8289a4fab7f6064601b811d447c2f9d3d432a17f689efe4d68988450"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:3f5dc7aac52c58e82510217e3c7efd80765c134c097c2815d59e40face0d1fe6"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e131b4dbe798c391090c6407641d6ab12c0fa1bb952379dde45e5000e208dabb"},
-    {file = "ruff-0.0.269-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a374434e588e06550df0f8dcb74777290f285678de991fda4e1063c367ab2eb2"},
-    {file = "ruff-0.0.269-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:cec2f4b84a14b87f1b121488649eb5b4eaa06467a2387373f750da74bdcb5679"},
-    {file = "ruff-0.0.269-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:374b161753a247904aec7a32d45e165302b76b6e83d22d099bf3ff7c232c888f"},
-    {file = "ruff-0.0.269-py3-none-musllinux_1_2_i686.whl", hash = "sha256:9ca0a1ddb1d835b5f742db9711c6cf59f213a1ad0088cb1e924a005fd399e7d8"},
-    {file = "ruff-0.0.269-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:5a20658f0b97d207c7841c13d528f36d666bf445b00b01139f28a8ccb80093bb"},
-    {file = "ruff-0.0.269-py3-none-win32.whl", hash = "sha256:03ff42bc91ceca58e0f0f072cb3f9286a9208f609812753474e799a997cdad1a"},
-    {file = "ruff-0.0.269-py3-none-win_amd64.whl", hash = "sha256:f3b59ccff57b21ef0967ea8021fd187ec14c528ec65507d8bcbe035912050776"},
-    {file = "ruff-0.0.269-py3-none-win_arm64.whl", hash = "sha256:bbeb857b1e508a4487bdb02ca1e6d41dd8d5ac5335a5246e25de8a3dff38c1ff"},
-    {file = "ruff-0.0.269.tar.gz", hash = "sha256:11ddcfbab32cf5c420ea9dd5531170ace5a3e59c16d9251c7bd2581f7b16f602"},
+    {file = "ruff-0.0.255-py3-none-macosx_10_7_x86_64.whl", hash = "sha256:b2d71fb6a7e50501a2473864acffc85dee6b750c25db198f7e71fe1dbbff1aad"},
+    {file = "ruff-0.0.255-py3-none-macosx_10_9_x86_64.macosx_11_0_arm64.macosx_10_9_universal2.whl", hash = "sha256:6c97d746861a6010f941179e84bba9feb8a871815667471d9ed6beb98d45c252"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9a7fa60085079b91a298b963361be9b1b1c724582af6c84be954cbabdbd9309a"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:c089f7141496334ab5a127b54ce55e41f0d6714e68a4453a1e09d2204cdea8c3"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:0423908caa7d437a416b853214565b9c33bbd1106c4f88147982216dddcbbd96"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_ppc64.manylinux2014_ppc64.whl", hash = "sha256:981493e92547cacbb8e0874904ec049fe744507ee890dc8736caf89a8864f9a7"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:59d5193d2aedb35db180824462b374dbcfc306b2e76076245088afa6e5837df2"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:dd5e00733c9d160c8a34a22e62b390da9d1e9f326676402421cb8c1236beefc3"},
+    {file = "ruff-0.0.255-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:694418cf41838bd19c6229e4e1b2d04505b1e6b86fe3ab81165484fc96d36f01"},
+    {file = "ruff-0.0.255-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:5d0408985c9777369daebb5d3340a99e9f7294bdd7120642239261508185cf89"},
+    {file = "ruff-0.0.255-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:abd6376ef9d12f370d95a8c7c98682fbb9bfedfba59f40e84a816fef8ddcb8de"},
+    {file = "ruff-0.0.255-py3-none-musllinux_1_2_i686.whl", hash = "sha256:f9b1a5df0bc09193cbef58a6f78e4a9a0b058a4f9733c0442866d078006d1bb9"},
+    {file = "ruff-0.0.255-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:6a25c5f4ff087445b2e1bbcb9963f2ae7c868d65e4a8d5f84c36c12f71571179"},
+    {file = "ruff-0.0.255-py3-none-win32.whl", hash = "sha256:1ff87a8310354f9f1a099625e54a27fdd6756d9cd2a40b45922f2e943daf982d"},
+    {file = "ruff-0.0.255-py3-none-win_amd64.whl", hash = "sha256:f3d8416be618f023f93ec4fd6ee3048585ef85dba9563b2a7e38fc7e5131d5b1"},
+    {file = "ruff-0.0.255-py3-none-win_arm64.whl", hash = "sha256:8ba124819624145d7b6b53add40c367c44318893215ffc1bfe3d72e0225a1c9c"},
+    {file = "ruff-0.0.255.tar.gz", hash = "sha256:f9eb1d3b2eecbeedae419fa494c4e2a5e4484baf93a1ce0f81eddb005e1919c5"},
 ]
 
 [[package]]
@@ -2262,7 +2271,7 @@ files = [
 name = "tomli"
 version = "2.0.1"
 description = "A lil' TOML parser"
-category = "main"
+category = "dev"
 optional = false
 python-versions = ">=3.7"
 files = [
@@ -2272,54 +2281,42 @@ files = [
 
 [[package]]
 name = "types-psutil"
-version = "5.9.5.12"
+version = "5.9.5.4"
 description = "Typing stubs for psutil"
 category = "main"
 optional = false
 python-versions = "*"
 files = [
-    {file = "types-psutil-5.9.5.12.tar.gz", hash = "sha256:61a91679d3fe737250013b624dca09375e7cc3ad77dcc734553746c429c02aca"},
-    {file = "types_psutil-5.9.5.12-py3-none-any.whl", hash = "sha256:e9a147b8561235c6afcce5aa1adb973fad9ab2c50cf89820697687f53510358f"},
+    {file = "types-psutil-5.9.5.4.tar.gz", hash = "sha256:aa09102b80c65a3b4573216614372398dab78972d650488eaff1ff05482cc18f"},
+    {file = "types_psutil-5.9.5.4-py3-none-any.whl", hash = "sha256:28e59764630187e462d43788efa16d59d5e77b510115f9e25901b2d4007fca62"},
 ]
 
 [[package]]
 name = "types-psycopg2"
-version = "2.9.21.10"
+version = "2.9.18"
 description = "Typing stubs for psycopg2"
 category = "main"
 optional = false
 python-versions = "*"
 files = [
-    {file = "types-psycopg2-2.9.21.10.tar.gz", hash = "sha256:c2600892312ae1c34e12f145749795d93dc4eac3ef7dbf8a9c1bfd45385e80d7"},
-    {file = "types_psycopg2-2.9.21.10-py3-none-any.whl", hash = "sha256:918224a0731a3650832e46633e720703b5beef7693a064e777d9748654fcf5e5"},
-]
-
-[[package]]
-name = "types-pytest-lazy-fixture"
-version = "0.6.3.3"
-description = "Typing stubs for pytest-lazy-fixture"
-category = "main"
-optional = false
-python-versions = "*"
-files = [
-    {file = "types-pytest-lazy-fixture-0.6.3.3.tar.gz", hash = "sha256:2ef79d66bcde0e50acdac8dc55074b9ae0d4cfaeabdd638f5522f4cac7c8a2c7"},
-    {file = "types_pytest_lazy_fixture-0.6.3.3-py3-none-any.whl", hash = "sha256:a56a55649147ff960ff79d4b2c781a4f769351abc1876873f3116d0bd0c96353"},
+    {file = "types-psycopg2-2.9.18.tar.gz", hash = "sha256:9b0e9e1f097b15cd9fa8aad2596a9e3082fd72f8d9cfe52b190cfa709105b6c0"},
+    {file = "types_psycopg2-2.9.18-py3-none-any.whl", hash = "sha256:14c779dcab18c31453fa1cad3cf4b1601d33540a344adead3c47a6b8091cd2fa"},
 ]
 
 [[package]]
 name = "types-requests"
-version = "2.31.0.0"
+version = "2.28.5"
 description = "Typing stubs for requests"
 category = "main"
 optional = false
 python-versions = "*"
 files = [
-    {file = "types-requests-2.31.0.0.tar.gz", hash = "sha256:c1c29d20ab8d84dff468d7febfe8e0cb0b4664543221b386605e14672b44ea25"},
-    {file = "types_requests-2.31.0.0-py3-none-any.whl", hash = "sha256:7c5cea7940f8e92ec560bbc468f65bf684aa3dcf0554a6f8c4710f5f708dc598"},
+    {file = "types-requests-2.28.5.tar.gz", hash = "sha256:ac618bfefcb3742eaf97c961e13e9e5a226e545eda4a3dbe293b898d40933ad1"},
+    {file = "types_requests-2.28.5-py3-none-any.whl", hash = "sha256:98ab647ae88b5e2c41d6d20cfcb5117da1bea561110000b6fdeeea07b3e89877"},
 ]
 
 [package.dependencies]
-types-urllib3 = "*"
+types-urllib3 = "<1.27"
 
