Compare commits

..

1 Commits

Author SHA1 Message Date
Heikki Linnakangas
8dc1c48605 Fix race condition in XLogWaitForReplayOf()
ConditionVariablePrepareToSleep has this comment on it:

> * Caution: "before entering the loop" means you *must* test the exit
> * condition between calling ConditionVariablePrepareToSleep and calling
> * ConditionVariableSleep.  If that is inconvenient, omit calling
> * ConditionVariablePrepareToSleep.

We were not obeying that: we did not test the exit condition correctly
between the ConditionVariablePrepareToSleep and ConditionVariableSleep
calls, because the test that we had in between them only checked the
local 'replayRecPtr' variable, without updating it from shared memory.

That wasn't too serious, because the loop includes a 10 second
timeout, and would retry and succeed if the original update was
missed. Still, better fix it.

To fix, just remove the ConditionVariablePrepareToSleep() call. As the
comment says, that's also correct, and even more efficient if we
assume that sleeping is rare.
2024-05-19 21:07:29 +03:00
289 changed files with 5177 additions and 15199 deletions

View File

@@ -8,7 +8,6 @@
!scripts/combine_control_files.py
!scripts/ninstall.sh
!vm-cgconfig.conf
!docker-compose/run-tests.sh
# Directories
!.cargo/
@@ -18,10 +17,9 @@
!libs/
!neon_local/
!pageserver/
!patches/
!pgxn/
!proxy/
!storage_scrubber/
!s3_scrubber/
!safekeeper/
!storage_broker/
!storage_controller/

View File

@@ -3,13 +3,13 @@ description: 'Create Branch using API'
inputs:
api_key:
description: 'Neon API key'
desctiption: 'Neon API key'
required: true
project_id:
description: 'ID of the Project to create Branch in'
desctiption: 'ID of the Project to create Branch in'
required: true
api_host:
description: 'Neon API host'
desctiption: 'Neon API host'
default: console-stage.neon.build
outputs:
dsn:

View File

@@ -3,16 +3,16 @@ description: 'Delete Branch using API'
inputs:
api_key:
description: 'Neon API key'
desctiption: 'Neon API key'
required: true
project_id:
description: 'ID of the Project which should be deleted'
desctiption: 'ID of the Project which should be deleted'
required: true
branch_id:
description: 'ID of the branch to delete'
desctiption: 'ID of the branch to delete'
required: true
api_host:
description: 'Neon API host'
desctiption: 'Neon API host'
default: console-stage.neon.build
runs:

View File

@@ -3,22 +3,22 @@ description: 'Create Neon Project using API'
inputs:
api_key:
description: 'Neon API key'
desctiption: 'Neon API key'
required: true
region_id:
description: 'Region ID, if not set the project will be created in the default region'
desctiption: 'Region ID, if not set the project will be created in the default region'
default: aws-us-east-2
postgres_version:
description: 'Postgres version; default is 15'
default: '15'
desctiption: 'Postgres version; default is 15'
default: 15
api_host:
description: 'Neon API host'
desctiption: 'Neon API host'
default: console-stage.neon.build
provisioner:
description: 'k8s-pod or k8s-neonvm'
desctiption: 'k8s-pod or k8s-neonvm'
default: 'k8s-pod'
compute_units:
description: '[Min, Max] compute units; Min and Max are used for k8s-neonvm with autoscaling, for k8s-pod values Min and Max should be equal'
desctiption: '[Min, Max] compute units; Min and Max are used for k8s-neonvm with autoscaling, for k8s-pod values Min and Max should be equal'
default: '[1, 1]'
outputs:

View File

@@ -3,13 +3,13 @@ description: 'Delete Neon Project using API'
inputs:
api_key:
description: 'Neon API key'
desctiption: 'Neon API key'
required: true
project_id:
description: 'ID of the Project to delete'
desctiption: 'ID of the Project to delete'
required: true
api_host:
description: 'Neon API host'
desctiption: 'Neon API host'
default: console-stage.neon.build
runs:

View File

@@ -24,7 +24,7 @@ jobs:
actionlint:
needs: [ check-permissions ]
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: reviewdog/action-actionlint@v1
@@ -36,15 +36,3 @@ jobs:
fail_on_error: true
filter_mode: nofilter
level: error
- run: |
PAT='^\s*runs-on:.*-latest'
if grep -ERq $PAT .github/workflows
then
grep -ERl $PAT .github/workflows |\
while read -r f
do
l=$(grep -nE $PAT .github/workflows/release.yml | awk -F: '{print $1}' | head -1)
echo "::error file=$f,line=$l::Please, do not use ubuntu-latest images to run on, use LTS instead."
done
exit 1
fi

View File

@@ -44,7 +44,7 @@ jobs:
contains(fromJSON('["opened", "synchronize", "reopened", "closed"]'), github.event.action) &&
contains(github.event.pull_request.labels.*.name, 'approved-for-ci-run')
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- run: gh pr --repo "${GITHUB_REPOSITORY}" edit "${PR_NUMBER}" --remove-label "approved-for-ci-run"
@@ -60,7 +60,7 @@ jobs:
github.event.action == 'labeled' &&
contains(github.event.pull_request.labels.*.name, 'approved-for-ci-run')
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- run: gh pr --repo "${GITHUB_REPOSITORY}" edit "${PR_NUMBER}" --remove-label "approved-for-ci-run"
@@ -69,41 +69,15 @@ jobs:
with:
ref: main
token: ${{ secrets.CI_ACCESS_TOKEN }}
- name: Look for existing PR
id: get-pr
env:
GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
run: |
ALREADY_CREATED="$(gh pr --repo ${GITHUB_REPOSITORY} list --head ${BRANCH} --base main --json number --jq '.[].number')"
echo "ALREADY_CREATED=${ALREADY_CREATED}" >> ${GITHUB_OUTPUT}
- name: Get changed labels
id: get-labels
if: steps.get-pr.outputs.ALREADY_CREATED != ''
env:
ALREADY_CREATED: ${{ steps.get-pr.outputs.ALREADY_CREATED }}
GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
run: |
LABELS_TO_REMOVE=$(comm -23 <(gh pr --repo ${GITHUB_REPOSITORY} view ${ALREADY_CREATED} --json labels --jq '.labels.[].name'| ( grep -E '^run' || true ) | sort) \
<(gh pr --repo ${GITHUB_REPOSITORY} view ${PR_NUMBER} --json labels --jq '.labels.[].name' | ( grep -E '^run' || true ) | sort ) |\
( grep -v run-e2e-tests-in-draft || true ) | paste -sd , -)
LABELS_TO_ADD=$(comm -13 <(gh pr --repo ${GITHUB_REPOSITORY} view ${ALREADY_CREATED} --json labels --jq '.labels.[].name'| ( grep -E '^run' || true ) |sort) \
<(gh pr --repo ${GITHUB_REPOSITORY} view ${PR_NUMBER} --json labels --jq '.labels.[].name' | ( grep -E '^run' || true ) | sort ) |\
paste -sd , -)
echo "LABELS_TO_ADD=${LABELS_TO_ADD}" >> ${GITHUB_OUTPUT}
echo "LABELS_TO_REMOVE=${LABELS_TO_REMOVE}" >> ${GITHUB_OUTPUT}
- run: gh pr checkout "${PR_NUMBER}"
- run: git checkout -b "${BRANCH}"
- run: git push --force origin "${BRANCH}"
if: steps.get-pr.outputs.ALREADY_CREATED == ''
- name: Create a Pull Request for CI run (if required)
if: steps.get-pr.outputs.ALREADY_CREATED == ''
env:
env:
GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
run: |
cat << EOF > body.md
@@ -114,33 +88,16 @@ jobs:
Feel free to review/comment/discuss the original PR #${PR_NUMBER}.
EOF
LABELS=$( (gh pr --repo "${GITHUB_REPOSITORY}" view ${PR_NUMBER} --json labels --jq '.labels.[].name'; echo run-e2e-tests-in-draft )| \
grep -E '^run' | paste -sd , -)
gh pr --repo "${GITHUB_REPOSITORY}" create --title "CI run for PR #${PR_NUMBER}" \
ALREADY_CREATED="$(gh pr --repo ${GITHUB_REPOSITORY} list --head ${BRANCH} --base main --json number --jq '.[].number')"
if [ -z "${ALREADY_CREATED}" ]; then
gh pr --repo "${GITHUB_REPOSITORY}" create --title "CI run for PR #${PR_NUMBER}" \
--body-file "body.md" \
--head "${BRANCH}" \
--base "main" \
--label ${LABELS} \
--label "run-e2e-tests-in-draft" \
--draft
- name: Modify the existing pull request (if required)
if: steps.get-pr.outputs.ALREADY_CREATED != ''
env:
GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
LABELS_TO_ADD: ${{ steps.get-labels.outputs.LABELS_TO_ADD }}
LABELS_TO_REMOVE: ${{ steps.get-labels.outputs.LABELS_TO_REMOVE }}
ALREADY_CREATED: ${{ steps.get-pr.outputs.ALREADY_CREATED }}
run: |
ADD_CMD=
REMOVE_CMD=
[ -z "${LABELS_TO_ADD}" ] || ADD_CMD="--add-label ${LABELS_TO_ADD}"
[ -z "${LABELS_TO_REMOVE}" ] || REMOVE_CMD="--remove-label ${LABELS_TO_REMOVE}"
if [ -n "${ADD_CMD}" ] || [ -n "${REMOVE_CMD}" ]; then
gh pr --repo "${GITHUB_REPOSITORY}" edit ${ALREADY_CREATED} ${ADD_CMD} ${REMOVE_CMD}
fi
- run: git push --force origin "${BRANCH}"
if: steps.get-pr.outputs.ALREADY_CREATED != ''
cleanup:
# Close PRs and delete branchs if the original PR is closed.
@@ -152,7 +109,7 @@ jobs:
github.event.action == 'closed' &&
github.event.pull_request.head.repo.full_name != github.repository
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- name: Close PR and delete `ci-run/pr-${{ env.PR_NUMBER }}` branch

View File

@@ -38,11 +38,6 @@ on:
description: 'AWS-RDS and AWS-AURORA normally only run on Saturday. Set this to true to run them on every workflow_dispatch'
required: false
default: false
run_only_pgvector_tests:
type: boolean
description: 'Run pgvector tests but no other tests. If not set, all tests including pgvector tests will be run'
required: false
default: false
defaults:
run:
@@ -55,7 +50,6 @@ concurrency:
jobs:
bench:
if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
env:
TEST_PG_BENCH_DURATIONS_MATRIX: "300"
TEST_PG_BENCH_SCALES_MATRIX: "10,100"
@@ -99,7 +93,7 @@ jobs:
# Set --sparse-ordering option of pytest-order plugin
# to ensure tests are running in order of appears in the file.
# It's important for test_perf_pgbench.py::test_pgbench_remote_* tests
extra_params: -m remote_cluster --sparse-ordering --timeout 5400 --ignore test_runner/performance/test_perf_olap.py --ignore test_runner/performance/test_perf_pgvector_queries.py
extra_params: -m remote_cluster --sparse-ordering --timeout 5400 --ignore test_runner/performance/test_perf_olap.py
env:
BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
@@ -126,7 +120,6 @@ jobs:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
generate-matrices:
if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
# Create matrices for the benchmarking jobs, so we run benchmarks on rds only once a week (on Saturday)
#
# Available platforms:
@@ -137,7 +130,7 @@ jobs:
# - rds-postgres: RDS Postgres db.m5.large instance (2 vCPU, 8 GiB) with gp3 EBS storage
env:
RUN_AWS_RDS_AND_AURORA: ${{ github.event.inputs.run_AWS_RDS_AND_AURORA || 'false' }}
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
outputs:
pgbench-compare-matrix: ${{ steps.pgbench-compare-matrix.outputs.matrix }}
olap-compare-matrix: ${{ steps.olap-compare-matrix.outputs.matrix }}
@@ -204,7 +197,6 @@ jobs:
echo "matrix=$(echo "$matrix" | jq --compact-output '.')" >> $GITHUB_OUTPUT
pgbench-compare:
if: ${{ github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null }}
needs: [ generate-matrices ]
strategy:
@@ -351,92 +343,6 @@ jobs:
env:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
pgbench-pgvector:
env:
TEST_PG_BENCH_DURATIONS_MATRIX: "15m"
TEST_PG_BENCH_SCALES_MATRIX: "1"
POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
DEFAULT_PG_VERSION: 16
TEST_OUTPUT: /tmp/test_output
BUILD_TYPE: remote
SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref_name == 'main' ) }}
PLATFORM: "neon-captest-pgvector"
runs-on: [ self-hosted, us-east-2, x64 ]
container:
image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/build-tools:pinned
options: --init
steps:
- uses: actions/checkout@v4
- name: Download Neon artifact
uses: ./.github/actions/download
with:
name: neon-${{ runner.os }}-release-artifact
path: /tmp/neon/
prefix: latest
- name: Add Postgres binaries to PATH
run: |
${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/pgbench --version
echo "${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin" >> $GITHUB_PATH
- name: Set up Connection String
id: set-up-connstr
run: |
CONNSTR=${{ secrets.BENCHMARK_PGVECTOR_CONNSTR }}
echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
QUERIES=("SELECT version()")
QUERIES+=("SHOW neon.tenant_id")
QUERIES+=("SHOW neon.timeline_id")
for q in "${QUERIES[@]}"; do
psql ${CONNSTR} -c "${q}"
done
- name: Benchmark pgvector hnsw indexing
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ env.BUILD_TYPE }}
test_selection: performance/test_perf_olap.py
run_in_parallel: false
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgvector_indexing
env:
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
- name: Benchmark pgvector queries
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ env.BUILD_TYPE }}
test_selection: performance/test_perf_pgvector_queries.py
run_in_parallel: false
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600
env:
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
- name: Create Allure report
if: ${{ !cancelled() }}
uses: ./.github/actions/allure-report-generate
- name: Post to a Slack channel
if: ${{ github.event.schedule && failure() }}
uses: slackapi/slack-github-action@v1
with:
channel-id: "C033QLM5P7D" # dev-staging-stream
slack-message: "Periodic perf testing neon-captest-pgvector: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
env:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
clickbench-compare:
# ClichBench DB for rds-aurora and rds-Postgres deployed to the same clusters
# we use for performance testing in pgbench-compare.
@@ -445,7 +351,7 @@ jobs:
#
# *_CLICKBENCH_CONNSTR: Genuine ClickBench DB with ~100M rows
# *_CLICKBENCH_10M_CONNSTR: DB with the first 10M rows of ClickBench DB
if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
if: ${{ !cancelled() }}
needs: [ generate-matrices, pgbench-compare ]
strategy:
@@ -549,7 +455,7 @@ jobs:
# We might change it after https://github.com/neondatabase/neon/issues/2900.
#
# *_TPCH_S10_CONNSTR: DB generated with scale factor 10 (~10 GB)
if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
if: ${{ !cancelled() }}
needs: [ generate-matrices, clickbench-compare ]
strategy:
@@ -651,7 +557,7 @@ jobs:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
user-examples-compare:
if: ${{ !cancelled() && (github.event.inputs.run_only_pgvector_tests == 'false' || github.event.inputs.run_only_pgvector_tests == null) }}
if: ${{ !cancelled() }}
needs: [ generate-matrices, tpch-compare ]
strategy:

View File

@@ -55,7 +55,7 @@ jobs:
exit 1
fi
- uses: actions/checkout@v4
- uses: actions/checkout@v3
# Use custom DOCKER_CONFIG directory to avoid conflicts with default settings
# The default value is ~/.docker
@@ -88,7 +88,7 @@ jobs:
merge-images:
needs: [ build-image ]
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
env:
IMAGE_TAG: ${{ inputs.image-tag }}

View File

@@ -35,7 +35,7 @@ jobs:
cancel-previous-e2e-tests:
needs: [ check-permissions ]
if: github.event_name == 'pull_request'
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- name: Cancel previous e2e-tests runs for this PR
@@ -548,8 +548,8 @@ jobs:
report-benchmarks-failures:
needs: [ benchmarks, create-test-report ]
if: github.ref_name == 'main' && failure() && needs.benchmarks.result == 'failure'
runs-on: ubuntu-22.04
if: github.ref_name == 'main' && needs.benchmarks.result == 'failure'
runs-on: ubuntu-latest
steps:
- uses: slackapi/slack-github-action@v1
@@ -723,13 +723,9 @@ jobs:
uses: ./.github/workflows/trigger-e2e-tests.yml
secrets: inherit
neon-image-arch:
neon-image:
needs: [ check-permissions, build-build-tools-image, tag ]
strategy:
matrix:
arch: [ x64, arm64 ]
runs-on: ${{ fromJson(format('["self-hosted", "gen3", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
runs-on: [ self-hosted, gen3, large ]
steps:
- name: Checkout
@@ -751,6 +747,12 @@ jobs:
username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
- uses: docker/login-action@v3
with:
registry: 369495373322.dkr.ecr.eu-central-1.amazonaws.com
username: ${{ secrets.AWS_ACCESS_KEY_DEV }}
password: ${{ secrets.AWS_SECRET_KEY_DEV }}
- uses: docker/build-push-action@v5
with:
context: .
@@ -762,52 +764,25 @@ jobs:
push: true
pull: true
file: Dockerfile
cache-from: type=registry,ref=neondatabase/neon:cache-${{ matrix.arch }}
cache-to: type=registry,ref=neondatabase/neon:cache-${{ matrix.arch }},mode=max
cache-from: type=registry,ref=neondatabase/neon:cache
cache-to: type=registry,ref=neondatabase/neon:cache,mode=max
tags: |
neondatabase/neon:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}
369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}}
neondatabase/neon:${{needs.tag.outputs.build-tag}}
- name: Remove custom docker config directory
if: always()
run: |
rm -rf .docker-custom
neon-image:
needs: [ neon-image-arch, tag ]
runs-on: ubuntu-22.04
steps:
- uses: docker/login-action@v3
with:
username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
- name: Create multi-arch image
run: |
docker buildx imagetools create -t neondatabase/neon:${{ needs.tag.outputs.build-tag }} \
neondatabase/neon:${{ needs.tag.outputs.build-tag }}-x64 \
neondatabase/neon:${{ needs.tag.outputs.build-tag }}-arm64
- uses: docker/login-action@v3
with:
registry: 369495373322.dkr.ecr.eu-central-1.amazonaws.com
username: ${{ secrets.AWS_ACCESS_KEY_DEV }}
password: ${{ secrets.AWS_SECRET_KEY_DEV }}
- name: Push multi-arch image to ECR
run: |
docker buildx imagetools create -t 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{ needs.tag.outputs.build-tag }} \
neondatabase/neon:${{ needs.tag.outputs.build-tag }}
compute-node-image-arch:
compute-node-image:
needs: [ check-permissions, build-build-tools-image, tag ]
runs-on: [ self-hosted, gen3, large ]
strategy:
fail-fast: false
matrix:
version: [ v14, v15, v16 ]
arch: [ x64, arm64 ]
runs-on: ${{ fromJson(format('["self-hosted", "gen3", "{0}"]', matrix.arch == 'arm64' && 'large-arm64' || 'large')) }}
steps:
- name: Checkout
@@ -854,34 +829,15 @@ jobs:
push: true
pull: true
file: Dockerfile.compute-node
cache-from: type=registry,ref=neondatabase/compute-node-${{ matrix.version }}:cache-${{ matrix.arch }}
cache-to: type=registry,ref=neondatabase/compute-node-${{ matrix.version }}:cache-${{ matrix.arch }},mode=max
cache-from: type=registry,ref=neondatabase/compute-node-${{ matrix.version }}:cache
cache-to: type=registry,ref=neondatabase/compute-node-${{ matrix.version }}:cache,mode=max
tags: |
neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}
- name: Build neon extensions test image
if: matrix.version == 'v16'
uses: docker/build-push-action@v5
with:
context: .
build-args: |
GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
PG_VERSION=${{ matrix.version }}
BUILD_TAG=${{ needs.tag.outputs.build-tag }}
TAG=${{ needs.build-build-tools-image.outputs.image-tag }}
provenance: false
push: true
pull: true
file: Dockerfile.compute-node
target: neon-pg-ext-test
cache-from: type=registry,ref=neondatabase/neon-test-extensions-${{ matrix.version }}:cache-${{ matrix.arch }}
cache-to: type=registry,ref=neondatabase/neon-test-extensions-${{ matrix.version }}:cache-${{ matrix.arch }},mode=max
tags: |
neondatabase/neon-test-extensions-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}-${{ matrix.arch }}
369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
neondatabase/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
- name: Build compute-tools image
# compute-tools are Postgres independent, so build it only once
if: matrix.version == 'v16'
if: ${{ matrix.version == 'v16' }}
uses: docker/build-push-action@v5
with:
target: compute-tools-image
@@ -895,64 +851,14 @@ jobs:
pull: true
file: Dockerfile.compute-node
tags: |
neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}-${{ matrix.arch }}
369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{ needs.tag.outputs.build-tag }}
neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}
- name: Remove custom docker config directory
if: always()
run: |
rm -rf .docker-custom
compute-node-image:
needs: [ compute-node-image-arch, tag ]
runs-on: ubuntu-22.04
strategy:
matrix:
version: [ v14, v15, v16 ]
steps:
- uses: docker/login-action@v3
with:
username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
- name: Create multi-arch compute-node image
run: |
docker buildx imagetools create -t neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }} \
neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-x64 \
neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-arm64
- name: Create multi-arch neon-test-extensions image
if: matrix.version == 'v16'
run: |
docker buildx imagetools create -t neondatabase/neon-test-extensions-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }} \
neondatabase/neon-test-extensions-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-x64 \
neondatabase/neon-test-extensions-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}-arm64
- name: Create multi-arch compute-tools image
if: matrix.version == 'v16'
run: |
docker buildx imagetools create -t neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }} \
neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}-x64 \
neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}-arm64
- uses: docker/login-action@v3
with:
registry: 369495373322.dkr.ecr.eu-central-1.amazonaws.com
username: ${{ secrets.AWS_ACCESS_KEY_DEV }}
password: ${{ secrets.AWS_SECRET_KEY_DEV }}
- name: Push multi-arch compute-node-${{ matrix.version }} image to ECR
run: |
docker buildx imagetools create -t 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }} \
neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}
- name: Push multi-arch compute-tools image to ECR
if: matrix.version == 'v16'
run: |
docker buildx imagetools create -t 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{ needs.tag.outputs.build-tag }} \
neondatabase/compute-tools:${{ needs.tag.outputs.build-tag }}
vm-compute-node-image:
needs: [ check-permissions, tag, compute-node-image ]
runs-on: [ self-hosted, gen3, large ]
@@ -960,12 +866,15 @@ jobs:
fail-fast: false
matrix:
version: [ v14, v15, v16 ]
defaults:
run:
shell: sh -eu {0}
env:
VM_BUILDER_VERSION: v0.29.3
VM_BUILDER_VERSION: v0.28.1
steps:
- name: Checkout
uses: actions/checkout@v4
uses: actions/checkout@v1
with:
fetch-depth: 0
@@ -974,48 +883,26 @@ jobs:
curl -fL https://github.com/neondatabase/autoscaling/releases/download/$VM_BUILDER_VERSION/vm-builder -o vm-builder
chmod +x vm-builder
# Use custom DOCKER_CONFIG directory to avoid conflicts with default settings
# The default value is ~/.docker
- name: Set custom docker config directory
run: |
mkdir -p .docker-custom
echo DOCKER_CONFIG=$(pwd)/.docker-custom >> $GITHUB_ENV
- uses: docker/login-action@v3
with:
username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
# Note: we need a separate pull step here because otherwise vm-builder will try to pull, and
# it won't have the proper authentication (written at v0.6.0)
- name: Pulling compute-node image
run: |
docker pull neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}
docker pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
- name: Build vm image
run: |
./vm-builder \
-spec=vm-image-spec.yaml \
-src=neondatabase/compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }} \
-dst=neondatabase/vm-compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}
-src=369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}} \
-dst=369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
- name: Pushing vm-compute-node image
run: |
docker push neondatabase/vm-compute-node-${{ matrix.version }}:${{ needs.tag.outputs.build-tag }}
- name: Remove custom docker config directory
if: always()
run: |
rm -rf .docker-custom
docker push 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-${{ matrix.version }}:${{needs.tag.outputs.build-tag}}
test-images:
needs: [ check-permissions, tag, neon-image, compute-node-image ]
strategy:
fail-fast: false
matrix:
arch: [ x64, arm64 ]
runs-on: ${{ fromJson(format('["self-hosted", "gen3", "{0}"]', matrix.arch == 'arm64' && 'small-arm64' || 'small')) }}
runs-on: [ self-hosted, gen3, small ]
steps:
- name: Checkout
@@ -1033,7 +920,7 @@ jobs:
- name: Verify image versions
shell: bash # ensure no set -e for better error messages
run: |
pageserver_version=$(docker run --rm neondatabase/neon:${{ needs.tag.outputs.build-tag }} "/bin/sh" "-c" "/usr/local/bin/pageserver --version")
pageserver_version=$(docker run --rm 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}} "/bin/sh" "-c" "/usr/local/bin/pageserver --version")
echo "Pageserver version string: $pageserver_version"
@@ -1047,7 +934,7 @@ jobs:
exit 1
fi
- name: Verify docker-compose example and test extensions
- name: Verify docker-compose example
timeout-minutes: 20
run: env TAG=${{needs.tag.outputs.build-tag}} ./docker-compose/docker_compose_test.sh
@@ -1059,54 +946,82 @@ jobs:
promote-images:
needs: [ check-permissions, tag, test-images, vm-compute-node-image ]
runs-on: ubuntu-22.04
env:
VERSIONS: v14 v15 v16
runs-on: [ self-hosted, gen3, small ]
container: golang:1.19-bullseye
# Don't add if-condition here.
# The job should always be run because we have dependant other jobs that shouldn't be skipped
steps:
- uses: docker/login-action@v3
with:
username: ${{ secrets.NEON_DOCKERHUB_USERNAME }}
password: ${{ secrets.NEON_DOCKERHUB_PASSWORD }}
- uses: docker/login-action@v3
with:
registry: 369495373322.dkr.ecr.eu-central-1.amazonaws.com
username: ${{ secrets.AWS_ACCESS_KEY_DEV }}
password: ${{ secrets.AWS_SECRET_KEY_DEV }}
- name: Copy vm-compute-node images to ECR
- name: Install Crane & ECR helper
run: |
for version in ${VERSIONS}; do
docker buildx imagetools create -t 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-${version}:${{ needs.tag.outputs.build-tag }} \
neondatabase/vm-compute-node-${version}:${{ needs.tag.outputs.build-tag }}
done
go install github.com/google/go-containerregistry/cmd/crane@31786c6cbb82d6ec4fb8eb79cd9387905130534e # v0.11.0
go install github.com/awslabs/amazon-ecr-credential-helper/ecr-login/cli/docker-credential-ecr-login@69c85dc22db6511932bbf119e1a0cc5c90c69a7f # v0.6.0
- name: Configure ECR login
run: |
mkdir /github/home/.docker/
echo "{\"credsStore\":\"ecr-login\"}" > /github/home/.docker/config.json
- name: Copy vm-compute-node images to Docker Hub
run: |
crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v14:${{needs.tag.outputs.build-tag}} vm-compute-node-v14
crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v15:${{needs.tag.outputs.build-tag}} vm-compute-node-v15
crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v16:${{needs.tag.outputs.build-tag}} vm-compute-node-v16
- name: Add latest tag to images
if: github.ref_name == 'main'
if: github.ref_name == 'main' || github.ref_name == 'release' || github.ref_name == 'release-proxy'
run: |
for repo in neondatabase 369495373322.dkr.ecr.eu-central-1.amazonaws.com; do
docker buildx imagetools create -t $repo/neon:latest \
$repo/neon:${{ needs.tag.outputs.build-tag }}
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v14:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v15:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v16:${{needs.tag.outputs.build-tag}} latest
crane tag 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v16:${{needs.tag.outputs.build-tag}} latest
docker buildx imagetools create -t $repo/compute-tools:latest \
$repo/compute-tools:${{ needs.tag.outputs.build-tag }}
- name: Push images to production ECR
if: github.ref_name == 'main' || github.ref_name == 'release'|| github.ref_name == 'release-proxy'
run: |
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/neon:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v14:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v14:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v15:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v15:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v16:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v16:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v16:${{needs.tag.outputs.build-tag}} 093970136003.dkr.ecr.eu-central-1.amazonaws.com/vm-compute-node-v16:latest
for version in ${VERSIONS}; do
docker buildx imagetools create -t $repo/compute-node-${version}:latest \
$repo/compute-node-${version}:${{ needs.tag.outputs.build-tag }}
- name: Configure Docker Hub login
run: |
# ECR Credential Helper & Docker Hub don't work together in config, hence reset
echo "" > /github/home/.docker/config.json
crane auth login -u ${{ secrets.NEON_DOCKERHUB_USERNAME }} -p ${{ secrets.NEON_DOCKERHUB_PASSWORD }} index.docker.io
docker buildx imagetools create -t $repo/vm-compute-node-${version}:latest \
$repo/vm-compute-node-${version}:${{ needs.tag.outputs.build-tag }}
done
done
docker buildx imagetools create -t neondatabase/neon-test-extensions-v16:latest \
neondatabase/neon-test-extensions-v16:${{ needs.tag.outputs.build-tag }}
- name: Push vm-compute-node to Docker Hub
run: |
crane push vm-compute-node-v14 neondatabase/vm-compute-node-v14:${{needs.tag.outputs.build-tag}}
crane push vm-compute-node-v15 neondatabase/vm-compute-node-v15:${{needs.tag.outputs.build-tag}}
crane push vm-compute-node-v16 neondatabase/vm-compute-node-v16:${{needs.tag.outputs.build-tag}}
- name: Push latest tags to Docker Hub
if: github.ref_name == 'main' || github.ref_name == 'release'|| github.ref_name == 'release-proxy'
run: |
crane tag neondatabase/neon:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-tools:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v14:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/vm-compute-node-v14:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v15:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/vm-compute-node-v15:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v16:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/vm-compute-node-v16:${{needs.tag.outputs.build-tag}} latest
- name: Cleanup ECR folder
run: rm -rf ~/.ecr
trigger-custom-extensions-build-and-wait:
needs: [ check-permissions, tag ]
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- name: Set PR's status to pending and request a remote CI test
run: |

View File

@@ -19,7 +19,7 @@ permissions: {}
jobs:
check-image:
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
outputs:
tag: ${{ steps.get-build-tools-tag.outputs.image-tag }}
found: ${{ steps.check-image.outputs.found }}

View File

@@ -16,7 +16,7 @@ permissions: {}
jobs:
check-permissions:
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- name: Disallow CI runs on PRs from forks
if: |

View File

@@ -9,7 +9,7 @@ on:
jobs:
cleanup:
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- name: Cleanup
run: |

View File

@@ -20,7 +20,7 @@ concurrency:
jobs:
test-postgres-client-libs:
# TODO: switch to gen2 runner, requires docker
runs-on: ubuntu-22.04
runs-on: [ ubuntu-latest ]
env:
DEFAULT_PG_VERSION: 14

View File

@@ -26,7 +26,7 @@ permissions: {}
jobs:
tag-image:
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
env:
FROM_TAG: ${{ inputs.from-tag }}

View File

@@ -19,7 +19,7 @@ on:
jobs:
notify:
runs-on: ubuntu-22.04
runs-on: [ ubuntu-latest ]
steps:
- uses: neondatabase/dev-actions/release-pr-notify@main

View File

@@ -26,7 +26,7 @@ defaults:
jobs:
create-storage-release-branch:
if: ${{ github.event.schedule == '0 6 * * MON' || format('{0}', inputs.create-storage-release-branch) == 'true' }}
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
permissions:
contents: write # for `git push`
@@ -53,7 +53,7 @@ jobs:
GH_TOKEN: ${{ secrets.CI_ACCESS_TOKEN }}
run: |
cat << EOF > body.md
## Storage & Compute release ${RELEASE_DATE}
## Release ${RELEASE_DATE}
**Please merge this Pull Request using 'Create a merge commit' button**
EOF
@@ -65,7 +65,7 @@ jobs:
create-proxy-release-branch:
if: ${{ github.event.schedule == '0 6 * * THU' || format('{0}', inputs.create-proxy-release-branch) == 'true' }}
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
permissions:
contents: write # for `git push`

View File

@@ -19,7 +19,7 @@ env:
jobs:
cancel-previous-e2e-tests:
if: github.event_name == 'pull_request'
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
steps:
- name: Cancel previous e2e-tests runs for this PR
@@ -31,7 +31,7 @@ jobs:
--field concurrency_group="${{ env.E2E_CONCURRENCY_GROUP }}"
tag:
runs-on: ubuntu-22.04
runs-on: [ ubuntu-latest ]
outputs:
build-tag: ${{ steps.build-tag.outputs.tag }}
@@ -62,7 +62,7 @@ jobs:
trigger-e2e-tests:
needs: [ tag ]
runs-on: ubuntu-22.04
runs-on: ubuntu-latest
env:
TAG: ${{ needs.tag.outputs.build-tag }}
steps:

306
Cargo.lock generated
View File

@@ -776,6 +776,7 @@ dependencies = [
"pin-project",
"serde",
"time",
"tz-rs",
"url",
"uuid",
]
@@ -1071,9 +1072,9 @@ dependencies = [
[[package]]
name = "chrono"
version = "0.4.38"
version = "0.4.31"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a21f936df1771bf62b77f047b726c4625ff2e8aa607c01ec06e5a05bd8463401"
checksum = "7f2c685bad3eb3d45a01354cedb7d5faa66194d1d58ba6e267a8de788f79db38"
dependencies = [
"android-tzdata",
"iana-time-zone",
@@ -1081,7 +1082,7 @@ dependencies = [
"num-traits",
"serde",
"wasm-bindgen",
"windows-targets 0.52.4",
"windows-targets 0.48.0",
]
[[package]]
@@ -1108,7 +1109,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "defaa24ecc093c77630e6c15e17c51f5e187bf35ee514f4e2d67baaa96dae22b"
dependencies = [
"ciborium-io",
"half 1.8.2",
"half",
]
[[package]]
@@ -1290,6 +1291,12 @@ dependencies = [
"tiny-keccak",
]
[[package]]
name = "const_fn"
version = "0.4.9"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "fbdcdcb6d86f71c5e97409ad45898af11cbc995b4ee8112d59095a28d376c935"
[[package]]
name = "const_format"
version = "0.2.30"
@@ -1464,21 +1471,26 @@ dependencies = [
[[package]]
name = "crossbeam-deque"
version = "0.8.5"
version = "0.8.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "613f8cc01fe9cf1a3eb3d7f488fd2fa8388403e97039e2f73692932e291a770d"
checksum = "ce6fd6f855243022dcecf8702fef0c297d4338e226845fe067f6341ad9fa0cef"
dependencies = [
"cfg-if",
"crossbeam-epoch",
"crossbeam-utils",
]
[[package]]
name = "crossbeam-epoch"
version = "0.9.18"
version = "0.9.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5b82ac4a3c2ca9c3460964f020e1402edd5753411d7737aa39c3714ad1b5420e"
checksum = "46bd5f3f85273295a9d14aedfb86f6aadbff6d8f5295c4a9edb08e819dcf5695"
dependencies = [
"autocfg",
"cfg-if",
"crossbeam-utils",
"memoffset 0.8.0",
"scopeguard",
]
[[package]]
@@ -1969,6 +1981,21 @@ version = "1.0.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3f9eec918d3f24069decb9af1554cad7c880e2da24a9afd88aca000531ab82c1"
[[package]]
name = "foreign-types"
version = "0.3.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f6f339eb8adc052cd2ca78910fda869aefa38d22d5cb648e6485e4d3fc06f3b1"
dependencies = [
"foreign-types-shared",
]
[[package]]
name = "foreign-types-shared"
version = "0.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "00b0228411908ca8685dba7fc2cdd70ec9990a6e753e89b6ac91a84c40fbaf4b"
[[package]]
name = "form_urlencoded"
version = "1.1.0"
@@ -2251,17 +2278,6 @@ version = "1.8.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "eabb4a44450da02c90444cf74558da904edde8fb4e9035a9a6a4e15445af0bd7"
[[package]]
name = "half"
version = "2.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6dd08c532ae367adf81c312a4580bc67f1d0fe8bc9c460520283f4c0ff277888"
dependencies = [
"cfg-if",
"crunchy",
"num-traits",
]
[[package]]
name = "hash32"
version = "0.3.1"
@@ -2598,6 +2614,19 @@ dependencies = [
"tokio-io-timeout",
]
[[package]]
name = "hyper-tls"
version = "0.5.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d6183ddfa99b85da61a140bea0efc93fdf56ceaa041b37d553518030827f9905"
dependencies = [
"bytes",
"hyper 0.14.26",
"native-tls",
"tokio",
"tokio-native-tls",
]
[[package]]
name = "hyper-util"
version = "0.1.3"
@@ -2915,12 +2944,6 @@ version = "0.4.13"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "01cda141df6706de531b6c46c3a33ecca755538219bd484262fa09410c13539c"
[[package]]
name = "linux-raw-sys"
version = "0.6.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f0b5399f6804fbab912acbd8878ed3532d506b7c951b8f9f164ef90fef39e3f4"
[[package]]
name = "lock_api"
version = "0.4.10"
@@ -3139,6 +3162,24 @@ version = "0.8.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e5ce46fe64a9d73be07dcbe690a38ce1b293be448fd8ce1e6c1b8062c9f72c6a"
[[package]]
name = "native-tls"
version = "0.2.11"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "07226173c32f2926027b63cce4bcd8076c3552846cbe7925f3aaffeac0a3b92e"
dependencies = [
"lazy_static",
"libc",
"log",
"openssl",
"openssl-probe",
"openssl-sys",
"schannel",
"security-framework",
"security-framework-sys",
"tempfile",
]
[[package]]
name = "nix"
version = "0.25.1"
@@ -3309,6 +3350,15 @@ dependencies = [
"libc",
]
[[package]]
name = "num_threads"
version = "0.1.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2819ce041d2ee131036f4fc9d6ae7ae125a3a40e97ba64d04fe799ad9dabbb44"
dependencies = [
"libc",
]
[[package]]
name = "oauth2"
version = "4.4.2"
@@ -3358,12 +3408,50 @@ version = "11.1.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0ab1bc2a289d34bd04a330323ac98a1b4bc82c9d9fcb1e66b63caa84da26b575"
[[package]]
name = "openssl"
version = "0.10.60"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "79a4c6c3a2b158f7f8f2a2fc5a969fa3a068df6fc9dbb4a43845436e3af7c800"
dependencies = [
"bitflags 2.4.1",
"cfg-if",
"foreign-types",
"libc",
"once_cell",
"openssl-macros",
"openssl-sys",
]
[[package]]
name = "openssl-macros"
version = "0.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a948666b637a0f465e8564c73e89d4dde00d72d4d473cc972f390fc3dcee7d9c"
dependencies = [
"proc-macro2",
"quote",
"syn 2.0.52",
]
[[package]]
name = "openssl-probe"
version = "0.1.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ff011a302c396a5197692431fc1948019154afc178baf7d8e37367442a4601cf"
[[package]]
name = "openssl-sys"
version = "0.9.96"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3812c071ba60da8b5677cc12bcb1d42989a65553772897a7e0355545a819838f"
dependencies = [
"cc",
"libc",
"pkg-config",
"vcpkg",
]
[[package]]
name = "opentelemetry"
version = "0.20.0"
@@ -3570,7 +3658,6 @@ dependencies = [
"serde",
"serde_json",
"svg_fmt",
"thiserror",
"tokio",
"tokio-util",
"toml_edit",
@@ -3815,13 +3902,12 @@ dependencies = [
[[package]]
name = "parquet"
version = "51.0.0"
source = "git+https://github.com/apache/arrow-rs?branch=master#2534976a564be3d2d56312dc88fb1b6ed4cef829"
version = "49.0.0"
source = "git+https://github.com/neondatabase/arrow-rs?branch=neon-fix-bugs#8a0bc58aa67b98aabbd8eee7c6ca4281967ff9e9"
dependencies = [
"ahash",
"bytes",
"chrono",
"half 2.4.1",
"hashbrown 0.14.5",
"num",
"num-bigint",
@@ -3830,13 +3916,12 @@ dependencies = [
"thrift",
"twox-hash",
"zstd",
"zstd-sys",
]
[[package]]
name = "parquet_derive"
version = "51.0.0"
source = "git+https://github.com/apache/arrow-rs?branch=master#2534976a564be3d2d56312dc88fb1b6ed4cef829"
version = "49.0.0"
source = "git+https://github.com/neondatabase/arrow-rs?branch=neon-fix-bugs#8a0bc58aa67b98aabbd8eee7c6ca4281967ff9e9"
dependencies = [
"parquet",
"proc-macro2",
@@ -3863,9 +3948,9 @@ checksum = "de3145af08024dea9fa9914f381a17b8fc6034dfb00f3a84013f7ff43f29ed4c"
[[package]]
name = "pbkdf2"
version = "0.12.2"
version = "0.12.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f8ed6a7761f76e3b9f92dfb0a60a6a6477c61024b775147ff0973a02653abaf2"
checksum = "f0ca0b5a68607598bf3bad68f32227a8164f6254833f84eafaac409cd6746c31"
dependencies = [
"digest",
"hmac",
@@ -4012,6 +4097,17 @@ dependencies = [
"tokio-postgres",
]
[[package]]
name = "postgres-native-tls"
version = "0.5.0"
source = "git+https://github.com/neondatabase/rust-postgres.git?branch=neon#20031d7a9ee1addeae6e0968e3899ae6bf01cee2"
dependencies = [
"native-tls",
"tokio",
"tokio-native-tls",
"tokio-postgres",
]
[[package]]
name = "postgres-protocol"
version = "0.6.4"
@@ -4120,7 +4216,6 @@ version = "0.1.0"
dependencies = [
"byteorder",
"bytes",
"itertools",
"pin-project-lite",
"postgres-protocol",
"rand 0.8.5",
@@ -4278,7 +4373,6 @@ dependencies = [
name = "proxy"
version = "0.1.0"
dependencies = [
"ahash",
"anyhow",
"async-compression",
"async-trait",
@@ -4295,7 +4389,6 @@ dependencies = [
"chrono",
"clap",
"consumption_metrics",
"crossbeam-deque",
"dashmap",
"env_logger",
"fallible-iterator",
@@ -4310,7 +4403,6 @@ dependencies = [
"http 1.1.0",
"http-body-util",
"humantime",
"humantime-serde",
"hyper 0.14.26",
"hyper 1.2.0",
"hyper-util",
@@ -4321,6 +4413,7 @@ dependencies = [
"md5",
"measured",
"metrics",
"native-tls",
"once_cell",
"opentelemetry",
"parking_lot 0.12.1",
@@ -4328,6 +4421,7 @@ dependencies = [
"parquet_derive",
"pbkdf2",
"pin-project-lite",
"postgres-native-tls",
"postgres-protocol",
"postgres_backend",
"pq_proto",
@@ -4346,7 +4440,6 @@ dependencies = [
"rstest",
"rustc-hash",
"rustls 0.22.4",
"rustls-native-certs 0.7.0",
"rustls-pemfile 2.1.1",
"scopeguard",
"serde",
@@ -4376,6 +4469,7 @@ dependencies = [
"utils",
"uuid",
"walkdir",
"webpki-roots 0.25.2",
"workspace_hack",
"x509-parser",
]
@@ -4682,21 +4776,20 @@ dependencies = [
"http 0.2.9",
"http-body 0.4.5",
"hyper 0.14.26",
"hyper-rustls 0.24.0",
"hyper-tls",
"ipnet",
"js-sys",
"log",
"mime",
"native-tls",
"once_cell",
"percent-encoding",
"pin-project-lite",
"rustls 0.21.11",
"rustls-pemfile 1.0.2",
"serde",
"serde_json",
"serde_urlencoded",
"tokio",
"tokio-rustls 0.24.0",
"tokio-native-tls",
"tokio-util",
"tower-service",
"url",
@@ -4704,7 +4797,6 @@ dependencies = [
"wasm-bindgen-futures",
"wasm-streams 0.3.0",
"web-sys",
"webpki-roots 0.25.2",
"winreg 0.50.0",
]
@@ -5109,6 +5201,51 @@ version = "1.0.13"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f91339c0467de62360649f8d3e185ca8de4224ff281f66000de5eb2a77a79041"
[[package]]
name = "s3_scrubber"
version = "0.1.0"
dependencies = [
"anyhow",
"async-stream",
"aws-config",
"aws-sdk-s3",
"aws-smithy-async",
"bincode",
"bytes",
"camino",
"chrono",
"clap",
"crc32c",
"either",
"futures",
"futures-util",
"hex",
"histogram",
"itertools",
"native-tls",
"pageserver",
"pageserver_api",
"postgres-native-tls",
"postgres_ffi",
"rand 0.8.5",
"remote_storage",
"reqwest 0.12.4",
"serde",
"serde_json",
"serde_with",
"thiserror",
"tokio",
"tokio-postgres",
"tokio-rustls 0.25.0",
"tokio-stream",
"tokio-util",
"tracing",
"tracing-appender",
"tracing-subscriber",
"utils",
"workspace_hack",
]
[[package]]
name = "safekeeper"
version = "0.1.0"
@@ -5765,54 +5902,6 @@ dependencies = [
"workspace_hack",
]
[[package]]
name = "storage_scrubber"
version = "0.1.0"
dependencies = [
"anyhow",
"async-stream",
"aws-config",
"aws-sdk-s3",
"aws-smithy-async",
"bincode",
"bytes",
"camino",
"chrono",
"clap",
"crc32c",
"either",
"futures",
"futures-util",
"hex",
"histogram",
"humantime",
"itertools",
"once_cell",
"pageserver",
"pageserver_api",
"postgres_ffi",
"rand 0.8.5",
"remote_storage",
"reqwest 0.12.4",
"rustls 0.22.4",
"rustls-native-certs 0.7.0",
"serde",
"serde_json",
"serde_with",
"thiserror",
"tokio",
"tokio-postgres",
"tokio-postgres-rustls",
"tokio-rustls 0.25.0",
"tokio-stream",
"tokio-util",
"tracing",
"tracing-appender",
"tracing-subscriber",
"utils",
"workspace_hack",
]
[[package]]
name = "storcon_cli"
version = "0.1.0"
@@ -5820,8 +5909,6 @@ dependencies = [
"anyhow",
"clap",
"comfy-table",
"futures",
"humantime",
"hyper 0.14.26",
"pageserver_api",
"pageserver_client",
@@ -6092,6 +6179,8 @@ checksum = "8f3403384eaacbca9923fa06940178ac13e4edb725486d70e8e15881d0c836cc"
dependencies = [
"itoa",
"js-sys",
"libc",
"num_threads",
"serde",
"time-core",
"time-macros",
@@ -6167,7 +6256,7 @@ dependencies = [
[[package]]
name = "tokio-epoll-uring"
version = "0.1.0"
source = "git+https://github.com/neondatabase/tokio-epoll-uring.git?branch=main#08ccfa94ff5507727bf4d8d006666b5b192e04c6"
source = "git+https://github.com/neondatabase/tokio-epoll-uring.git?branch=main#342ddd197a060a8354e8f11f4d12994419fff939"
dependencies = [
"futures",
"nix 0.26.4",
@@ -6201,6 +6290,16 @@ dependencies = [
"syn 2.0.52",
]
[[package]]
name = "tokio-native-tls"
version = "0.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bbae76ab933c85776efabc971569dd6119c580d8f5d448769dec1764bf796ef2"
dependencies = [
"native-tls",
"tokio",
]
[[package]]
name = "tokio-postgres"
version = "0.7.7"
@@ -6607,6 +6706,15 @@ version = "1.16.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "497961ef93d974e23eb6f433eb5fe1b7930b659f06d12dec6fc44a8f554c0bba"
[[package]]
name = "tz-rs"
version = "0.6.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "33851b15c848fad2cf4b105c6bb66eb9512b6f6c44a4b13f57c53c73c707e2b4"
dependencies = [
"const_fn",
]
[[package]]
name = "uname"
version = "0.1.1"
@@ -6679,12 +6787,11 @@ dependencies = [
[[package]]
name = "uring-common"
version = "0.1.0"
source = "git+https://github.com/neondatabase/tokio-epoll-uring.git?branch=main#08ccfa94ff5507727bf4d8d006666b5b192e04c6"
source = "git+https://github.com/neondatabase/tokio-epoll-uring.git?branch=main#342ddd197a060a8354e8f11f4d12994419fff939"
dependencies = [
"bytes",
"io-uring",
"libc",
"linux-raw-sys 0.6.4",
]
[[package]]
@@ -7353,7 +7460,6 @@ dependencies = [
name = "workspace_hack"
version = "0.1.0"
dependencies = [
"ahash",
"anyhow",
"aws-config",
"aws-runtime",
@@ -7512,9 +7618,9 @@ dependencies = [
[[package]]
name = "zeroize"
version = "1.7.0"
version = "1.6.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "525b4ec142c6b68a2d10f01f7bbf6755599ca3f81ea53b8431b7dd348f5fdb2d"
checksum = "2a0956f1ba7c7909bfb66c2e9e4124ab6f6482560f6628b5aaeba39207c9aad9"
dependencies = [
"zeroize_derive",
]

View File

@@ -13,7 +13,7 @@ members = [
"safekeeper",
"storage_broker",
"storage_controller",
"storage_scrubber",
"s3_scrubber",
"workspace_hack",
"trace",
"libs/compute_api",
@@ -41,15 +41,14 @@ license = "Apache-2.0"
## All dependency versions, used in the project
[workspace.dependencies]
ahash = "0.8"
anyhow = { version = "1.0", features = ["backtrace"] }
arc-swap = "1.6"
async-compression = { version = "0.4.0", features = ["tokio", "gzip", "zstd"] }
atomic-take = "1.1.0"
azure_core = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls", "hmac_rust"] }
azure_identity = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls"] }
azure_storage = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls"] }
azure_storage_blobs = { version = "0.19", default-features = false, features = ["enable_reqwest_rustls"] }
azure_core = "0.19"
azure_identity = "0.19"
azure_storage = "0.19"
azure_storage_blobs = "0.19"
flate2 = "1.0.26"
async-stream = "0.3"
async-trait = "0.1"
@@ -75,7 +74,6 @@ clap = { version = "4.0", features = ["derive"] }
comfy-table = "6.1"
const_format = "0.2"
crc32c = "0.6"
crossbeam-deque = "0.8.5"
crossbeam-utils = "0.8.5"
dashmap = { version = "5.5.0", features = ["raw-api"] }
either = "1.8"
@@ -114,21 +112,22 @@ md5 = "0.7.0"
measured = { version = "0.0.21", features=["lasso"] }
measured-process = { version = "0.0.21" }
memoffset = "0.8"
native-tls = "0.2"
nix = { version = "0.27", features = ["fs", "process", "socket", "signal", "poll"] }
notify = "6.0.0"
num_cpus = "1.15"
num-traits = "0.2.15"
once_cell = "1.13"
opentelemetry = "0.20.0"
opentelemetry-otlp = { version = "0.13.0", default-features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
opentelemetry-otlp = { version = "0.13.0", default_features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
opentelemetry-semantic-conventions = "0.12.0"
parking_lot = "0.12"
parquet = { version = "51.0.0", default-features = false, features = ["zstd"] }
parquet_derive = "51.0.0"
parquet = { version = "49.0.0", default-features = false, features = ["zstd"] }
parquet_derive = "49.0.0"
pbkdf2 = { version = "0.12.1", features = ["simple", "std"] }
pin-project-lite = "0.2"
procfs = "0.14"
prometheus = {version = "0.13", default-features=false, features = ["process"]} # removes protobuf dependency
prometheus = {version = "0.13", default_features=false, features = ["process"]} # removes protobuf dependency
prost = "0.11"
rand = "0.8"
redis = { version = "0.25.2", features = ["tokio-rustls-comp", "keep-alive"] }
@@ -184,13 +183,13 @@ tower-service = "0.3.2"
tracing = "0.1"
tracing-error = "0.2.0"
tracing-opentelemetry = "0.21.0"
tracing-subscriber = { version = "0.3", default-features = false, features = ["smallvec", "fmt", "tracing-log", "std", "env-filter", "json", "ansi"] }
tracing-subscriber = { version = "0.3", default_features = false, features = ["smallvec", "fmt", "tracing-log", "std", "env-filter", "json", "ansi"] }
twox-hash = { version = "1.6.3", default-features = false }
url = "2.2"
urlencoding = "2.1"
uuid = { version = "1.6.1", features = ["v4", "v7", "serde"] }
walkdir = "2.3.2"
rustls-native-certs = "0.7"
webpki-roots = "0.25"
x509-parser = "0.15"
## TODO replace this with tracing
@@ -199,6 +198,7 @@ log = "0.4"
## Libraries from neondatabase/ git forks, ideally with changes to be upstreamed
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
postgres-native-tls = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
postgres-protocol = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
postgres-types = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", branch="neon" }
@@ -239,12 +239,13 @@ tonic-build = "0.9"
[patch.crates-io]
# Needed to get `tokio-postgres-rustls` to depend on our fork.
# 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", branch="neon" }
# bug fixes for UUID
parquet = { git = "https://github.com/apache/arrow-rs", branch = "master" }
parquet_derive = { git = "https://github.com/apache/arrow-rs", branch = "master" }
parquet = { git = "https://github.com/neondatabase/arrow-rs", branch = "neon-fix-bugs" }
parquet_derive = { git = "https://github.com/neondatabase/arrow-rs", branch = "neon-fix-bugs" }
################# Binary contents sections

View File

@@ -141,7 +141,7 @@ WORKDIR /home/nonroot
# Rust
# Please keep the version of llvm (installed above) in sync with rust llvm (`rustc --version --verbose | grep LLVM`)
ENV RUSTC_VERSION=1.79.0
ENV RUSTC_VERSION=1.78.0
ENV RUSTUP_HOME="/home/nonroot/.rustup"
ENV PATH="/home/nonroot/.cargo/bin:${PATH}"
RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux-gnu/rustup-init && whoami && \

View File

@@ -89,7 +89,7 @@ RUN apt update && \
# SFCGAL > 1.3 requires CGAL > 5.2, Bullseye's libcgal-dev is 5.2
RUN wget https://gitlab.com/Oslandia/SFCGAL/-/archive/v1.3.10/SFCGAL-v1.3.10.tar.gz -O SFCGAL.tar.gz && \
echo "4e39b3b2adada6254a7bdba6d297bb28e1a9835a9f879b74f37e2dab70203232 SFCGAL.tar.gz" | sha256sum --check && \
mkdir sfcgal-src && cd sfcgal-src && tar xzf ../SFCGAL.tar.gz --strip-components=1 -C . && \
mkdir sfcgal-src && cd sfcgal-src && tar xvzf ../SFCGAL.tar.gz --strip-components=1 -C . && \
cmake -DCMAKE_BUILD_TYPE=Release . && make -j $(getconf _NPROCESSORS_ONLN) && \
DESTDIR=/sfcgal make install -j $(getconf _NPROCESSORS_ONLN) && \
make clean && cp -R /sfcgal/* /
@@ -98,7 +98,7 @@ ENV PATH "/usr/local/pgsql/bin:$PATH"
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.3.tar.gz -O postgis.tar.gz && \
echo "74eb356e3f85f14233791013360881b6748f78081cc688ff9d6f0f673a762d13 postgis.tar.gz" | sha256sum --check && \
mkdir postgis-src && cd postgis-src && tar xzf ../postgis.tar.gz --strip-components=1 -C . && \
mkdir postgis-src && cd postgis-src && tar xvzf ../postgis.tar.gz --strip-components=1 -C . && \
find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
./autogen.sh && \
./configure --with-sfcgal=/usr/local/bin/sfcgal-config && \
@@ -124,7 +124,7 @@ RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.3.tar.gz -O postg
RUN wget https://github.com/pgRouting/pgrouting/archive/v3.4.2.tar.gz -O pgrouting.tar.gz && \
echo "cac297c07d34460887c4f3b522b35c470138760fe358e351ad1db4edb6ee306e pgrouting.tar.gz" | sha256sum --check && \
mkdir pgrouting-src && cd pgrouting-src && tar xzf ../pgrouting.tar.gz --strip-components=1 -C . && \
mkdir pgrouting-src && cd pgrouting-src && tar xvzf ../pgrouting.tar.gz --strip-components=1 -C . && \
mkdir build && cd build && \
cmake -DCMAKE_BUILD_TYPE=Release .. && \
make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -149,7 +149,7 @@ RUN apt update && \
RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.10.tar.gz -O plv8.tar.gz && \
echo "7096c3290928561f0d4901b7a52794295dc47f6303102fae3f8e42dd575ad97d plv8.tar.gz" | sha256sum --check && \
mkdir plv8-src && cd plv8-src && tar xzf ../plv8.tar.gz --strip-components=1 -C . && \
mkdir plv8-src && cd plv8-src && tar xvzf ../plv8.tar.gz --strip-components=1 -C . && \
# generate and copy upgrade scripts
mkdir -p upgrade && ./generate_upgrade.sh 3.1.10 && \
cp upgrade/* /usr/local/pgsql/share/extension/ && \
@@ -194,7 +194,7 @@ RUN case "$(uname -m)" in \
RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz && \
echo "ec99f1f5974846bde64f4513cf8d2ea1b8d172d2218ab41803bf6a63532272bc h3.tar.gz" | sha256sum --check && \
mkdir h3-src && cd h3-src && tar xzf ../h3.tar.gz --strip-components=1 -C . && \
mkdir h3-src && cd h3-src && tar xvzf ../h3.tar.gz --strip-components=1 -C . && \
mkdir build && cd build && \
cmake .. -DCMAKE_BUILD_TYPE=Release && \
make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -204,7 +204,7 @@ RUN wget https://github.com/uber/h3/archive/refs/tags/v4.1.0.tar.gz -O h3.tar.gz
RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.1.3.tar.gz -O h3-pg.tar.gz && \
echo "5c17f09a820859ffe949f847bebf1be98511fb8f1bd86f94932512c00479e324 h3-pg.tar.gz" | sha256sum --check && \
mkdir h3-pg-src && cd h3-pg-src && tar xzf ../h3-pg.tar.gz --strip-components=1 -C . && \
mkdir h3-pg-src && cd h3-pg-src && tar xvzf ../h3-pg.tar.gz --strip-components=1 -C . && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
@@ -222,7 +222,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -O postgresql-unit.tar.gz && \
echo "411d05beeb97e5a4abf17572bfcfbb5a68d98d1018918feff995f6ee3bb03e79 postgresql-unit.tar.gz" | sha256sum --check && \
mkdir postgresql-unit-src && cd postgresql-unit-src && tar xzf ../postgresql-unit.tar.gz --strip-components=1 -C . && \
mkdir postgresql-unit-src && cd postgresql-unit-src && tar xvzf ../postgresql-unit.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
# unit extension's "create extension" script relies on absolute install path to fill some reference tables.
@@ -241,22 +241,11 @@ RUN wget https://github.com/df7cb/postgresql-unit/archive/refs/tags/7.7.tar.gz -
FROM build-deps AS vector-pg-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
COPY patches/pgvector.patch /pgvector.patch
# By default, pgvector Makefile uses `-march=native`. We don't want that,
# because we build the images on different machines than where we run them.
# Pass OPTFLAGS="" to remove it.
RUN if [ "$(uname -m)" = "x86_64" ]; then \
OPTFLAGS=" -march=x86-64 "; \
elif [ "$(uname -m)" = "aarch64" ]; then \
OPTFLAGS=""; \
fi && \
wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.7.2.tar.gz -O pgvector.tar.gz && \
echo "617fba855c9bcb41a2a9bc78a78567fd2e147c72afd5bf9d37b31b9591632b30 pgvector.tar.gz" | sha256sum --check && \
mkdir pgvector-src && cd pgvector-src && tar xzf ../pgvector.tar.gz --strip-components=1 -C . && \
patch -p1 < /pgvector.patch && \
make -j $(getconf _NPROCESSORS_ONLN) OPTFLAGS="$OPTFLAGS" PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) OPTFLAGS="$OPTFLAGS" install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
RUN wget https://github.com/pgvector/pgvector/archive/refs/tags/v0.5.1.tar.gz -O pgvector.tar.gz && \
echo "cc7a8e034a96e30a819911ac79d32f6bc47bdd1aa2de4d7d4904e26b83209dc8 pgvector.tar.gz" | sha256sum --check && \
mkdir pgvector-src && cd pgvector-src && tar xvzf ../pgvector.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/vector.control
#########################################################################################
@@ -271,7 +260,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
# 9742dab1b2f297ad3811120db7b21451bca2d3c9 made on 13/11/2021
RUN wget https://github.com/michelp/pgjwt/archive/9742dab1b2f297ad3811120db7b21451bca2d3c9.tar.gz -O pgjwt.tar.gz && \
echo "cfdefb15007286f67d3d45510f04a6a7a495004be5b3aecb12cda667e774203f pgjwt.tar.gz" | sha256sum --check && \
mkdir pgjwt-src && cd pgjwt-src && tar xzf ../pgjwt.tar.gz --strip-components=1 -C . && \
mkdir pgjwt-src && cd pgjwt-src && tar xvzf ../pgjwt.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgjwt.control
@@ -286,7 +275,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/HypoPG/hypopg/archive/refs/tags/1.4.0.tar.gz -O hypopg.tar.gz && \
echo "0821011743083226fc9b813c1f2ef5897a91901b57b6bea85a78e466187c6819 hypopg.tar.gz" | sha256sum --check && \
mkdir hypopg-src && cd hypopg-src && tar xzf ../hypopg.tar.gz --strip-components=1 -C . && \
mkdir hypopg-src && cd hypopg-src && tar xvzf ../hypopg.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/hypopg.control
@@ -302,7 +291,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/iCyberon/pg_hashids/archive/refs/tags/v1.2.1.tar.gz -O pg_hashids.tar.gz && \
echo "74576b992d9277c92196dd8d816baa2cc2d8046fe102f3dcd7f3c3febed6822a pg_hashids.tar.gz" | sha256sum --check && \
mkdir pg_hashids-src && cd pg_hashids-src && tar xzf ../pg_hashids.tar.gz --strip-components=1 -C . && \
mkdir pg_hashids-src && cd pg_hashids-src && tar xvzf ../pg_hashids.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_hashids.control
@@ -318,7 +307,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/postgrespro/rum/archive/refs/tags/1.3.13.tar.gz -O rum.tar.gz && \
echo "6ab370532c965568df6210bd844ac6ba649f53055e48243525b0b7e5c4d69a7d rum.tar.gz" | sha256sum --check && \
mkdir rum-src && cd rum-src && tar xzf ../rum.tar.gz --strip-components=1 -C . && \
mkdir rum-src && cd rum-src && tar xvzf ../rum.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/rum.control
@@ -334,7 +323,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/theory/pgtap/archive/refs/tags/v1.2.0.tar.gz -O pgtap.tar.gz && \
echo "9c7c3de67ea41638e14f06da5da57bac6f5bd03fea05c165a0ec862205a5c052 pgtap.tar.gz" | sha256sum --check && \
mkdir pgtap-src && cd pgtap-src && tar xzf ../pgtap.tar.gz --strip-components=1 -C . && \
mkdir pgtap-src && cd pgtap-src && tar xvzf ../pgtap.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/pgtap.control
@@ -350,7 +339,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/RhodiumToad/ip4r/archive/refs/tags/2.4.2.tar.gz -O ip4r.tar.gz && \
echo "0f7b1f159974f49a47842a8ab6751aecca1ed1142b6d5e38d81b064b2ead1b4b ip4r.tar.gz" | sha256sum --check && \
mkdir ip4r-src && cd ip4r-src && tar xzf ../ip4r.tar.gz --strip-components=1 -C . && \
mkdir ip4r-src && cd ip4r-src && tar xvzf ../ip4r.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/ip4r.control
@@ -366,7 +355,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/dimitri/prefix/archive/refs/tags/v1.2.10.tar.gz -O prefix.tar.gz && \
echo "4342f251432a5f6fb05b8597139d3ccde8dcf87e8ca1498e7ee931ca057a8575 prefix.tar.gz" | sha256sum --check && \
mkdir prefix-src && cd prefix-src && tar xzf ../prefix.tar.gz --strip-components=1 -C . && \
mkdir prefix-src && cd prefix-src && tar xvzf ../prefix.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/prefix.control
@@ -382,7 +371,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/citusdata/postgresql-hll/archive/refs/tags/v2.18.tar.gz -O hll.tar.gz && \
echo "e2f55a6f4c4ab95ee4f1b4a2b73280258c5136b161fe9d059559556079694f0e hll.tar.gz" | sha256sum --check && \
mkdir hll-src && cd hll-src && tar xzf ../hll.tar.gz --strip-components=1 -C . && \
mkdir hll-src && cd hll-src && tar xvzf ../hll.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/hll.control
@@ -398,7 +387,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN wget https://github.com/okbob/plpgsql_check/archive/refs/tags/v2.5.3.tar.gz -O plpgsql_check.tar.gz && \
echo "6631ec3e7fb3769eaaf56e3dfedb829aa761abf163d13dba354b4c218508e1c0 plpgsql_check.tar.gz" | sha256sum --check && \
mkdir plpgsql_check-src && cd plpgsql_check-src && tar xzf ../plpgsql_check.tar.gz --strip-components=1 -C . && \
mkdir plpgsql_check-src && cd plpgsql_check-src && tar xvzf ../plpgsql_check.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config USE_PGXS=1 && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/plpgsql_check.control
@@ -429,7 +418,7 @@ RUN case "${PG_VERSION}" in \
apt-get install -y cmake && \
wget https://github.com/timescale/timescaledb/archive/refs/tags/${TIMESCALEDB_VERSION}.tar.gz -O timescaledb.tar.gz && \
echo "${TIMESCALEDB_CHECKSUM} timescaledb.tar.gz" | sha256sum --check && \
mkdir timescaledb-src && cd timescaledb-src && tar xzf ../timescaledb.tar.gz --strip-components=1 -C . && \
mkdir timescaledb-src && cd timescaledb-src && tar xvzf ../timescaledb.tar.gz --strip-components=1 -C . && \
./bootstrap -DSEND_TELEMETRY_DEFAULT:BOOL=OFF -DUSE_TELEMETRY:BOOL=OFF -DAPACHE_ONLY:BOOL=ON -DCMAKE_BUILD_TYPE=Release && \
cd build && \
make -j $(getconf _NPROCESSORS_ONLN) && \
@@ -467,7 +456,7 @@ RUN case "${PG_VERSION}" in \
esac && \
wget https://github.com/ossc-db/pg_hint_plan/archive/refs/tags/REL${PG_HINT_PLAN_VERSION}.tar.gz -O pg_hint_plan.tar.gz && \
echo "${PG_HINT_PLAN_CHECKSUM} pg_hint_plan.tar.gz" | sha256sum --check && \
mkdir pg_hint_plan-src && cd pg_hint_plan-src && tar xzf ../pg_hint_plan.tar.gz --strip-components=1 -C . && \
mkdir pg_hint_plan-src && cd pg_hint_plan-src && tar xvzf ../pg_hint_plan.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make install -j $(getconf _NPROCESSORS_ONLN) && \
echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_hint_plan.control
@@ -486,7 +475,7 @@ RUN apt-get update && \
apt-get install -y git libgtk2.0-dev libpq-dev libpam-dev libxslt-dev libkrb5-dev cmake && \
wget https://github.com/ketteq-neon/postgres-exts/archive/e0bd1a9d9313d7120c1b9c7bb15c48c0dede4c4e.tar.gz -O kq_imcx.tar.gz && \
echo "dc93a97ff32d152d32737ba7e196d9687041cda15e58ab31344c2f2de8855336 kq_imcx.tar.gz" | sha256sum --check && \
mkdir kq_imcx-src && cd kq_imcx-src && tar xzf ../kq_imcx.tar.gz --strip-components=1 -C . && \
mkdir kq_imcx-src && cd kq_imcx-src && tar xvzf ../kq_imcx.tar.gz --strip-components=1 -C . && \
find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
mkdir build && cd build && \
cmake -DCMAKE_BUILD_TYPE=Release .. && \
@@ -510,7 +499,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/citusdata/pg_cron/archive/refs/tags/v1.6.0.tar.gz -O pg_cron.tar.gz && \
echo "383a627867d730222c272bfd25cd5e151c578d73f696d32910c7db8c665cc7db pg_cron.tar.gz" | sha256sum --check && \
mkdir pg_cron-src && cd pg_cron-src && tar xzf ../pg_cron.tar.gz --strip-components=1 -C . && \
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 $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_cron.control
@@ -536,7 +525,7 @@ RUN apt-get update && \
ENV PATH "/usr/local/pgsql/bin/:/usr/local/pgsql/:$PATH"
RUN wget https://github.com/rdkit/rdkit/archive/refs/tags/Release_2023_03_3.tar.gz -O rdkit.tar.gz && \
echo "bdbf9a2e6988526bfeb8c56ce3cdfe2998d60ac289078e2215374288185e8c8d rdkit.tar.gz" | sha256sum --check && \
mkdir rdkit-src && cd rdkit-src && tar xzf ../rdkit.tar.gz --strip-components=1 -C . && \
mkdir rdkit-src && cd rdkit-src && tar xvzf ../rdkit.tar.gz --strip-components=1 -C . && \
cmake \
-D RDK_BUILD_CAIRO_SUPPORT=OFF \
-D RDK_BUILD_INCHI_SUPPORT=ON \
@@ -576,7 +565,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/fboulnois/pg_uuidv7/archive/refs/tags/v1.0.1.tar.gz -O pg_uuidv7.tar.gz && \
echo "0d0759ab01b7fb23851ecffb0bce27822e1868a4a5819bfd276101c716637a7a pg_uuidv7.tar.gz" | sha256sum --check && \
mkdir pg_uuidv7-src && cd pg_uuidv7-src && tar xzf ../pg_uuidv7.tar.gz --strip-components=1 -C . && \
mkdir pg_uuidv7-src && cd pg_uuidv7-src && tar xvzf ../pg_uuidv7.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_uuidv7.control
@@ -593,7 +582,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/ChenHuajun/pg_roaringbitmap/archive/refs/tags/v0.5.4.tar.gz -O pg_roaringbitmap.tar.gz && \
echo "b75201efcb1c2d1b014ec4ae6a22769cc7a224e6e406a587f5784a37b6b5a2aa pg_roaringbitmap.tar.gz" | sha256sum --check && \
mkdir pg_roaringbitmap-src && cd pg_roaringbitmap-src && tar xzf ../pg_roaringbitmap.tar.gz --strip-components=1 -C . && \
mkdir pg_roaringbitmap-src && cd pg_roaringbitmap-src && tar xvzf ../pg_roaringbitmap.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/roaringbitmap.control
@@ -610,7 +599,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/theory/pg-semver/archive/refs/tags/v0.32.1.tar.gz -O pg_semver.tar.gz && \
echo "fbdaf7512026d62eec03fad8687c15ed509b6ba395bff140acd63d2e4fbe25d7 pg_semver.tar.gz" | sha256sum --check && \
mkdir pg_semver-src && cd pg_semver-src && tar xzf ../pg_semver.tar.gz --strip-components=1 -C . && \
mkdir pg_semver-src && cd pg_semver-src && tar xvzf ../pg_semver.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/semver.control
@@ -636,7 +625,7 @@ RUN case "${PG_VERSION}" in \
esac && \
wget https://github.com/neondatabase/pg_embedding/archive/refs/tags/${PG_EMBEDDING_VERSION}.tar.gz -O pg_embedding.tar.gz && \
echo "${PG_EMBEDDING_CHECKSUM} pg_embedding.tar.gz" | sha256sum --check && \
mkdir pg_embedding-src && cd pg_embedding-src && tar xzf ../pg_embedding.tar.gz --strip-components=1 -C . && \
mkdir pg_embedding-src && cd pg_embedding-src && tar xvzf ../pg_embedding.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install
@@ -652,7 +641,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/neondatabase/postgresql_anonymizer/archive/refs/tags/neon_1.1.1.tar.gz -O pg_anon.tar.gz && \
echo "321ea8d5c1648880aafde850a2c576e4a9e7b9933a34ce272efc839328999fa9 pg_anon.tar.gz" | sha256sum --check && \
mkdir pg_anon-src && cd pg_anon-src && tar xzf ../pg_anon.tar.gz --strip-components=1 -C . && \
mkdir pg_anon-src && cd pg_anon-src && tar xvzf ../pg_anon.tar.gz --strip-components=1 -C . && \
find /usr/local/pgsql -type f | sed 's|^/usr/local/pgsql/||' > /before.txt &&\
make -j $(getconf _NPROCESSORS_ONLN) install PG_CONFIG=/usr/local/pgsql/bin/pg_config && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/anon.control && \
@@ -701,7 +690,7 @@ ARG PG_VERSION
RUN wget https://github.com/supabase/pg_jsonschema/archive/refs/tags/v0.2.0.tar.gz -O pg_jsonschema.tar.gz && \
echo "9118fc508a6e231e7a39acaa6f066fcd79af17a5db757b47d2eefbe14f7794f0 pg_jsonschema.tar.gz" | sha256sum --check && \
mkdir pg_jsonschema-src && cd pg_jsonschema-src && tar xzf ../pg_jsonschema.tar.gz --strip-components=1 -C . && \
mkdir pg_jsonschema-src && cd pg_jsonschema-src && tar xvzf ../pg_jsonschema.tar.gz --strip-components=1 -C . && \
sed -i 's/pgrx = "0.10.2"/pgrx = { version = "0.10.2", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
cargo pgrx install --release && \
echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_jsonschema.control
@@ -718,7 +707,7 @@ ARG PG_VERSION
RUN wget https://github.com/supabase/pg_graphql/archive/refs/tags/v1.4.0.tar.gz -O pg_graphql.tar.gz && \
echo "bd8dc7230282b3efa9ae5baf053a54151ed0e66881c7c53750e2d0c765776edc pg_graphql.tar.gz" | sha256sum --check && \
mkdir pg_graphql-src && cd pg_graphql-src && tar xzf ../pg_graphql.tar.gz --strip-components=1 -C . && \
mkdir pg_graphql-src && cd pg_graphql-src && tar xvzf ../pg_graphql.tar.gz --strip-components=1 -C . && \
sed -i 's/pgrx = "=0.10.2"/pgrx = { version = "0.10.2", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
cargo pgrx install --release && \
# it's needed to enable extension because it uses untrusted C language
@@ -738,7 +727,7 @@ ARG PG_VERSION
# 26806147b17b60763039c6a6878884c41a262318 made on 26/09/2023
RUN wget https://github.com/kelvich/pg_tiktoken/archive/26806147b17b60763039c6a6878884c41a262318.tar.gz -O pg_tiktoken.tar.gz && \
echo "e64e55aaa38c259512d3e27c572da22c4637418cf124caba904cd50944e5004e pg_tiktoken.tar.gz" | sha256sum --check && \
mkdir pg_tiktoken-src && cd pg_tiktoken-src && tar xzf ../pg_tiktoken.tar.gz --strip-components=1 -C . && \
mkdir pg_tiktoken-src && cd pg_tiktoken-src && tar xvzf ../pg_tiktoken.tar.gz --strip-components=1 -C . && \
cargo pgrx install --release && \
echo "trusted = true" >> /usr/local/pgsql/share/extension/pg_tiktoken.control
@@ -754,7 +743,7 @@ ARG PG_VERSION
RUN wget https://github.com/pksunkara/pgx_ulid/archive/refs/tags/v0.1.3.tar.gz -O pgx_ulid.tar.gz && \
echo "ee5db82945d2d9f2d15597a80cf32de9dca67b897f605beb830561705f12683c pgx_ulid.tar.gz" | sha256sum --check && \
mkdir pgx_ulid-src && cd pgx_ulid-src && tar xzf ../pgx_ulid.tar.gz --strip-components=1 -C . && \
mkdir pgx_ulid-src && cd pgx_ulid-src && tar xvzf ../pgx_ulid.tar.gz --strip-components=1 -C . && \
echo "******************* Apply a patch for Postgres 16 support; delete in the next release ******************" && \
wget https://github.com/pksunkara/pgx_ulid/commit/f84954cf63fc8c80d964ac970d9eceed3c791196.patch && \
patch -p1 < f84954cf63fc8c80d964ac970d9eceed3c791196.patch && \
@@ -776,7 +765,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/eulerto/wal2json/archive/refs/tags/wal2json_2_5.tar.gz && \
echo "b516653575541cf221b99cf3f8be9b6821f6dbcfc125675c85f35090f824f00e wal2json_2_5.tar.gz" | sha256sum --check && \
mkdir wal2json-src && cd wal2json-src && tar xzf ../wal2json_2_5.tar.gz --strip-components=1 -C . && \
mkdir wal2json-src && cd wal2json-src && tar xvzf ../wal2json_2_5.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install
@@ -792,7 +781,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/sraoss/pg_ivm/archive/refs/tags/v1.7.tar.gz -O pg_ivm.tar.gz && \
echo "ebfde04f99203c7be4b0e873f91104090e2e83e5429c32ac242d00f334224d5e pg_ivm.tar.gz" | sha256sum --check && \
mkdir pg_ivm-src && cd pg_ivm-src && tar xzf ../pg_ivm.tar.gz --strip-components=1 -C . && \
mkdir pg_ivm-src && cd pg_ivm-src && tar xvzf ../pg_ivm.tar.gz --strip-components=1 -C . && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/pg_ivm.control
@@ -809,7 +798,7 @@ COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
ENV PATH "/usr/local/pgsql/bin/:$PATH"
RUN wget https://github.com/pgpartman/pg_partman/archive/refs/tags/v5.0.1.tar.gz -O pg_partman.tar.gz && \
echo "75b541733a9659a6c90dbd40fccb904a630a32880a6e3044d0c4c5f4c8a65525 pg_partman.tar.gz" | sha256sum --check && \
mkdir pg_partman-src && cd pg_partman-src && tar xzf ../pg_partman.tar.gz --strip-components=1 -C . && \
mkdir pg_partman-src && cd pg_partman-src && tar xvzf ../pg_partman.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_partman.control
@@ -933,69 +922,6 @@ RUN rm -r /usr/local/pgsql/include
# if they were to be used by other libraries.
RUN rm /usr/local/pgsql/lib/lib*.a
#########################################################################################
#
# Layer neon-pg-ext-test
#
#########################################################################################
FROM neon-pg-ext-build AS neon-pg-ext-test
ARG PG_VERSION
RUN mkdir /ext-src
#COPY --from=postgis-build /postgis.tar.gz /ext-src/
#COPY --from=postgis-build /sfcgal/* /usr
COPY --from=plv8-build /plv8.tar.gz /ext-src/
COPY --from=h3-pg-build /h3-pg.tar.gz /ext-src/
COPY --from=unit-pg-build /postgresql-unit.tar.gz /ext-src/
COPY --from=vector-pg-build /pgvector.tar.gz /ext-src/
COPY --from=vector-pg-build /pgvector.patch /ext-src/
COPY --from=pgjwt-pg-build /pgjwt.tar.gz /ext-src
#COPY --from=pg-jsonschema-pg-build /home/nonroot/pg_jsonschema.tar.gz /ext-src
#COPY --from=pg-graphql-pg-build /home/nonroot/pg_graphql.tar.gz /ext-src
#COPY --from=pg-tiktoken-pg-build /home/nonroot/pg_tiktoken.tar.gz /ext-src
COPY --from=hypopg-pg-build /hypopg.tar.gz /ext-src
COPY --from=pg-hashids-pg-build /pg_hashids.tar.gz /ext-src
#COPY --from=rum-pg-build /rum.tar.gz /ext-src
#COPY --from=pgtap-pg-build /pgtap.tar.gz /ext-src
COPY --from=ip4r-pg-build /ip4r.tar.gz /ext-src
COPY --from=prefix-pg-build /prefix.tar.gz /ext-src
COPY --from=hll-pg-build /hll.tar.gz /ext-src
COPY --from=plpgsql-check-pg-build /plpgsql_check.tar.gz /ext-src
#COPY --from=timescaledb-pg-build /timescaledb.tar.gz /ext-src
COPY --from=pg-hint-plan-pg-build /pg_hint_plan.tar.gz /ext-src
COPY patches/pg_hintplan.patch /ext-src
#COPY --from=kq-imcx-pg-build /kq_imcx.tar.gz /ext-src
COPY --from=pg-cron-pg-build /pg_cron.tar.gz /ext-src
COPY patches/pg_cron.patch /ext-src
#COPY --from=pg-pgx-ulid-build /home/nonroot/pgx_ulid.tar.gz /ext-src
COPY --from=rdkit-pg-build /rdkit.tar.gz /ext-src
COPY --from=pg-uuidv7-pg-build /pg_uuidv7.tar.gz /ext-src
COPY --from=pg-roaringbitmap-pg-build /pg_roaringbitmap.tar.gz /ext-src
COPY --from=pg-semver-pg-build /pg_semver.tar.gz /ext-src
#COPY --from=pg-embedding-pg-build /home/nonroot/pg_embedding-src/ /ext-src
#COPY --from=wal2json-pg-build /wal2json_2_5.tar.gz /ext-src
COPY --from=pg-anon-pg-build /pg_anon.tar.gz /ext-src
COPY patches/pg_anon.patch /ext-src
COPY --from=pg-ivm-build /pg_ivm.tar.gz /ext-src
COPY --from=pg-partman-build /pg_partman.tar.gz /ext-src
RUN cd /ext-src/ && for f in *.tar.gz; \
do echo $f; dname=$(echo $f | sed 's/\.tar.*//')-src; \
rm -rf $dname; mkdir $dname; tar xzf $f --strip-components=1 -C $dname \
|| exit 1; rm -f $f; done
RUN cd /ext-src/pgvector-src && patch -p1 <../pgvector.patch
# cmake is required for the h3 test
RUN apt-get update && apt-get install -y cmake
RUN patch -p1 < /ext-src/pg_hintplan.patch
COPY --chmod=755 docker-compose/run-tests.sh /run-tests.sh
RUN patch -p1 </ext-src/pg_anon.patch
RUN patch -p1 </ext-src/pg_cron.patch
ENV PATH=/usr/local/pgsql/bin:$PATH
ENV PGHOST=compute
ENV PGPORT=55433
ENV PGUSER=cloud_admin
ENV PGDATABASE=postgres
#########################################################################################
#
# Final layer