 [[package]]
 name = "types-s3transfer"
@@ -2335,14 +2332,14 @@ files = [
 
 [[package]]
 name = "types-toml"
-version = "0.10.8.6"
+version = "0.10.8"
 description = "Typing stubs for toml"
 category = "main"
 optional = false
 python-versions = "*"
 files = [
-    {file = "types-toml-0.10.8.6.tar.gz", hash = "sha256:6d3ac79e36c9ee593c5d4fb33a50cca0e3adceb6ef5cff8b8e5aef67b4c4aaf2"},
-    {file = "types_toml-0.10.8.6-py3-none-any.whl", hash = "sha256:de7b2bb1831d6f7a4b554671ffe5875e729753496961b3e9b202745e4955dafa"},
+    {file = "types-toml-0.10.8.tar.gz", hash = "sha256:b7e7ea572308b1030dc86c3ba825c5210814c2825612ec679eb7814f8dd9295a"},
+    {file = "types_toml-0.10.8-py3-none-any.whl", hash = "sha256:8300fd093e5829eb9c1fba69cee38130347d4b74ddf32d0a7df650ae55c2b599"},
 ]
 
 [[package]]
@@ -2359,14 +2356,14 @@ files = [
 
 [[package]]
 name = "typing-extensions"
-version = "4.6.1"
+version = "4.3.0"
 description = "Backported and Experimental Type Hints for Python 3.7+"
 category = "main"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "typing_extensions-4.6.1-py3-none-any.whl", hash = "sha256:6bac751f4789b135c43228e72de18637e9a6c29d12777023a703fd1a6858469f"},
-    {file = "typing_extensions-4.6.1.tar.gz", hash = "sha256:558bc0c4145f01e6405f4a5fdbd82050bd221b119f4bf72a961a1cfd471349d6"},
+    {file = "typing_extensions-4.3.0-py3-none-any.whl", hash = "sha256:25642c956049920a5aa49edcdd6ab1e06d7e5d467fc00e0506c44ac86fbfca02"},
+    {file = "typing_extensions-4.3.0.tar.gz", hash = "sha256:e6d2677a32f47fc7eb2795db1dd15c1f34eff616bcaf2cfb5e997f854fa1c4a6"},
 ]
 
 [[package]]
@@ -2614,4 +2611,4 @@ testing = ["func-timeout", "jaraco.itertools", "pytest (>=6)", "pytest-black (>=
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.9"
-content-hash = "c6c217033f50430c31b0979b74db222e6bab2301abd8b9f0cce5a9d5bccc578f"
+content-hash = "b689ffd6eae32b966f1744b5ac3343fe0dd26b31ee1f50e13daf5045ee0623e1"

proxy/README.md

@@ -1,6 +1,6 @@
 # Proxy
 
-Proxy binary accepts `--auth-backend` CLI option, which determines auth scheme and cluster routing method. Following routing backends are currently implemented:
+Proxy binary accepts `--auth-backend` CLI option, which determines auth scheme and cluster routing method. Following backends are currently implemented:
 
 * console
   new SCRAM-based console API; uses SNI info to select the destination project (endpoint soon)
@@ -9,90 +9,6 @@ Proxy binary accepts `--auth-backend` CLI option, which determines auth scheme a
 * link
   sends login link for all usernames
 
-Also proxy can expose following services to the external world:
-
-* postgres protocol over TCP -- usual postgres endpoint compatible with usual
-  postgres drivers
-* postgres protocol over WebSockets -- same protocol tunneled over websockets
-  for environments where TCP connection is not available. We have our own
-  implementation of a client that uses node-postgres and tunnels traffic through
-  websockets: https://github.com/neondatabase/serverless
-* SQL over HTTP -- service that accepts POST requests with SQL text over HTTP
-  and responds with JSON-serialised results.
-
-
-## SQL over HTTP
-
-Contrary to the usual postgres proto over TCP and WebSockets using plain
-one-shot HTTP request achieves smaller amortized latencies in edge setups due to
-fewer round trips and an enhanced open connection reuse by the v8 engine. Also
-such endpoint could be used directly without any driver.
-
-To play with it locally one may start proxy over a local postgres installation
-(see end of this page on how to generate certs with openssl):
-
-```
-./target/debug/proxy -c server.crt -k server.key --auth-backend=postgres --auth-endpoint=postgres://stas@127.0.0.1:5432/stas --wss 0.0.0.0:4444
-```
-
-If both postgres and proxy are running you may send a SQL query:
-```json
-curl -k -X POST 'https://proxy.localtest.me:4444/sql' \
-  -H 'Neon-Connection-String: postgres://stas:pass@proxy.localtest.me:4444/postgres' \
-  -H 'Content-Type: application/json' \
-  --data '{
-    "query":"SELECT $1::int[] as arr, $2::jsonb as obj, 42 as num",
-    "params":[ "{{1,2},{\"3\",4}}", {"key":"val", "ikey":4242}]
-  }' | jq
-
-{
-  "command": "SELECT",
-  "fields": [
-    { "dataTypeID": 1007, "name": "arr" },
-    { "dataTypeID": 3802, "name": "obj" },
-    { "dataTypeID": 23, "name": "num" }
-  ],
-  "rowCount": 1,
-  "rows": [
-    {
-      "arr": [[1,2],[3,4]],
-      "num": 42,
-      "obj": {
-        "ikey": 4242,
-        "key": "val"
-      }
-    }
-  ]
-}
-```
-
-
-With the current approach we made the following design decisions:
-
-1. SQL injection protection: We employed the extended query protocol, modifying
-   the rust-postgres driver to send queries in one roundtrip using a text
-   protocol rather than binary, bypassing potential issues like those identified
-   in sfackler/rust-postgres#1030.
-
-2. Postgres type compatibility: As not all postgres types have binary
-   representations (e.g., acl's in pg_class), we adjusted rust-postgres to
-   respond with text protocol, simplifying serialization and fixing queries with
-   text-only types in response.
-
-3. Data type conversion: Considering JSON supports fewer data types than
-   Postgres, we perform conversions where possible, passing all other types as
-   strings. Key conversions include:
-   - postgres int2, int4, float4, float8 -> json number (NaN and Inf remain
-     text)
-   - postgres bool, null, text -> json bool, null, string
-   - postgres array -> json array
-   - postgres json and jsonb -> json object
-
-4. Alignment with node-postgres: To facilitate integration with js libraries,
-   we've matched the response structure of node-postgres, returning command tags
-   and column oids. Command tag capturing was added to the rust-postgres
-   functionality as part of this change.
-
 ## Using SNI-based routing on localhost
 
 Now proxy determines project name from the subdomain, request to the `round-rice-566201.somedomain.tld` will be routed to the project named `round-rice-566201`. Unfortunately, `/etc/hosts` does not support domain wildcards, so I usually use `*.localtest.me` which resolves to `127.0.0.1`. Now we can create self-signed certificate and play with proxy:

proxy/src/auth.rs

@@ -7,7 +7,6 @@ mod credentials;
 pub use credentials::ClientCredentials;
 
 mod password_hack;
-pub use password_hack::parse_endpoint_param;
 use password_hack::PasswordHackPayload;
 
 mod flow;
@@ -45,10 +44,10 @@ pub enum AuthErrorImpl {
     #[error(
         "Endpoint ID is not specified. \
         Either please upgrade the postgres client library (libpq) for SNI support \
-        or pass the endpoint ID (first part of the domain name) as a parameter: '?options=endpoint%3D<endpoint-id>'. \
+        or pass the endpoint ID (first part of the domain name) as a parameter: '?options=project%3D<endpoint-id>'. \
         See more at https://neon.tech/sni"
     )]
-    MissingEndpointName,
+    MissingProjectName,
 
     #[error("password authentication failed for user '{0}'")]
     AuthFailed(Box<str>),
@@ -89,7 +88,7 @@ impl UserFacingError for AuthError {
             AuthFailed(_) => self.to_string(),
             BadAuthMethod(_) => self.to_string(),
             MalformedPassword(_) => self.to_string(),
-            MissingEndpointName => self.to_string(),
+            MissingProjectName => self.to_string(),
             Io(_) => "Internal error".to_string(),
         }
     }

proxy/src/auth/backend.rs

@@ -139,16 +139,6 @@ async fn auth_quirks(
 }
 
 impl BackendType<'_, ClientCredentials<'_>> {
-    /// Get compute endpoint name from the credentials.
-    pub fn get_endpoint(&self) -> Option<String> {
-        use BackendType::*;
-
-        match self {
-            Console(_, creds) => creds.project.clone(),
-            Postgres(_, creds) => creds.project.clone(),
-            Link(_) => Some("link".to_owned()),
-        }
-    }
     /// Authenticate the client via the requested backend, possibly using credentials.
     #[tracing::instrument(fields(allow_cleartext = allow_cleartext), skip_all)]
     pub async fn authenticate(

proxy/src/auth/backend/hacks.rs

@@ -52,8 +52,8 @@ pub async fn password_hack(
         .authenticate()
         .await?;
 
-    info!(project = &payload.endpoint, "received missing parameter");
-    creds.project = Some(payload.endpoint);
+    info!(project = &payload.project, "received missing parameter");
+    creds.project = Some(payload.project);
 
     let mut node = api.wake_compute(extra, creds).await?;
     node.config.password(payload.password);

proxy/src/auth/credentials.rs

@@ -1,7 +1,6 @@
 //! User credentials used in authentication.
 