View File

@@ -124,8 +124,6 @@ postgres-%: postgres-configure-% \
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pageinspect install
+@echo "Compiling amcheck $*"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/amcheck install
+@echo "Compiling test_decoding $*"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/test_decoding install
.PHONY: postgres-clean-%
postgres-clean-%:

View File

@@ -1,6 +1,4 @@
[![Neon](https://github.com/neondatabase/neon/assets/11527560/f15a17f0-836e-40c5-b35d-030606a6b660)](https://neon.tech)
[![Neon](https://user-images.githubusercontent.com/13738772/236813940-dcfdcb5b-69d3-449b-a686-013febe834d4.png)](https://neon.tech)
# Neon

View File

@@ -735,7 +735,7 @@ fn cli() -> clap::Command {
Arg::new("filecache-connstr")
.long("filecache-connstr")
.default_value(
"host=localhost port=5432 dbname=postgres user=cloud_admin sslmode=disable application_name=vm-monitor",
"host=localhost port=5432 dbname=postgres user=cloud_admin sslmode=disable",
)
.value_name("FILECACHE_CONNSTR"),
)

View File

@@ -1 +0,0 @@
ALTER ROLE neon_superuser BYPASSRLS;

View File

@@ -1,18 +0,0 @@
DO $$
DECLARE
role_name text;
BEGIN
FOR role_name IN SELECT rolname FROM pg_roles WHERE pg_has_role(rolname, 'neon_superuser', 'member')
LOOP
RAISE NOTICE 'EXECUTING ALTER ROLE % INHERIT', quote_ident(role_name);
EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' INHERIT';
END LOOP;
FOR role_name IN SELECT rolname FROM pg_roles
WHERE
NOT pg_has_role(rolname, 'neon_superuser', 'member') AND NOT starts_with(rolname, 'pg_')
LOOP
RAISE NOTICE 'EXECUTING ALTER ROLE % NOBYPASSRLS', quote_ident(role_name);
EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' NOBYPASSRLS';
END LOOP;
END $$;

View File

@@ -1,6 +0,0 @@
DO $$
BEGIN
IF (SELECT setting::numeric >= 160000 FROM pg_settings WHERE name = 'server_version_num') THEN
EXECUTE 'GRANT pg_create_subscription TO neon_superuser';
END IF;
END $$;

View File

@@ -1 +0,0 @@
GRANT pg_monitor TO neon_superuser WITH ADMIN OPTION;

View File

@@ -1,4 +0,0 @@
-- SKIP: Deemed insufficient for allowing relations created by extensions to be
-- interacted with by neon_superuser without permission issues.
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON TABLES TO neon_superuser;

View File

@@ -1,4 +0,0 @@
-- SKIP: Deemed insufficient for allowing relations created by extensions to be
-- interacted with by neon_superuser without permission issues.
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON SEQUENCES TO neon_superuser;

View File

@@ -1,3 +0,0 @@
-- SKIP: Moved inline to the handle_grants() functions.
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON TABLES TO neon_superuser WITH GRANT OPTION;

View File

@@ -1,3 +0,0 @@
-- SKIP: Moved inline to the handle_grants() functions.
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL ON SEQUENCES TO neon_superuser WITH GRANT OPTION;

View File

@@ -1,13 +0,0 @@
-- SKIP: The original goal of this migration was to prevent creating
-- subscriptions, but this migration was insufficient.
DO $$
DECLARE
role_name TEXT;
BEGIN
FOR role_name IN SELECT rolname FROM pg_roles WHERE rolreplication IS TRUE
LOOP
RAISE NOTICE 'EXECUTING ALTER ROLE % NOREPLICATION', quote_ident(role_name);
EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' NOREPLICATION';
END LOOP;
END $$;

View File

@@ -774,21 +774,44 @@ pub fn handle_migrations(client: &mut Client) -> Result<()> {
// !BE SURE TO ONLY ADD MIGRATIONS TO THE END OF THIS ARRAY. IF YOU DO NOT, VERY VERY BAD THINGS MAY HAPPEN!
// !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
// Add new migrations in numerical order.
let migrations = [
include_str!("./migrations/0000-neon_superuser_bypass_rls.sql"),
include_str!("./migrations/0001-alter_roles.sql"),
include_str!("./migrations/0002-grant_pg_create_subscription_to_neon_superuser.sql"),
include_str!("./migrations/0003-grant_pg_monitor_to_neon_superuser.sql"),
include_str!("./migrations/0004-grant_all_on_tables_to_neon_superuser.sql"),
include_str!("./migrations/0005-grant_all_on_sequences_to_neon_superuser.sql"),
include_str!(
"./migrations/0006-grant_all_on_tables_to_neon_superuser_with_grant_option.sql"
),
include_str!(
"./migrations/0007-grant_all_on_sequences_to_neon_superuser_with_grant_option.sql"
),
include_str!("./migrations/0008-revoke_replication_for_previously_allowed_roles.sql"),
"ALTER ROLE neon_superuser BYPASSRLS",
r#"
DO $$
DECLARE
role_name text;
BEGIN
FOR role_name IN SELECT rolname FROM pg_roles WHERE pg_has_role(rolname, 'neon_superuser', 'member')
LOOP
RAISE NOTICE 'EXECUTING ALTER ROLE % INHERIT', quote_ident(role_name);
EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' INHERIT';
END LOOP;
FOR role_name IN SELECT rolname FROM pg_roles
WHERE
NOT pg_has_role(rolname, 'neon_superuser', 'member') AND NOT starts_with(rolname, 'pg_')
LOOP
RAISE NOTICE 'EXECUTING ALTER ROLE % NOBYPASSRLS', quote_ident(role_name);
EXECUTE 'ALTER ROLE ' || quote_ident(role_name) || ' NOBYPASSRLS';
END LOOP;
END $$;
"#,
r#"
DO $$
BEGIN
IF (SELECT setting::numeric >= 160000 FROM pg_settings WHERE name = 'server_version_num') THEN
EXECUTE 'GRANT pg_create_subscription TO neon_superuser';
END IF;
END
$$;"#,
"GRANT pg_monitor TO neon_superuser WITH ADMIN OPTION",
// Don't remove: these are some SQLs that we originally applied in migrations but turned out to execute somewhere else.
"",
"",
"",
"",
"",
// Add new migrations below.
];
let mut func = || {
@@ -824,13 +847,10 @@ pub fn handle_migrations(client: &mut Client) -> Result<()> {
while current_migration < migrations.len() {
let migration = &migrations[current_migration];
if migration.starts_with("-- SKIP") {
info!("Skipping migration id={}", current_migration);
if migration.is_empty() {
info!("Skip migration id={}", current_migration);
} else {
info!(
"Running migration id={}:\n{}\n",
current_migration, migration
);
info!("Running migration:\n{}\n", migration);
client.simple_query(migration).with_context(|| {
format!("handle_migrations current_migration={}", current_migration)
})?;

View File

@@ -1,5 +1,3 @@
use std::path::Path;
use anyhow::{anyhow, Context};
use tracing::warn;
@@ -19,24 +17,17 @@ pub fn resize_swap(size_bytes: u64) -> anyhow::Result<()> {
.arg(size_bytes.to_string())
.spawn();
if matches!(&child_result, Err(e) if e.kind() == std::io::ErrorKind::NotFound) {
warn!("ignoring \"not found\" error from resize-swap to avoid swapoff while compute is running");
return Ok(());
}
child_result
.context("spawn() failed")
.and_then(|mut child| child.wait().context("wait() failed"))
.and_then(|status| match status.success() {
true => Ok(()),
false => {
// The command failed. Maybe it was because the resize-swap file doesn't exist?
// The --once flag causes it to delete itself on success so we don't disable swap
// while postgres is running; maybe this is fine.
match Path::new(RESIZE_SWAP_BIN).try_exists() {
Err(_) | Ok(true) => Err(anyhow!("process exited with {status}")),
// The path doesn't exist; we're actually ok
Ok(false) => {
warn!("ignoring \"not found\" error from resize-swap to avoid swapoff while compute is running");
Ok(())
},
}
}
false => Err(anyhow!("process exited with {status}")),
})
// wrap any prior error with the overall context that we couldn't run the command
.with_context(|| {

View File

@@ -243,13 +243,9 @@ impl StorageController {
anyhow::bail!("initdb failed with status {status}");
}
// Write a minimal config file:
// - Specify the port, since this is chosen dynamically
// - Switch off fsync, since we're running on lightweight test environments and when e.g. scale testing
// the storage controller we don't want a slow local disk to interfere with that.
tokio::fs::write(
&pg_data_path.join("postgresql.conf"),
format!("port = {}\nfsync=off\n", self.postgres_port),
format!("port = {}", self.postgres_port),
)
.await?;
};

View File

@@ -9,8 +9,6 @@ license.workspace = true
anyhow.workspace = true
clap.workspace = true
comfy-table.workspace = true
futures.workspace = true
humantime.workspace = true
hyper.workspace = true
pageserver_api.workspace = true
pageserver_client.workspace = true

View File

@@ -1,4 +1,3 @@
use futures::StreamExt;
use std::{collections::HashMap, str::FromStr, time::Duration};
use clap::{Parser, Subcommand};
@@ -8,9 +7,8 @@ use pageserver_api::{
TenantDescribeResponse, TenantPolicyRequest,
},
models::{
EvictionPolicy, EvictionPolicyLayerAccessThreshold, LocationConfigSecondary,
ShardParameters, TenantConfig, TenantConfigRequest, TenantCreateRequest,
TenantShardSplitRequest, TenantShardSplitResponse,
LocationConfigSecondary, ShardParameters, TenantConfig, TenantConfigRequest,
TenantCreateRequest, TenantShardSplitRequest, TenantShardSplitResponse,
},
shard::{ShardStripeSize, TenantShardId},
};
@@ -127,44 +125,6 @@ enum Command {
#[arg(long)]
tenant_id: TenantId,
},
/// Uncleanly drop a tenant from the storage controller: this doesn't delete anything from pageservers. Appropriate
/// if you e.g. used `tenant-warmup` by mistake on a tenant ID that doesn't really exist, or is in some other region.
TenantDrop {
#[arg(long)]
tenant_id: TenantId,
#[arg(long)]
unclean: bool,
},
NodeDrop {
#[arg(long)]
node_id: NodeId,
#[arg(long)]
unclean: bool,
},
TenantSetTimeBasedEviction {
#[arg(long)]
tenant_id: TenantId,
#[arg(long)]
period: humantime::Duration,
#[arg(long)]
threshold: humantime::Duration,
},
// Drain a set of specified pageservers by moving the primary attachments to pageservers
// outside of the specified set.
Drain {
// Set of pageserver node ids to drain.
#[arg(long)]
nodes: Vec<NodeId>,
// Optional: migration concurrency (default is 8)
#[arg(long)]
concurrency: Option<usize>,
// Optional: maximum number of shards to migrate
#[arg(long)]
max_shards: Option<usize>,
// Optional: when set to true, nothing is migrated, but the plan is printed to stdout
#[arg(long)]
dry_run: Option<bool>,
},
}
#[derive(Parser)]
@@ -714,234 +674,6 @@ async fn main() -> anyhow::Result<()> {
}
}
}
Command::TenantDrop { tenant_id, unclean } => {
if !unclean {
anyhow::bail!("This command is not a tenant deletion, and uncleanly drops all controller state for the tenant. If you know what you're doing, add `--unclean` to proceed.")
}
storcon_client
.dispatch::<(), ()>(
Method::POST,
format!("debug/v1/tenant/{tenant_id}/drop"),
None,
)
.await?;
}
Command::NodeDrop { node_id, unclean } => {
if !unclean {
anyhow::bail!("This command is not a clean node decommission, and uncleanly drops all controller state for the node, without checking if any tenants still refer to it. If you know what you're doing, add `--unclean` to proceed.")
}
storcon_client
.dispatch::<(), ()>(Method::POST, format!("debug/v1/node/{node_id}/drop"), None)
.await?;
}
Command::TenantSetTimeBasedEviction {
tenant_id,
period,
threshold,
} => {
vps_client
.tenant_config(&TenantConfigRequest {
tenant_id,
config: TenantConfig {
eviction_policy: Some(EvictionPolicy::LayerAccessThreshold(
EvictionPolicyLayerAccessThreshold {
period: period.into(),
threshold: threshold.into(),
},
)),
..Default::default()
},
})
.await?;
}
Command::Drain {
nodes,
concurrency,
max_shards,
dry_run,
} => {
// Load the list of nodes, split them up into the drained and filled sets,
// and validate that draining is possible.
let node_descs = storcon_client
.dispatch::<(), Vec<NodeDescribeResponse>>(
Method::GET,
"control/v1/node".to_string(),
None,
)
.await?;
let mut node_to_drain_descs = Vec::new();
let mut node_to_fill_descs = Vec::new();
for desc in node_descs {
let to_drain = nodes.iter().any(|id| *id == desc.id);
if to_drain {
node_to_drain_descs.push(desc);
} else {
node_to_fill_descs.push(desc);
}
}
if nodes.len() != node_to_drain_descs.len() {
anyhow::bail!("Drain requested for node which doesn't exist.")
}
node_to_fill_descs.retain(|desc| {
matches!(desc.availability, NodeAvailabilityWrapper::Active)
&& matches!(
desc.scheduling,
NodeSchedulingPolicy::Active | NodeSchedulingPolicy::Filling
)
});
if node_to_fill_descs.is_empty() {
anyhow::bail!("There are no nodes to drain to")
}
// Set the node scheduling policy to draining for the nodes which
// we plan to drain.
for node_desc in node_to_drain_descs.iter() {
let req = NodeConfigureRequest {
node_id: node_desc.id,
availability: None,
scheduling: Some(NodeSchedulingPolicy::Draining),
};
storcon_client
.dispatch::<_, ()>(
Method::PUT,
format!("control/v1/node/{}/config", node_desc.id),
Some(req),
)
.await?;
}
// Perform the drain: move each tenant shard scheduled on a node to
// be drained to a node which is being filled. A simple round robin
// strategy is used to pick the new node.
let tenants = storcon_client
.dispatch::<(), Vec<TenantDescribeResponse>>(
Method::GET,
"control/v1/tenant".to_string(),
None,
)
.await?;
let mut selected_node_idx = 0;
struct DrainMove {
tenant_shard_id: TenantShardId,
from: NodeId,
to: NodeId,
}
let mut moves: Vec<DrainMove> = Vec::new();
let shards = tenants
.into_iter()
.flat_map(|tenant| tenant.shards.into_iter());
for shard in shards {
if let Some(max_shards) = max_shards {
if moves.len() >= max_shards {
println!(
"Stop planning shard moves since the requested maximum was reached"
);
break;
}
}
let should_migrate = {
if let Some(attached_to) = shard.node_attached {
node_to_drain_descs
.iter()
.map(|desc| desc.id)
.any(|id| id == attached_to)
} else {
false
}
};
if !should_migrate {
continue;
}
moves.push(DrainMove {
tenant_shard_id: shard.tenant_shard_id,
from: shard
.node_attached
.expect("We only migrate attached tenant shards"),
to: node_to_fill_descs[selected_node_idx].id,
});
selected_node_idx = (selected_node_idx + 1) % node_to_fill_descs.len();
}
let total_moves = moves.len();
if dry_run == Some(true) {
println!("Dryrun requested. Planned {total_moves} moves:");
for mv in &moves {
println!("{}: {} -> {}", mv.tenant_shard_id, mv.from, mv.to)
}
return Ok(());
}
const DEFAULT_MIGRATE_CONCURRENCY: usize = 8;
let mut stream = futures::stream::iter(moves)
.map(|mv| {
let client = Client::new(cli.api.clone(), cli.jwt.clone());
async move {
client
.dispatch::<TenantShardMigrateRequest, TenantShardMigrateResponse>(
Method::PUT,
format!("control/v1/tenant/{}/migrate", mv.tenant_shard_id),
Some(TenantShardMigrateRequest {
tenant_shard_id: mv.tenant_shard_id,
node_id: mv.to,
}),
)
.await
.map_err(|e| (mv.tenant_shard_id, mv.from, mv.to, e))
}
})
.buffered(concurrency.unwrap_or(DEFAULT_MIGRATE_CONCURRENCY));
let mut success = 0;
let mut failure = 0;
while let Some(res) = stream.next().await {
match res {
Ok(_) => {
success += 1;
}
Err((tenant_shard_id, from, to, error)) => {
failure += 1;
println!(
"Failed to migrate {} from node {} to node {}: {}",
tenant_shard_id, from, to, error
);
}
}
if (success + failure) % 20 == 0 {
println!(
"Processed {}/{} shards: {} succeeded, {} failed",
success + failure,
total_moves,
success,
failure
);
}
}
println!(
"Processed {}/{} shards: {} succeeded, {} failed",
success + failure,
total_moves,
success,
failure
);
}
}
Ok(())

View File

@@ -99,13 +99,6 @@ name = "async-executor"
[[bans.deny]]
name = "smol"
[[bans.deny]]
# We want to use rustls instead of the platform's native tls implementation.
name = "native-tls"
[[bans.deny]]
name = "openssl"
# This section is considered when running `cargo deny check sources`.
# More documentation about the 'sources' section can be found here:
# https://embarkstudios.github.io/cargo-deny/checks/sources/cfg.html

View File

@@ -1,4 +1,4 @@
ARG REPOSITORY=neondatabase
ARG REPOSITORY=369495373322.dkr.ecr.eu-central-1.amazonaws.com
ARG COMPUTE_IMAGE=compute-node-v14
ARG TAG=latest
@@ -8,11 +8,6 @@ USER root
RUN apt-get update && \
apt-get install -y curl \
jq \
python3-pip \
netcat
#Faker is required for the pg_anon test
RUN pip3 install Faker
#This is required for the pg_hintplan test
RUN mkdir -p /ext-src/pg_hint_plan-src && chown postgres /ext-src/pg_hint_plan-src
USER postgres
USER postgres

View File

@@ -95,7 +95,7 @@
},
{
"name": "shared_preload_libraries",
"value": "neon,pg_cron,timescaledb,pg_stat_statements",
"value": "neon",
"vartype": "string"
},
{
@@ -127,16 +127,6 @@
"name": "max_replication_flush_lag",
"value": "10GB",
"vartype": "string"
},
{
"name": "cron.database",
"value": "postgres",
"vartype": "string"
},
{
"name": "session_preload_libraries",
"value": "anon",
"vartype": "string"
}
]
},

View File

@@ -1,3 +1,5 @@
version: '3'
services:
minio:
restart: always
@@ -159,12 +161,12 @@ services:
context: ./compute_wrapper/
args:
- REPOSITORY=${REPOSITORY:-neondatabase}
- COMPUTE_IMAGE=compute-node-v${PG_VERSION:-16}
- COMPUTE_IMAGE=compute-node-v${PG_VERSION:-14}
- TAG=${TAG:-latest}
- http_proxy=$http_proxy
- https_proxy=$https_proxy
environment:
- PG_VERSION=${PG_VERSION:-16}
- PG_VERSION=${PG_VERSION:-14}
#- RUST_BACKTRACE=1
# Mount the test files directly, for faster editing cycle.
volumes:
@@ -192,14 +194,3 @@ services:
done"
depends_on:
- compute
neon-test-extensions:
profiles: ["test-extensions"]
image: ${REPOSITORY:-neondatabase}/neon-test-extensions-v${PG_TEST_VERSION:-16}:${TAG:-latest}
entrypoint:
- "/bin/bash"
- "-c"
command:
- sleep 1800
depends_on:
- compute