-use crate::{auth::password_hack::parse_endpoint_param, error::UserFacingError};
-use itertools::Itertools;
+use crate::error::UserFacingError;
 use pq_proto::StartupMessageParams;
 use std::collections::HashSet;
 use thiserror::Error;
@@ -62,15 +61,7 @@ impl<'a> ClientCredentials<'a> {
         // Project name might be passed via PG's command-line options.
         let project_option = params
             .options_raw()
-            .and_then(|options| {
-                // We support both `project` (deprecated) and `endpoint` options for backward compatibility.
-                // However, if both are present, we don't exactly know which one to use.
-                // Therefore we require that only one of them is present.
-                options
-                    .filter_map(parse_endpoint_param)
-                    .at_most_one()
-                    .ok()?
-            })
+            .and_then(|mut options| options.find_map(|opt| opt.strip_prefix("project=")))
             .map(|name| name.to_string());
 
         let project_from_domain = if let Some(sni_str) = sni {
@@ -186,51 +177,6 @@ mod tests {
         Ok(())
     }
 
-    #[test]
-    fn parse_endpoint_from_options() -> anyhow::Result<()> {
-        let options = StartupMessageParams::new([
-            ("user", "john_doe"),
-            ("options", "-ckey=1 endpoint=bar -c geqo=off"),
-        ]);
-
-        let creds = ClientCredentials::parse(&options, None, None)?;
-        assert_eq!(creds.user, "john_doe");
-        assert_eq!(creds.project.as_deref(), Some("bar"));
-
-        Ok(())
-    }
-
-    #[test]
-    fn parse_three_endpoints_from_options() -> anyhow::Result<()> {
-        let options = StartupMessageParams::new([
-            ("user", "john_doe"),
-            (
-                "options",
-                "-ckey=1 endpoint=one endpoint=two endpoint=three -c geqo=off",
-            ),
-        ]);
-
-        let creds = ClientCredentials::parse(&options, None, None)?;
-        assert_eq!(creds.user, "john_doe");
-        assert!(creds.project.is_none());
-
-        Ok(())
-    }
-
-    #[test]
-    fn parse_when_endpoint_and_project_are_in_options() -> anyhow::Result<()> {
-        let options = StartupMessageParams::new([
-            ("user", "john_doe"),
-            ("options", "-ckey=1 endpoint=bar project=foo -c geqo=off"),
-        ]);
-
-        let creds = ClientCredentials::parse(&options, None, None)?;
-        assert_eq!(creds.user, "john_doe");
-        assert!(creds.project.is_none());
-
-        Ok(())
-    }
-
     #[test]
     fn parse_projects_identical() -> anyhow::Result<()> {
         let options = StartupMessageParams::new([("user", "john_doe"), ("options", "project=baz")]);

proxy/src/auth/flow.rs

@@ -91,7 +91,7 @@ impl<S: AsyncRead + AsyncWrite + Unpin> AuthFlow<'_, S, PasswordHack> {
             // the user neither enabled SNI nor resorted to any other method
             // for passing the project name we rely on. We should show them
             // the most helpful error message and point to the documentation.
-            .ok_or(AuthErrorImpl::MissingEndpointName)?;
+            .ok_or(AuthErrorImpl::MissingProjectName)?;
 
         Ok(payload)
     }

proxy/src/auth/password_hack.rs

@@ -6,55 +6,27 @@
 use bstr::ByteSlice;
 
 pub struct PasswordHackPayload {
-    pub endpoint: String,
+    pub project: String,
     pub password: Vec<u8>,
 }
 
 impl PasswordHackPayload {
     pub fn parse(bytes: &[u8]) -> Option<Self> {
         // The format is `project=<utf-8>;<password-bytes>`.
-        let mut iter = bytes.splitn_str(2, ";");
-        let endpoint = iter.next()?.to_str().ok()?;
-        let endpoint = parse_endpoint_param(endpoint)?.to_owned();
+        let mut iter = bytes.strip_prefix(b"project=")?.splitn_str(2, ";");
+        let project = iter.next()?.to_str().ok()?.to_owned();
         let password = iter.next()?.to_owned();
 
-        Some(Self { endpoint, password })
+        Some(Self { project, password })
     }
 }
 
-pub fn parse_endpoint_param(bytes: &str) -> Option<&str> {
-    bytes
-        .strip_prefix("project=")
-        .or_else(|| bytes.strip_prefix("endpoint="))
-}
-
 #[cfg(test)]
 mod tests {
     use super::*;
 
     #[test]
-    fn parse_endpoint_param_fn() {
-        let input = "";
-        assert!(parse_endpoint_param(input).is_none());
-
-        let input = "project=";
-        assert_eq!(parse_endpoint_param(input), Some(""));
-
-        let input = "project=foobar";
-        assert_eq!(parse_endpoint_param(input), Some("foobar"));
-
-        let input = "endpoint=";
-        assert_eq!(parse_endpoint_param(input), Some(""));
-
-        let input = "endpoint=foobar";
-        assert_eq!(parse_endpoint_param(input), Some("foobar"));
-
-        let input = "other_option=foobar";
-        assert!(parse_endpoint_param(input).is_none());
-    }
-
-    #[test]
-    fn parse_password_hack_payload_project() {
+    fn parse_password_hack_payload() {
         let bytes = b"";
         assert!(PasswordHackPayload::parse(bytes).is_none());
 
@@ -62,33 +34,13 @@ mod tests {
         assert!(PasswordHackPayload::parse(bytes).is_none());
 
         let bytes = b"project=;";
-        let payload: PasswordHackPayload =
-            PasswordHackPayload::parse(bytes).expect("parsing failed");
-        assert_eq!(payload.endpoint, "");
+        let payload = PasswordHackPayload::parse(bytes).expect("parsing failed");
+        assert_eq!(payload.project, "");
         assert_eq!(payload.password, b"");
 
         let bytes = b"project=foobar;pass;word";
         let payload = PasswordHackPayload::parse(bytes).expect("parsing failed");
-        assert_eq!(payload.endpoint, "foobar");
-        assert_eq!(payload.password, b"pass;word");
-    }
-
-    #[test]
-    fn parse_password_hack_payload_endpoint() {
-        let bytes = b"";
-        assert!(PasswordHackPayload::parse(bytes).is_none());
-
-        let bytes = b"endpoint=";
-        assert!(PasswordHackPayload::parse(bytes).is_none());
-
-        let bytes = b"endpoint=;";
-        let payload = PasswordHackPayload::parse(bytes).expect("parsing failed");
-        assert_eq!(payload.endpoint, "");
-        assert_eq!(payload.password, b"");
-
-        let bytes = b"endpoint=foobar;pass;word";
-        let payload = PasswordHackPayload::parse(bytes).expect("parsing failed");
-        assert_eq!(payload.endpoint, "foobar");
+        assert_eq!(payload.project, "foobar");
         assert_eq!(payload.password, b"pass;word");
     }
 }

proxy/src/compute.rs

@@ -1,4 +1,4 @@
-use crate::{auth::parse_endpoint_param, cancellation::CancelClosure, error::UserFacingError};
+use crate::{cancellation::CancelClosure, error::UserFacingError};
 use futures::{FutureExt, TryFutureExt};
 use itertools::Itertools;
 use pq_proto::StartupMessageParams;
@@ -279,7 +279,7 @@ fn filtered_options(params: &StartupMessageParams) -> Option<String> {
     #[allow(unstable_name_collisions)]
     let options: String = params
         .options_raw()?
-        .filter(|opt| parse_endpoint_param(opt).is_none())
+        .filter(|opt| !opt.starts_with("project="))
         .intersperse(" ") // TODO: use impl from std once it's stabilized
         .collect();
 

proxy/src/config.rs

@@ -100,10 +100,9 @@ impl CertResolver {
         is_default: bool,
     ) -> anyhow::Result<()> {
         let priv_key = {
-            let key_bytes = std::fs::read(key_path)
-                .context(format!("Failed to read TLS keys at '{key_path}'"))?;
+            let key_bytes = std::fs::read(key_path).context("TLS key file")?;
             let mut keys = rustls_pemfile::pkcs8_private_keys(&mut &key_bytes[..])
-                .context(format!("Failed to parse TLS keys at '{key_path}'"))?;
+                .context(format!("Failed to read TLS keys at '{key_path}'"))?;
 
             ensure!(keys.len() == 1, "keys.len() = {} (should be 1)", keys.len());
             keys.pop().map(rustls::PrivateKey).unwrap()

proxy/src/http.rs

@@ -3,7 +3,6 @@
 //! directly relying on deps like `reqwest` (think loose coupling).
 