View File

@@ -7,94 +7,54 @@
# Implicitly accepts `REPOSITORY` and `TAG` env vars that are passed into the compose file
# Their defaults point at DockerHub `neondatabase/neon:latest` image.`,
# to verify custom image builds (e.g pre-published ones).
#
# A test script for postgres extensions
# Currently supports only v16
#
# XXX: Current does not work on M1 macs due to x86_64 Docker images compiled only, and no seccomp support in M1 Docker emulation layer.
set -eux -o pipefail
COMPOSE_FILE='docker-compose.yml'
cd $(dirname $0)
SCRIPT_DIR="$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
COMPOSE_FILE=$SCRIPT_DIR/docker-compose.yml
COMPUTE_CONTAINER_NAME=docker-compose-compute-1
TEST_CONTAINER_NAME=docker-compose-neon-test-extensions-1
PSQL_OPTION="-h localhost -U cloud_admin -p 55433 -d postgres"
: ${http_proxy:=}
: ${https_proxy:=}
export http_proxy https_proxy
SQL="CREATE TABLE t(key int primary key, value text); insert into t values(1,1); select * from t;"
PSQL_OPTION="-h localhost -U cloud_admin -p 55433 -c '$SQL' postgres"
cleanup() {
echo "show container information"
docker ps
docker compose --profile test-extensions -f $COMPOSE_FILE logs
docker compose -f $COMPOSE_FILE logs
echo "stop containers..."
docker compose --profile test-extensions -f $COMPOSE_FILE down
docker compose -f $COMPOSE_FILE down
}
echo "clean up containers if exists"
cleanup
for pg_version in 14 15 16; do
echo "clean up containers if exists"
cleanup
PG_TEST_VERSION=$(($pg_version < 16 ? 16 : $pg_version))
PG_VERSION=$pg_version PG_TEST_VERSION=$PG_TEST_VERSION docker compose --profile test-extensions -f $COMPOSE_FILE up --build -d
echo "start containers (pg_version=$pg_version)."
PG_VERSION=$pg_version docker compose -f $COMPOSE_FILE up --build -d
echo "wait until the compute is ready. timeout after 60s. "
cnt=0
while sleep 3; do
while sleep 1; do
# check timeout
cnt=`expr $cnt + 3`
cnt=`expr $cnt + 1`
if [ $cnt -gt 60 ]; then
echo "timeout before the compute is ready."
cleanup
exit 1
fi
if docker compose --profile test-extensions -f $COMPOSE_FILE logs "compute_is_ready" | grep -q "accepting connections"; then
# check if the compute is ready
set +o pipefail
result=`docker compose -f $COMPOSE_FILE logs "compute_is_ready" | grep "accepting connections" | wc -l`
set -o pipefail
if [ $result -eq 1 ]; then
echo "OK. The compute is ready to connect."
echo "execute simple queries."
docker exec $COMPUTE_CONTAINER_NAME /bin/bash -c "psql $PSQL_OPTION"
cleanup
break
fi
done
if [ $pg_version -ge 16 ]
then
echo Enabling trust connection
docker exec $COMPUTE_CONTAINER_NAME bash -c "sed -i '\$d' /var/db/postgres/compute/pg_hba.conf && echo -e 'host\t all\t all\t all\t trust' >> /var/db/postgres/compute/pg_hba.conf && psql $PSQL_OPTION -c 'select pg_reload_conf()' "
echo Adding postgres role
docker exec $COMPUTE_CONTAINER_NAME psql $PSQL_OPTION -c "CREATE ROLE postgres SUPERUSER LOGIN"
# This is required for the pg_hint_plan test, to prevent flaky log message causing the test to fail
# It cannot be moved to Dockerfile now because the database directory is created after the start of the container
echo Adding dummy config
docker exec $COMPUTE_CONTAINER_NAME touch /var/db/postgres/compute/compute_ctl_temp_override.conf
# This block is required for the pg_anon extension test.
# The test assumes that it is running on the same host with the postgres engine.
# In our case it's not true, that's why we are copying files to the compute node
TMPDIR=$(mktemp -d)
docker cp $TEST_CONTAINER_NAME:/ext-src/pg_anon-src/data $TMPDIR/data
echo -e '1\t too \t many \t tabs' > $TMPDIR/data/bad.csv
docker cp $TMPDIR/data $COMPUTE_CONTAINER_NAME:/tmp/tmp_anon_alternate_data
rm -rf $TMPDIR
TMPDIR=$(mktemp -d)
# The following block does the same for the pg_hintplan test
docker cp $TEST_CONTAINER_NAME:/ext-src/pg_hint_plan-src/data $TMPDIR/data
docker cp $TMPDIR/data $COMPUTE_CONTAINER_NAME:/ext-src/pg_hint_plan-src/
rm -rf $TMPDIR
# We are running tests now
if docker exec -e SKIP=rum-src,timescaledb-src,rdkit-src,postgis-src,pgx_ulid-src,pgtap-src,pg_tiktoken-src,pg_jsonschema-src,pg_graphql-src,kq_imcx-src,wal2json_2_5-src \
$TEST_CONTAINER_NAME /run-tests.sh | tee testout.txt
then
cleanup
else
FAILED=$(tail -1 testout.txt)
for d in $FAILED
do
mkdir $d
docker cp $TEST_CONTAINER_NAME:/ext-src/$d/regression.diffs $d || true
docker cp $TEST_CONTAINER_NAME:/ext-src/$d/regression.out $d || true
cat $d/regression.out $d/regression.diffs || true
done
rm -rf $FAILED
cleanup
exit 1
fi
fi
cleanup
done

View File

@@ -1,15 +0,0 @@
#!/bin/bash
set -x
cd /ext-src
FAILED=
LIST=$((echo ${SKIP} | sed 's/,/\n/g'; ls -d *-src) | sort | uniq -u)
for d in ${LIST}
do
[ -d ${d} ] || continue
psql -c "select 1" >/dev/null || break
make -C ${d} installcheck || FAILED="${d} ${FAILED}"
done
[ -z "${FAILED}" ] && exit 0
echo ${FAILED}
exit 1

View File

@@ -4,18 +4,18 @@
Currently we build two main images:
- [neondatabase/neon](https://hub.docker.com/repository/docker/neondatabase/neon) — image with pre-built `pageserver`, `safekeeper` and `proxy` binaries and all the required runtime dependencies. Built from [/Dockerfile](/Dockerfile).
- [neondatabase/compute-node-v16](https://hub.docker.com/repository/docker/neondatabase/compute-node-v16) — compute node image with pre-built Postgres binaries from [neondatabase/postgres](https://github.com/neondatabase/postgres). Similar images exist for v15 and v14.
- [neondatabase/neon](https://hub.docker.com/repository/docker/zenithdb/zenith) — image with pre-built `pageserver`, `safekeeper` and `proxy` binaries and all the required runtime dependencies. Built from [/Dockerfile](/Dockerfile).
- [neondatabase/compute-node](https://hub.docker.com/repository/docker/zenithdb/compute-node) — compute node image with pre-built Postgres binaries from [neondatabase/postgres](https://github.com/neondatabase/postgres).
And additional intermediate image:
- [neondatabase/compute-tools](https://hub.docker.com/repository/docker/neondatabase/compute-tools) — compute node configuration management tools.
## Build pipeline
## Building pipeline
We build all images after a successful `release` tests run and push automatically to Docker Hub with two parallel CI jobs
1. `neondatabase/compute-tools` and `neondatabase/compute-node-v16` (and -v15 and -v14)
1. `neondatabase/compute-tools` and `neondatabase/compute-node`
2. `neondatabase/neon`
@@ -34,12 +34,12 @@ You can see a [docker compose](https://docs.docker.com/compose/) example to crea
1. create containers
You can specify version of neon cluster using following environment values.
- PG_VERSION: postgres version for compute (default is 16 as of this writing)
- TAG: the tag version of [docker image](https://registry.hub.docker.com/r/neondatabase/neon/tags), which is tagged in [CI test](/.github/workflows/build_and_test.yml). Default is 'latest'
- PG_VERSION: postgres version for compute (default is 14)
- TAG: the tag version of [docker image](https://registry.hub.docker.com/r/neondatabase/neon/tags) (default is latest), which is tagged in [CI test](/.github/workflows/build_and_test.yml)
```
$ cd docker-compose/
$ docker-compose down # remove the containers if exists
$ PG_VERSION=16 TAG=latest docker-compose up --build -d # You can specify the postgres and image version
$ PG_VERSION=15 TAG=2937 docker-compose up --build -d # You can specify the postgres and image version
Creating network "dockercompose_default" with the default driver
Creating docker-compose_storage_broker_1 ... done
(...omit...)
@@ -47,31 +47,29 @@ Creating docker-compose_storage_broker_1 ... done
2. connect compute node
```
$ psql postgresql://cloud_admin:cloud_admin@localhost:55433/postgres
psql (16.3)
Type "help" for help.
$ echo "localhost:55433:postgres:cloud_admin:cloud_admin" >> ~/.pgpass
$ chmod 600 ~/.pgpass
$ psql -h localhost -p 55433 -U cloud_admin
postgres=# CREATE TABLE t(key int primary key, value text);
CREATE TABLE
postgres=# insert into t values(1, 1);
postgres=# insert into t values(1,1);
INSERT 0 1
postgres=# select * from t;
key | value
key | value
-----+-------
1 | 1
(1 row)
```
3. If you want to see the log, you can use `docker-compose logs` command.
```
# check the container name you want to see
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
3582f6d76227 docker-compose_compute "/shell/compute.sh" 2 minutes ago Up 2 minutes 0.0.0.0:3080->3080/tcp, :::3080->3080/tcp, 0.0.0.0:55433->55433/tcp, :::55433->55433/tcp docker-compose_compute_1
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
d6968a5ae912 dockercompose_compute "/shell/compute.sh" 5 minutes ago Up 5 minutes 0.0.0.0:3080->3080/tcp, 0.0.0.0:55433->55433/tcp dockercompose_compute_1
(...omit...)
$ docker logs -f docker-compose_compute_1
$ docker logs -f dockercompose_compute_1
2022-10-21 06:15:48.757 GMT [56] LOG: connection authorized: user=cloud_admin database=postgres application_name=psql
2022-10-21 06:17:00.307 GMT [56] LOG: [NEON_SMGR] libpagestore: connected to 'host=pageserver port=6400'
(...omit...)

View File

@@ -1,7 +1,7 @@
use anyhow::{bail, Result};
use byteorder::{ByteOrder, BE};
use bytes::BufMut;
use postgres_ffi::relfile_utils::{FSM_FORKNUM, VISIBILITYMAP_FORKNUM};
use postgres_ffi::RepOriginId;
use postgres_ffi::{Oid, TransactionId};
use serde::{Deserialize, Serialize};
use std::{fmt, ops::Range};
@@ -39,9 +39,6 @@ pub const RELATION_SIZE_PREFIX: u8 = 0x61;
/// The key prefix of AUX file keys.
pub const AUX_KEY_PREFIX: u8 = 0x62;
/// The key prefix of ReplOrigin keys.
pub const REPL_ORIGIN_KEY_PREFIX: u8 = 0x63;
/// Check if the key falls in the range of metadata keys.
pub const fn is_metadata_key_slice(key: &[u8]) -> bool {
key[0] >= METADATA_KEY_BEGIN_PREFIX && key[0] < METADATA_KEY_END_PREFIX
@@ -56,8 +53,14 @@ impl Key {
/// Encode a metadata key to a storage key.
pub fn from_metadata_key_fixed_size(key: &[u8; METADATA_KEY_SIZE]) -> Self {
assert!(is_metadata_key_slice(key), "key not in metadata key range");
// Metadata key space ends at 0x7F so it's fine to directly convert it to i128.
Self::from_i128(i128::from_be_bytes(*key))
Key {
field1: key[0],
field2: u16::from_be_bytes(key[1..3].try_into().unwrap()) as u32,
field3: u32::from_be_bytes(key[3..7].try_into().unwrap()),
field4: u32::from_be_bytes(key[7..11].try_into().unwrap()),
field5: key[11],
field6: u32::from_be_bytes(key[12..16].try_into().unwrap()),
}
}
/// Encode a metadata key to a storage key.
@@ -65,6 +68,17 @@ impl Key {
Self::from_metadata_key_fixed_size(key.try_into().expect("expect 16 byte metadata key"))
}
/// Extract a metadata key to a writer. The result should always be 16 bytes.
pub fn extract_metadata_key_to_writer(&self, mut writer: impl BufMut) {
writer.put_u8(self.field1);
assert!(self.field2 <= 0xFFFF);
writer.put_u16(self.field2 as u16);
writer.put_u32(self.field3);
writer.put_u32(self.field4);
writer.put_u8(self.field5);
writer.put_u32(self.field6);
}
/// Get the range of metadata keys.
pub const fn metadata_key_range() -> Range<Self> {
Key {
@@ -107,7 +121,7 @@ impl Key {
/// As long as Neon does not support tablespace (because of lack of access to local file system),
/// we can assume that only some predefined namespace OIDs are used which can fit in u16
pub fn to_i128(&self) -> i128 {
assert!(self.field2 <= 0xFFFF || self.field2 == 0xFFFFFFFF || self.field2 == 0x22222222);
assert!(self.field2 < 0xFFFF || self.field2 == 0xFFFFFFFF || self.field2 == 0x22222222);
(((self.field1 & 0x7F) as i128) << 120)
| (((self.field2 & 0xFFFF) as i128) << 104)
| ((self.field3 as i128) << 72)
@@ -161,7 +175,7 @@ impl Key {
}
/// Convert a 18B slice to a key. This function should not be used for metadata keys because field2 is handled differently.
/// Use [`Key::from_i128`] instead if you want to handle 16B keys (i.e., metadata keys).
/// Use [`Key::from_metadata_key`] instead.
pub fn from_slice(b: &[u8]) -> Self {
Key {
field1: b[0],
@@ -174,7 +188,7 @@ impl Key {
}
/// Convert a key to a 18B slice. This function should not be used for metadata keys because field2 is handled differently.
/// Use [`Key::to_i128`] instead if you want to get a 16B key (i.e., metadata keys).
/// Use [`Key::extract_metadata_key_to_writer`] instead.
pub fn write_to_byte_slice(&self, buf: &mut [u8]) {
buf[0] = self.field1;
BE::write_u32(&mut buf[1..5], self.field2);
@@ -385,14 +399,7 @@ pub fn rel_size_to_key(rel: RelTag) -> Key {
field3: rel.dbnode,
field4: rel.relnode,
field5: rel.forknum,
field6: 0xffff_ffff,
}
}
impl Key {
#[inline(always)]
pub fn is_rel_size_key(&self) -> bool {
self.field1 == 0 && self.field6 == u32::MAX
field6: 0xffffffff,
}
}
@@ -433,25 +440,6 @@ pub fn slru_dir_to_key(kind: SlruKind) -> Key {
}
}
#[inline(always)]
pub fn slru_dir_kind(key: &Key) -> Option<Result<SlruKind, u32>> {
if key.field1 == 0x01
&& key.field3 == 0
&& key.field4 == 0
&& key.field5 == 0
&& key.field6 == 0
{
match key.field2 {
0 => Some(Ok(SlruKind::Clog)),
1 => Some(Ok(SlruKind::MultiXactMembers)),
2 => Some(Ok(SlruKind::MultiXactOffsets)),
x => Some(Err(x)),
}
} else {
None
}
}
#[inline(always)]
pub fn slru_block_to_key(kind: SlruKind, segno: u32, blknum: BlockNumber) -> Key {
Key {
@@ -480,17 +468,7 @@ pub fn slru_segment_size_to_key(kind: SlruKind, segno: u32) -> Key {
field3: 1,
field4: segno,
field5: 0,
field6: 0xffff_ffff,
}
}
impl Key {
pub fn is_slru_segment_size_key(&self) -> bool {
self.field1 == 0x01
&& self.field2 < 0x03
&& self.field3 == 0x01
&& self.field5 == 0
&& self.field6 == u32::MAX
field6: 0xffffffff,
}
}
@@ -591,37 +569,6 @@ pub const AUX_FILES_KEY: Key = Key {
field6: 2,
};
#[inline(always)]
pub fn repl_origin_key(origin_id: RepOriginId) -> Key {
Key {
field1: REPL_ORIGIN_KEY_PREFIX,
field2: 0,
field3: 0,
field4: 0,
field5: 0,
field6: origin_id as u32,
}
}
/// Get the range of replorigin keys.
pub fn repl_origin_key_range() -> Range<Key> {
Key {
field1: REPL_ORIGIN_KEY_PREFIX,
field2: 0,
field3: 0,
field4: 0,
field5: 0,
field6: 0,
}..Key {
field1: REPL_ORIGIN_KEY_PREFIX,
field2: 0,
field3: 0,
field4: 0,
field5: 0,
field6: 0x10000,
}
}
// Reverse mappings for a few Keys.
// These are needed by WAL redo manager.
@@ -630,78 +577,73 @@ pub const NON_INHERITED_RANGE: Range<Key> = AUX_FILES_KEY..AUX_FILES_KEY.next();
/// Sparse keyspace range for vectored get. Missing key error will be ignored for this range.
pub const NON_INHERITED_SPARSE_RANGE: Range<Key> = Key::metadata_key_range();
impl Key {
// AUX_FILES currently stores only data for logical replication (slots etc), and
// we don't preserve these on a branch because safekeepers can't follow timeline
// switch (and generally it likely should be optional), so ignore these.
#[inline(always)]
pub fn is_inherited_key(self) -> bool {
!NON_INHERITED_RANGE.contains(&self) && !NON_INHERITED_SPARSE_RANGE.contains(&self)
}
// AUX_FILES currently stores only data for logical replication (slots etc), and
// we don't preserve these on a branch because safekeepers can't follow timeline
// switch (and generally it likely should be optional), so ignore these.
#[inline(always)]
pub fn is_inherited_key(key: Key) -> bool {
!NON_INHERITED_RANGE.contains(&key) && !NON_INHERITED_SPARSE_RANGE.contains(&key)
}
#[inline(always)]
pub fn is_rel_fsm_block_key(self) -> bool {
self.field1 == 0x00
&& self.field4 != 0
&& self.field5 == FSM_FORKNUM
&& self.field6 != 0xffffffff
}
#[inline(always)]
pub fn is_rel_fsm_block_key(key: Key) -> bool {
key.field1 == 0x00 && key.field4 != 0 && key.field5 == FSM_FORKNUM && key.field6 != 0xffffffff
}
#[inline(always)]
pub fn is_rel_vm_block_key(self) -> bool {
self.field1 == 0x00
&& self.field4 != 0
&& self.field5 == VISIBILITYMAP_FORKNUM
&& self.field6 != 0xffffffff
}
#[inline(always)]
pub fn is_rel_vm_block_key(key: Key) -> bool {
key.field1 == 0x00
&& key.field4 != 0
&& key.field5 == VISIBILITYMAP_FORKNUM
&& key.field6 != 0xffffffff
}
#[inline(always)]
pub fn to_slru_block(self) -> anyhow::Result<(SlruKind, u32, BlockNumber)> {
Ok(match self.field1 {
0x01 => {
let kind = match self.field2 {
0x00 => SlruKind::Clog,
0x01 => SlruKind::MultiXactMembers,
0x02 => SlruKind::MultiXactOffsets,
_ => anyhow::bail!("unrecognized slru kind 0x{:02x}", self.field2),
};
let segno = self.field4;
let blknum = self.field6;
#[inline(always)]
pub fn key_to_slru_block(key: Key) -> anyhow::Result<(SlruKind, u32, BlockNumber)> {
Ok(match key.field1 {
0x01 => {
let kind = match key.field2 {
0x00 => SlruKind::Clog,
0x01 => SlruKind::MultiXactMembers,
0x02 => SlruKind::MultiXactOffsets,
_ => anyhow::bail!("unrecognized slru kind 0x{:02x}", key.field2),
};
let segno = key.field4;
let blknum = key.field6;
(kind, segno, blknum)
}
_ => anyhow::bail!("unexpected value kind 0x{:02x}", self.field1),
})
}
(kind, segno, blknum)
}
_ => anyhow::bail!("unexpected value kind 0x{:02x}", key.field1),
})
}
#[inline(always)]
pub fn is_slru_block_key(self) -> bool {
self.field1 == 0x01 // SLRU-related
&& self.field3 == 0x00000001 // but not SlruDir
&& self.field6 != 0xffffffff // and not SlruSegSize
}
#[inline(always)]
pub fn is_slru_block_key(key: Key) -> bool {
key.field1 == 0x01 // SLRU-related
&& key.field3 == 0x00000001 // but not SlruDir
&& key.field6 != 0xffffffff // and not SlruSegSize
}
#[inline(always)]
pub fn is_rel_block_key(&self) -> bool {
self.field1 == 0x00 && self.field4 != 0 && self.field6 != 0xffffffff
}
#[inline(always)]
pub fn is_rel_block_key(key: &Key) -> bool {
key.field1 == 0x00 && key.field4 != 0 && key.field6 != 0xffffffff
}
/// Guaranteed to return `Ok()` if [`Self::is_rel_block_key`] returns `true` for `key`.
#[inline(always)]
pub fn to_rel_block(self) -> anyhow::Result<(RelTag, BlockNumber)> {
Ok(match self.field1 {
0x00 => (
RelTag {
spcnode: self.field2,
dbnode: self.field3,
relnode: self.field4,
forknum: self.field5,
},
self.field6,
),
_ => anyhow::bail!("unexpected value kind 0x{:02x}", self.field1),
})
}
/// Guaranteed to return `Ok()` if [[is_rel_block_key]] returns `true` for `key`.
#[inline(always)]
pub fn key_to_rel_block(key: Key) -> anyhow::Result<(RelTag, BlockNumber)> {
Ok(match key.field1 {
0x00 => (
RelTag {
spcnode: key.field2,
dbnode: key.field3,
relnode: key.field4,
forknum: key.field5,
},
key.field6,
),
_ => anyhow::bail!("unexpected value kind 0x{:02x}", key.field1),
})
}
impl std::str::FromStr for Key {
@@ -745,15 +687,10 @@ mod tests {
let mut metadata_key = vec![AUX_KEY_PREFIX];
metadata_key.extend_from_slice(&[0xFF; 15]);
let encoded_key = Key::from_metadata_key(&metadata_key);
let output_key = encoded_key.to_i128().to_be_bytes();
let mut output_key = Vec::new();
encoded_key.extract_metadata_key_to_writer(&mut output_key);
assert_eq!(metadata_key, output_key);
assert!(encoded_key.is_metadata_key());
assert!(is_metadata_key_slice(&metadata_key));
}
#[test]
fn test_possible_largest_key() {
Key::from_i128(0x7FFF_FFFF_FFFF_FFFF_FFFF_FFFF_FFFF_FFFF);
// TODO: put this key into the system and see if anything breaks.
}
}

View File

@@ -307,7 +307,7 @@ impl KeySpace {
}
/// Merge another keyspace into the current one.
/// Note: the keyspaces must not overlap (enforced via assertions). To merge overlapping key ranges, use `KeySpaceRandomAccum`.
/// Note: the keyspaces must not ovelap (enforced via assertions)
pub fn merge(&mut self, other: &KeySpace) {
let all_ranges = self
.ranges
@@ -558,12 +558,6 @@ impl KeySpaceRandomAccum {
self.ranges.push(range);
}
pub fn add_keyspace(&mut self, keyspace: KeySpace) {
for range in keyspace.ranges {
self.add_range(range);
}
}
pub fn to_keyspace(mut self) -> KeySpace {
let mut ranges = Vec::new();
if !self.ranges.is_empty() {

View File

@@ -9,6 +9,7 @@ use std::{
collections::HashMap,
io::{BufRead, Read},
num::{NonZeroU64, NonZeroUsize},
str::FromStr,
sync::atomic::AtomicUsize,
time::{Duration, SystemTime},
};
@@ -161,22 +162,6 @@ impl std::fmt::Debug for TenantState {
}
}
/// A temporary lease to a specific lsn inside a timeline.
/// Access to the lsn is guaranteed by the pageserver until the expiration indicated by `valid_until`.
#[serde_as]
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
pub struct LsnLease {
#[serde_as(as = "SystemTimeAsRfc3339Millis")]
pub valid_until: SystemTime,
}
serde_with::serde_conv!(
SystemTimeAsRfc3339Millis,
SystemTime,
|time: &SystemTime| humantime::format_rfc3339_millis(*time).to_string(),
|value: String| -> Result<_, humantime::TimestampError> { humantime::parse_rfc3339(&value) }
);
/// The only [`TenantState`] variants we could be `TenantState::Activating` from.
#[derive(Clone, Copy, Debug, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub enum ActivatingFrom {
@@ -305,7 +290,7 @@ pub struct TenantConfig {
pub compaction_period: Option<String>,
pub compaction_threshold: Option<usize>,
// defer parsing compaction_algorithm, like eviction_policy
pub compaction_algorithm: Option<CompactionAlgorithmSettings>,
pub compaction_algorithm: Option<CompactionAlgorithm>,
pub gc_horizon: Option<u64>,
pub gc_period: Option<String>,
pub image_creation_threshold: Option<usize>,
@@ -333,28 +318,14 @@ pub struct TenantConfig {
/// Unset -> V1
/// -> V2
/// -> CrossValidation -> V2
#[derive(
Eq,
PartialEq,
Debug,
Copy,
Clone,
strum_macros::EnumString,
strum_macros::Display,
serde_with::DeserializeFromStr,
serde_with::SerializeDisplay,
)]
#[strum(serialize_all = "kebab-case")]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
pub enum AuxFilePolicy {
/// V1 aux file policy: store everything in AUX_FILE_KEY
#[strum(ascii_case_insensitive)]
V1,
/// V2 aux file policy: store in the AUX_FILE keyspace
#[strum(ascii_case_insensitive)]
V2,
/// Cross validation runs both formats on the write path and does validation
/// on the read path.
#[strum(ascii_case_insensitive)]
CrossValidation,
}
@@ -420,6 +391,23 @@ impl AuxFilePolicy {
}
}
impl FromStr for AuxFilePolicy {
type Err = anyhow::Error;
fn from_str(s: &str) -> Result<Self, Self::Err> {
let s = s.to_lowercase();
if s == "v1" {
Ok(Self::V1)
} else if s == "v2" {
Ok(Self::V2)
} else if s == "crossvalidation" || s == "cross_validation" {
Ok(Self::CrossValidation)
} else {
anyhow::bail!("cannot parse {} to aux file policy", s)
}
}
}
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
#[serde(tag = "kind")]
pub enum EvictionPolicy {
@@ -438,28 +426,13 @@ impl EvictionPolicy {
}
}
#[derive(
Eq,
PartialEq,
Debug,
Copy,
Clone,
strum_macros::EnumString,
strum_macros::Display,
serde_with::DeserializeFromStr,
serde_with::SerializeDisplay,
)]
#[strum(serialize_all = "kebab-case")]
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
#[serde(tag = "kind")]
pub enum CompactionAlgorithm {
Legacy,
Tiered,
}
#[derive(Eq, PartialEq, Debug, Clone, Serialize, Deserialize)]
pub struct CompactionAlgorithmSettings {
pub kind: CompactionAlgorithm,
}
#[derive(Debug, Clone, Copy, PartialEq, Eq, Serialize, Deserialize)]
pub struct EvictionPolicyLayerAccessThreshold {
#[serde(with = "humantime_serde")]
@@ -816,8 +789,6 @@ pub enum HistoricLayerInfo {
lsn_end: Lsn,
remote: bool,
access_stats: LayerAccessStats,
l0: bool,
},
Image {
layer_file_name: String,
@@ -1416,7 +1387,6 @@ impl PagestreamBeMessage {
#[cfg(test)]
mod tests {
use serde_json::json;
use std::str::FromStr;
use super::*;
@@ -1679,14 +1649,4 @@ mod tests {
AuxFilePolicy::V2
));
}
#[test]
fn test_aux_parse() {
assert_eq!(AuxFilePolicy::from_str("V2").unwrap(), AuxFilePolicy::V2);
assert_eq!(AuxFilePolicy::from_str("v2").unwrap(), AuxFilePolicy::V2);
assert_eq!(
AuxFilePolicy::from_str("cross-validation").unwrap(),
AuxFilePolicy::CrossValidation
);
}
}

View File

@@ -3,7 +3,7 @@ use std::cmp::Ordering;
use std::fmt;
use postgres_ffi::pg_constants::GLOBALTABLESPACE_OID;
use postgres_ffi::relfile_utils::{forkname_to_number, forknumber_to_name, MAIN_FORKNUM};
use postgres_ffi::relfile_utils::forknumber_to_name;
use postgres_ffi::Oid;
///
@@ -68,57 +68,6 @@ impl fmt::Display for RelTag {
}
}
#[derive(Debug, thiserror::Error)]
pub enum ParseRelTagError {
#[error("invalid forknum")]
InvalidForknum(#[source] std::num::ParseIntError),
#[error("missing triplet member {}", .0)]
MissingTripletMember(usize),
#[error("invalid triplet member {}", .0)]
InvalidTripletMember(usize, #[source] std::num::ParseIntError),
}
impl std::str::FromStr for RelTag {
type Err = ParseRelTagError;
fn from_str(s: &str) -> Result<Self, Self::Err> {
use ParseRelTagError::*;
// FIXME: in postgres logs this separator is dot
// Example:
// could not read block 2 in rel 1663/208101/2620.1 from page server at lsn 0/2431E6F0
// with a regex we could get this more painlessly
let (triplet, forknum) = match s.split_once('_').or_else(|| s.split_once('.')) {
Some((t, f)) => {
let forknum = forkname_to_number(Some(f));
let forknum = if let Ok(f) = forknum {
f
} else {
f.parse::<u8>().map_err(InvalidForknum)?
};
(t, Some(forknum))
}
None => (s, None),
};
let mut split = triplet
.splitn(3, '/')
.enumerate()
.map(|(i, s)| s.parse::<u32>().map_err(|e| InvalidTripletMember(i, e)));
let spcnode = split.next().ok_or(MissingTripletMember(0))??;
let dbnode = split.next().ok_or(MissingTripletMember(1))??;
let relnode = split.next().ok_or(MissingTripletMember(2))??;
Ok(RelTag {
spcnode,
forknum: forknum.unwrap_or(MAIN_FORKNUM),
dbnode,
relnode,
})
}
}
impl RelTag {
pub fn to_segfile_name(&self, segno: u32) -> String {
let mut name = if self.spcnode == GLOBALTABLESPACE_OID {

View File

@@ -1,6 +1,9 @@
use std::{ops::RangeInclusive, str::FromStr};
use crate::{key::Key, models::ShardParameters};
use crate::{
key::{is_rel_block_key, Key},
models::ShardParameters,
};
use hex::FromHex;
use postgres_ffi::relfile_utils::INIT_FORKNUM;
use serde::{Deserialize, Serialize};
@@ -425,12 +428,6 @@ impl<'de> Deserialize<'de> for TenantShardId {
#[derive(Clone, Copy, Serialize, Deserialize, Eq, PartialEq, Debug)]
pub struct ShardStripeSize(pub u32);
impl Default for ShardStripeSize {
fn default() -> Self {
DEFAULT_STRIPE_SIZE
}
}
/// Layout version: for future upgrades where we might change how the key->shard mapping works
#[derive(Clone, Copy, Serialize, Deserialize, Eq, PartialEq, Debug)]
pub struct ShardLayout(u8);
@@ -562,14 +559,6 @@ impl ShardIdentity {
}
}
/// Obtains the shard number and count combined into a `ShardIndex`.
pub fn shard_index(&self) -> ShardIndex {
ShardIndex {
shard_count: self.count,
shard_number: self.number,
}
}
pub fn shard_slug(&self) -> String {
if self.count > ShardCount(0) {
format!("-{:02x}{:02x}", self.number.0, self.count.0)
@@ -669,7 +658,7 @@ fn key_is_shard0(key: &Key) -> bool {
// because they must be included in basebackups.
let is_initfork = key.field5 == INIT_FORKNUM;
!key.is_rel_block_key() || is_initfork
!is_rel_block_key(key) || is_initfork
}
/// Provide the same result as the function in postgres `hashfn.h` with the same name
@@ -716,25 +705,6 @@ fn key_to_shard_number(count: ShardCount, stripe_size: ShardStripeSize, key: &Ke
ShardNumber((hash % count.0 as u32) as u8)
}
/// For debugging, while not exposing the internals.
#[derive(Debug)]
#[allow(unused)] // used by debug formatting by pagectl
struct KeyShardingInfo {
shard0: bool,
shard_number: ShardNumber,
}
pub fn describe(
key: &Key,
shard_count: ShardCount,
stripe_size: ShardStripeSize,
) -> impl std::fmt::Debug {
KeyShardingInfo {
shard0: key_is_shard0(key),
shard_number: key_to_shard_number(shard_count, stripe_size, key),
}
}
#[cfg(test)]
mod tests {
use utils::Hex;

View File

@@ -820,11 +820,10 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
Ok(ProcessMsgResult::Continue)
}
/// - Log as info/error result of handling COPY stream and send back
/// ErrorResponse if that makes sense.
/// - Shutdown the stream if we got Terminate.
/// - Then close the connection because we don't handle exiting from COPY
/// stream normally.
/// Log as info/error result of handling COPY stream and send back
/// ErrorResponse if that makes sense. Shutdown the stream if we got
/// Terminate. TODO: transition into waiting for Sync msg if we initiate the
/// close.
pub async fn handle_copy_stream_end(&mut self, end: CopyStreamHandlerEnd) {
use CopyStreamHandlerEnd::*;
@@ -850,6 +849,10 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
}
}
if let Terminate = &end {
self.state = ProtoState::Closed;
}
let err_to_send_and_errcode = match &end {
ServerInitiated(_) => Some((end.to_string(), SQLSTATE_SUCCESSFUL_COMPLETION)),
Other(_) => Some((format!("{end:#}"), SQLSTATE_INTERNAL_ERROR)),
@@ -879,12 +882,6 @@ impl<IO: AsyncRead + AsyncWrite + Unpin> PostgresBackend<IO> {
error!("failed to send ErrorResponse: {}", ee);
}
}
// Proper COPY stream finishing to continue using the connection is not
// implemented at the server side (we don't need it so far). To prevent
// further usages of the connection, close it.
self.framed.shutdown().await.ok();
self.state = ProtoState::Closed;
}
}

View File

@@ -178,13 +178,6 @@ impl PgConnectionConfig {
}
}
impl fmt::Display for PgConnectionConfig {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
// The password is intentionally hidden and not part of this display string.
write!(f, "postgresql://{}:{}", self.host, self.port)
}
}
impl fmt::Debug for PgConnectionConfig {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
// We want `password: Some(REDACTED-STRING)`, not `password: Some("REDACTED-STRING")`

View File

@@ -126,7 +126,6 @@ fn main() -> anyhow::Result<()> {
.allowlist_type("PageHeaderData")
.allowlist_type("DBState")
.allowlist_type("RelMapFile")
.allowlist_type("RepOriginId")
// Because structs are used for serialization, tell bindgen to emit
// explicit padding fields.
.explicit_padding(true)