 pub mod server;
-pub mod sql_over_http;
 pub mod websocket;
 
 pub use reqwest::{Request, Response, StatusCode};

proxy/src/http/sql_over_http.rs

@@ -1,603 +0,0 @@
-use futures::pin_mut;
-use futures::StreamExt;
-use hyper::body::HttpBody;
-use hyper::{Body, HeaderMap, Request};
-use pq_proto::StartupMessageParams;
-use serde_json::json;
-use serde_json::Map;
-use serde_json::Value;
-use tokio_postgres::types::Kind;
-use tokio_postgres::types::Type;
-use tokio_postgres::Row;
-use url::Url;
-
-use crate::{auth, config::ProxyConfig, console};
-
-#[derive(serde::Deserialize)]
-struct QueryData {
-    query: String,
-    params: Vec<serde_json::Value>,
-}
-
-const APP_NAME: &str = "sql_over_http";
-const MAX_RESPONSE_SIZE: usize = 1024 * 1024; // 1 MB
-const MAX_REQUEST_SIZE: u64 = 1024 * 1024; // 1 MB
-
-//
-// Convert json non-string types to strings, so that they can be passed to Postgres
-// as parameters.
-//
-fn json_to_pg_text(json: Vec<Value>) -> Result<Vec<String>, serde_json::Error> {
-    json.iter()
-        .map(|value| {
-            match value {
-                Value::Null => serde_json::to_string(value),
-                Value::Bool(_) => serde_json::to_string(value),
-                Value::Number(_) => serde_json::to_string(value),
-                Value::Object(_) => serde_json::to_string(value),
-
-                // no need to escape
-                Value::String(s) => Ok(s.to_string()),
-
-                // special care for arrays
-                Value::Array(_) => json_array_to_pg_array(value),
-            }
-        })
-        .collect()
-}
-
-//
-// Serialize a JSON array to a Postgres array. Contrary to the strings in the params
-// in the array we need to escape the strings. Postgres is okay with arrays of form
-// '{1,"2",3}'::int[], so we don't check that array holds values of the same type, leaving
-// it for Postgres to check.
-//
-// Example of the same escaping in node-postgres: packages/pg/lib/utils.js
-//
-fn json_array_to_pg_array(value: &Value) -> Result<String, serde_json::Error> {
-    match value {
-        // same
-        Value::Null => serde_json::to_string(value),
-        Value::Bool(_) => serde_json::to_string(value),
-        Value::Number(_) => serde_json::to_string(value),
-        Value::Object(_) => serde_json::to_string(value),
-
-        // now needs to be escaped, as it is part of the array
-        Value::String(_) => serde_json::to_string(value),
-
-        // recurse into array
-        Value::Array(arr) => {
-            let vals = arr
-                .iter()
-                .map(json_array_to_pg_array)
-                .collect::<Result<Vec<_>, _>>()?
-                .join(",");
-            Ok(format!("{{{}}}", vals))
-        }
-    }
-}
-
-fn get_conn_info(
-    headers: &HeaderMap,
-    sni_hostname: Option<String>,
-) -> Result<(String, String, String, String), anyhow::Error> {
-    let connection_string = headers
-        .get("Neon-Connection-String")
-        .ok_or(anyhow::anyhow!("missing connection string"))?
-        .to_str()?;
-
-    let connection_url = Url::parse(connection_string)?;
-
-    let protocol = connection_url.scheme();
-    if protocol != "postgres" && protocol != "postgresql" {
-        return Err(anyhow::anyhow!(
-            "connection string must start with postgres: or postgresql:"
-        ));
-    }
-
-    let mut url_path = connection_url
-        .path_segments()
-        .ok_or(anyhow::anyhow!("missing database name"))?;
-
-    let dbname = url_path
-        .next()
-        .ok_or(anyhow::anyhow!("invalid database name"))?;
-
-    let username = connection_url.username();
-    if username.is_empty() {
-        return Err(anyhow::anyhow!("missing username"));
-    }
-
-    let password = connection_url
-        .password()
-        .ok_or(anyhow::anyhow!("no password"))?;
-
-    // TLS certificate selector now based on SNI hostname, so if we are running here
-    // we are sure that SNI hostname is set to one of the configured domain names.
-    let sni_hostname = sni_hostname.ok_or(anyhow::anyhow!("no SNI hostname set"))?;
-
-    let hostname = connection_url
-        .host_str()
-        .ok_or(anyhow::anyhow!("no host"))?;
-
-    let host_header = headers
-        .get("host")
-        .and_then(|h| h.to_str().ok())
-        .and_then(|h| h.split(':').next());
-
-    if hostname != sni_hostname {
-        return Err(anyhow::anyhow!("mismatched SNI hostname and hostname"));
-    } else if let Some(h) = host_header {
-        if h != hostname {
-            return Err(anyhow::anyhow!("mismatched host header and hostname"));
-        }
-    }
-
-    Ok((
-        username.to_owned(),
-        dbname.to_owned(),
-        hostname.to_owned(),
-        password.to_owned(),
-    ))
-}
-
-// TODO: return different http error codes
-pub async fn handle(
-    config: &'static ProxyConfig,
-    request: Request<Body>,
-    sni_hostname: Option<String>,
-) -> anyhow::Result<Value> {
-    //
-    // Determine the destination and connection params
-    //
-    let headers = request.headers();
-    let (username, dbname, hostname, password) = get_conn_info(headers, sni_hostname)?;
-    let credential_params = StartupMessageParams::new([
-        ("user", &username),
-        ("database", &dbname),
-        ("application_name", APP_NAME),
-    ]);
-
-    //
-    // Wake up the destination if needed. Code here is a bit involved because
-    // we reuse the code from the usual proxy and we need to prepare few structures
-    // that this code expects.
-    //
-    let tls = config.tls_config.as_ref();
-    let common_names = tls.and_then(|tls| tls.common_names.clone());
-    let creds = config
-        .auth_backend
-        .as_ref()
-        .map(|_| auth::ClientCredentials::parse(&credential_params, Some(&hostname), common_names))
-        .transpose()?;
-    let extra = console::ConsoleReqExtra {
-        session_id: uuid::Uuid::new_v4(),
-        application_name: Some(APP_NAME),
-    };
-    let node = creds.wake_compute(&extra).await?.expect("msg");
-    let conf = node.value.config;
-    let port = *conf.get_ports().first().expect("no port");
-    let host = match conf.get_hosts().first().expect("no host") {
-        tokio_postgres::config::Host::Tcp(host) => host,
-        tokio_postgres::config::Host::Unix(_) => {
-            return Err(anyhow::anyhow!("unix socket is not supported"));
-        }
-    };
-
-    let request_content_length = match request.body().size_hint().upper() {
-        Some(v) => v,
-        None => MAX_REQUEST_SIZE + 1,
-    };
-
-    if request_content_length > MAX_REQUEST_SIZE {
-        return Err(anyhow::anyhow!(
-            "request is too large (max {MAX_REQUEST_SIZE} bytes)"
-        ));
-    }
-
-    //
-    // Read the query and query params from the request body
-    //
-    let body = hyper::body::to_bytes(request.into_body()).await?;
-    let QueryData { query, params } = serde_json::from_slice(&body)?;
-    let query_params = json_to_pg_text(params)?;
-
-    //
-    // Connenct to the destination
-    //
-    let (client, connection) = tokio_postgres::Config::new()
-        .host(host)
-        .port(port)
-        .user(&username)
-        .password(&password)
-        .dbname(&dbname)
-        .max_backend_message_size(MAX_RESPONSE_SIZE)
-        .connect(tokio_postgres::NoTls)
-        .await?;
-
-    tokio::spawn(async move {
-        if let Err(e) = connection.await {
-            eprintln!("connection error: {}", e);
-        }
-    });
-
-    //
-    // Now execute the query and return the result
-    //
-    let row_stream = client.query_raw_txt(query, query_params).await?;
-
-    // Manually drain the stream into a vector to leave row_stream hanging
-    // around to get a command tag. Also check that the response is not too
-    // big.
-    pin_mut!(row_stream);
-    let mut rows: Vec<tokio_postgres::Row> = Vec::new();
-    let mut curret_size = 0;
-    while let Some(row) = row_stream.next().await {
-        let row = row?;
-        curret_size += row.body_len();
-        rows.push(row);
-        if curret_size > MAX_RESPONSE_SIZE {
-            return Err(anyhow::anyhow!("response too large"));
-        }
-    }
-
-    // grab the command tag and number of rows affected
-    let command_tag = row_stream.command_tag().unwrap_or_default();
-    let mut command_tag_split = command_tag.split(' ');
-    let command_tag_name = command_tag_split.next().unwrap_or_default();
-    let command_tag_count = if command_tag_name == "INSERT" {
-        // INSERT returns OID first and then number of rows
-        command_tag_split.nth(1)
-    } else {
-        // other commands return number of rows (if any)
-        command_tag_split.next()
-    }
-    .and_then(|s| s.parse::<i64>().ok());
-
-    let fields = if !rows.is_empty() {
-        rows[0]
-            .columns()
-            .iter()
-            .map(|c| {
-                json!({
-                    "name": Value::String(c.name().to_owned()),
-                    "dataTypeID": Value::Number(c.type_().oid().into()),
-                })
-            })
-            .collect::<Vec<_>>()
-    } else {
-        Vec::new()
-    };
-
-    // convert rows to JSON
-    let rows = rows
-        .iter()
-        .map(pg_text_row_to_json)
-        .collect::<Result<Vec<_>, _>>()?;
-
-    // resulting JSON format is based on the format of node-postgres result
-    Ok(json!({
-        "command": command_tag_name,
-        "rowCount": command_tag_count,
-        "rows": rows,
-        "fields": fields,
-    }))
-}
-
-//
-// Convert postgres row with text-encoded values to JSON object
-//
-pub fn pg_text_row_to_json(row: &Row) -> Result<Value, anyhow::Error> {
-    let res = row
-        .columns()
-        .iter()
-        .enumerate()
-        .map(|(i, column)| {
-            let name = column.name();
-            let pg_value = row.as_text(i)?;
-            let json_value = pg_text_to_json(pg_value, column.type_())?;
-            Ok((name.to_string(), json_value))