View File

@@ -110,7 +110,6 @@ pub mod pg_constants;
pub mod relfile_utils;
// Export some widely used datatypes that are unlikely to change across Postgres versions
pub use v14::bindings::RepOriginId;
pub use v14::bindings::{uint32, uint64, Oid};
pub use v14::bindings::{BlockNumber, OffsetNumber};
pub use v14::bindings::{MultiXactId, TransactionId};

View File

@@ -102,7 +102,7 @@ pub const XACT_XINFO_HAS_SUBXACTS: u32 = 1u32 << 1;
pub const XACT_XINFO_HAS_RELFILENODES: u32 = 1u32 << 2;
pub const XACT_XINFO_HAS_INVALS: u32 = 1u32 << 3;
pub const XACT_XINFO_HAS_TWOPHASE: u32 = 1u32 << 4;
pub const XACT_XINFO_HAS_ORIGIN: u32 = 1u32 << 5;
// pub const XACT_XINFO_HAS_ORIGIN: u32 = 1u32 << 5;
// pub const XACT_XINFO_HAS_AE_LOCKS: u32 = 1u32 << 6;
// pub const XACT_XINFO_HAS_GID: u32 = 1u32 << 7;
@@ -167,7 +167,6 @@ 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_REPLORIGIN_ID: u8 = 19;
pub const RM_LOGICALMSG_ID: u8 = 21;
// from neon_rmgr.h
@@ -224,10 +223,6 @@ pub const XLOG_CHECKPOINT_ONLINE: u8 = 0x10;
pub const XLP_FIRST_IS_CONTRECORD: u16 = 0x0001;
pub const XLP_LONG_HEADER: u16 = 0x0002;
/* From xlog.h */
pub const XLOG_REPLORIGIN_SET: u8 = 0x00;
pub const XLOG_REPLORIGIN_DROP: u8 = 0x10;
/* From replication/slot.h */
pub const REPL_SLOT_ON_DISK_OFFSETOF_RESTART_LSN: usize = 4*4 /* offset of `slotdata` in ReplicationSlotOnDisk */
+ 64 /* NameData */ + 4*4;
@@ -242,9 +237,6 @@ pub const SLOTS_PER_FSM_PAGE: u32 = FSM_LEAF_NODES_PER_PAGE as u32;
pub const VM_HEAPBLOCKS_PER_PAGE: u32 =
(BLCKSZ as usize - SIZEOF_PAGE_HEADER_DATA) as u32 * (8 / 2); // MAPSIZE * (BITS_PER_BYTE / BITS_PER_HEAPBLOCK)
/* From origin.c */
pub const REPLICATION_STATE_MAGIC: u32 = 0x1257DADE;
// List of subdirectories inside pgdata.
// Copied from src/bin/initdb/initdb.c
pub const PGDATA_SUBDIRS: [&str; 22] = [

View File

@@ -359,7 +359,7 @@ impl Crafter for LastWalRecordXlogSwitchEndsOnPageBoundary {
// Is there enough space on the page for another logical message and an
// XLOG_SWITCH? If not, start over.
let page_remain = XLOG_BLCKSZ as u64 - u64::from(after_lsn) % XLOG_BLCKSZ as u64;
if page_remain < base_size + XLOG_SIZE_OF_XLOG_RECORD as u64 {
if page_remain < base_size - XLOG_SIZE_OF_XLOG_RECORD as u64 {
continue;
}
@@ -373,29 +373,31 @@ impl Crafter for LastWalRecordXlogSwitchEndsOnPageBoundary {
"SELECT pg_logical_emit_message(false, 'swch', REPEAT('a', $1))",
&[&(repeats as i32)],
)?;
info!(
"current_wal_insert_lsn={}, XLOG_SIZE_OF_XLOG_RECORD={}",
client.pg_current_wal_insert_lsn()?,
XLOG_SIZE_OF_XLOG_RECORD
);
// Emit the XLOG_SWITCH
let before_xlog_switch = client.pg_current_wal_insert_lsn()?;
let xlog_switch_record_end: PgLsn =
client.query_one("SELECT pg_switch_wal()", &[])?.get(0);
if u64::from(xlog_switch_record_end) as usize % XLOG_BLCKSZ
!= XLOG_SIZE_OF_XLOG_SHORT_PHD
{
warn!(
"XLOG_SWITCH message ended not on page boundary: {}, offset = {}, repeating",
xlog_switch_record_end,
u64::from(xlog_switch_record_end) as usize % XLOG_BLCKSZ
);
continue;
}
return Ok(vec![before_xlog_switch, xlog_switch_record_end]);
break;
}
info!(
"current_wal_insert_lsn={}, XLOG_SIZE_OF_XLOG_RECORD={}",
client.pg_current_wal_insert_lsn()?,
XLOG_SIZE_OF_XLOG_RECORD
);
// Emit the XLOG_SWITCH
let before_xlog_switch = client.pg_current_wal_insert_lsn()?;
let xlog_switch_record_end: PgLsn = client.query_one("SELECT pg_switch_wal()", &[])?.get(0);
let next_segment = PgLsn::from(0x0200_0000);
ensure!(
xlog_switch_record_end < next_segment,
"XLOG_SWITCH record ended on or after the expected segment boundary: {} > {}",
xlog_switch_record_end,
next_segment
);
ensure!(
u64::from(xlog_switch_record_end) as usize % XLOG_BLCKSZ == XLOG_SIZE_OF_XLOG_SHORT_PHD,
"XLOG_SWITCH message ended not on page boundary: {}, offset = {}",
xlog_switch_record_end,
u64::from(xlog_switch_record_end) as usize % XLOG_BLCKSZ
);
Ok(vec![before_xlog_switch, xlog_switch_record_end])
}
}

View File

@@ -7,7 +7,6 @@ license.workspace = true
[dependencies]
bytes.workspace = true
byteorder.workspace = true
itertools.workspace = true
pin-project-lite.workspace = true
postgres-protocol.workspace = true
rand.workspace = true

View File

@@ -7,9 +7,8 @@ pub mod framed;
use byteorder::{BigEndian, ReadBytesExt};
use bytes::{Buf, BufMut, Bytes, BytesMut};
use itertools::Itertools;
use serde::{Deserialize, Serialize};
use std::{borrow::Cow, fmt, io, str};
use std::{borrow::Cow, collections::HashMap, fmt, io, str};
// re-export for use in utils pageserver_feedback.rs
pub use postgres_protocol::PG_EPOCH;
@@ -51,37 +50,15 @@ pub enum FeStartupPacket {
},
}
#[derive(Debug, Clone, Default)]
pub struct StartupMessageParamsBuilder {
params: BytesMut,
}
impl StartupMessageParamsBuilder {
/// Set parameter's value by its name.
/// name and value must not contain a \0 byte
pub fn insert(&mut self, name: &str, value: &str) {
self.params.put(name.as_bytes());
self.params.put(&b"\0"[..]);
self.params.put(value.as_bytes());
self.params.put(&b"\0"[..]);
}
pub fn freeze(self) -> StartupMessageParams {
StartupMessageParams {
params: self.params.freeze(),
}
}
}
#[derive(Debug, Clone, Default)]
#[derive(Debug)]
pub struct StartupMessageParams {
params: Bytes,
params: HashMap<String, String>,
}
impl StartupMessageParams {
/// Get parameter's value by its name.
pub fn get(&self, name: &str) -> Option<&str> {
self.iter().find_map(|(k, v)| (k == name).then_some(v))
self.params.get(name).map(|s| s.as_str())
}
/// Split command-line options according to PostgreSQL's logic,
@@ -135,19 +112,15 @@ impl StartupMessageParams {
/// Iterate through key-value pairs in an arbitrary order.
pub fn iter(&self) -> impl Iterator<Item = (&str, &str)> {
let params =
std::str::from_utf8(&self.params).expect("should be validated as utf8 already");
params.split_terminator('\0').tuples()
self.params.iter().map(|(k, v)| (k.as_str(), v.as_str()))
}
// This function is mostly useful in tests.
#[doc(hidden)]
pub fn new<'a, const N: usize>(pairs: [(&'a str, &'a str); N]) -> Self {
let mut b = StartupMessageParamsBuilder::default();
for (k, v) in pairs {
b.insert(k, v)
Self {
params: pairs.map(|(k, v)| (k.to_owned(), v.to_owned())).into(),
}
b.freeze()
}
}
@@ -372,21 +345,35 @@ impl FeStartupPacket {
(major_version, minor_version) => {
// StartupMessage
let s = str::from_utf8(&msg).map_err(|_e| {
ProtocolError::BadMessage("StartupMessage params: invalid utf-8".to_owned())
})?;
let s = s.strip_suffix('\0').ok_or_else(|| {
ProtocolError::Protocol(
"StartupMessage params: missing null terminator".to_string(),
)
})?;
// Parse pairs of null-terminated strings (key, value).
// See `postgres: ProcessStartupPacket, build_startup_packet`.
let mut tokens = str::from_utf8(&msg)
.map_err(|_e| {
ProtocolError::BadMessage("StartupMessage params: invalid utf-8".to_owned())
})?
.strip_suffix('\0') // drop packet's own null
.ok_or_else(|| {
ProtocolError::Protocol(
"StartupMessage params: missing null terminator".to_string(),
)
})?
.split_terminator('\0');
let mut params = HashMap::new();
while let Some(name) = tokens.next() {
let value = tokens.next().ok_or_else(|| {
ProtocolError::Protocol(
"StartupMessage params: key without value".to_string(),
)
})?;
params.insert(name.to_owned(), value.to_owned());
}
FeStartupPacket::StartupMessage {
major_version,
minor_version,
params: StartupMessageParams {
params: msg.slice_ref(s.as_bytes()),
},
params: StartupMessageParams { params },
}
}
};

View File

@@ -3,7 +3,6 @@
use std::borrow::Cow;
use std::collections::HashMap;
use std::env;
use std::fmt::Display;
use std::io;
use std::num::NonZeroU32;
use std::pin::Pin;
@@ -27,15 +26,14 @@ use futures::stream::Stream;
use futures_util::StreamExt;
use futures_util::TryStreamExt;
use http_types::{StatusCode, Url};
use scopeguard::ScopeGuard;
use tokio_util::sync::CancellationToken;
use tracing::debug;
use utils::backoff;
use crate::metrics::{start_measuring_requests, AttemptOutcome, RequestKind};
use crate::RemoteStorageActivity;
use crate::{
error::Cancelled, AzureConfig, ConcurrencyLimiter, Download, DownloadError, Listing,
ListingMode, RemotePath, RemoteStorage, StorageMetadata, TimeTravelError, TimeoutOrCancel,
error::Cancelled, s3_bucket::RequestKind, AzureConfig, ConcurrencyLimiter, Download,
DownloadError, Listing, ListingMode, RemotePath, RemoteStorage, StorageMetadata,
TimeTravelError, TimeoutOrCancel,
};
pub struct AzureBlobStorage {
@@ -140,8 +138,6 @@ impl AzureBlobStorage {
let mut last_modified = None;
let mut metadata = HashMap::new();
let started_at = start_measuring_requests(kind);
let download = async {
let response = builder
// convert to concrete Pageable
@@ -205,22 +201,13 @@ impl AzureBlobStorage {
})
};
let download = tokio::select! {
tokio::select! {
bufs = download => bufs,
cancel_or_timeout = cancel_or_timeout => match cancel_or_timeout {
TimeoutOrCancel::Timeout => return Err(DownloadError::Timeout),
TimeoutOrCancel::Cancel => return Err(DownloadError::Cancelled),
TimeoutOrCancel::Timeout => Err(DownloadError::Timeout),
TimeoutOrCancel::Cancel => Err(DownloadError::Cancelled),
},
};
let started_at = ScopeGuard::into_inner(started_at);
let outcome = match &download {
Ok(_) => AttemptOutcome::Ok,
Err(_) => AttemptOutcome::Err,
};
crate::metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, outcome, started_at);
download
}
}
async fn permit(
@@ -354,10 +341,7 @@ impl RemoteStorage for AzureBlobStorage {
metadata: Option<StorageMetadata>,
cancel: &CancellationToken,
) -> anyhow::Result<()> {
let kind = RequestKind::Put;
let _permit = self.permit(kind, cancel).await?;
let started_at = start_measuring_requests(kind);
let _permit = self.permit(RequestKind::Put, cancel).await?;
let op = async {
let blob_client = self.client.blob_client(self.relative_path_to_name(to));
@@ -381,25 +365,14 @@ impl RemoteStorage for AzureBlobStorage {
match fut.await {
Ok(Ok(_response)) => Ok(()),
Ok(Err(azure)) => Err(azure.into()),
Err(_timeout) => Err(TimeoutOrCancel::Timeout.into()),
Err(_timeout) => Err(TimeoutOrCancel::Cancel.into()),
}
};
let res = tokio::select! {
tokio::select! {
res = op => res,
_ = cancel.cancelled() => return Err(TimeoutOrCancel::Cancel.into()),
};
let outcome = match res {
Ok(_) => AttemptOutcome::Ok,
Err(_) => AttemptOutcome::Err,
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, outcome, started_at);
res
_ = cancel.cancelled() => Err(TimeoutOrCancel::Cancel.into()),
}
}
async fn download(
@@ -445,79 +418,40 @@ impl RemoteStorage for AzureBlobStorage {
paths: &'a [RemotePath],
cancel: &CancellationToken,
) -> anyhow::Result<()> {
let kind = RequestKind::Delete;
let _permit = self.permit(kind, cancel).await?;
let started_at = start_measuring_requests(kind);
let _permit = self.permit(RequestKind::Delete, cancel).await?;
let op = async {
// TODO batch requests are not supported by the SDK
// TODO batch requests are also not supported by the SDK
// https://github.com/Azure/azure-sdk-for-rust/issues/1068
// https://github.com/Azure/azure-sdk-for-rust/issues/1249
for path in paths {
#[derive(Debug)]
enum AzureOrTimeout {
AzureError(azure_core::Error),
Timeout,
Cancel,
}
impl Display for AzureOrTimeout {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(f, "{self:?}")
}
}
let warn_threshold = 3;
let max_retries = 5;
backoff::retry(
|| async {
let blob_client = self.client.blob_client(self.relative_path_to_name(path));
let blob_client = self.client.blob_client(self.relative_path_to_name(path));
let request = blob_client.delete().into_future();
let request = blob_client.delete().into_future();
let res = tokio::time::timeout(self.timeout, request).await;
let res = tokio::time::timeout(self.timeout, request).await;
match res {
Ok(Ok(_v)) => Ok(()),
Ok(Err(azure_err)) => {
if let Some(http_err) = azure_err.as_http_error() {
if http_err.status() == StatusCode::NotFound {
return Ok(());
}
}
Err(AzureOrTimeout::AzureError(azure_err))
match res {
Ok(Ok(_response)) => continue,
Ok(Err(e)) => {
if let Some(http_err) = e.as_http_error() {
if http_err.status() == StatusCode::NotFound {
continue;
}
Err(_elapsed) => Err(AzureOrTimeout::Timeout),
}
},
|err| match err {
AzureOrTimeout::AzureError(_) | AzureOrTimeout::Timeout => false,
AzureOrTimeout::Cancel => true,
},
warn_threshold,
max_retries,
"deleting remote object",
cancel,
)
.await
.ok_or_else(|| AzureOrTimeout::Cancel)
.and_then(|x| x)
.map_err(|e| match e {
AzureOrTimeout::AzureError(err) => anyhow::Error::from(err),
AzureOrTimeout::Timeout => TimeoutOrCancel::Timeout.into(),
AzureOrTimeout::Cancel => TimeoutOrCancel::Cancel.into(),
})?;
return Err(e.into());
}
Err(_elapsed) => return Err(TimeoutOrCancel::Timeout.into()),
}
}
Ok(())
};
let res = tokio::select! {
tokio::select! {
res = op => res,
_ = cancel.cancelled() => return Err(TimeoutOrCancel::Cancel.into()),
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, &res, started_at);
res
_ = cancel.cancelled() => Err(TimeoutOrCancel::Cancel.into()),
}
}
async fn copy(
@@ -526,9 +460,7 @@ impl RemoteStorage for AzureBlobStorage {
to: &RemotePath,
cancel: &CancellationToken,
) -> anyhow::Result<()> {
let kind = RequestKind::Copy;
let _permit = self.permit(kind, cancel).await?;
let started_at = start_measuring_requests(kind);
let _permit = self.permit(RequestKind::Copy, cancel).await?;
let timeout = tokio::time::sleep(self.timeout);
@@ -572,21 +504,15 @@ impl RemoteStorage for AzureBlobStorage {
}
};
let res = tokio::select! {
tokio::select! {
res = op => res,
_ = cancel.cancelled() => return Err(anyhow::Error::new(TimeoutOrCancel::Cancel)),
_ = cancel.cancelled() => Err(anyhow::Error::new(TimeoutOrCancel::Cancel)),
_ = timeout => {
let e = anyhow::Error::new(TimeoutOrCancel::Timeout);
let e = e.context(format!("Timeout, last status: {copy_status:?}"));
Err(e)
},
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, &res, started_at);
res
}
}
async fn time_travel_recover(
@@ -600,6 +526,10 @@ impl RemoteStorage for AzureBlobStorage {
// https://learn.microsoft.com/en-us/azure/storage/blobs/point-in-time-restore-overview
Err(TimeTravelError::Unimplemented)
}
fn activity(&self) -> RemoteStorageActivity {
self.concurrency_limiter.activity()
}
}
pin_project_lite::pin_project! {

View File

@@ -12,7 +12,6 @@
mod azure_blob;
mod error;
mod local_fs;
mod metrics;
mod s3_bucket;
mod simulate_failures;
mod support;
@@ -122,8 +121,8 @@ impl RemotePath {
self.0.file_name()
}
pub fn join(&self, path: impl AsRef<Utf8Path>) -> Self {
Self(self.0.join(path))
pub fn join(&self, segment: &Utf8Path) -> Self {
Self(self.0.join(segment))
}
pub fn get_path(&self) -> &Utf8PathBuf {
@@ -264,6 +263,17 @@ pub trait RemoteStorage: Send + Sync + 'static {
done_if_after: SystemTime,
cancel: &CancellationToken,
) -> Result<(), TimeTravelError>;
/// Query how busy we currently are: may be used by callers which wish to politely
/// back off if there are already a lot of operations underway.
fn activity(&self) -> RemoteStorageActivity;
}
pub struct RemoteStorageActivity {
pub read_available: usize,
pub read_total: usize,
pub write_available: usize,
pub write_total: usize,
}
/// DownloadStream is sensitive to the timeout and cancellation used with the original
@@ -445,6 +455,15 @@ impl<Other: RemoteStorage> GenericRemoteStorage<Arc<Other>> {
}
}
}
pub fn activity(&self) -> RemoteStorageActivity {
match self {
Self::LocalFs(s) => s.activity(),
Self::AwsS3(s) => s.activity(),
Self::AzureBlob(s) => s.activity(),
Self::Unreliable(s) => s.activity(),
}
}
}
impl GenericRemoteStorage {
@@ -775,6 +794,9 @@ struct ConcurrencyLimiter {
// The helps to ensure we don't exceed the thresholds.
write: Arc<Semaphore>,
read: Arc<Semaphore>,
write_total: usize,
read_total: usize,
}
impl ConcurrencyLimiter {
@@ -803,10 +825,21 @@ impl ConcurrencyLimiter {
Arc::clone(self.for_kind(kind)).acquire_owned().await
}
fn activity(&self) -> RemoteStorageActivity {
RemoteStorageActivity {
read_available: self.read.available_permits(),
read_total: self.read_total,
write_available: self.write.available_permits(),
write_total: self.write_total,
}
}
fn new(limit: usize) -> ConcurrencyLimiter {
Self {
read: Arc::new(Semaphore::new(limit)),
write: Arc::new(Semaphore::new(limit)),
read_total: limit,
write_total: limit,
}
}
}

View File

@@ -23,8 +23,8 @@ use tokio_util::{io::ReaderStream, sync::CancellationToken};
use utils::crashsafe::path_with_suffix_extension;
use crate::{
Download, DownloadError, Listing, ListingMode, RemotePath, TimeTravelError, TimeoutOrCancel,
REMOTE_STORAGE_PREFIX_SEPARATOR,
Download, DownloadError, Listing, ListingMode, RemotePath, RemoteStorageActivity,
TimeTravelError, TimeoutOrCancel, REMOTE_STORAGE_PREFIX_SEPARATOR,
};
use super::{RemoteStorage, StorageMetadata};
@@ -605,6 +605,16 @@ impl RemoteStorage for LocalFs {
) -> Result<(), TimeTravelError> {
Err(TimeTravelError::Unimplemented)
}
fn activity(&self) -> RemoteStorageActivity {
// LocalFS has no concurrency limiting: give callers the impression that plenty of units are available
RemoteStorageActivity {
read_available: 16,
read_total: 16,
write_available: 16,
write_total: 16,
}
}
}
fn storage_metadata_path(original_path: &Utf8Path) -> Utf8PathBuf {

View File

@@ -46,16 +46,15 @@ use utils::backoff;
use super::StorageMetadata;
use crate::{
error::Cancelled,
metrics::{start_counting_cancelled_wait, start_measuring_requests},
support::PermitCarrying,
ConcurrencyLimiter, Download, DownloadError, Listing, ListingMode, RemotePath, RemoteStorage,
S3Config, TimeTravelError, TimeoutOrCancel, MAX_KEYS_PER_DELETE,
REMOTE_STORAGE_PREFIX_SEPARATOR,
error::Cancelled, support::PermitCarrying, ConcurrencyLimiter, Download, DownloadError,
Listing, ListingMode, RemotePath, RemoteStorage, RemoteStorageActivity, S3Config,
TimeTravelError, TimeoutOrCancel, MAX_KEYS_PER_DELETE, REMOTE_STORAGE_PREFIX_SEPARATOR,
};
use crate::metrics::AttemptOutcome;
pub(super) use crate::metrics::RequestKind;
pub(super) mod metrics;
use self::metrics::AttemptOutcome;
pub(super) use self::metrics::RequestKind;
/// AWS S3 storage.
pub struct S3Bucket {
@@ -228,7 +227,7 @@ impl S3Bucket {
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.wait_seconds
.observe_elapsed(kind, started_at);
@@ -249,7 +248,7 @@ impl S3Bucket {
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.wait_seconds
.observe_elapsed(kind, started_at);
Ok(permit)
@@ -288,7 +287,7 @@ impl S3Bucket {
// Count this in the AttemptOutcome::Ok bucket, because 404 is not
// an error: we expect to sometimes fetch an object and find it missing,
// e.g. when probing for timeline indices.
crate::metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
kind,
AttemptOutcome::Ok,
started_at,
@@ -296,7 +295,7 @@ impl S3Bucket {
return Err(DownloadError::NotFound);
}
Err(e) => {
crate::metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
kind,
AttemptOutcome::Err,
started_at,
@@ -372,12 +371,12 @@ impl S3Bucket {
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, &resp, started_at);
let resp = resp.context("request deletion")?;
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.deleted_objects_total
.inc_by(chunk.len() as u64);
@@ -436,14 +435,14 @@ pin_project_lite::pin_project! {
/// Times and tracks the outcome of the request.
struct TimedDownload<S> {
started_at: std::time::Instant,
outcome: AttemptOutcome,
outcome: metrics::AttemptOutcome,
#[pin]
inner: S
}
impl<S> PinnedDrop for TimedDownload<S> {
fn drop(mut this: Pin<&mut Self>) {
crate::metrics::BUCKET_METRICS.req_seconds.observe_elapsed(RequestKind::Get, this.outcome, this.started_at);
metrics::BUCKET_METRICS.req_seconds.observe_elapsed(RequestKind::Get, this.outcome, this.started_at);
}
}
}
@@ -452,7 +451,7 @@ impl<S> TimedDownload<S> {
fn new(started_at: std::time::Instant, inner: S) -> Self {
TimedDownload {
started_at,
outcome: AttemptOutcome::Cancelled,
outcome: metrics::AttemptOutcome::Cancelled,
inner,
}
}
@@ -469,8 +468,8 @@ impl<S: Stream<Item = std::io::Result<Bytes>>> Stream for TimedDownload<S> {
let res = ready!(this.inner.poll_next(cx));
match &res {
Some(Ok(_)) => {}
Some(Err(_)) => *this.outcome = AttemptOutcome::Err,
None => *this.outcome = AttemptOutcome::Ok,
Some(Err(_)) => *this.outcome = metrics::AttemptOutcome::Err,
None => *this.outcome = metrics::AttemptOutcome::Ok,
}
Poll::Ready(res)
@@ -544,7 +543,7 @@ impl RemoteStorage for S3Bucket {
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, &response, started_at);
@@ -626,7 +625,7 @@ impl RemoteStorage for S3Bucket {
if let Ok(inner) = &res {
// do not incl. timeouts as errors in metrics but cancellations
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, inner, started_at);
}
@@ -674,7 +673,7 @@ impl RemoteStorage for S3Bucket {
};
let started_at = ScopeGuard::into_inner(started_at);
crate::metrics::BUCKET_METRICS
metrics::BUCKET_METRICS
.req_seconds
.observe_elapsed(kind, &res, started_at);
@@ -976,6 +975,32 @@ impl RemoteStorage for S3Bucket {
}
Ok(())
}
fn activity(&self) -> RemoteStorageActivity {
self.concurrency_limiter.activity()
}
}
/// On drop (cancellation) count towards [`metrics::BucketMetrics::cancelled_waits`].
fn start_counting_cancelled_wait(
kind: RequestKind,
) -> ScopeGuard<std::time::Instant, impl FnOnce(std::time::Instant), scopeguard::OnSuccess> {
scopeguard::guard_on_success(std::time::Instant::now(), move |_| {
metrics::BUCKET_METRICS.cancelled_waits.get(kind).inc()
})
}
/// On drop (cancellation) add time to [`metrics::BucketMetrics::req_seconds`].
fn start_measuring_requests(
kind: RequestKind,
) -> ScopeGuard<std::time::Instant, impl FnOnce(std::time::Instant), scopeguard::OnSuccess> {
scopeguard::guard_on_success(std::time::Instant::now(), move |started_at| {
metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
kind,
AttemptOutcome::Cancelled,
started_at,
)
})
}
// Save RAM and only store the needed data instead of the entire ObjectVersion/DeleteMarkerEntry

View File

@@ -15,7 +15,6 @@ pub(crate) enum RequestKind {
TimeTravel = 5,
}
use scopeguard::ScopeGuard;
use RequestKind::*;
impl RequestKind {
@@ -34,10 +33,10 @@ impl RequestKind {
}
}
pub(crate) struct RequestTyped<C>([C; 6]);
pub(super) struct RequestTyped<C>([C; 6]);
impl<C> RequestTyped<C> {
pub(crate) fn get(&self, kind: RequestKind) -> &C {
pub(super) fn get(&self, kind: RequestKind) -> &C {
&self.0[kind.as_index()]
}
@@ -59,19 +58,19 @@ impl<C> RequestTyped<C> {
}
impl RequestTyped<Histogram> {
pub(crate) fn observe_elapsed(&self, kind: RequestKind, started_at: std::time::Instant) {
pub(super) fn observe_elapsed(&self, kind: RequestKind, started_at: std::time::Instant) {
self.get(kind).observe(started_at.elapsed().as_secs_f64())
}
}
pub(crate) struct PassFailCancelledRequestTyped<C> {
pub(super) struct PassFailCancelledRequestTyped<C> {
success: RequestTyped<C>,
fail: RequestTyped<C>,
cancelled: RequestTyped<C>,
}
#[derive(Debug, Clone, Copy)]
pub(crate) enum AttemptOutcome {
pub(super) enum AttemptOutcome {
Ok,
Err,
Cancelled,
@@ -87,7 +86,7 @@ impl<T, E> From<&Result<T, E>> for AttemptOutcome {
}
impl AttemptOutcome {
pub(crate) fn as_str(&self) -> &'static str {
pub(super) fn as_str(&self) -> &'static str {
match self {
AttemptOutcome::Ok => "ok",
AttemptOutcome::Err => "err",
@@ -97,7 +96,7 @@ impl AttemptOutcome {
}
impl<C> PassFailCancelledRequestTyped<C> {
pub(crate) fn get(&self, kind: RequestKind, outcome: AttemptOutcome) -> &C {
pub(super) fn get(&self, kind: RequestKind, outcome: AttemptOutcome) -> &C {
let target = match outcome {
AttemptOutcome::Ok => &self.success,
AttemptOutcome::Err => &self.fail,
@@ -120,7 +119,7 @@ impl<C> PassFailCancelledRequestTyped<C> {
}
impl PassFailCancelledRequestTyped<Histogram> {
pub(crate) fn observe_elapsed(
pub(super) fn observe_elapsed(
&self,
kind: RequestKind,
outcome: impl Into<AttemptOutcome>,
@@ -131,44 +130,19 @@ impl PassFailCancelledRequestTyped<Histogram> {
}
}
/// On drop (cancellation) count towards [`BucketMetrics::cancelled_waits`].
pub(crate) fn start_counting_cancelled_wait(
kind: RequestKind,
) -> ScopeGuard<std::time::Instant, impl FnOnce(std::time::Instant), scopeguard::OnSuccess> {
scopeguard::guard_on_success(std::time::Instant::now(), move |_| {
crate::metrics::BUCKET_METRICS
.cancelled_waits
.get(kind)
.inc()
})
}
/// On drop (cancellation) add time to [`BucketMetrics::req_seconds`].
pub(crate) fn start_measuring_requests(
kind: RequestKind,
) -> ScopeGuard<std::time::Instant, impl FnOnce(std::time::Instant), scopeguard::OnSuccess> {
scopeguard::guard_on_success(std::time::Instant::now(), move |started_at| {
crate::metrics::BUCKET_METRICS.req_seconds.observe_elapsed(
kind,
AttemptOutcome::Cancelled,
started_at,
)
})
}
pub(crate) struct BucketMetrics {
pub(super) struct BucketMetrics {
/// Full request duration until successful completion, error or cancellation.
pub(crate) req_seconds: PassFailCancelledRequestTyped<Histogram>,
pub(super) req_seconds: PassFailCancelledRequestTyped<Histogram>,
/// Total amount of seconds waited on queue.
pub(crate) wait_seconds: RequestTyped<Histogram>,
pub(super) wait_seconds: RequestTyped<Histogram>,
/// Track how many semaphore awaits were cancelled per request type.
///
/// This is in case cancellations are happening more than expected.
pub(crate) cancelled_waits: RequestTyped<IntCounter>,
pub(super) cancelled_waits: RequestTyped<IntCounter>,
/// Total amount of deleted objects in batches or single requests.
pub(crate) deleted_objects_total: IntCounter,
pub(super) deleted_objects_total: IntCounter,
}
impl Default for BucketMetrics {

View File

@@ -12,7 +12,7 @@ use tokio_util::sync::CancellationToken;
use crate::{
Download, DownloadError, GenericRemoteStorage, Listing, ListingMode, RemotePath, RemoteStorage,
StorageMetadata, TimeTravelError,
RemoteStorageActivity, StorageMetadata, TimeTravelError,
};
pub struct UnreliableWrapper {
@@ -213,4 +213,8 @@ impl RemoteStorage for UnreliableWrapper {
.time_travel_recover(prefix, timestamp, done_if_after, cancel)
.await
}
fn activity(&self) -> RemoteStorageActivity {
self.inner.activity()
}
}

View File

@@ -78,10 +78,6 @@ where
let e = Err(std::io::Error::from(e));
return Poll::Ready(Some(e));
}
} else {
// this would be perfectly valid behaviour for doing a graceful completion on the
// download for example, but not one we expect to do right now.
tracing::warn!("continuing polling after having cancelled or timeouted");
}
this.inner.poll_next(cx)
@@ -93,22 +89,13 @@ where
}
/// Fires only on the first cancel or timeout, not on both.
pub(crate) fn cancel_or_timeout(
pub(crate) async fn cancel_or_timeout(
timeout: Duration,
cancel: CancellationToken,
) -> impl std::future::Future<Output = TimeoutOrCancel> + 'static {
// futures are lazy, they don't do anything before being polled.
//
// "precalculate" the wanted deadline before returning the future, so that we can use pause
// failpoint to trigger a timeout in test.
let deadline = tokio::time::Instant::now() + timeout;
async move {
tokio::select! {
_ = tokio::time::sleep_until(deadline) => TimeoutOrCancel::Timeout,
_ = cancel.cancelled() => {
TimeoutOrCancel::Cancel
},
}
) -> TimeoutOrCancel {
tokio::select! {
_ = tokio::time::sleep(timeout) => TimeoutOrCancel::Timeout,
_ = cancel.cancelled() => TimeoutOrCancel::Cancel,
}
}
@@ -185,31 +172,4 @@ mod tests {
_ = tokio::time::sleep(Duration::from_secs(121)) => {},
}
}
#[tokio::test]
async fn notified_but_pollable_after() {
let inner = futures::stream::once(futures::future::ready(Ok(bytes::Bytes::from_static(
b"hello world",
))));
let timeout = Duration::from_secs(120);
let cancel = CancellationToken::new();
cancel.cancel();
let stream = DownloadStream::new(cancel_or_timeout(timeout, cancel.clone()), inner);
let mut stream = std::pin::pin!(stream);
let next = stream.next().await;
let ioe = next.unwrap().unwrap_err();
assert!(
matches!(
ioe.get_ref().unwrap().downcast_ref::<DownloadError>(),
Some(&DownloadError::Cancelled)
),
"{ioe:?}"
);
let next = stream.next().await;
let bytes = next.unwrap().unwrap();
assert_eq!(&b"hello world"[..], bytes);
}
}

View File

@@ -50,9 +50,6 @@ pub struct SkTimelineInfo {
pub safekeeper_connstr: Option<String>,
#[serde(default)]
pub http_connstr: Option<String>,
// Minimum of all active RO replicas flush LSN
#[serde(default = "lsn_invalid")]
pub standby_horizon: Lsn,
}
#[derive(Debug, Clone, Deserialize, Serialize)]

View File

@@ -7,7 +7,7 @@ license.workspace = true
[dependencies]
hyper.workspace = true
opentelemetry = { workspace = true, features=["rt-tokio"] }
opentelemetry-otlp = { workspace = true, default-features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
opentelemetry-otlp = { workspace = true, default_features=false, features = ["http-proto", "trace", "http", "reqwest-client"] }
opentelemetry-semantic-conventions.workspace = true
reqwest = { workspace = true, default-features = false, features = ["rustls-tls"] }
tokio = { workspace = true, features = ["rt", "rt-multi-thread"] }

View File

@@ -9,33 +9,6 @@ use serde::{Deserialize, Serialize};
use tokio_util::sync::CancellationToken;
use tracing::*;
/// Declare a failpoint that can use the `pause` failpoint action.
/// We don't want to block the executor thread, hence, spawn_blocking + await.
#[macro_export]
macro_rules! pausable_failpoint {
($name:literal) => {
if cfg!(feature = "testing") {
tokio::task::spawn_blocking({
let current = tracing::Span::current();
move || {
let _entered = current.entered();
tracing::info!("at failpoint {}", $name);
fail::fail_point!($name);
}
})
.await
.expect("spawn_blocking");
}
};
($name:literal, $cond:expr) => {
if cfg!(feature = "testing") {
if $cond {
pausable_failpoint!($name)
}
}
};
}
/// use with fail::cfg("$name", "return(2000)")
///
/// The effect is similar to a "sleep(2000)" action, i.e. we sleep for the

View File

@@ -3,9 +3,6 @@ use std::{fs, io, path::Path};
use anyhow::Context;
mod rename_noreplace;
pub use rename_noreplace::rename_noreplace;
pub trait PathExt {
/// Returns an error if `self` is not a directory.
fn is_empty_dir(&self) -> io::Result<bool>;

View File

@@ -1,109 +0,0 @@
use nix::NixPath;
/// Rename a file without replacing an existing file.
///
/// This is a wrapper around platform-specific APIs.
pub fn rename_noreplace<P1: ?Sized + NixPath, P2: ?Sized + NixPath>(
src: &P1,
dst: &P2,
) -> nix::Result<()> {
{
#[cfg(target_os = "linux")]
{
nix::fcntl::renameat2(
None,
src,
None,
dst,
nix::fcntl::RenameFlags::RENAME_NOREPLACE,
)
}
#[cfg(target_os = "macos")]
{
let res = src.with_nix_path(|src| {
dst.with_nix_path(|dst|
// SAFETY: `src` and `dst` are valid C strings as per the NixPath trait and they outlive the call to renamex_np.
unsafe {
nix::libc::renamex_np(src.as_ptr(), dst.as_ptr(), nix::libc::RENAME_EXCL)
})
})??;
nix::errno::Errno::result(res).map(drop)
}
#[cfg(not(any(target_os = "linux", target_os = "macos")))]
{
std::compile_error!("OS does not support no-replace renames");
}
}
}
#[cfg(test)]
mod test {
use std::{fs, path::PathBuf};
use super::*;
fn testdir() -> camino_tempfile::Utf8TempDir {
match crate::env::var("NEON_UTILS_RENAME_NOREPLACE_TESTDIR") {
Some(path) => {
let path: camino::Utf8PathBuf = path;
camino_tempfile::tempdir_in(path).unwrap()
}
None => camino_tempfile::tempdir().unwrap(),
}
}
#[test]
fn test_absolute_paths() {
let testdir = testdir();
println!("testdir: {}", testdir.path());
let src = testdir.path().join("src");
let dst = testdir.path().join("dst");
fs::write(&src, b"").unwrap();
fs::write(&dst, b"").unwrap();
let src = src.canonicalize().unwrap();
assert!(src.is_absolute());
let dst = dst.canonicalize().unwrap();
assert!(dst.is_absolute());
let result = rename_noreplace(&src, &dst);
assert_eq!(result.unwrap_err(), nix::Error::EEXIST);
}
#[test]
fn test_relative_paths() {
let testdir = testdir();
println!("testdir: {}", testdir.path());
// this is fine because we run in nextest => process per test
std::env::set_current_dir(testdir.path()).unwrap();
let src = PathBuf::from("src");
let dst = PathBuf::from("dst");
fs::write(&src, b"").unwrap();
fs::write(&dst, b"").unwrap();
let result = rename_noreplace(&src, &dst);
assert_eq!(result.unwrap_err(), nix::Error::EEXIST);
}
#[test]
fn test_works_when_not_exists() {
let testdir = testdir();
println!("testdir: {}", testdir.path());
let src = testdir.path().join("src");
let dst = testdir.path().join("dst");
fs::write(&src, b"content").unwrap();
rename_noreplace(src.as_std_path(), dst.as_std_path()).unwrap();
assert_eq!(
"content",
String::from_utf8(std::fs::read(&dst).unwrap()).unwrap()
);
}
}

View File

@@ -19,13 +19,13 @@
/// // right: [0x68; 1]
/// # fn serialize_something() -> Vec<u8> { "hello world".as_bytes().to_vec() }
/// ```
pub struct Hex<S>(pub S);
#[derive(PartialEq)]
pub struct Hex<'a>(pub &'a [u8]);
impl<S: AsRef<[u8]>> std::fmt::Debug for Hex<S> {
impl std::fmt::Debug for Hex<'_> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(f, "[")?;
let chunks = self.0.as_ref().chunks(16);
for (i, c) in chunks.enumerate() {
for (i, c) in self.0.chunks(16).enumerate() {
if i > 0 && !c.is_empty() {
writeln!(f, ", ")?;
}
@@ -36,15 +36,6 @@ impl<S: AsRef<[u8]>> std::fmt::Debug for Hex<S> {
write!(f, "0x{b:02x}")?;
}
}
write!(f, "; {}]", self.0.as_ref().len())
}
}
impl<R: AsRef<[u8]>, L: AsRef<[u8]>> PartialEq<Hex<R>> for Hex<L> {
fn eq(&self, other: &Hex<R>) -> bool {
let left = self.0.as_ref();
let right = other.0.as_ref();
left == right
write!(f, "; {}]", self.0.len())
}
}

View File

@@ -34,9 +34,6 @@ pub enum ApiError {
#[error("Timeout")]
Timeout(Cow<'static, str>),
#[error("Request cancelled")]
Cancelled,
#[error(transparent)]
InternalServerError(anyhow::Error),
}
@@ -77,10 +74,6 @@ impl ApiError {
err.to_string(),
StatusCode::REQUEST_TIMEOUT,
),
ApiError::Cancelled => HttpErrorBody::response_from_msg_and_status(
self.to_string(),
StatusCode::INTERNAL_SERVER_ERROR,
),
ApiError::InternalServerError(err) => HttpErrorBody::response_from_msg_and_status(
err.to_string(),
StatusCode::INTERNAL_SERVER_ERROR,
@@ -140,7 +133,6 @@ pub fn api_error_handler(api_error: ApiError) -> Response<Body> {
ApiError::InternalServerError(_) => error!("Error processing HTTP request: {api_error:?}"),
ApiError::ShuttingDown => info!("Shut down while processing HTTP request"),
ApiError::Timeout(_) => info!("Timeout while processing HTTP request: {api_error:#}"),
ApiError::Cancelled => info!("Request cancelled while processing HTTP request"),
_ => info!("Error processing HTTP request: {api_error:#}"),
}

View File

@@ -135,8 +135,7 @@ impl Gate {
let started_at = std::time::Instant::now();
let mut do_close = std::pin::pin!(self.do_close());
// with 1s we rarely saw anything, let's try if we get more gate closing reasons with 100ms
let nag_after = Duration::from_millis(100);
let nag_after = Duration::from_secs(1);
let Err(_timeout) = tokio::time::timeout(nag_after, &mut do_close).await else {
return;

View File

@@ -496,9 +496,9 @@ mod tests {
// TODO: When updating Postgres versions, this test will cause
// problems. Postgres version in message needs updating.
//
// Greeting(ProposerGreeting { protocol_version: 2, pg_version: 160003, proposer_id: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], system_id: 0, timeline_id: 9e4c8f36063c6c6e93bc20d65a820f3d, tenant_id: 9e4c8f36063c6c6e93bc20d65a820f3d, tli: 1, wal_seg_size: 16777216 })
// Greeting(ProposerGreeting { protocol_version: 2, pg_version: 160002, proposer_id: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], system_id: 0, timeline_id: 9e4c8f36063c6c6e93bc20d65a820f3d, tenant_id: 9e4c8f36063c6c6e93bc20d65a820f3d, tli: 1, wal_seg_size: 16777216 })
vec![
103, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 3, 113, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
103, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 2, 113, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 158, 76, 143, 54, 6, 60, 108, 110,
147, 188, 32, 214, 90, 130, 15, 61, 158, 76, 143, 54, 6, 60, 108, 110, 147,
188, 32, 214, 90, 130, 15, 61, 1, 0, 0, 0, 0, 0, 0, 1,

View File

@@ -380,8 +380,8 @@ impl interface::CompactionLayer<Key> for MockLayer {
}
fn file_size(&self) -> u64 {
match self {
MockLayer::Delta(this) => this.file_size,
MockLayer::Image(this) => this.file_size,
MockLayer::Delta(this) => this.file_size(),
MockLayer::Image(this) => this.file_size(),
}
}
fn short_id(&self) -> String {

View File

@@ -17,7 +17,6 @@ pageserver = { path = ".." }
pageserver_api.workspace = true
remote_storage = { path = "../../libs/remote_storage" }
postgres_ffi.workspace = true
thiserror.workspace = true
tokio.workspace = true
tokio-util.workspace = true
toml_edit.workspace = true

View File

@@ -1,6 +1,11 @@
use std::collections::HashMap;
use anyhow::Context;
use camino::Utf8PathBuf;
use pageserver::tenant::IndexPart;
use pageserver::tenant::remote_timeline_client::index::IndexLayerMetadata;
use pageserver::tenant::storage_layer::LayerName;
use pageserver::tenant::{metadata::TimelineMetadata, IndexPart};
use utils::lsn::Lsn;
#[derive(clap::Subcommand)]
pub(crate) enum IndexPartCmd {
@@ -12,7 +17,20 @@ pub(crate) async fn main(cmd: &IndexPartCmd) -> anyhow::Result<()> {
IndexPartCmd::Dump { path } => {
let bytes = tokio::fs::read(path).await.context("read file")?;
let des: IndexPart = IndexPart::from_s3_bytes(&bytes).context("deserialize")?;
let output = serde_json::to_string_pretty(&des).context("serialize output")?;
#[derive(serde::Serialize)]
struct Output<'a> {
layer_metadata: &'a HashMap<LayerName, IndexLayerMetadata>,
disk_consistent_lsn: Lsn,
timeline_metadata: &'a TimelineMetadata,
}
let output = Output {
layer_metadata: &des.layer_metadata,
disk_consistent_lsn: des.get_disk_consistent_lsn(),
timeline_metadata: &des.metadata,
};
let output = serde_json::to_string_pretty(&output).context("serialize output")?;
println!("{output}");
Ok(())
}

View File

@@ -1,475 +0,0 @@
use anyhow::Context;
use clap::Parser;
use pageserver_api::{
key::Key,
reltag::{BlockNumber, RelTag, SlruKind},
shard::{ShardCount, ShardStripeSize},
};
use std::str::FromStr;
#[derive(Parser)]
pub(super) struct DescribeKeyCommand {
/// Key material in one of the forms: hex, span attributes captured from log, reltag blocknum
input: Vec<String>,
/// The number of shards to calculate what Keys placement would be.
#[arg(long)]
shard_count: Option<CustomShardCount>,
/// The sharding stripe size.
///
/// The default is hardcoded. It makes no sense to provide this without providing
/// `--shard-count`.
#[arg(long, requires = "shard_count")]
stripe_size: Option<u32>,
}
/// Sharded shard count without unsharded count, which the actual ShardCount supports.
#[derive(Clone, Copy)]
pub(super) struct CustomShardCount(std::num::NonZeroU8);
#[derive(Debug, thiserror::Error)]
pub(super) enum InvalidShardCount {
#[error(transparent)]
ParsingFailed(#[from] std::num::ParseIntError),
#[error("too few shards")]
TooFewShards,
}
impl FromStr for CustomShardCount {
type Err = InvalidShardCount;
fn from_str(s: &str) -> Result<Self, Self::Err> {
let inner: std::num::NonZeroU8 = s.parse()?;
if inner.get() < 2 {
Err(InvalidShardCount::TooFewShards)
} else {
Ok(CustomShardCount(inner))
}
}
}
impl From<CustomShardCount> for ShardCount {
fn from(value: CustomShardCount) -> Self {
ShardCount::new(value.0.get())
}
}
impl DescribeKeyCommand {
pub(super) fn execute(self) {
let DescribeKeyCommand {
input,
shard_count,
stripe_size,
} = self;
let material = KeyMaterial::try_from(input.as_slice()).unwrap();
let kind = material.kind();
let key = Key::from(material);
println!("parsed from {kind}: {key}:");
println!();
println!("{key:?}");
macro_rules! kind_query {
([$($name:ident),*$(,)?]) => {{[$(kind_query!($name)),*]}};
($name:ident) => {{
let s: &'static str = stringify!($name);
let s = s.strip_prefix("is_").unwrap_or(s);
let s = s.strip_suffix("_key").unwrap_or(s);
#[allow(clippy::needless_borrow)]
(s, key.$name())
}};
}
// the current characterization is a mess of these boolean queries and separate
// "recognization". I think it accurately represents how strictly we model the Key
// right now, but could of course be made less confusing.
let queries = kind_query!([
is_rel_block_key,
is_rel_vm_block_key,
is_rel_fsm_block_key,
is_slru_block_key,
is_inherited_key,
is_rel_size_key,
is_slru_segment_size_key,
]);
let recognized_kind = "recognized kind";
let metadata_key = "metadata key";
let shard_placement = "shard placement";
let longest = queries
.iter()
.map(|t| t.0)
.chain([recognized_kind, metadata_key, shard_placement])
.map(|s| s.len())
.max()
.unwrap();
let colon = 1;
let padding = 1;
for (name, is) in queries {
let width = longest - name.len() + colon + padding;
println!("{}{:width$}{}", name, ":", is);
}
let width = longest - recognized_kind.len() + colon + padding;
println!(
"{}{:width$}{:?}",
recognized_kind,
":",
RecognizedKeyKind::new(key),
);
if let Some(shard_count) = shard_count {
// seeing the sharding placement might be confusing, so leave it out unless shard
// count was given.
let stripe_size = stripe_size.map(ShardStripeSize).unwrap_or_default();
println!(
"# placement with shard_count: {} and stripe_size: {}:",
shard_count.0, stripe_size.0
);
let width = longest - shard_placement.len() + colon + padding;
println!(
"{}{:width$}{:?}",
shard_placement,
":",
pageserver_api::shard::describe(&key, shard_count.into(), stripe_size)
);
}
}
}
/// Hand-wavy "inputs we accept" for a key.
#[derive(Debug)]
pub(super) enum KeyMaterial {
Hex(Key),
String(SpanAttributesFromLogs),
Split(RelTag, BlockNumber),
}
impl KeyMaterial {
fn kind(&self) -> &'static str {
match self {
KeyMaterial::Hex(_) => "hex",
KeyMaterial::String(_) | KeyMaterial::Split(_, _) => "split",
}
}
}
impl From<KeyMaterial> for Key {
fn from(value: KeyMaterial) -> Self {
match value {
KeyMaterial::Hex(key) => key,
KeyMaterial::String(SpanAttributesFromLogs(rt, blocknum))
| KeyMaterial::Split(rt, blocknum) => {
pageserver_api::key::rel_block_to_key(rt, blocknum)
}
}
}
}
impl<S: AsRef<str>> TryFrom<&[S]> for KeyMaterial {
type Error = anyhow::Error;
fn try_from(value: &[S]) -> Result<Self, Self::Error> {
match value {
[] => anyhow::bail!(
"need 1..N positional arguments describing the key, try hex or a log line"
),
[one] => {
let one = one.as_ref();
let key = Key::from_hex(one).map(KeyMaterial::Hex);
let attrs = SpanAttributesFromLogs::from_str(one).map(KeyMaterial::String);
match (key, attrs) {
(Ok(key), _) => Ok(key),
(_, Ok(s)) => Ok(s),
(Err(e1), Err(e2)) => anyhow::bail!(
"failed to parse {one:?} as hex or span attributes:\n- {e1:#}\n- {e2:#}"
),
}
}
more => {
// assume going left to right one of these is a reltag and then we find a blocknum
// this works, because we don't have plain numbers at least right after reltag in
// logs. for some definition of "works".
let Some((reltag_at, reltag)) = more
.iter()
.map(AsRef::as_ref)
.enumerate()
.find_map(|(i, s)| {
s.split_once("rel=")
.map(|(_garbage, actual)| actual)
.unwrap_or(s)
.parse::<RelTag>()
.ok()
.map(|rt| (i, rt))
})
else {
anyhow::bail!("found no RelTag in arguments");
};
let Some(blocknum) = more
.iter()
.map(AsRef::as_ref)
.skip(reltag_at)
.find_map(|s| {
s.split_once("blkno=")
.map(|(_garbage, actual)| actual)
.unwrap_or(s)
.parse::<BlockNumber>()
.ok()
})
else {
anyhow::bail!("found no blocknum in arguments");
};
Ok(KeyMaterial::Split(reltag, blocknum))
}
}
}
}
#[derive(Debug)]
pub(super) struct SpanAttributesFromLogs(RelTag, BlockNumber);
impl std::str::FromStr for SpanAttributesFromLogs {
type Err = anyhow::Error;
fn from_str(s: &str) -> Result<Self, Self::Err> {
// accept the span separator but do not require or fail if either is missing
// "whatever{rel=1663/16389/24615 blkno=1052204 req_lsn=FFFFFFFF/FFFFFFFF}"
let (_, reltag) = s
.split_once("rel=")
.ok_or_else(|| anyhow::anyhow!("cannot find 'rel='"))?;
let reltag = reltag.split_whitespace().next().unwrap();
let (_, blocknum) = s
.split_once("blkno=")
.ok_or_else(|| anyhow::anyhow!("cannot find 'blkno='"))?;
let blocknum = blocknum.split_whitespace().next().unwrap();
let reltag = reltag
.parse()
.with_context(|| format!("parse reltag from {reltag:?}"))?;
let blocknum = blocknum
.parse()
.with_context(|| format!("parse blocknum from {blocknum:?}"))?;
Ok(Self(reltag, blocknum))
}
}
#[derive(Debug)]
#[allow(dead_code)] // debug print is used
enum RecognizedKeyKind {
DbDir,
ControlFile,
Checkpoint,
AuxFilesV1,
SlruDir(Result<SlruKind, u32>),
RelMap(RelTagish<2>),
RelDir(RelTagish<2>),
AuxFileV2(Result<AuxFileV2, utils::Hex<[u8; 16]>>),
}
#[derive(Debug, PartialEq)]
#[allow(unused)]
enum AuxFileV2 {
Recognized(&'static str, utils::Hex<[u8; 13]>),
OtherWithPrefix(&'static str, utils::Hex<[u8; 13]>),
Other(utils::Hex<[u8; 13]>),
}
impl RecognizedKeyKind {
fn new(key: Key) -> Option<Self> {
use RecognizedKeyKind::{
AuxFilesV1, Checkpoint, ControlFile, DbDir, RelDir, RelMap, SlruDir,
};
let slru_dir_kind = pageserver_api::key::slru_dir_kind(&key);
Some(match key {
pageserver_api::key::DBDIR_KEY => DbDir,
pageserver_api::key::CONTROLFILE_KEY => ControlFile,
pageserver_api::key::CHECKPOINT_KEY => Checkpoint,
pageserver_api::key::AUX_FILES_KEY => AuxFilesV1,
_ if slru_dir_kind.is_some() => SlruDir(slru_dir_kind.unwrap()),
_ if key.field1 == 0 && key.field4 == 0 && key.field5 == 0 && key.field6 == 0 => {
RelMap([key.field2, key.field3].into())
}
_ if key.field1 == 0 && key.field4 == 0 && key.field5 == 0 && key.field6 == 1 => {
RelDir([key.field2, key.field3].into())
}
_ if key.is_metadata_key() => RecognizedKeyKind::AuxFileV2(
AuxFileV2::new(key).ok_or_else(|| utils::Hex(key.to_i128().to_be_bytes())),
),
_ => return None,
})
}
}
impl AuxFileV2 {
fn new(key: Key) -> Option<AuxFileV2> {
const EMPTY_HASH: [u8; 13] = {
let mut out = [0u8; 13];
let hash = pageserver::aux_file::fnv_hash(b"").to_be_bytes();
let mut i = 3;
while i < 16 {
out[i - 3] = hash[i];
i += 1;
}
out
};
let bytes = key.to_i128().to_be_bytes();
let hash = utils::Hex(<[u8; 13]>::try_from(&bytes[3..]).unwrap());
assert_eq!(EMPTY_HASH.len(), hash.0.len());
// TODO: we could probably find the preimages for the hashes
Some(match (bytes[1], bytes[2]) {
(1, 1) => AuxFileV2::Recognized("pg_logical/mappings/", hash),
(1, 2) => AuxFileV2::Recognized("pg_logical/snapshots/", hash),
(1, 3) if hash.0 == EMPTY_HASH => {
AuxFileV2::Recognized("pg_logical/replorigin_checkpoint", hash)
}
(2, 1) => AuxFileV2::Recognized("pg_replslot/", hash),
(1, 0xff) => AuxFileV2::OtherWithPrefix("pg_logical/", hash),
(0xff, 0xff) => AuxFileV2::Other(hash),
_ => return None,
})
}
}
/// Prefix of RelTag, currently only known use cases are the two item versions.
///
/// Renders like a reltag with `/`, nothing else.
struct RelTagish<const N: usize>([u32; N]);
impl<const N: usize> From<[u32; N]> for RelTagish<N> {
fn from(val: [u32; N]) -> Self {
RelTagish(val)
}
}
impl<const N: usize> std::fmt::Debug for RelTagish<N> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
use std::fmt::Write as _;
let mut first = true;
self.0.iter().try_for_each(|x| {
if !first {
f.write_char('/')?;
}
first = false;
write!(f, "{}", x)
})
}
}
#[cfg(test)]
mod tests {
use pageserver::aux_file::encode_aux_file_key;
use super::*;
#[test]
fn hex_is_key_material() {
let m = KeyMaterial::try_from(&["000000067F0000400200DF927900FFFFFFFF"][..]).unwrap();
assert!(matches!(m, KeyMaterial::Hex(_)), "{m:?}");
}
#[test]
fn single_positional_spanalike_is_key_material() {
// why is this needed? if you are checking many, then copypaste starts to appeal
let strings = [
(line!(), "2024-05-15T15:33:49.873906Z ERROR page_service_conn_main{peer_addr=A:B}:process_query{tenant_id=C timeline_id=D}:handle_pagerequests:handle_get_page_at_lsn_request{rel=1663/208101/2620_fsm blkno=2 req_lsn=0/238D98C8}: error reading relation or page version: Read error: could not find data for key 000000067F00032CE5000000000000000001 (shard ShardNumber(0)) at LSN 0/1D0A16C1, request LSN 0/238D98C8, ancestor 0/0"),
(line!(), "rel=1663/208101/2620_fsm blkno=2"),
(line!(), "rel=1663/208101/2620.1 blkno=2"),
];
let mut first: Option<Key> = None;
for (line, example) in strings {
let m = KeyMaterial::try_from(&[example][..])
.unwrap_or_else(|e| panic!("failed to parse example from line {line}: {e:?}"));
let key = Key::from(m);
if let Some(first) = first {
assert_eq!(first, key);
} else {
first = Some(key);
}
}
// not supporting this is rather accidential, but I think the input parsing is lenient
// enough already
KeyMaterial::try_from(&["1663/208101/2620_fsm 2"][..]).unwrap_err();
}
#[test]
fn multiple_spanlike_args() {
let strings = [
(line!(), &["process_query{tenant_id=C", "timeline_id=D}:handle_pagerequests:handle_get_page_at_lsn_request{rel=1663/208101/2620_fsm", "blkno=2", "req_lsn=0/238D98C8}"][..]),
(line!(), &["rel=1663/208101/2620_fsm", "blkno=2"][..]),
(line!(), &["1663/208101/2620_fsm", "2"][..]),
];
let mut first: Option<Key> = None;
for (line, example) in strings {
let m = KeyMaterial::try_from(example)
.unwrap_or_else(|e| panic!("failed to parse example from line {line}: {e:?}"));
let key = Key::from(m);
if let Some(first) = first {
assert_eq!(first, key);
} else {
first = Some(key);
}
}
}
#[test]
fn recognized_auxfiles() {
use AuxFileV2::*;
let empty = [
0x2e, 0x07, 0xbb, 0x01, 0x42, 0x62, 0xb8, 0x21, 0x75, 0x62, 0x95, 0xc5, 0x8d,
];
let foobar = [
0x62, 0x79, 0x3c, 0x64, 0xbf, 0x6f, 0x0d, 0x35, 0x97, 0xba, 0x44, 0x6f, 0x18,
];
#[rustfmt::skip]
let examples = [
(line!(), "pg_logical/mappings/foobar", Recognized("pg_logical/mappings/", utils::Hex(foobar))),
(line!(), "pg_logical/snapshots/foobar", Recognized("pg_logical/snapshots/", utils::Hex(foobar))),
(line!(), "pg_logical/replorigin_checkpoint", Recognized("pg_logical/replorigin_checkpoint", utils::Hex(empty))),
(line!(), "pg_logical/foobar", OtherWithPrefix("pg_logical/", utils::Hex(foobar))),
(line!(), "pg_replslot/foobar", Recognized("pg_replslot/", utils::Hex(foobar))),
(line!(), "foobar", Other(utils::Hex(foobar))),
];
for (line, path, expected) in examples {
let key = encode_aux_file_key(path);
let recognized =
AuxFileV2::new(key).unwrap_or_else(|| panic!("line {line} example failed"));
assert_eq!(recognized, expected);
}
assert_eq!(
AuxFileV2::new(Key::from_hex("600000102000000000000000000000000000").unwrap()),
None,
"example key has one too few 0 after 6 before 1"
);
}
}

View File

@@ -6,7 +6,6 @@
mod draw_timeline_dir;
mod index_part;
mod key;
mod layer_map_analyzer;
mod layers;
@@ -62,8 +61,6 @@ enum Commands {
AnalyzeLayerMap(AnalyzeLayerMapCmd),
#[command(subcommand)]
Layer(LayerCmd),
/// Debug print a hex key found from logs
Key(key::DescribeKeyCommand),
}
/// Read and update pageserver metadata file
@@ -186,7 +183,6 @@ async fn main() -> anyhow::Result<()> {
.time_travel_recover(Some(&prefix), timestamp, done_if_after, &cancel)
.await?;
}
Commands::Key(dkc) => dkc.execute(),
};
Ok(())
}

View File

@@ -5,7 +5,6 @@ use utils::lsn::Lsn;
use std::collections::HashMap;
use std::sync::Arc;
use std::time::Instant;
/// Ingest aux files into the pageserver.
#[derive(clap::Parser)]
@@ -89,17 +88,11 @@ async fn main_impl(args: Args) -> anyhow::Result<()> {
println!("ingested {file_cnt} files");
}
for _ in 0..100 {
let start = Instant::now();
let files = mgmt_api_client
.list_aux_files(tenant_shard_id, timeline_id, Lsn(Lsn::MAX.0 - 1))
.await?;
println!(
"{} files found in {}s",
files.len(),
start.elapsed().as_secs_f64()
);
}
let files = mgmt_api_client
.list_aux_files(tenant_shard_id, timeline_id, Lsn(Lsn::MAX.0 - 1))
.await?;
println!("{} files found", files.len());
anyhow::Ok(())
}

View File

@@ -1,6 +1,6 @@
use anyhow::Context;
use camino::Utf8PathBuf;
use pageserver_api::key::Key;
use pageserver_api::key::{is_rel_block_key, key_to_rel_block, Key};
use pageserver_api::keyspace::KeySpaceAccum;
use pageserver_api::models::PagestreamGetPageRequest;
@@ -187,7 +187,7 @@ async fn main_impl(
for r in partitioning.keys.ranges.iter() {
let mut i = r.start;
while i != r.end {
if i.is_rel_block_key() {
if is_rel_block_key(&i) {
filtered.add_key(i);
}
i = i.next();
@@ -308,10 +308,9 @@ async fn main_impl(
let r = &ranges[weights.sample(&mut rng)];
let key: i128 = rng.gen_range(r.start..r.end);
let key = Key::from_i128(key);
assert!(key.is_rel_block_key());
let (rel_tag, block_no) = key
.to_rel_block()
.expect("we filter non-rel-block keys out above");
assert!(is_rel_block_key(&key));
let (rel_tag, block_no) =
key_to_rel_block(key).expect("we filter non-rel-block keys out above");
PagestreamGetPageRequest {
request_lsn: if rng.gen_bool(args.req_latest_probability) {
Lsn::MAX

View File

@@ -178,8 +178,7 @@ impl AuxFileSizeEstimator {
}
}
/// When generating base backup or doing initial logical size calculation
pub fn on_initial(&self, new_size: usize) {
pub fn on_base_backup(&self, new_size: usize) {
let mut guard = self.size.lock().unwrap();
*guard = Some(new_size as isize);
self.report(new_size as isize);

View File

@@ -13,7 +13,7 @@
use anyhow::{anyhow, Context};
use bytes::{BufMut, Bytes, BytesMut};
use fail::fail_point;
use pageserver_api::key::Key;
use pageserver_api::key::{key_to_slru_block, Key};
use postgres_ffi::pg_constants;
use std::fmt::Write as FmtWrite;
use std::time::SystemTime;
@@ -170,7 +170,7 @@ where
}
async fn add_block(&mut self, key: &Key, block: Bytes) -> Result<(), BasebackupError> {
let (kind, segno, _) = key.to_slru_block()?;
let (kind, segno, _) = key_to_slru_block(*key)?;
match kind {
SlruKind::Clog => {
@@ -362,13 +362,6 @@ where
));
info!("Replication slot {} restart LSN={}", path, restart_lsn);
min_restart_lsn = Lsn::min(min_restart_lsn, restart_lsn);
} else if path == "pg_logical/replorigin_checkpoint" {
// replorigin_checkoint is written only on compute shutdown, so it contains
// deteriorated values. So we generate our own version of this file for the particular LSN
// based on information about replorigins extracted from transaction commit records.
// In future we will not generate AUX record for "pg_logical/replorigin_checkpoint" at all,
// but now we should handle (skip) it for backward compatibility.
continue;
}
let header = new_tar_header(&path, content.len() as u64)?;
self.ar
@@ -397,32 +390,6 @@ where
{
self.add_twophase_file(xid).await?;
}
let repl_origins = self
.timeline
.get_replorigins(self.lsn, self.ctx)
.await
.map_err(|e| BasebackupError::Server(e.into()))?;
let n_origins = repl_origins.len();
if n_origins != 0 {
//
// Construct "pg_logical/replorigin_checkpoint" file based on information about replication origins
// extracted from transaction commit record. We are using this file to pass information about replication
// origins to compute to allow logical replication to restart from proper point.
//
let mut content = Vec::with_capacity(n_origins * 16 + 8);
content.extend_from_slice(&pg_constants::REPLICATION_STATE_MAGIC.to_le_bytes());
for (origin_id, origin_lsn) in repl_origins {
content.extend_from_slice(&origin_id.to_le_bytes());
content.extend_from_slice(&[0u8; 6]); // align to 8 bytes
content.extend_from_slice(&origin_lsn.0.to_le_bytes());
}
let crc32 = crc32c::crc32c(&content);
content.extend_from_slice(&crc32.to_le_bytes());
let header = new_tar_header("pg_logical/replorigin_checkpoint", content.len() as u64)?;
self.ar.append(&header, &*content).await.context(
"could not add pg_logical/replorigin_checkpoint file to basebackup tarball",
)?;
}
fail_point!("basebackup-before-control-file", |_| {
Err(BasebackupError::Server(anyhow!(

View File

@@ -99,6 +99,8 @@ pub mod defaults {
pub const DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB: usize = 0;
pub const DEFAULT_WALREDO_PROCESS_KIND: &str = "async";
///
/// Default built-in configuration file.
///
@@ -144,6 +146,8 @@ pub mod defaults {
#validate_vectored_get = '{DEFAULT_VALIDATE_VECTORED_GET}'
#walredo_process_kind = '{DEFAULT_WALREDO_PROCESS_KIND}'
[tenant_config]
#checkpoint_distance = {DEFAULT_CHECKPOINT_DISTANCE} # in bytes
#checkpoint_timeout = {DEFAULT_CHECKPOINT_TIMEOUT}
@@ -296,6 +300,8 @@ pub struct PageServerConf {
///
/// Setting this to zero disables limits on total ephemeral layer size.
pub ephemeral_bytes_per_memory_kb: usize,
pub walredo_process_kind: crate::walredo::ProcessKind,
}
/// We do not want to store this in a PageServerConf because the latter may be logged
@@ -401,6 +407,8 @@ struct PageServerConfigBuilder {
validate_vectored_get: BuilderValue<bool>,
ephemeral_bytes_per_memory_kb: BuilderValue<usize>,
walredo_process_kind: BuilderValue<crate::walredo::ProcessKind>,
}
impl PageServerConfigBuilder {
@@ -489,6 +497,8 @@ impl PageServerConfigBuilder {
)),
validate_vectored_get: Set(DEFAULT_VALIDATE_VECTORED_GET),
ephemeral_bytes_per_memory_kb: Set(DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB),
walredo_process_kind: Set(DEFAULT_WALREDO_PROCESS_KIND.parse().unwrap()),
}
}
}
@@ -676,6 +686,10 @@ impl PageServerConfigBuilder {
self.ephemeral_bytes_per_memory_kb = BuilderValue::Set(value);
}
pub fn get_walredo_process_kind(&mut self, value: crate::walredo::ProcessKind) {
self.walredo_process_kind = BuilderValue::Set(value);
}
pub fn build(self) -> anyhow::Result<PageServerConf> {
let default = Self::default_values();
@@ -733,6 +747,7 @@ impl PageServerConfigBuilder {
max_vectored_read_bytes,
validate_vectored_get,
ephemeral_bytes_per_memory_kb,
walredo_process_kind,
}
CUSTOM LOGIC
{
@@ -1029,6 +1044,9 @@ impl PageServerConf {
"ephemeral_bytes_per_memory_kb" => {
builder.get_ephemeral_bytes_per_memory_kb(parse_toml_u64("ephemeral_bytes_per_memory_kb", item)? as usize)
}
"walredo_process_kind" => {
builder.get_walredo_process_kind(parse_toml_from_str("walredo_process_kind", item)?)
}
_ => bail!("unrecognized pageserver option '{key}'"),
}
}
@@ -1112,6 +1130,7 @@ impl PageServerConf {
),
validate_vectored_get: defaults::DEFAULT_VALIDATE_VECTORED_GET,
ephemeral_bytes_per_memory_kb: defaults::DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB,
walredo_process_kind: defaults::DEFAULT_WALREDO_PROCESS_KIND.parse().unwrap(),
}
}
}
@@ -1351,6 +1370,7 @@ background_task_maximum_delay = '334 s'
),
validate_vectored_get: defaults::DEFAULT_VALIDATE_VECTORED_GET,
ephemeral_bytes_per_memory_kb: defaults::DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB,
walredo_process_kind: defaults::DEFAULT_WALREDO_PROCESS_KIND.parse().unwrap(),
},
"Correct defaults should be used when no config values are provided"
);
@@ -1424,6 +1444,7 @@ background_task_maximum_delay = '334 s'
),
validate_vectored_get: defaults::DEFAULT_VALIDATE_VECTORED_GET,
ephemeral_bytes_per_memory_kb: defaults::DEFAULT_EPHEMERAL_BYTES_PER_MEMORY_KB,
walredo_process_kind: defaults::DEFAULT_WALREDO_PROCESS_KIND.parse().unwrap(),
},
"Should be able to parse all basic config values correctly"
);