-        })
-        .collect::<Result<Map<String, Value>, anyhow::Error>>()?;
-
-    Ok(Value::Object(res))
-}
-
-//
-// Convert postgres text-encoded value to JSON value
-//
-pub fn pg_text_to_json(pg_value: Option<&str>, pg_type: &Type) -> Result<Value, anyhow::Error> {
-    if let Some(val) = pg_value {
-        if val == "NULL" {
-            return Ok(Value::Null);
-        }
-
-        if let Kind::Array(elem_type) = pg_type.kind() {
-            return pg_array_parse(val, elem_type);
-        }
-
-        match *pg_type {
-            Type::BOOL => Ok(Value::Bool(val == "t")),
-            Type::INT2 | Type::INT4 => {
-                let val = val.parse::<i32>()?;
-                Ok(Value::Number(serde_json::Number::from(val)))
-            }
-            Type::FLOAT4 | Type::FLOAT8 => {
-                let fval = val.parse::<f64>()?;
-                let num = serde_json::Number::from_f64(fval);
-                if let Some(num) = num {
-                    Ok(Value::Number(num))
-                } else {
-                    // Pass Nan, Inf, -Inf as strings
-                    // JS JSON.stringify() does converts them to null, but we
-                    // want to preserve them, so we pass them as strings
-                    Ok(Value::String(val.to_string()))
-                }
-            }
-            Type::JSON | Type::JSONB => Ok(serde_json::from_str(val)?),
-            _ => Ok(Value::String(val.to_string())),
-        }
-    } else {
-        Ok(Value::Null)
-    }
-}
-
-//
-// Parse postgres array into JSON array.
-//
-// This is a bit involved because we need to handle nested arrays and quoted
-// values. Unlike postgres we don't check that all nested arrays have the same
-// dimensions, we just return them as is.
-//
-fn pg_array_parse(pg_array: &str, elem_type: &Type) -> Result<Value, anyhow::Error> {
-    _pg_array_parse(pg_array, elem_type, false).map(|(v, _)| v)
-}
-
-fn _pg_array_parse(
-    pg_array: &str,
-    elem_type: &Type,
-    nested: bool,
-) -> Result<(Value, usize), anyhow::Error> {
-    let mut pg_array_chr = pg_array.char_indices();
-    let mut level = 0;
-    let mut quote = false;
-    let mut entries: Vec<Value> = Vec::new();
-    let mut entry = String::new();
-
-    // skip bounds decoration
-    if let Some('[') = pg_array.chars().next() {
-        for (_, c) in pg_array_chr.by_ref() {
-            if c == '=' {
-                break;
-            }
-        }
-    }
-
-    while let Some((mut i, mut c)) = pg_array_chr.next() {
-        let mut escaped = false;
-
-        if c == '\\' {
-            escaped = true;
-            (i, c) = pg_array_chr.next().unwrap();
-        }
-
-        match c {
-            '{' if !quote => {
-                level += 1;
-                if level > 1 {
-                    let (res, off) = _pg_array_parse(&pg_array[i..], elem_type, true)?;
-                    entries.push(res);
-                    for _ in 0..off - 1 {
-                        pg_array_chr.next();
-                    }
-                }
-            }
-            '}' => {
-                level -= 1;
-                if level == 0 {
-                    if !entry.is_empty() {
-                        entries.push(pg_text_to_json(Some(&entry), elem_type)?);
-                    }
-                    if nested {
-                        return Ok((Value::Array(entries), i));
-                    }
-                }
-            }
-            '"' if !escaped => {
-                if quote {
-                    // push even if empty
-                    entries.push(pg_text_to_json(Some(&entry), elem_type)?);
-                    entry = String::new();
-                }
-                quote = !quote;
-            }
-            ',' if !quote => {
-                if !entry.is_empty() {
-                    entries.push(pg_text_to_json(Some(&entry), elem_type)?);
-                    entry = String::new();
-                }
-            }
-            _ => {
-                entry.push(c);
-            }
-        }
-    }
-
-    if level != 0 {
-        return Err(anyhow::anyhow!("unbalanced array"));
-    }
-
-    Ok((Value::Array(entries), 0))
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use serde_json::json;
-
-    #[test]
-    fn test_atomic_types_to_pg_params() {
-        let json = vec![Value::Bool(true), Value::Bool(false)];
-        let pg_params = json_to_pg_text(json).unwrap();
-        assert_eq!(pg_params, vec!["true", "false"]);
-
-        let json = vec![Value::Number(serde_json::Number::from(42))];
-        let pg_params = json_to_pg_text(json).unwrap();
-        assert_eq!(pg_params, vec!["42"]);
-
-        let json = vec![Value::String("foo\"".to_string())];
-        let pg_params = json_to_pg_text(json).unwrap();
-        assert_eq!(pg_params, vec!["foo\""]);
-
-        let json = vec![Value::Null];
-        let pg_params = json_to_pg_text(json).unwrap();
-        assert_eq!(pg_params, vec!["null"]);
-    }
-
-    #[test]
-    fn test_json_array_to_pg_array() {
-        // atoms and escaping
-        let json = "[true, false, null, 42, \"foo\", \"bar\\\"-\\\\\"]";
-        let json: Value = serde_json::from_str(json).unwrap();
-        let pg_params = json_to_pg_text(vec![json]).unwrap();
-        assert_eq!(
-            pg_params,
-            vec!["{true,false,null,42,\"foo\",\"bar\\\"-\\\\\"}"]
-        );
-
-        // nested arrays
-        let json = "[[true, false], [null, 42], [\"foo\", \"bar\\\"-\\\\\"]]";
-        let json: Value = serde_json::from_str(json).unwrap();
-        let pg_params = json_to_pg_text(vec![json]).unwrap();
-        assert_eq!(
-            pg_params,
-            vec!["{{true,false},{null,42},{\"foo\",\"bar\\\"-\\\\\"}}"]
-        );
-    }
-
-    #[test]
-    fn test_atomic_types_parse() {
-        assert_eq!(
-            pg_text_to_json(Some("foo"), &Type::TEXT).unwrap(),
-            json!("foo")
-        );
-        assert_eq!(pg_text_to_json(None, &Type::TEXT).unwrap(), json!(null));
-        assert_eq!(pg_text_to_json(Some("42"), &Type::INT4).unwrap(), json!(42));
-        assert_eq!(pg_text_to_json(Some("42"), &Type::INT2).unwrap(), json!(42));
-        assert_eq!(
-            pg_text_to_json(Some("42"), &Type::INT8).unwrap(),
-            json!("42")
-        );
-        assert_eq!(
-            pg_text_to_json(Some("42.42"), &Type::FLOAT8).unwrap(),
-            json!(42.42)
-        );
-        assert_eq!(
-            pg_text_to_json(Some("42.42"), &Type::FLOAT4).unwrap(),
-            json!(42.42)
-        );
-        assert_eq!(
-            pg_text_to_json(Some("NaN"), &Type::FLOAT4).unwrap(),
-            json!("NaN")
-        );
-        assert_eq!(
-            pg_text_to_json(Some("Infinity"), &Type::FLOAT4).unwrap(),
-            json!("Infinity")
-        );
-        assert_eq!(
-            pg_text_to_json(Some("-Infinity"), &Type::FLOAT4).unwrap(),
-            json!("-Infinity")
-        );
-
-        let json: Value =
-            serde_json::from_str("{\"s\":\"str\",\"n\":42,\"f\":4.2,\"a\":[null,3,\"a\"]}")
-                .unwrap();
-        assert_eq!(
-            pg_text_to_json(
-                Some(r#"{"s":"str","n":42,"f":4.2,"a":[null,3,"a"]}"#),
-                &Type::JSONB
-            )
-            .unwrap(),
-            json
-        );
-    }
-
-    #[test]
-    fn test_pg_array_parse_text() {
-        fn pt(pg_arr: &str) -> Value {
-            pg_array_parse(pg_arr, &Type::TEXT).unwrap()
-        }
-        assert_eq!(
-            pt(r#"{"aa\"\\\,a",cha,"bbbb"}"#),
-            json!(["aa\"\\,a", "cha", "bbbb"])
-        );
-        assert_eq!(
-            pt(r#"{{"foo","bar"},{"bee","bop"}}"#),
-            json!([["foo", "bar"], ["bee", "bop"]])
-        );
-        assert_eq!(
-            pt(r#"{{{{"foo",NULL,"bop",bup}}}}"#),
-            json!([[[["foo", null, "bop", "bup"]]]])
-        );
-        assert_eq!(
-            pt(r#"{{"1",2,3},{4,NULL,6},{NULL,NULL,NULL}}"#),
-            json!([["1", "2", "3"], ["4", null, "6"], [null, null, null]])
-        );
-    }
-
-    #[test]
-    fn test_pg_array_parse_bool() {
-        fn pb(pg_arr: &str) -> Value {
-            pg_array_parse(pg_arr, &Type::BOOL).unwrap()
-        }
-        assert_eq!(pb(r#"{t,f,t}"#), json!([true, false, true]));
-        assert_eq!(pb(r#"{{t,f,t}}"#), json!([[true, false, true]]));
-        assert_eq!(
-            pb(r#"{{t,f},{f,t}}"#),
-            json!([[true, false], [false, true]])
-        );
-        assert_eq!(
-            pb(r#"{{t,NULL},{NULL,f}}"#),
-            json!([[true, null], [null, false]])
-        );
-    }
-
-    #[test]
-    fn test_pg_array_parse_numbers() {
-        fn pn(pg_arr: &str, ty: &Type) -> Value {
-            pg_array_parse(pg_arr, ty).unwrap()
-        }
-        assert_eq!(pn(r#"{1,2,3}"#, &Type::INT4), json!([1, 2, 3]));
-        assert_eq!(pn(r#"{1,2,3}"#, &Type::INT2), json!([1, 2, 3]));
-        assert_eq!(pn(r#"{1,2,3}"#, &Type::INT8), json!(["1", "2", "3"]));
-        assert_eq!(pn(r#"{1,2,3}"#, &Type::FLOAT4), json!([1.0, 2.0, 3.0]));
-        assert_eq!(pn(r#"{1,2,3}"#, &Type::FLOAT8), json!([1.0, 2.0, 3.0]));
-        assert_eq!(
-            pn(r#"{1.1,2.2,3.3}"#, &Type::FLOAT4),
-            json!([1.1, 2.2, 3.3])
-        );
-        assert_eq!(
-            pn(r#"{1.1,2.2,3.3}"#, &Type::FLOAT8),
-            json!([1.1, 2.2, 3.3])
-        );
-        assert_eq!(
-            pn(r#"{NaN,Infinity,-Infinity}"#, &Type::FLOAT4),
-            json!(["NaN", "Infinity", "-Infinity"])
-        );
-        assert_eq!(
-            pn(r#"{NaN,Infinity,-Infinity}"#, &Type::FLOAT8),
-            json!(["NaN", "Infinity", "-Infinity"])
-        );
-    }
-
-    #[test]
-    fn test_pg_array_with_decoration() {
-        fn p(pg_arr: &str) -> Value {
-            pg_array_parse(pg_arr, &Type::INT2).unwrap()
-        }
-        assert_eq!(
-            p(r#"[1:1][-2:-1][3:5]={{{1,2,3},{4,5,6}}}"#),
-            json!([[[1, 2, 3], [4, 5, 6]]])
-        );
-    }
-}

proxy/src/http/websocket.rs

@@ -4,17 +4,12 @@ use crate::{
 use bytes::{Buf, Bytes};
 use futures::{Sink, Stream, StreamExt};
 use hyper::{
-    server::{
-        accept,
-        conn::{AddrIncoming, AddrStream},
-    },
+    server::{accept, conn::AddrIncoming},
     upgrade::Upgraded,
-    Body, Method, Request, Response, StatusCode,
+    Body, Request, Response, StatusCode,
 };
 use hyper_tungstenite::{tungstenite::Message, HyperWebsocket, WebSocketStream};
 use pin_project_lite::pin_project;
-use serde_json::{json, Value};
-
 use std::{
     convert::Infallible,
     future::ready,
@@ -26,7 +21,6 @@ use tls_listener::TlsListener;
 use tokio::{
     io::{self, AsyncBufRead, AsyncRead, AsyncWrite, ReadBuf},
     net::TcpListener,
-    select,
 };
 use tokio_util::sync::CancellationToken;
 use tracing::{error, info, info_span, warn, Instrument};
@@ -36,8 +30,6 @@ use utils::http::{error::ApiError, json::json_response};
 // Tracking issue: https://github.com/rust-lang/rust/issues/98407.
 use sync_wrapper::SyncWrapper;
 
-use super::sql_over_http;
-
 pin_project! {
     /// This is a wrapper around a [`WebSocketStream`] that
     /// implements [`AsyncRead`] and [`AsyncWrite`].
@@ -167,7 +159,6 @@ async fn ws_handler(
     config: &'static ProxyConfig,
     cancel_map: Arc<CancelMap>,
     session_id: uuid::Uuid,
-    sni_hostname: Option<String>,
 ) -> Result<Response<Body>, ApiError> {
     let host = request
         .headers()
@@ -190,44 +181,8 @@ async fn ws_handler(
 
         // Return the response so the spawned future can continue.
         Ok(response)
-    // TODO: that deserves a refactor as now this function also handles http json client besides websockets.
-    // Right now I don't want to blow up sql-over-http patch with file renames and do that as a follow up instead.
-    } else if request.uri().path() == "/sql" && request.method() == Method::POST {
-        let result = select! {
-            _ = tokio::time::sleep(std::time::Duration::from_secs(10)) => {
-                Err(anyhow::anyhow!("Query timed out"))
-            }
-            response = sql_over_http::handle(config, request, sni_hostname) => {
-                response
-            }
-        };
-        let status_code = match result {
-            Ok(_) => StatusCode::OK,
-            Err(_) => StatusCode::BAD_REQUEST,
-        };
-        let json = match result {
-            Ok(r) => r,
-            Err(e) => {
-                let message = format!("{:?}", e);
-                let code = match e.downcast_ref::<tokio_postgres::Error>() {
-                    Some(e) => match e.code() {
-                        Some(e) => serde_json::to_value(e.code()).unwrap(),
-                        None => Value::Null,
-                    },
-                    None => Value::Null,
-                };
-                json!({ "message": message, "code": code })
-            }
-        };
-        json_response(status_code, json).map(|mut r| {
-            r.headers_mut().insert(
-                "Access-Control-Allow-Origin",
-                hyper::http::HeaderValue::from_static("*"),
-            );
-            r
-        })
     } else {
-        json_response(StatusCode::BAD_REQUEST, "query is not supported")
+        json_response(StatusCode::OK, "Connect with a websocket client")
     }
 }
 
@@ -261,27 +216,20 @@ pub async fn task_main(
         }
     });
 
-    let make_svc =
-        hyper::service::make_service_fn(|stream: &tokio_rustls::server::TlsStream<AddrStream>| {
-            let sni_name = stream.get_ref().1.sni_hostname().map(|s| s.to_string());
-
-            async move {
-                Ok::<_, Infallible>(hyper::service::service_fn(move |req: Request<Body>| {
-                    let sni_name = sni_name.clone();
-                    async move {
-                        let cancel_map = Arc::new(CancelMap::default());
-                        let session_id = uuid::Uuid::new_v4();
-
-                        ws_handler(req, config, cancel_map, session_id, sni_name)
-                            .instrument(info_span!(
-                                "ws-client",
-                                session = format_args!("{session_id}")
-                            ))
-                            .await
-                    }
-                }))
-            }
-        });
+    let make_svc = hyper::service::make_service_fn(|_stream| async move {
+        Ok::<_, Infallible>(hyper::service::service_fn(
+            move |req: Request<Body>| async move {
+                let cancel_map = Arc::new(CancelMap::default());
+                let session_id = uuid::Uuid::new_v4();
+                ws_handler(req, config, cancel_map, session_id)
+                    .instrument(info_span!(
+                        "ws-client",
+                        session = format_args!("{session_id}")
+                    ))
+                    .await
+            },
+        ))
+    });
 
     hyper::Server::builder(accept::from_stream(tls_listener))
         .serve(make_svc)

proxy/src/proxy.rs

@@ -455,9 +455,6 @@ impl<'a, S> Client<'a, S> {
 
 impl<S: AsyncRead + AsyncWrite + Unpin> Client<'_, S> {
     /// Let the client authenticate and connect to the designated compute node.
-    // Instrumentation logs endpoint name everywhere. Doesn't work for link
-    // auth; strictly speaking we don't know endpoint name in its case.
-    #[tracing::instrument(name = "", fields(ep = self.creds.get_endpoint().unwrap_or("".to_owned())), skip_all)]
     async fn connect_to_db(
         self,
         session: cancellation::Session<'_>,

pyproject.toml

@@ -6,41 +6,40 @@ authors = []
 