View File

@@ -2,9 +2,10 @@
//! and push them to a HTTP endpoint.
use crate::context::{DownloadBehavior, RequestContext};
use crate::task_mgr::{self, TaskKind, BACKGROUND_RUNTIME};
use crate::tenant::size::CalculateSyntheticSizeError;
use crate::tenant::tasks::BackgroundLoopKind;
use crate::tenant::{mgr::TenantManager, LogicalSizeCalculationCause, Tenant};
use crate::tenant::{
mgr::TenantManager, LogicalSizeCalculationCause, PageReconstructError, Tenant,
};
use camino::Utf8PathBuf;
use consumption_metrics::EventType;
use pageserver_api::models::TenantState;
@@ -349,12 +350,19 @@ async fn calculate_and_log(tenant: &Tenant, cancel: &CancellationToken, ctx: &Re
// Same for the loop that fetches computed metrics.
// By using the same limiter, we centralize metrics collection for "start" and "finished" counters,
// which turns out is really handy to understand the system.
match tenant.calculate_synthetic_size(CAUSE, cancel, ctx).await {
Ok(_) => {}
Err(CalculateSyntheticSizeError::Cancelled) => {}
Err(e) => {
let tenant_shard_id = tenant.tenant_shard_id();
error!("failed to calculate synthetic size for tenant {tenant_shard_id}: {e:#}");
}
let Err(e) = tenant.calculate_synthetic_size(CAUSE, cancel, ctx).await else {
return;
};
// this error can be returned if timeline is shutting down, but it does not
// mean the synthetic size worker should terminate.
let shutting_down = matches!(
e.downcast_ref::<PageReconstructError>(),
Some(PageReconstructError::Cancelled | PageReconstructError::AncestorStopping(_))
);
if !shutting_down {
let tenant_shard_id = tenant.tenant_shard_id();
error!("failed to calculate synthetic size for tenant {tenant_shard_id}: {e:#}");
}
}

View File

@@ -311,7 +311,7 @@ impl DeletionList {
result.extend(
timeline_layers
.into_iter()
.map(|l| timeline_remote_path.join(Utf8PathBuf::from(l))),
.map(|l| timeline_remote_path.join(&Utf8PathBuf::from(l))),
);
}
}

View File

@@ -534,7 +534,7 @@ pub(crate) async fn disk_usage_eviction_task_iteration_impl<U: Usage>(
});
}
EvictionLayer::Secondary(layer) => {
let file_size = layer.metadata.file_size;
let file_size = layer.metadata.file_size();
js.spawn(async move {
layer
@@ -641,7 +641,7 @@ impl EvictionLayer {
pub(crate) fn get_file_size(&self) -> u64 {
match self {
Self::Attached(l) => l.layer_desc().file_size,
Self::Secondary(sl) => sl.metadata.file_size,
Self::Secondary(sl) => sl.metadata.file_size(),
}
}
}

View File

@@ -81,10 +81,8 @@ paths:
Attempts to delete specified tenant. 500, 503 and 409 errors should be retried until 404 is retrieved.
404 means that deletion successfully finished"
responses:
"200":
description: Tenant was successfully deleted, or was already not found.
"404":
description: Tenant not found. This is a success result, equivalent to 200.
description: Tenant not found. This is the success path.
content:
application/json:
schema:
@@ -259,37 +257,6 @@ paths:
schema:
$ref: "#/components/schemas/LsnByTimestampResponse"
/v1/tenant/{tenant_shard_id}/timeline/{timeline_id}/lsn_lease:
parameters:
- name: tenant_shard_id
in: path
required: true
schema:
type: string
- name: timeline_id
in: path
required: true
schema:
type: string
format: hex
post:
description: Obtain lease for the given LSN
parameters:
- name: lsn
in: query
required: true
schema:
type: string
format: hex
description: A LSN to obtain the lease for
responses:
"200":
description: OK
content:
application/json:
schema:
$ref: "#/components/schemas/LsnLease"
/v1/tenant/{tenant_id}/timeline/{timeline_id}/do_gc:
parameters:
- name: tenant_id
@@ -614,80 +581,6 @@ paths:
schema:
$ref: "#/components/schemas/Error"
/v1/tenant/{tenant_shard_id}/timeline/{timeline_id}/detach_ancestor:
parameters:
- name: tenant_shard_id
in: path
required: true
schema:
type: string
- name: timeline_id
in: path
ŕequired: true
schema:
type: string
put:
description: |
Detach a timeline from its ancestor and reparent all ancestors timelines with lower `ancestor_lsn`.
Current implementation might not be retryable across failure cases, but will be enhanced in future.
Detaching should be expected to be expensive operation. Timeouts should be retried.
responses:
"200":
description: |
The timeline has been detached from it's ancestor (now or earlier), and at least the returned timelines have been reparented.
If any timelines were deleted after reparenting, they might not be on this list.
content:
application/json:
schema:
$ref: "#/components/schemas/AncestorDetached"
"400":
description: |
Number of early checks meaning the timeline cannot be detached now:
- the ancestor of timeline has an ancestor: not supported, see RFC
content:
application/json:
schema:
$ref: "#/components/schemas/Error"
"404":
description: Tenant or timeline not found.
content:
application/json:
schema:
$ref: "#/components/schemas/NotFoundError"
"409":
description: |
The timeline can never be detached:
- timeline has no ancestor, implying that the timeline has never had an ancestor
content:
application/json:
schema:
$ref: "#/components/schemas/ConflictError"
"500":
description: |
Transient error, for example, pageserver shutdown happened while
processing the request but we were unable to distinguish that. Must
be retried.
content:
application/json:
schema:
$ref: "#/components/schemas/Error"
"503":
description: |
Temporarily unavailable, please retry. Possible reasons:
- another timeline detach for the same tenant is underway, please retry later
- detected shutdown error
content:
application/json:
schema:
$ref: "#/components/schemas/ServiceUnavailableError"
/v1/tenant/:
get:
description: Get tenants list
@@ -1087,15 +980,6 @@ components:
type: string
enum: [past, present, future, nodata]
LsnLease:
type: object
required:
- valid_until
properties:
valid_until:
type: string
format: date-time
PageserverUtilization:
type: object
required:
@@ -1153,19 +1037,6 @@ components:
format: int64
description: How many bytes of layer content were in the latest layer heatmap
AncestorDetached:
type: object
required:
- reparented_timelines
properties:
reparented_timelines:
type: array
description: Set of reparented timeline ids
properties:
type: string
format: hex
description: TimelineId
Error:
type: object

View File

@@ -16,7 +16,6 @@ use hyper::header;
use hyper::StatusCode;
use hyper::{Body, Request, Response, Uri};
use metrics::launch_timestamp::LaunchTimestamp;
use pageserver_api::models::AuxFilePolicy;
use pageserver_api::models::IngestAuxFilesRequest;
use pageserver_api::models::ListAuxFilesRequest;
use pageserver_api::models::LocationConfig;
@@ -74,9 +73,7 @@ use crate::tenant::size::ModelInputs;
use crate::tenant::storage_layer::LayerAccessStatsReset;
use crate::tenant::storage_layer::LayerName;
use crate::tenant::timeline::CompactFlags;
use crate::tenant::timeline::CompactionError;
use crate::tenant::timeline::Timeline;
use crate::tenant::GetTimelineError;
use crate::tenant::SpawnMode;
use crate::tenant::{LogicalSizeCalculationCause, PageReconstructError};
use crate::{config::PageServerConf, tenant::mgr};
@@ -181,7 +178,12 @@ impl From<PageReconstructError> for ApiError {
PageReconstructError::MissingKey(e) => {
ApiError::InternalServerError(anyhow::anyhow!("{e}"))
}
PageReconstructError::Cancelled => ApiError::Cancelled,
PageReconstructError::Cancelled => {
ApiError::InternalServerError(anyhow::anyhow!("request was cancelled"))
}
PageReconstructError::AncestorStopping(_) => {
ApiError::ResourceUnavailable(format!("{pre}").into())
}
PageReconstructError::AncestorLsnTimeout(e) => ApiError::Timeout(format!("{e}").into()),
PageReconstructError::WalRedo(pre) => ApiError::InternalServerError(pre),
}
@@ -277,13 +279,6 @@ impl From<GetTenantError> for ApiError {
}
}
impl From<GetTimelineError> for ApiError {
fn from(gte: GetTimelineError) -> Self {
// Rationale: tenant is activated only after eligble timelines activate
ApiError::NotFound(gte.into())
}
}
impl From<GetActiveTenantError> for ApiError {
fn from(e: GetActiveTenantError) -> ApiError {
match e {
@@ -391,7 +386,7 @@ async fn build_timeline_info_common(
let guard = timeline.last_received_wal.lock().unwrap();
if let Some(info) = guard.as_ref() {
(
Some(format!("{}", info.wal_source_connconf)), // Password is hidden, but it's for statistics only.
Some(format!("{:?}", info.wal_source_connconf)), // Password is hidden, but it's for statistics only.
Some(info.last_received_msg_lsn),
Some(info.last_received_msg_ts),
)
@@ -648,7 +643,9 @@ async fn timeline_preserve_initdb_handler(
.tenant_manager
.get_attached_tenant_shard(tenant_shard_id)?;
let timeline = tenant.get_timeline(timeline_id, false)?;
let timeline = tenant
.get_timeline(timeline_id, false)
.map_err(|e| ApiError::NotFound(e.into()))?;
timeline
.preserve_initdb_archive()
@@ -690,7 +687,9 @@ async fn timeline_detail_handler(
tenant.wait_to_become_active(ACTIVE_TENANT_TIMEOUT).await?;
let timeline = tenant.get_timeline(timeline_id, false)?;
let timeline = tenant
.get_timeline(timeline_id, false)
.map_err(|e| ApiError::NotFound(e.into()))?;
let timeline_info = build_timeline_info(
&timeline,
@@ -1071,7 +1070,7 @@ async fn tenant_delete_handler(
let state = get_state(&request);
let status = state
state
.tenant_manager
.delete_tenant(tenant_shard_id, ACTIVE_TENANT_TIMEOUT)
.instrument(info_span!("tenant_delete_handler",
@@ -1080,14 +1079,7 @@ async fn tenant_delete_handler(
))
.await?;
// Callers use 404 as success for deletions, for historical reasons.
if status == StatusCode::NOT_FOUND {
return Err(ApiError::NotFound(
anyhow::anyhow!("Deletion complete").into(),
));
}
json_response(status, ())
json_response(StatusCode::ACCEPTED, ())
}
/// HTTP endpoint to query the current tenant_size of a tenant.
@@ -1135,10 +1127,7 @@ async fn tenant_size_handler(
&ctx,
)
.await
.map_err(|e| match e {
crate::tenant::size::CalculateSyntheticSizeError::Cancelled => ApiError::ShuttingDown,
other => ApiError::InternalServerError(anyhow::anyhow!(other)),
})?;
.map_err(ApiError::InternalServerError)?;
let mut sizes = None;
let accepts_html = headers
@@ -1146,7 +1135,9 @@ async fn tenant_size_handler(
.map(|v| v == "text/html")
.unwrap_or_default();
if !inputs_only.unwrap_or(false) {
let storage_model = inputs.calculate_model();
let storage_model = inputs
.calculate_model()
.map_err(ApiError::InternalServerError)?;
let size = storage_model.calculate();
// If request header expects html, return html
@@ -1710,32 +1701,6 @@ async fn handle_tenant_break(
json_response(StatusCode::OK, ())
}
// Obtains an lsn lease on the given timeline.
async fn lsn_lease_handler(
request: Request<Body>,
_cancel: CancellationToken,
) -> Result<Response<Body>, ApiError> {
let tenant_shard_id: TenantShardId = parse_request_param(&request, "tenant_shard_id")?;
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
check_permission(&request, Some(tenant_shard_id.tenant_id))?;
let lsn: Lsn = parse_query_param(&request, "lsn")?
.ok_or_else(|| ApiError::BadRequest(anyhow!("missing 'lsn' query parameter")))?;
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
let state = get_state(&request);
let timeline =
active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
.await?;
let result = timeline
.make_lsn_lease(lsn, &ctx)
.map_err(|e| ApiError::InternalServerError(e.context("lsn lease http handler")))?;
json_response(StatusCode::OK, result)
}
// Run GC immediately on given timeline.
async fn timeline_gc_handler(
mut request: Request<Body>,
@@ -1771,8 +1736,6 @@ async fn timeline_compact_handler(
if Some(true) == parse_query_param::<_, bool>(&request, "force_image_layer_creation")? {
flags |= CompactFlags::ForceImageLayerCreation;
}
let wait_until_uploaded =
parse_query_param::<_, bool>(&request, "wait_until_uploaded")?.unwrap_or(false);
async {
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
@@ -1781,9 +1744,6 @@ async fn timeline_compact_handler(
.compact(&cancel, flags, &ctx)
.await
.map_err(|e| ApiError::InternalServerError(e.into()))?;
if wait_until_uploaded {
timeline.remote_client.wait_completion().await.map_err(ApiError::InternalServerError)?;
}
json_response(StatusCode::OK, ())
}
.instrument(info_span!("manual_compaction", tenant_id = %tenant_shard_id.tenant_id, shard_id = %tenant_shard_id.shard_slug(), %timeline_id))
@@ -1808,8 +1768,6 @@ async fn timeline_checkpoint_handler(
if Some(true) == parse_query_param::<_, bool>(&request, "force_image_layer_creation")? {
flags |= CompactFlags::ForceImageLayerCreation;
}
let wait_until_uploaded =
parse_query_param::<_, bool>(&request, "wait_until_uploaded")?.unwrap_or(false);
async {
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
@@ -1817,26 +1775,11 @@ async fn timeline_checkpoint_handler(
timeline
.freeze_and_flush()
.await
.map_err(|e| {
match e {
tenant::timeline::FlushLayerError::Cancelled => ApiError::ShuttingDown,
other => ApiError::InternalServerError(other.into()),
}
})?;
.map_err(ApiError::InternalServerError)?;
timeline
.compact(&cancel, flags, &ctx)
.await
.map_err(|e|
match e {
CompactionError::ShuttingDown => ApiError::ShuttingDown,
CompactionError::Other(e) => ApiError::InternalServerError(e)
}
)?;
if wait_until_uploaded {
timeline.remote_client.wait_completion().await.map_err(ApiError::InternalServerError)?;
}
.map_err(|e| ApiError::InternalServerError(e.into()))?;
json_response(StatusCode::OK, ())
}
@@ -1921,11 +1864,14 @@ async fn timeline_detach_ancestor_handler(
let ctx = RequestContext::new(TaskKind::DetachAncestor, DownloadBehavior::Download);
let ctx = &ctx;
let timeline = tenant.get_timeline(timeline_id, true)?;
let timeline = tenant
.get_timeline(timeline_id, true)
.map_err(|e| ApiError::NotFound(e.into()))?;
let (_guard, prepared) = timeline
.prepare_to_detach_from_ancestor(&tenant, options, ctx)
.await?;
.await
.map_err(|e| ApiError::InternalServerError(e.into()))?;
let res = state
.tenant_manager
@@ -2059,7 +2005,9 @@ async fn active_timeline_of_active_tenant(
tenant.wait_to_become_active(ACTIVE_TENANT_TIMEOUT).await?;
Ok(tenant.get_timeline(timeline_id, true)?)
tenant
.get_timeline(timeline_id, true)
.map_err(|e| ApiError::NotFound(e.into()))
}
async fn always_panic_handler(
@@ -2188,7 +2136,7 @@ async fn tenant_scan_remote_handler(
{
Ok((index_part, index_generation)) => {
tracing::info!("Found timeline {tenant_shard_id}/{timeline_id} metadata (gen {index_generation:?}, {} layers, {} consistent LSN)",
index_part.layer_metadata.len(), index_part.metadata.disk_consistent_lsn());
index_part.layer_metadata.len(), index_part.get_disk_consistent_lsn());
generation = std::cmp::max(generation, index_generation);
}
Err(DownloadError::NotFound) => {
@@ -2323,31 +2271,6 @@ async fn post_tracing_event_handler(
json_response(StatusCode::OK, ())
}
async fn force_aux_policy_switch_handler(
mut r: Request<Body>,
_cancel: CancellationToken,
) -> Result<Response<Body>, ApiError> {
check_permission(&r, None)?;
let tenant_shard_id: TenantShardId = parse_request_param(&r, "tenant_shard_id")?;
let timeline_id: TimelineId = parse_request_param(&r, "timeline_id")?;
let policy: AuxFilePolicy = json_request(&mut r).await?;
let state = get_state(&r);
let tenant = state
.tenant_manager
.get_attached_tenant_shard(tenant_shard_id)?;
tenant.wait_to_become_active(ACTIVE_TENANT_TIMEOUT).await?;
let timeline =
active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
.await?;
timeline
.do_switch_aux_policy(policy)
.map_err(ApiError::InternalServerError)?;
json_response(StatusCode::OK, ())
}
async fn put_io_engine_handler(
mut r: Request<Body>,
_cancel: CancellationToken,
@@ -2425,28 +2348,19 @@ async fn list_aux_files(
active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
.await?;
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
let files = timeline.list_aux_files(body.lsn, &ctx).await?;
json_response(StatusCode::OK, files)
}
let process = || async move {
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
let files = timeline.list_aux_files(body.lsn, &ctx).await?;
Ok::<_, anyhow::Error>(files)
};
async fn perf_info(
request: Request<Body>,
_cancel: CancellationToken,
) -> Result<Response<Body>, ApiError> {
let tenant_shard_id: TenantShardId = parse_request_param(&request, "tenant_shard_id")?;
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
check_permission(&request, Some(tenant_shard_id.tenant_id))?;
let state = get_state(&request);
let timeline =
active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
.await?;
let result = timeline.perf_info().await;
json_response(StatusCode::OK, result)
match process().await {
Ok(st) => json_response(StatusCode::OK, st),
Err(err) => json_response(
StatusCode::INTERNAL_SERVER_ERROR,
ApiError::InternalServerError(err).to_string(),
),
}
}
async fn ingest_aux_files(
@@ -2464,22 +2378,24 @@ async fn ingest_aux_files(
active_timeline_of_active_tenant(&state.tenant_manager, tenant_shard_id, timeline_id)
.await?;
let mut modification = timeline.begin_modification(
Lsn(timeline.get_last_record_lsn().0 + 8), /* advance LSN by 8 */
);
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
for (fname, content) in body.aux_files {
modification
.put_file(&fname, content.as_bytes(), &ctx)
.await
.map_err(ApiError::InternalServerError)?;
}
modification
.commit(&ctx)
.await
.map_err(ApiError::InternalServerError)?;
let process = || async move {
let mut modification = timeline.begin_modification(Lsn(
timeline.get_last_record_lsn().0 + 8
) /* advance LSN by 8 */);
let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Download);
for (fname, content) in body.aux_files {
modification
.put_file(&fname, content.as_bytes(), &ctx)
.await?;
}
modification.commit(&ctx).await?;
Ok::<_, anyhow::Error>(())
};
json_response(StatusCode::OK, ())
match process().await {
Ok(st) => json_response(StatusCode::OK, st),
Err(err) => Err(ApiError::InternalServerError(err)),
}
}
/// Report on the largest tenants on this pageserver, for the storage controller to identify
@@ -2785,10 +2701,6 @@ pub fn make_router(
"/v1/tenant/:tenant_shard_id/timeline/:timeline_id/get_timestamp_of_lsn",
|r| api_handler(r, get_timestamp_of_lsn_handler),
)
.post(
"/v1/tenant/:tenant_shard_id/timeline/:timeline_id/lsn_lease",
|r| api_handler(r, lsn_lease_handler),
)
.put(
"/v1/tenant/:tenant_shard_id/timeline/:timeline_id/do_gc",
|r| api_handler(r, timeline_gc_handler),
@@ -2862,10 +2774,6 @@ pub fn make_router(
|r| api_handler(r, timeline_collect_keyspace),
)
.put("/v1/io_engine", |r| api_handler(r, put_io_engine_handler))
.put(
"/v1/tenant/:tenant_shard_id/timeline/:timeline_id/force_aux_policy_switch",
|r| api_handler(r, force_aux_policy_switch_handler),
)
.get("/v1/utilization", |r| api_handler(r, get_utilization))
.post(
"/v1/tenant/:tenant_shard_id/timeline/:timeline_id/ingest_aux_files",
@@ -2876,9 +2784,5 @@ pub fn make_router(
|r| testing_api_handler("list_aux_files", r, list_aux_files),
)
.post("/v1/top_tenants", |r| api_handler(r, post_top_tenants))
.post(
"/v1/tenant/:tenant_shard_id/timeline/:timeline_id/perf_info",
|r| testing_api_handler("perf_info", r, perf_info),
)
.any(handler_404))
}

View File

@@ -525,15 +525,6 @@ static LAST_RECORD_LSN: Lazy<IntGaugeVec> = Lazy::new(|| {
.expect("failed to define a metric")
});
static STANDBY_HORIZON: Lazy<IntGaugeVec> = Lazy::new(|| {
register_int_gauge_vec!(
"pageserver_standby_horizon",
"Standby apply LSN for which GC is hold off, by timeline.",
&["tenant_id", "shard_id", "timeline_id"]
)
.expect("failed to define a metric")
});
static RESIDENT_PHYSICAL_SIZE: Lazy<UIntGaugeVec> = Lazy::new(|| {
register_uint_gauge_vec!(
"pageserver_resident_physical_size",
@@ -1867,6 +1858,7 @@ pub(crate) struct WalIngestMetrics {
pub(crate) records_received: IntCounter,
pub(crate) records_committed: IntCounter,
pub(crate) records_filtered: IntCounter,
pub(crate) time_spent_on_ingest: Histogram,
}
pub(crate) static WAL_INGEST: Lazy<WalIngestMetrics> = Lazy::new(|| WalIngestMetrics {
@@ -1890,6 +1882,12 @@ pub(crate) static WAL_INGEST: Lazy<WalIngestMetrics> = Lazy::new(|| WalIngestMet
"Number of WAL records filtered out due to sharding"
)
.expect("failed to define a metric"),
time_spent_on_ingest: register_histogram!(
"pageserver_wal_ingest_put_value_seconds",
"Actual time spent on ingesting a record",
redo_histogram_time_buckets!(),
)
.expect("failed to define a metric"),
});
pub(crate) static WAL_REDO_TIME: Lazy<Histogram> = Lazy::new(|| {
@@ -2100,7 +2098,6 @@ pub(crate) struct TimelineMetrics {
pub garbage_collect_histo: StorageTimeMetrics,
pub find_gc_cutoffs_histo: StorageTimeMetrics,
pub last_record_gauge: IntGauge,
pub standby_horizon_gauge: IntGauge,
pub resident_physical_size_gauge: UIntGauge,
/// copy of LayeredTimeline.current_logical_size
pub current_logical_size_gauge: UIntGauge,
@@ -2108,7 +2105,6 @@ pub(crate) struct TimelineMetrics {
pub directory_entries_count_gauge: Lazy<UIntGauge, Box<dyn Send + Fn() -> UIntGauge>>,
pub evictions: IntCounter,
pub evictions_with_low_residence_duration: std::sync::RwLock<EvictionsWithLowResidenceDuration>,
shutdown: std::sync::atomic::AtomicBool,
}
impl TimelineMetrics {
@@ -2171,9 +2167,6 @@ impl TimelineMetrics {
let last_record_gauge = LAST_RECORD_LSN
.get_metric_with_label_values(&[&tenant_id, &shard_id, &timeline_id])
.unwrap();
let standby_horizon_gauge = STANDBY_HORIZON
.get_metric_with_label_values(&[&tenant_id, &shard_id, &timeline_id])
.unwrap();
let resident_physical_size_gauge = RESIDENT_PHYSICAL_SIZE
.get_metric_with_label_values(&[&tenant_id, &shard_id, &timeline_id])
.unwrap();
@@ -2219,7 +2212,6 @@ impl TimelineMetrics {
find_gc_cutoffs_histo,
load_layer_map_histo,
last_record_gauge,
standby_horizon_gauge,
resident_physical_size_gauge,
current_logical_size_gauge,
aux_file_size_gauge,
@@ -2228,7 +2220,6 @@ impl TimelineMetrics {
evictions_with_low_residence_duration: std::sync::RwLock::new(
evictions_with_low_residence_duration,
),
shutdown: std::sync::atomic::AtomicBool::default(),
}
}
@@ -2251,22 +2242,10 @@ impl TimelineMetrics {
}
pub(crate) fn shutdown(&self) {
let was_shutdown = self
.shutdown
.swap(true, std::sync::atomic::Ordering::Relaxed);
if was_shutdown {
// this happens on tenant deletion because tenant first shuts down timelines, then
// invokes timeline deletion which first shuts down the timeline again.
// TODO: this can be removed once https://github.com/neondatabase/neon/issues/5080
return;
}
let tenant_id = &self.tenant_id;
let timeline_id = &self.timeline_id;
let shard_id = &self.shard_id;
let _ = LAST_RECORD_LSN.remove_label_values(&[tenant_id, shard_id, timeline_id]);
let _ = STANDBY_HORIZON.remove_label_values(&[tenant_id, shard_id, timeline_id]);
{
RESIDENT_PHYSICAL_SIZE_GLOBAL.sub(self.resident_physical_size_get());
let _ = RESIDENT_PHYSICAL_SIZE.remove_label_values(&[tenant_id, shard_id, timeline_id]);

View File

@@ -19,7 +19,6 @@ use pageserver_api::models::{
};
use pageserver_api::shard::ShardIndex;
use pageserver_api::shard::ShardNumber;
use pageserver_api::shard::TenantShardId;
use postgres_backend::{is_expected_io_error, AuthType, PostgresBackend, QueryError};
use pq_proto::framed::ConnectionError;
use pq_proto::FeStartupPacket;
@@ -34,7 +33,6 @@ use std::str::FromStr;
use std::sync::Arc;
use std::time::Duration;
use std::time::Instant;
use std::time::SystemTime;
use tokio::io::AsyncWriteExt;
use tokio::io::{AsyncRead, AsyncWrite};
use tokio_util::io::StreamReader;
@@ -66,7 +64,6 @@ use crate::tenant::mgr::GetTenantError;
use crate::tenant::mgr::ShardResolveResult;
use crate::tenant::mgr::ShardSelector;
use crate::tenant::mgr::TenantManager;
use crate::tenant::timeline::FlushLayerError;
use crate::tenant::timeline::WaitLsnError;
use crate::tenant::GetTimelineError;
use crate::tenant::PageReconstructError;
@@ -261,8 +258,6 @@ async fn page_service_conn_main(
socket.set_timeout(Some(std::time::Duration::from_millis(socket_timeout_ms)));
let socket = std::pin::pin!(socket);
fail::fail_point!("ps::connection-start::pre-login");
// 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
@@ -373,7 +368,7 @@ impl From<WaitLsnError> for PageStreamError {
match value {
e @ WaitLsnError::Timeout(_) => Self::LsnTimeout(e),
WaitLsnError::Shutdown => Self::Shutdown,
e @ WaitLsnError::BadState { .. } => Self::Reconnect(format!("{e}").into()),
WaitLsnError::BadState => Self::Reconnect("Timeline is not active".into()),
}
}
}
@@ -383,7 +378,7 @@ impl From<WaitLsnError> for QueryError {
match value {
e @ WaitLsnError::Timeout(_) => Self::Other(anyhow::Error::new(e)),
WaitLsnError::Shutdown => Self::Shutdown,
WaitLsnError::BadState { .. } => Self::Reconnect,
WaitLsnError::BadState => Self::Reconnect,
}
}
}
@@ -606,7 +601,6 @@ impl PageServerHandler {
};
trace!("query: {copy_data_bytes:?}");
fail::fail_point!("ps::handle-pagerequest-message");
// Trace request if needed
if let Some(t) = tracer.as_mut() {
@@ -621,7 +615,6 @@ impl PageServerHandler {
let (response, span) = match neon_fe_msg {
PagestreamFeMessage::Exists(req) => {
fail::fail_point!("ps::handle-pagerequest-message::exists");
let span = tracing::info_span!("handle_get_rel_exists_request", rel = %req.rel, req_lsn = %req.request_lsn);
(
self.handle_get_rel_exists_request(tenant_id, timeline_id, &req, &ctx)
@@ -631,7 +624,6 @@ impl PageServerHandler {
)
}
PagestreamFeMessage::Nblocks(req) => {
fail::fail_point!("ps::handle-pagerequest-message::nblocks");
let span = tracing::info_span!("handle_get_nblocks_request", rel = %req.rel, req_lsn = %req.request_lsn);
(
self.handle_get_nblocks_request(tenant_id, timeline_id, &req, &ctx)
@@ -641,7 +633,6 @@ impl PageServerHandler {
)
}
PagestreamFeMessage::GetPage(req) => {
fail::fail_point!("ps::handle-pagerequest-message::getpage");
// shard_id is filled in by the handler
let span = tracing::info_span!("handle_get_page_at_lsn_request", rel = %req.rel, blkno = %req.blkno, req_lsn = %req.request_lsn);
(
@@ -652,7 +643,6 @@ impl PageServerHandler {
)
}
PagestreamFeMessage::DbSize(req) => {
fail::fail_point!("ps::handle-pagerequest-message::dbsize");
let span = tracing::info_span!("handle_db_size_request", dbnode = %req.dbnode, req_lsn = %req.request_lsn);
(
self.handle_db_size_request(tenant_id, timeline_id, &req, &ctx)
@@ -662,7 +652,6 @@ impl PageServerHandler {
)
}
PagestreamFeMessage::GetSlruSegment(req) => {
fail::fail_point!("ps::handle-pagerequest-message::slrusegment");
let span = tracing::info_span!("handle_get_slru_segment_request", kind = %req.kind, segno = %req.segno, req_lsn = %req.request_lsn);
(
self.handle_get_slru_segment_request(tenant_id, timeline_id, &req, &ctx)
@@ -831,10 +820,7 @@ impl PageServerHandler {
// We only want to persist the data, and it doesn't matter if it's in the
// shape of deltas or images.
info!("flushing layers");
timeline.freeze_and_flush().await.map_err(|e| match e {
FlushLayerError::Cancelled => QueryError::Shutdown,
other => QueryError::Other(other.into()),
})?;
timeline.freeze_and_flush().await?;
info!("done");
Ok(())
@@ -919,39 +905,6 @@ impl PageServerHandler {
}
}
#[instrument(skip_all, fields(shard_id, %lsn))]
async fn handle_make_lsn_lease<IO>(
&self,
pgb: &mut PostgresBackend<IO>,
tenant_shard_id: TenantShardId,
timeline_id: TimelineId,
lsn: Lsn,
ctx: &RequestContext,
) -> Result<(), QueryError>
where
IO: AsyncRead + AsyncWrite + Send + Sync + Unpin,
{
let shard_selector = ShardSelector::Known(tenant_shard_id.to_index());
let timeline = self
.get_active_tenant_timeline(tenant_shard_id.tenant_id, timeline_id, shard_selector)
.await?;
let lease = timeline.make_lsn_lease(lsn, ctx)?;
let valid_until = lease
.valid_until
.duration_since(SystemTime::UNIX_EPOCH)
.map_err(|e| QueryError::Other(e.into()))?;
pgb.write_message_noflush(&BeMessage::RowDescription(&[RowDescriptor::text_col(
b"valid_until",
)]))?
.write_message_noflush(&BeMessage::DataRow(&[Some(
&valid_until.as_millis().to_be_bytes(),
)]))?
.write_message_noflush(&BeMessage::CommandComplete(b"SELECT 1"))?;
Ok(())
}
#[instrument(skip_all, fields(shard_id))]
async fn handle_get_rel_exists_request(
&mut self,
@@ -1517,7 +1470,6 @@ where
_pgb: &mut PostgresBackend<IO>,
_sm: &FeStartupPacket,
) -> Result<(), QueryError> {
fail::fail_point!("ps::connection-start::startup-packet");
Ok(())
}
@@ -1532,12 +1484,11 @@ where
Err(QueryError::SimulatedConnectionError)
});
fail::fail_point!("ps::connection-start::process-query");
let ctx = self.connection_ctx.attached_child();
debug!("process query {query_string:?}");
let parts = query_string.split_whitespace().collect::<Vec<_>>();
if let Some(params) = parts.strip_prefix(&["pagestream_v2"]) {
if query_string.starts_with("pagestream_v2 ") {
let (_, params_raw) = query_string.split_at("pagestream_v2 ".len());
let params = params_raw.split(' ').collect::<Vec<_>>();
if params.len() != 2 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for pagestream command"
@@ -1562,7 +1513,9 @@ where
ctx,
)
.await?;
} else if let Some(params) = parts.strip_prefix(&["pagestream"]) {
} else if query_string.starts_with("pagestream ") {
let (_, params_raw) = query_string.split_at("pagestream ".len());
let params = params_raw.split(' ').collect::<Vec<_>>();
if params.len() != 2 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for pagestream command"
@@ -1587,7 +1540,10 @@ where
ctx,
)
.await?;
} else if let Some(params) = parts.strip_prefix(&["basebackup"]) {
} else if query_string.starts_with("basebackup ") {
let (_, params_raw) = query_string.split_at("basebackup ".len());
let params = params_raw.split_whitespace().collect::<Vec<_>>();
if params.len() < 2 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for basebackup command"
@@ -1605,23 +1561,26 @@ where
self.check_permission(Some(tenant_id))?;
let lsn = if let Some(lsn_str) = params.get(2) {
let lsn = if params.len() >= 3 {
Some(
Lsn::from_str(lsn_str)
.with_context(|| format!("Failed to parse Lsn from {lsn_str}"))?,
Lsn::from_str(params[2])
.with_context(|| format!("Failed to parse Lsn from {}", params[2]))?,
)
} else {
None
};
let gzip = match params.get(3) {
Some(&"--gzip") => true,
None => false,
Some(third_param) => {
let gzip = if params.len() >= 4 {
if params[3] == "--gzip" {
true
} else {
return Err(QueryError::Other(anyhow::anyhow!(
"Parameter in position 3 unknown {third_param}",
)))
"Parameter in position 3 unknown {}",
params[3],
)));
}
} else {
false
};
let metric_recording = metrics::BASEBACKUP_QUERY_TIME.start_recording(&ctx);
@@ -1645,7 +1604,10 @@ where
res?;
}
// return pair of prev_lsn and last_lsn
else if let Some(params) = parts.strip_prefix(&["get_last_record_rlsn"]) {
else if query_string.starts_with("get_last_record_rlsn ") {
let (_, params_raw) = query_string.split_at("get_last_record_rlsn ".len());
let params = params_raw.split_whitespace().collect::<Vec<_>>();
if params.len() != 2 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for get_last_record_rlsn command"
@@ -1687,7 +1649,10 @@ where
.await?;
}
// same as basebackup, but result includes relational data as well
else if let Some(params) = parts.strip_prefix(&["fullbackup"]) {
else if query_string.starts_with("fullbackup ") {
let (_, params_raw) = query_string.split_at("fullbackup ".len());
let params = params_raw.split_whitespace().collect::<Vec<_>>();
if params.len() < 2 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for fullbackup command"
@@ -1704,18 +1669,18 @@ where
.record("timeline_id", field::display(timeline_id));
// The caller is responsible for providing correct lsn and prev_lsn.
let lsn = if let Some(lsn_str) = params.get(2) {
let lsn = if params.len() > 2 {
Some(
Lsn::from_str(lsn_str)
.with_context(|| format!("Failed to parse Lsn from {lsn_str}"))?,
Lsn::from_str(params[2])
.with_context(|| format!("Failed to parse Lsn from {}", params[2]))?,
)
} else {
None
};
let prev_lsn = if let Some(prev_lsn_str) = params.get(3) {
let prev_lsn = if params.len() > 3 {
Some(
Lsn::from_str(prev_lsn_str)
.with_context(|| format!("Failed to parse Lsn from {prev_lsn_str}"))?,
Lsn::from_str(params[3])
.with_context(|| format!("Failed to parse Lsn from {}", params[3]))?,
)
} else {
None
@@ -1748,7 +1713,8 @@ where
// 2. Run:
// cat my_backup/base.tar | psql -h $PAGESERVER \
// -c "import basebackup $TENANT $TIMELINE $START_LSN $END_LSN $PG_VERSION"
let params = &parts[2..];
let (_, params_raw) = query_string.split_at("import basebackup ".len());
let params = params_raw.split_whitespace().collect::<Vec<_>>();
if params.len() != 5 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for import basebackup command"
@@ -1797,7 +1763,8 @@ where
//
// Files are scheduled to be persisted to remote storage, and the
// caller should poll the http api to check when that is done.
let params = &parts[2..];
let (_, params_raw) = query_string.split_at("import wal ".len());
let params = params_raw.split_whitespace().collect::<Vec<_>>();
if params.len() != 4 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for import wal command"
@@ -1835,45 +1802,10 @@ where
// important because psycopg2 executes "SET datestyle TO 'ISO'"
// on connect
pgb.write_message_noflush(&BeMessage::CommandComplete(b"SELECT 1"))?;
} else if query_string.starts_with("lease lsn ") {
let params = &parts[2..];
if params.len() != 3 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number {} for lease lsn command",
params.len()
)));
}
let tenant_shard_id = TenantShardId::from_str(params[0])
.with_context(|| format!("Failed to parse tenant id from {}", params[0]))?;
let timeline_id = TimelineId::from_str(params[1])
.with_context(|| format!("Failed to parse timeline id from {}", params[1]))?;
tracing::Span::current()
.record("tenant_id", field::display(tenant_shard_id))
.record("timeline_id", field::display(timeline_id));
self.check_permission(Some(tenant_shard_id.tenant_id))?;
// The caller is responsible for providing correct lsn.
let lsn = Lsn::from_str(params[2])
.with_context(|| format!("Failed to parse Lsn from {}", params[2]))?;
match self
.handle_make_lsn_lease(pgb, tenant_shard_id, timeline_id, lsn, &ctx)
.await
{
Ok(()) => pgb.write_message_noflush(&BeMessage::CommandComplete(b"SELECT 1"))?,
Err(e) => {
error!("error obtaining lsn lease for {lsn}: {e:?}");
pgb.write_message_noflush(&BeMessage::ErrorResponse(
&e.to_string(),
Some(e.pg_error_code()),
))?
}
};
} else if let Some(params) = parts.strip_prefix(&["show"]) {
} else if query_string.starts_with("show ") {
// show <tenant_id>
let (_, params_raw) = query_string.split_at("show ".len());
let params = params_raw.split(' ').collect::<Vec<_>>();
if params.len() != 1 {
return Err(QueryError::Other(anyhow::anyhow!(
"invalid param number for config command"