 [tool.poetry.dependencies]
 python = "^3.9"
-pytest = "^7.3.1"
+pytest = "^6.2.5"
 psycopg2-binary = "^2.9.1"
-typing-extensions = "^4.6.1"
+typing-extensions = "^4.1.0"
 PyJWT = {version = "^2.1.0", extras = ["crypto"]}
-requests = "^2.31.0"
-pytest-xdist = "^3.3.1"
+requests = "^2.26.0"
+pytest-xdist = "^3.0.2"
 asyncpg = "^0.27.0"
 aiopg = "^1.3.1"
 Jinja2 = "^3.0.2"
-types-requests = "^2.31.0.0"
-types-psycopg2 = "^2.9.21.10"
+types-requests = "^2.28.5"
+types-psycopg2 = "^2.9.18"
 boto3 = "^1.26.16"
 boto3-stubs = {extras = ["s3"], version = "^1.26.16"}
 moto = {extras = ["server"], version = "^4.1.2"}
-backoff = "^2.2.1"
+backoff = "^1.11.1"
 pytest-lazy-fixture = "^0.6.3"
 prometheus-client = "^0.14.1"
 pytest-timeout = "^2.1.0"
 Werkzeug = "^2.2.3"
-pytest-order = "^1.1.0"
-allure-pytest = "^2.13.2"
-pytest-asyncio = "^0.21.0"
+pytest-order = "^1.0.1"
+allure-pytest = "^2.13.1"
+pytest-asyncio = "^0.19.0"
 toml = "^0.10.2"
 psutil = "^5.9.4"
-types-psutil = "^5.9.5.12"
-types-toml = "^0.10.8.6"
-pytest-httpserver = "^1.0.8"
+types-psutil = "^5.9.5.4"
+types-toml = "^0.10.8"
+pytest-httpserver = "^1.0.6"
 aiohttp = "3.7.4"
 pytest-rerunfailures = "^11.1.2"
-types-pytest-lazy-fixture = "^0.6.3.3"
 
 [tool.poetry.group.dev.dependencies]
-black = "^23.3.0"
-mypy = "==1.3.0"
-ruff = "^0.0.269"
+black = "^23.1.0"
+mypy = "==1.1.1"
+ruff = "^0.0.255"
 
 [build-system]
 requires = ["poetry-core>=1.0.0"]

scripts/export_import_between_pageservers.py

@@ -162,7 +162,7 @@ class PgProtocol:
         Returns psycopg2's connection object.
         This method passes all extra params to connstr.
         """
-        conn: PgConnection = psycopg2.connect(**self.conn_options(**kwargs))
+        conn = psycopg2.connect(**self.conn_options(**kwargs))
 
         # WARNING: this setting affects *all* tests!
         conn.autocommit = autocommit
@@ -535,8 +535,8 @@ def export_timeline(
 
 
 def main(args: argparse.Namespace):
-    # any psql version will do here. use current DEFAULT_PG_VERSION = 15
-    psql_path = str(Path(args.pg_distrib_dir) / "v15" / "bin" / "psql")
+    # any psql version will do here. use current DEFAULT_PG_VERSION = 14
+    psql_path = str(Path(args.pg_distrib_dir) / "v14" / "bin" / "psql")
 
     old_pageserver_host = args.old_pageserver_host
     new_pageserver_host = args.new_pageserver_host

scripts/flaky_tests.py

@@ -21,7 +21,6 @@ FLAKY_TESTS_QUERY = """
                 jsonb_array_elements(jsonb_array_elements(data -> 'children') -> 'children') -> 'name' as suite,
                 jsonb_array_elements(jsonb_array_elements(jsonb_array_elements(data -> 'children') -> 'children') -> 'children') -> 'name' as test,
                 jsonb_array_elements(jsonb_array_elements(jsonb_array_elements(data -> 'children') -> 'children') -> 'children') -> 'status' as status,
-                jsonb_array_elements(jsonb_array_elements(jsonb_array_elements(data -> 'children') -> 'children') -> 'children') -> 'retriesStatusChange' as retries_status_change,
                 to_timestamp((jsonb_array_elements(jsonb_array_elements(jsonb_array_elements(data -> 'children') -> 'children') -> 'children') -> 'time' -> 'start')::bigint / 1000)::date as timestamp
             FROM
                 regress_test_results
@@ -30,7 +29,7 @@ FLAKY_TESTS_QUERY = """
         ) data
     WHERE
         timestamp > CURRENT_DATE - INTERVAL '%s' day
-        AND (status::text IN ('"failed"', '"broken"') OR retries_status_change::boolean)
+        AND status::text IN ('"failed"', '"broken"')
     ;
 """
 

scripts/ingest_perf_test_result.py

@@ -35,7 +35,7 @@ def get_connection_cursor():
     connstr = os.getenv("DATABASE_URL")
     if not connstr:
         err("DATABASE_URL environment variable is not set")
-    with psycopg2.connect(connstr, connect_timeout=30) as conn:
+    with psycopg2.connect(connstr) as conn:
         with conn.cursor() as cur:
             yield cur
 

scripts/pr-comment-test-report.js

@@ -1,5 +1,6 @@
 //
 // The script parses Allure reports and posts a comment with a summary of the test results to the PR.
+// It accepts an array of items and creates a comment with a summary for each one (for "release" and "debug", together or separately if any of them failed to be generated).
 //
 // The comment is updated on each run with the latest results.
 //
@@ -12,35 +13,19 @@
 //         github,
 //         context,
 //         fetch,
-//         report: {
-//           reportUrl: "...",
-//           reportJsonUrl: "...",
-//         },
+//         reports: [{...}, ...], // each report is expected to have "buildType", "reportUrl", and "jsonUrl" properties
 //       })
 //
 
-// Analog of Python's defaultdict.
-//
-// const dm = new DefaultMap(() => new DefaultMap(() => []))
-// dm["firstKey"]["secondKey"].push("value")
-//
-class DefaultMap extends Map {
-    constructor(getDefaultValue) {
-        return new Proxy({}, {
-            get: (target, name) => name in target ? target[name] : (target[name] = getDefaultValue(name))
-        })
-    }
-}
-
-module.exports = async ({ github, context, fetch, report }) => {
+module.exports = async ({ github, context, fetch, reports }) => {
     // Marker to find the comment in the subsequent runs
     const startMarker = `<!--AUTOMATIC COMMENT START #${context.payload.number}-->`
-    // Let users know that the comment is updated automatically
-    const autoupdateNotice = `<div align="right"><sub>The comment gets automatically updated with the latest test results<br>${context.payload.pull_request.head.sha} at ${new Date().toISOString()} :recycle:</sub></div>`
     // GitHub bot id taken from (https://api.github.com/users/github-actions[bot])
     const githubActionsBotId = 41898282
+    // The latest commit in the PR URL
+    const commitUrl = `${context.serverUrl}/${context.repo.owner}/${context.repo.repo}/pull/${context.payload.number}/commits/${context.payload.pull_request.head.sha}`
     // Commend body itself
-    let commentBody = `${startMarker}\n`
+    let commentBody = `${startMarker}\n### Test results for ${commitUrl}:\n___\n`
 
     // Common parameters for GitHub API requests
     const ownerRepoParams = {
@@ -48,124 +33,76 @@ module.exports = async ({ github, context, fetch, report }) => {
         repo: context.repo.repo,
     }
 
-    const {reportUrl, reportJsonUrl} = report
+    for (const report of reports) {
+        const {buildType, reportUrl, jsonUrl} = report
 
-    if (!reportUrl || !reportJsonUrl) {
-        commentBody += `#### No tests were run or test report is not available\n`
-        commentBody += autoupdateNotice
-        return
-    }
+        if (!reportUrl || !jsonUrl) {
+            commentBody += `#### ${buildType} build: no tests were run or test report is not available\n`
+            continue
+        }
 
-    const suites = await (await fetch(reportJsonUrl)).json()
+        const suites = await (await fetch(jsonUrl)).json()
 
-    // Allure distinguishes "failed" (with an assertion error) and "broken" (with any other error) tests.
-    // For this report it's ok to treat them in the same way (as failed).
-    const failedTests = new DefaultMap(() => new DefaultMap(() => []))
-    const passedTests = new DefaultMap(() => new DefaultMap(() => []))
-    const skippedTests = new DefaultMap(() => new DefaultMap(() => []))
-    const retriedTests = new DefaultMap(() => new DefaultMap(() => []))
-    const flakyTests = new DefaultMap(() => new DefaultMap(() => []))
+        // Allure distinguishes "failed" (with an assertion error) and "broken" (with any other error) tests.
+        // For this report it's ok to treat them in the same way (as failed).
+        failedTests = []
+        passedTests = []
+        skippedTests = []
 
-    let failedTestsCount = 0
-    let passedTestsCount = 0
-    let skippedTestsCount = 0
-    let flakyTestsCount = 0
+        retriedTests = []
+        retriedStatusChangedTests = []
 
-    const pgVersions = new Set()
+        for (const parentSuite of suites.children) {
+            for (const suite of parentSuite.children) {
+                for (const test of suite.children) {
+                    pytestName = `${parentSuite.name.replace(".", "/")}/${suite.name}.py::${test.name}`
+                    test.pytestName = pytestName
 
-    for (const parentSuite of suites.children) {
-        for (const suite of parentSuite.children) {
-            for (const test of suite.children) {
-                let buildType, pgVersion
-                const match = test.name.match(/[\[-](?<buildType>debug|release)-pg(?<pgVersion>\d+)[-\]]/)?.groups
-                if (match) {
-                    ({buildType, pgVersion} = match)
-                } else {
-                    // It's ok, we embed BUILD_TYPE and Postgres Version into the test name only for regress suite and do not for other suites (like performance).
-                    console.info(`Cannot get BUILD_TYPE and Postgres Version from test name: "${test.name}", defaulting to "release" and "14"`)
+                    if (test.status === "passed") {
+                        passedTests.push(test);
+                    } else if (test.status === "failed" || test.status === "broken") {
+                        failedTests.push(test);
+                    } else if (test.status === "skipped") {
+                        skippedTests.push(test);
+                    }
 
-                    buildType = "release"
-                    pgVersion = "14"
+                    if (test.retriesCount > 0) {
+                        retriedTests.push(test);
+
+                        if (test.retriedStatusChangedTests) {
+                            retriedStatusChangedTests.push(test);
+                        }
+                    }
                 }
+            }
+        }
 
-                pgVersions.add(pgVersion)
-
-                // Removing build type and PostgreSQL version from the test name to make it shorter
-                const testName = test.name.replace(new RegExp(`${buildType}-pg${pgVersion}-?`), "").replace("[]", "")
-                test.pytestName = `${parentSuite.name.replace(".", "/")}/${suite.name}.py::${testName}`
-                test.pgVersion = pgVersion
-                test.buildType = buildType
-
-                if (test.status === "passed") {
-                    passedTests[pgVersion][testName].push(test)
-                    passedTestsCount += 1
-                } else if (test.status === "failed" || test.status === "broken") {
-                    failedTests[pgVersion][testName].push(test)
-                    failedTestsCount += 1
-                } else if (test.status === "skipped") {
-                    skippedTests[pgVersion][testName].push(test)
-                    skippedTestsCount += 1
-                }
+        const totalTestsCount = failedTests.length + passedTests.length + skippedTests.length
+        commentBody += `#### ${buildType} build: ${totalTestsCount} tests run: ${passedTests.length} passed, ${failedTests.length} failed, ${skippedTests.length} skipped ([full report](${reportUrl}))\n`
+        if (failedTests.length > 0) {
+            commentBody += `Failed tests:\n`
+            for (const test of failedTests) {
+                const allureLink = `${reportUrl}#suites/${test.parentUid}/${test.uid}`
 
+                commentBody += `- [\`${test.pytestName}\`](${allureLink})`
                 if (test.retriesCount > 0) {
-                    retriedTests[pgVersion][testName].push(test)
-
-                    if (test.retriesStatusChange) {
-                        flakyTests[pgVersion][testName].push(test)
-                        flakyTestsCount += 1
-                    }
+                    commentBody += ` (ran [${test.retriesCount + 1} times](${allureLink}/retries))`
                 }
+                commentBody += "\n"
             }
+            commentBody += "\n"
         }
-    }
-
-    const totalTestsCount = failedTestsCount + passedTestsCount + skippedTestsCount
-    commentBody += `### ${totalTestsCount} tests run: ${passedTestsCount} passed, ${failedTestsCount} failed, ${skippedTestsCount} skipped ([full report](${reportUrl}))\n___\n`
-
-    // Print test resuls from the newest to the oldest Postgres version for release and debug builds.
-    for (const pgVersion of Array.from(pgVersions).sort().reverse()) {
-        if (Object.keys(failedTests[pgVersion]).length > 0) {
-            commentBody += `#### Failures on Posgres ${pgVersion}\n\n`
-            for (const [testName, tests] of Object.entries(failedTests[pgVersion])) {
-                const links = []
-                for (const test of tests) {
-                    const allureLink = `${reportUrl}#suites/${test.parentUid}/${test.uid}`
-                    links.push(`[${test.buildType}](${allureLink})`)
-                }
-                commentBody += `- \`${testName}\`: ${links.join(", ")}\n`
+        if (retriedStatusChangedTests > 0) {
+            commentBody += `Flaky tests:\n`
+            for (const test of retriedStatusChangedTests) {
+                const status = test.status === "passed" ? ":white_check_mark:" : ":x:"
+                commentBody += `- ${status} [\`${test.pytestName}\`](${reportUrl}#suites/${test.parentUid}/${test.uid}/retries)\n`
             }
-
-            const testsToRerun = Object.values(failedTests[pgVersion]).map(x => x[0].name)
-            const command = `DEFAULT_PG_VERSION=${pgVersion} scripts/pytest -k "${testsToRerun.join(" or ")}"`
-
-            commentBody += "```\n"
-            commentBody += `# Run failed on Postgres ${pgVersion} tests locally:\n`
-            commentBody += `${command}\n`
-            commentBody += "```\n"
+            commentBody += "\n"
         }
+        commentBody += "___\n"
     }
 
-    if (flakyTestsCount > 0) {
-        commentBody += `<details>\n<summary>Flaky tests (${flakyTestsCount})</summary>\n\n`
-        for (const pgVersion of Array.from(pgVersions).sort().reverse()) {
-            if (Object.keys(flakyTests[pgVersion]).length > 0) {
-                commentBody += `#### Postgres ${pgVersion}\n\n`
-                for (const [testName, tests] of Object.entries(flakyTests[pgVersion])) {
-                    const links = []
-                    for (const test of tests) {
-                        const allureLink = `${reportUrl}#suites/${test.parentUid}/${test.uid}/retries`
-                        const status = test.status === "passed" ? ":white_check_mark:" : ":x:"
-                        links.push(`[${status} ${test.buildType}](${allureLink})`)
-                    }
-                    commentBody += `- \`${testName}\`: ${links.join(", ")}\n`
-                }
-            }
-        }
-        commentBody += "\n</details>\n"
-    }
-
-    commentBody += autoupdateNotice
-
     const { data: comments } = await github.rest.issues.listComments({
         issue_number: context.payload.number,
         ...ownerRepoParams,

storage_broker/src/lib.rs

@@ -40,9 +40,6 @@ pub type BrokerClientChannel = BrokerServiceClient<Channel>;
 // Create connection object configured to run TLS if schema starts with https://
 // and plain text otherwise. Connection is lazy, only endpoint sanity is
 // validated here.
-//
-// NB: this function is not async, but still must be run on a tokio runtime thread
-// because that's a requirement of tonic_endpoint.connect_lazy()'s Channel::new call.
 pub fn connect<U>(endpoint: U, keepalive_interval: Duration) -> anyhow::Result<BrokerClientChannel>
 where
     U: std::convert::TryInto<Uri>,

test_runner/conftest.py

@@ -1,6 +1,5 @@
 pytest_plugins = (
     "fixtures.pg_version",
-    "fixtures.allure",
     "fixtures.neon_fixtures",
     "fixtures.benchmark_fixture",
     "fixtures.pg_stats",

test_runner/fixtures/allure.py

@@ -1,25 +0,0 @@
-import os
-
-import pytest
-
-from fixtures.pg_version import DEFAULT_VERSION, PgVersion
-
-"""
-Set of utilities to make Allure report more informative.
-
-- It adds BUILD_TYPE and DEFAULT_PG_VERSION to the test names (only in test_runner/regress)
-to make tests distinguishable in Allure report.
-"""
-
-
-@pytest.fixture(scope="function", autouse=True)
-def allure_noop():
-    pass
-
-
-def pytest_generate_tests(metafunc):
-    if "test_runner/regress" in metafunc.definition._nodeid:
-        build_type = os.environ.get("BUILD_TYPE", "DEBUG").lower()
-        pg_version = PgVersion(os.environ.get("DEFAULT_PG_VERSION", DEFAULT_VERSION))
-
-        metafunc.parametrize("allure_noop", [f"{build_type}-pg{pg_version}"])

test_runner/fixtures/benchmark_fixture.py

@@ -451,17 +451,13 @@ def pytest_terminal_summary(
     revision = os.getenv("GITHUB_SHA", "local")
     platform = os.getenv("PLATFORM", "local")
 
-    is_header_printed = False
+    terminalreporter.section("Benchmark results", "-")
 
     result = []
     for test_report in terminalreporter.stats.get("passed", []):
         result_entry = []
 
         for _, recorded_property in test_report.user_properties:
-            if not is_header_printed:
-                terminalreporter.section("Benchmark results", "-")
-                is_header_printed = True
-
             terminalreporter.write(
                 "{}.{}: ".format(test_report.head_line, recorded_property["name"])
             )
@@ -489,6 +485,7 @@ def pytest_terminal_summary(
 
     out_dir = config.getoption("out_dir")
     if out_dir is None:
+        warnings.warn("no out dir provided to store performance test results")
         return
 
     if not result:

test_runner/fixtures/compare_fixtures.py

@@ -312,6 +312,6 @@ def neon_with_baseline(request: FixtureRequest) -> PgCompare:
     implementation-specific logic is widely useful across multiple tests, it might
     make sense to add methods to the PgCompare class.
     """
-    fixture = request.getfixturevalue(request.param)
+    fixture = request.getfixturevalue(request.param)  # type: ignore
     assert isinstance(fixture, PgCompare), f"test error: fixture {fixture} is not PgCompare"
     return fixture