View File

@@ -9,6 +9,7 @@
use super::tenant::{PageReconstructError, Timeline};
use crate::context::RequestContext;
use crate::keyspace::{KeySpace, KeySpaceAccum};
use crate::metrics::WAL_INGEST;
use crate::span::debug_assert_current_span_has_tenant_and_timeline_id_no_shard_id;
use crate::walrecord::NeonWalRecord;
use crate::{aux_file, repository::*};
@@ -17,8 +18,8 @@ use bytes::{Buf, Bytes, BytesMut};
use enum_map::Enum;
use itertools::Itertools;
use pageserver_api::key::{
dbdir_key_range, rel_block_to_key, rel_dir_to_key, rel_key_range, rel_size_to_key,
relmap_file_key, repl_origin_key, repl_origin_key_range, slru_block_to_key, slru_dir_to_key,
dbdir_key_range, is_rel_block_key, is_slru_block_key, rel_block_to_key, rel_dir_to_key,
rel_key_range, rel_size_to_key, relmap_file_key, slru_block_to_key, slru_dir_to_key,
slru_segment_key_range, slru_segment_size_to_key, twophase_file_key, twophase_key_range,
AUX_FILES_KEY, CHECKPOINT_KEY, CONTROLFILE_KEY, DBDIR_KEY, TWOPHASEDIR_KEY,
};
@@ -27,7 +28,7 @@ use pageserver_api::models::AuxFilePolicy;
use pageserver_api::reltag::{BlockNumber, RelTag, SlruKind};
use postgres_ffi::relfile_utils::{FSM_FORKNUM, VISIBILITYMAP_FORKNUM};
use postgres_ffi::BLCKSZ;
use postgres_ffi::{Oid, RepOriginId, TimestampTz, TransactionId};
use postgres_ffi::{Oid, TimestampTz, TransactionId};
use serde::{Deserialize, Serialize};
use std::collections::{hash_map, HashMap, HashSet};
use std::ops::ControlFlow;
@@ -36,15 +37,10 @@ use strum::IntoEnumIterator;
use tokio_util::sync::CancellationToken;
use tracing::{debug, info, trace, warn};
use utils::bin_ser::DeserializeError;
use utils::pausable_failpoint;
use utils::vec_map::{VecMap, VecMapOrdering};
use utils::{bin_ser::BeSer, lsn::Lsn};
/// Max delta records appended to the AUX_FILES_KEY (for aux v1). The write path will write a full image once this threshold is reached.
pub const MAX_AUX_FILE_DELTAS: usize = 1024;
/// Max number of aux-file-related delta layers. The compaction will create a new image layer once this threshold is reached.
pub const MAX_AUX_FILE_V2_DELTAS: usize = 64;
const MAX_AUX_FILE_DELTAS: usize = 1024;
#[derive(Debug)]
pub enum LsnForTimestamp {
@@ -79,19 +75,11 @@ pub enum LsnForTimestamp {
}
#[derive(Debug, thiserror::Error)]
pub(crate) enum CalculateLogicalSizeError {
pub enum CalculateLogicalSizeError {
#[error("cancelled")]
Cancelled,
/// Something went wrong while reading the metadata we use to calculate logical size
/// Note that cancellation variants of `PageReconstructError` are transformed to [`Self::Cancelled`]
/// in the `From` implementation for this variant.
#[error(transparent)]
PageRead(PageReconstructError),
/// Something went wrong deserializing metadata that we read to calculate logical size
#[error("decode error: {0}")]
Decode(#[from] DeserializeError),
Other(#[from] anyhow::Error),
}
#[derive(Debug, thiserror::Error)]
@@ -116,8 +104,10 @@ impl From<PageReconstructError> for CollectKeySpaceError {
impl From<PageReconstructError> for CalculateLogicalSizeError {
fn from(pre: PageReconstructError) -> Self {
match pre {
PageReconstructError::Cancelled => Self::Cancelled,
_ => Self::PageRead(pre),
PageReconstructError::AncestorStopping(_) | PageReconstructError::Cancelled => {
Self::Cancelled
}
_ => Self::Other(pre.into()),
}
}
}
@@ -410,8 +400,6 @@ impl Timeline {
cancel: &CancellationToken,
ctx: &RequestContext,
) -> Result<LsnForTimestamp, PageReconstructError> {
pausable_failpoint!("find-lsn-for-timestamp-pausable");
let gc_cutoff_lsn_guard = self.get_latest_gc_cutoff_lsn();
// We use this method to figure out the branching LSN for the new branch, but the
// GC cutoff could be before the branching point and we cannot create a new branch
@@ -427,7 +415,6 @@ impl Timeline {
let mut found_smaller = false;
let mut found_larger = false;
while low < high {
if cancel.is_cancelled() {
return Err(PageReconstructError::Cancelled);
@@ -722,22 +709,10 @@ impl Timeline {
result.insert(fname, content);
}
}
self.aux_file_size_estimator.on_initial(sz);
self.aux_file_size_estimator.on_base_backup(sz);
Ok(result)
}
pub(crate) async fn trigger_aux_file_size_computation(
&self,
lsn: Lsn,
ctx: &RequestContext,
) -> Result<(), PageReconstructError> {
let current_policy = self.last_aux_file_policy.load();
if let Some(AuxFilePolicy::V2) | Some(AuxFilePolicy::CrossValidation) = current_policy {
self.list_aux_files_v2(lsn, ctx).await?;
}
Ok(())
}
pub(crate) async fn list_aux_files(
&self,
lsn: Lsn,
@@ -776,27 +751,6 @@ impl Timeline {
}
}
pub(crate) async fn get_replorigins(
&self,
lsn: Lsn,
ctx: &RequestContext,
) -> Result<HashMap<RepOriginId, Lsn>, PageReconstructError> {
let kv = self
.scan(KeySpace::single(repl_origin_key_range()), lsn, ctx)
.await
.context("scan")?;
let mut result = HashMap::new();
for (k, v) in kv {
let v = v.context("get value")?;
let origin_id = k.field6 as RepOriginId;
let origin_lsn = Lsn::des(&v).unwrap();
if origin_lsn != Lsn::INVALID {
result.insert(origin_id, origin_lsn);
}
}
Ok(result)
}
/// Does the same as get_current_logical_size but counted on demand.
/// Used to initialize the logical size tracking on startup.
///
@@ -806,7 +760,7 @@ impl Timeline {
/// # Cancel-Safety
///
/// This method is cancellation-safe.
pub(crate) async fn get_current_logical_size_non_incremental(
pub async fn get_current_logical_size_non_incremental(
&self,
lsn: Lsn,
ctx: &RequestContext,
@@ -815,7 +769,7 @@ impl Timeline {
// Fetch list of database dirs and iterate them
let buf = self.get(DBDIR_KEY, lsn, ctx).await?;
let dbdir = DbDirectory::des(&buf)?;
let dbdir = DbDirectory::des(&buf).context("deserialize db directory")?;
let mut total_size: u64 = 0;
for (spcnode, dbnode) in dbdir.dbdirs.keys() {
@@ -919,20 +873,10 @@ impl Timeline {
result.add_key(AUX_FILES_KEY);
}
#[cfg(test)]
{
let guard = self.extra_test_dense_keyspace.load();
for kr in &guard.ranges {
result.add_range(kr.clone());
}
}
Ok((
result.to_keyspace(),
/* AUX sparse key space */
SparseKeySpace(KeySpace {
ranges: vec![repl_origin_key_range(), Key::metadata_aux_key_range()],
}),
SparseKeySpace(KeySpace::single(Key::metadata_aux_key_range())),
))
}
@@ -1201,20 +1145,6 @@ impl<'a> DatadirModification<'a> {
Ok(())
}
pub async fn set_replorigin(
&mut self,
origin_id: RepOriginId,
origin_lsn: Lsn,
) -> anyhow::Result<()> {
let key = repl_origin_key(origin_id);
self.put(key, Value::Image(origin_lsn.ser().unwrap().into()));
Ok(())
}
pub async fn drop_replorigin(&mut self, origin_id: RepOriginId) -> anyhow::Result<()> {
self.set_replorigin(origin_id, Lsn::INVALID).await
}
pub fn put_control_file(&mut self, img: Bytes) -> anyhow::Result<()> {
self.put(CONTROLFILE_KEY, Value::Image(img));
Ok(())
@@ -1547,24 +1477,11 @@ impl<'a> DatadirModification<'a> {
// Allowed switch path:
// * no aux files -> v1/v2/cross-validation
// * cross-validation->v2
let current_policy = if current_policy.is_none() {
// This path will only be hit once per tenant: we will decide the final policy in this code block.
// The next call to `put_file` will always have `last_aux_file_policy != None`.
let lsn = Lsn::max(self.tline.get_last_record_lsn(), self.lsn);
let aux_files_key_v1 = self.tline.list_aux_files_v1(lsn, ctx).await?;
if aux_files_key_v1.is_empty() {
None
} else {
self.tline.do_switch_aux_policy(AuxFilePolicy::V1)?;
Some(AuxFilePolicy::V1)
}
} else {
current_policy
};
if AuxFilePolicy::is_valid_migration_path(current_policy, switch_policy) {
self.tline.do_switch_aux_policy(switch_policy)?;
self.tline.last_aux_file_policy.store(Some(switch_policy));
self.tline
.remote_client
.schedule_index_upload_for_aux_file_policy_update(Some(switch_policy))?;
info!(current=?current_policy, next=?switch_policy, "switching aux file policy");
switch_policy
} else {
@@ -1619,7 +1536,7 @@ impl<'a> DatadirModification<'a> {
self.tline.aux_file_size_estimator.on_add(content.len());
new_files.push((path, content));
}
(None, true) => warn!("removing non-existing aux file: {}", path),
(None, true) => anyhow::bail!("removing non-existing aux file: {}", path),
}
let new_val = aux_file::encode_file_value(&new_files)?;
self.put(key, Value::Image(new_val.into()));
@@ -1673,7 +1590,8 @@ impl<'a> DatadirModification<'a> {
aux_files.dir = Some(dir);
}
Err(
e @ (PageReconstructError::Cancelled
e @ (PageReconstructError::AncestorStopping(_)
| PageReconstructError::Cancelled
| PageReconstructError::AncestorLsnTimeout(_)),
) => {
// Important that we do not interpret a shutdown error as "not found" and thereby
@@ -1745,7 +1663,7 @@ impl<'a> DatadirModification<'a> {
let mut retained_pending_updates = HashMap::<_, Vec<_>>::new();
for (key, values) in self.pending_updates.drain() {
for (lsn, value) in values {
if key.is_rel_block_key() || key.is_slru_block_key() {
if is_rel_block_key(&key) || is_slru_block_key(key) {
// This bails out on first error without modifying pending_updates.
// That's Ok, cf this function's doc comment.
writer.put(key, lsn, &value, ctx).await?;
@@ -1780,6 +1698,8 @@ impl<'a> DatadirModification<'a> {
pub async fn commit(&mut self, ctx: &RequestContext) -> anyhow::Result<()> {
let mut writer = self.tline.writer().await;
let timer = WAL_INGEST.time_spent_on_ingest.start_timer();
let pending_nblocks = self.pending_nblocks;
self.pending_nblocks = 0;
@@ -1819,6 +1739,8 @@ impl<'a> DatadirModification<'a> {
writer.update_directory_entries_count(kind, count as u64);
}
timer.observe_duration();
Ok(())
}
@@ -1854,12 +1776,6 @@ impl<'a> DatadirModification<'a> {
self.tline.get(key, lsn, ctx).await
}
/// Only used during unit tests, force putting a key into the modification.
#[cfg(test)]
pub(crate) fn put_for_test(&mut self, key: Key, val: Value) {
self.put(key, val);
}
fn put(&mut self, key: Key, val: Value) {
let values = self.pending_updates.entry(key).or_default();
// Replace the previous value if it exists at the same lsn

File diff suppressed because it is too large Load Diff

View File

@@ -238,13 +238,10 @@ impl<const BUFFERED: bool> BlobWriter<BUFFERED> {
io_buf,
Err(Error::new(
ErrorKind::Other,
format!("blob too large ({len} bytes)"),
format!("blob too large ({} bytes)", len),
)),
);
}
if len > 0x0fff_ffff {
tracing::warn!("writing blob above future limit ({len} bytes)");
}
let mut len_buf = (len as u32).to_be_bytes();
len_buf[0] |= 0x80;
io_buf.extend_from_slice(&len_buf[..]);

View File

@@ -11,7 +11,6 @@
use anyhow::bail;
use pageserver_api::models::AuxFilePolicy;
use pageserver_api::models::CompactionAlgorithm;
use pageserver_api::models::CompactionAlgorithmSettings;
use pageserver_api::models::EvictionPolicy;
use pageserver_api::models::{self, ThrottleConfig};
use pageserver_api::shard::{ShardCount, ShardIdentity, ShardNumber, ShardStripeSize};
@@ -321,7 +320,7 @@ pub struct TenantConf {
pub compaction_period: Duration,
// Level0 delta layer threshold for compaction.
pub compaction_threshold: usize,
pub compaction_algorithm: CompactionAlgorithmSettings,
pub compaction_algorithm: CompactionAlgorithm,
// Determines how much history is retained, to allow
// branching and read replicas at an older point in time.
// The unit is #of bytes of WAL.
@@ -407,7 +406,7 @@ pub struct TenantConfOpt {
#[serde(skip_serializing_if = "Option::is_none")]
#[serde(default)]
pub compaction_algorithm: Option<CompactionAlgorithmSettings>,
pub compaction_algorithm: Option<CompactionAlgorithm>,
#[serde(skip_serializing_if = "Option::is_none")]
#[serde(default)]
@@ -498,9 +497,7 @@ impl TenantConfOpt {
.unwrap_or(global_conf.compaction_threshold),
compaction_algorithm: self
.compaction_algorithm
.as_ref()
.unwrap_or(&global_conf.compaction_algorithm)
.clone(),
.unwrap_or(global_conf.compaction_algorithm),
gc_horizon: self.gc_horizon.unwrap_or(global_conf.gc_horizon),
gc_period: self.gc_period.unwrap_or(global_conf.gc_period),
image_creation_threshold: self
@@ -553,9 +550,7 @@ impl Default for TenantConf {
compaction_period: humantime::parse_duration(DEFAULT_COMPACTION_PERIOD)
.expect("cannot parse default compaction period"),
compaction_threshold: DEFAULT_COMPACTION_THRESHOLD,
compaction_algorithm: CompactionAlgorithmSettings {
kind: DEFAULT_COMPACTION_ALGORITHM,
},
compaction_algorithm: DEFAULT_COMPACTION_ALGORITHM,
gc_horizon: DEFAULT_GC_HORIZON,
gc_period: humantime::parse_duration(DEFAULT_GC_PERIOD)
.expect("cannot parse default gc period"),

View File

@@ -8,7 +8,7 @@ use tokio::sync::OwnedMutexGuard;
use tokio_util::sync::CancellationToken;
use tracing::{error, instrument, Instrument};
use utils::{backoff, completion, crashsafe, fs_ext, id::TimelineId, pausable_failpoint};
use utils::{backoff, completion, crashsafe, fs_ext, id::TimelineId};
use crate::{
config::PageServerConf,
@@ -16,7 +16,6 @@ use crate::{
task_mgr::{self, TaskKind},
tenant::{
mgr::{TenantSlot, TenantsMapRemoveResult},
remote_timeline_client::remote_heatmap_path,
timeline::ShutdownMode,
},
};
@@ -532,25 +531,6 @@ impl DeleteTenantFlow {
}
}
// Remove top-level tenant objects that don't belong to a timeline, such as heatmap
let heatmap_path = remote_heatmap_path(&tenant.tenant_shard_id());
if let Some(Err(e)) = backoff::retry(
|| async {
remote_storage
.delete(&heatmap_path, &task_mgr::shutdown_token())
.await
},
TimeoutOrCancel::caused_by_cancel,
FAILED_UPLOAD_WARN_THRESHOLD,
FAILED_REMOTE_OP_RETRIES,
"remove_remote_tenant_heatmap",
&task_mgr::shutdown_token(),
)
.await
{
tracing::warn!("Failed to delete heatmap at {heatmap_path}: {e}");
}
let timelines_path = conf.timelines_path(&tenant.tenant_shard_id);
// May not exist if we fail in cleanup_remaining_fs_traces after removing it
if timelines_path.exists() {

View File

@@ -1,23 +1,15 @@
//! Describes the legacy now hopefully no longer modified per-timeline metadata stored in
//! `index_part.json` managed by [`remote_timeline_client`]. For many tenants and their timelines,
//! this struct and it's original serialization format is still needed because they were written a
//! long time ago.
//! Every image of a certain timeline from [`crate::tenant::Tenant`]
//! has a metadata that needs to be stored persistently.
//!
//! Instead of changing and adding versioning to this, just change [`IndexPart`] with soft json
//! versioning.
//! Later, the file gets used in [`remote_timeline_client`] as a part of
//! external storage import and export operations.
//!
//! To clean up this module we need to migrate all index_part.json files to a later version.
//! While doing this, we need to be mindful about s3 based recovery as well, so it might take
//! however long we keep the old versions to be able to delete the old code. After that, we can
//! remove everything else than [`TimelineMetadataBodyV2`], rename it as `TimelineMetadata` and
//! move it to `index.rs`. Before doing all of this, we need to keep the structures for backwards
//! compatibility.
//! The module contains all structs and related helper methods related to timeline metadata.
//!
//! [`remote_timeline_client`]: super::remote_timeline_client
//! [`IndexPart`]: super::remote_timeline_client::index::IndexPart
use anyhow::ensure;
use serde::{Deserialize, Serialize};
use serde::{de::Error, Deserialize, Serialize, Serializer};
use utils::bin_ser::SerializeError;
use utils::{bin_ser::BeSer, id::TimelineId, lsn::Lsn};
@@ -25,37 +17,17 @@ use utils::{bin_ser::BeSer, id::TimelineId, lsn::Lsn};
const METADATA_FORMAT_VERSION: u16 = 4;
/// Previous supported format versions.
///
/// In practice, none of these should remain, all are [`METADATA_FORMAT_VERSION`], but confirming
/// that requires a scrubber run which is yet to be done.
const METADATA_OLD_FORMAT_VERSION: u16 = 3;
/// When the file existed on disk we assumed that a write of up to METADATA_MAX_SIZE bytes is atomic.
/// We assume that a write of up to METADATA_MAX_SIZE bytes is atomic.
///
/// This is the same assumption that PostgreSQL makes with the control file,
///
/// see PG_CONTROL_MAX_SAFE_SIZE
const METADATA_MAX_SIZE: usize = 512;
/// Legacy metadata stored as a component of `index_part.json` per timeline.
/// Metadata stored on disk for each timeline
///
/// Do not make new changes to this type or the module. In production, we have two different kinds
/// of serializations of this type: bincode and json. Bincode version reflects what used to be
/// stored on disk in earlier versions and does internal crc32 checksumming.
///
/// This type should not implement `serde::Serialize` or `serde::Deserialize` because there would
/// be a confusion whether you want the old version ([`TimelineMetadata::from_bytes`]) or the modern
/// as-exists in `index_part.json` ([`self::modern_serde`]).
///
/// ```compile_fail
/// #[derive(serde::Serialize)]
/// struct DoNotDoThis(pageserver::tenant::metadata::TimelineMetadata);
/// ```
///
/// ```compile_fail
/// #[derive(serde::Deserialize)]
/// struct NeitherDoThis(pageserver::tenant::metadata::TimelineMetadata);
/// ```
/// The fields correspond to the values we hold in memory, in Timeline.
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct TimelineMetadata {
hdr: TimelineMetadataHeader,
@@ -68,49 +40,6 @@ struct TimelineMetadataHeader {
size: u16, // size of serialized metadata
format_version: u16, // metadata format version (used for compatibility checks)
}
impl TryFrom<&TimelineMetadataBodyV2> for TimelineMetadataHeader {
type Error = Crc32CalculationFailed;
fn try_from(value: &TimelineMetadataBodyV2) -> Result<Self, Self::Error> {
#[derive(Default)]
struct Crc32Sink {
crc: u32,
count: usize,
}
impl std::io::Write for Crc32Sink {
fn write(&mut self, buf: &[u8]) -> std::io::Result<usize> {
self.crc = crc32c::crc32c_append(self.crc, buf);
self.count += buf.len();
Ok(buf.len())
}
fn flush(&mut self) -> std::io::Result<()> {
Ok(())
}
}
// jump through hoops to calculate the crc32 so that TimelineMetadata::ne works
// across serialization versions
let mut sink = Crc32Sink::default();
<TimelineMetadataBodyV2 as utils::bin_ser::BeSer>::ser_into(value, &mut sink)
.map_err(Crc32CalculationFailed)?;
let size = METADATA_HDR_SIZE + sink.count;
Ok(TimelineMetadataHeader {
checksum: sink.crc,
size: size as u16,
format_version: METADATA_FORMAT_VERSION,
})
}
}
#[derive(thiserror::Error, Debug)]
#[error("re-serializing for crc32 failed")]
struct Crc32CalculationFailed(#[source] utils::bin_ser::SerializeError);
const METADATA_HDR_SIZE: usize = std::mem::size_of::<TimelineMetadataHeader>();
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
@@ -182,12 +111,6 @@ impl TimelineMetadata {
}
}
#[cfg(test)]
pub(crate) fn with_recalculated_checksum(mut self) -> anyhow::Result<Self> {
self.hdr = TimelineMetadataHeader::try_from(&self.body)?;
Ok(self)
}
fn upgrade_timeline_metadata(metadata_bytes: &[u8]) -> anyhow::Result<Self> {
let mut hdr = TimelineMetadataHeader::des(&metadata_bytes[0..METADATA_HDR_SIZE])?;
@@ -338,93 +261,25 @@ impl TimelineMetadata {
}
}
pub(crate) mod modern_serde {
use super::{TimelineMetadata, TimelineMetadataBodyV2, TimelineMetadataHeader};
use serde::{Deserialize, Serialize};
pub(crate) fn deserialize<'de, D>(deserializer: D) -> Result<TimelineMetadata, D::Error>
impl<'de> Deserialize<'de> for TimelineMetadata {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: serde::de::Deserializer<'de>,
D: serde::Deserializer<'de>,
{
// for legacy reasons versions 1-5 had TimelineMetadata serialized as a Vec<u8> field with
// BeSer.
struct Visitor;
impl<'d> serde::de::Visitor<'d> for Visitor {
type Value = TimelineMetadata;
fn expecting(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
f.write_str("BeSer bytes or json structure")
}
fn visit_seq<A>(self, seq: A) -> Result<Self::Value, A::Error>
where
A: serde::de::SeqAccess<'d>,
{
use serde::de::Error;
let de = serde::de::value::SeqAccessDeserializer::new(seq);
Vec::<u8>::deserialize(de)
.map(|v| TimelineMetadata::from_bytes(&v).map_err(A::Error::custom))?
}
fn visit_map<A>(self, map: A) -> Result<Self::Value, A::Error>
where
A: serde::de::MapAccess<'d>,
{
use serde::de::Error;
let de = serde::de::value::MapAccessDeserializer::new(map);
let body = TimelineMetadataBodyV2::deserialize(de)?;
let hdr = TimelineMetadataHeader::try_from(&body).map_err(A::Error::custom)?;
Ok(TimelineMetadata { hdr, body })
}
}
deserializer.deserialize_any(Visitor)
let bytes = Vec::<u8>::deserialize(deserializer)?;
Self::from_bytes(bytes.as_slice()).map_err(|e| D::Error::custom(format!("{e}")))
}
}
pub(crate) fn serialize<S>(
metadata: &TimelineMetadata,
serializer: S,
) -> Result<S::Ok, S::Error>
impl Serialize for TimelineMetadata {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where
S: serde::Serializer,
S: Serializer,
{
// header is not needed, upon reading we've upgraded all v1 to v2
metadata.body.serialize(serializer)
}
#[test]
fn deserializes_bytes_as_well_as_equivalent_body_v2() {
#[derive(serde::Deserialize, serde::Serialize)]
struct Wrapper(
#[serde(deserialize_with = "deserialize", serialize_with = "serialize")]
TimelineMetadata,
);
let too_many_bytes = "[216,111,252,208,0,54,0,4,0,0,0,0,1,73,253,144,1,0,0,0,0,1,73,253,24,0,0,0,0,0,0,0,0,0,0,0,0,0,1,73,253,24,0,0,0,0,1,73,253,24,0,0,0,15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]";
let wrapper_from_bytes = serde_json::from_str::<Wrapper>(too_many_bytes).unwrap();
let serialized = serde_json::to_value(&wrapper_from_bytes).unwrap();
assert_eq!(
serialized,
serde_json::json! {{
"disk_consistent_lsn": "0/149FD90",
"prev_record_lsn": "0/149FD18",
"ancestor_timeline": null,
"ancestor_lsn": "0/0",
"latest_gc_cutoff_lsn": "0/149FD18",
"initdb_lsn": "0/149FD18",
"pg_version": 15
}}
);
let wrapper_from_json = serde_json::value::from_value::<Wrapper>(serialized).unwrap();
assert_eq!(wrapper_from_bytes.0, wrapper_from_json.0);
let bytes = self
.to_bytes()
.map_err(|e| serde::ser::Error::custom(format!("{e}")))?;
bytes.serialize(serializer)
}
}
@@ -548,6 +403,59 @@ mod tests {
);
}
#[test]
fn test_metadata_bincode_serde() {
let original_metadata = TimelineMetadata::new(
Lsn(0x200),
Some(Lsn(0x100)),
Some(TIMELINE_ID),
Lsn(0),
Lsn(0),
Lsn(0),
// Any version will do here, so use the default
crate::DEFAULT_PG_VERSION,
);
let metadata_bytes = original_metadata
.to_bytes()
.expect("Cannot create bytes array from metadata");
let metadata_bincode_be_bytes = original_metadata
.ser()
.expect("Cannot serialize the metadata");
// 8 bytes for the length of the vector
assert_eq!(metadata_bincode_be_bytes.len(), 8 + metadata_bytes.len());
let expected_bincode_bytes = {
let mut temp = vec![];
let len_bytes = metadata_bytes.len().to_be_bytes();
temp.extend_from_slice(&len_bytes);
temp.extend_from_slice(&metadata_bytes);
temp
};
assert_eq!(metadata_bincode_be_bytes, expected_bincode_bytes);
let deserialized_metadata = TimelineMetadata::des(&metadata_bincode_be_bytes).unwrap();
// Deserialized metadata has the metadata header, which is different from the serialized one.
// Reference: TimelineMetaData::to_bytes()
let expected_metadata = {
let mut temp_metadata = original_metadata;
let body_bytes = temp_metadata
.body
.ser()
.expect("Cannot serialize the metadata body");
let metadata_size = METADATA_HDR_SIZE + body_bytes.len();
let hdr = TimelineMetadataHeader {
size: metadata_size as u16,
format_version: METADATA_FORMAT_VERSION,
checksum: crc32c::crc32c(&body_bytes),
};
temp_metadata.hdr = hdr;
temp_metadata
};
assert_eq!(deserialized_metadata, expected_metadata);
}
#[test]
fn test_metadata_bincode_serde_ensure_roundtrip() {
let original_metadata = TimelineMetadata::new(
@@ -561,6 +469,8 @@ mod tests {
crate::DEFAULT_PG_VERSION,
);
let expected_bytes = vec![
/* bincode length encoding bytes */
0, 0, 0, 0, 0, 0, 2, 0, // 8 bytes for the length of the serialized vector
/* TimelineMetadataHeader */
4, 37, 101, 34, 0, 70, 0, 4, // checksum, size, format_version (4 + 2 + 2)
/* TimelineMetadataBodyV2 */
@@ -590,7 +500,7 @@ mod tests {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0,
];
let metadata_ser_bytes = original_metadata.to_bytes().unwrap();
let metadata_ser_bytes = original_metadata.ser().unwrap();
assert_eq!(metadata_ser_bytes, expected_bytes);
let expected_metadata = {
@@ -608,7 +518,7 @@ mod tests {
temp_metadata.hdr = hdr;
temp_metadata
};
let des_metadata = TimelineMetadata::from_bytes(&metadata_ser_bytes).unwrap();
let des_metadata = TimelineMetadata::des(&metadata_ser_bytes).unwrap();
assert_eq!(des_metadata, expected_metadata);
}
}

View File

@@ -3,12 +3,11 @@
use camino::{Utf8DirEntry, Utf8Path, Utf8PathBuf};
use futures::StreamExt;
use hyper::StatusCode;
use itertools::Itertools;
use pageserver_api::key::Key;
use pageserver_api::models::LocationConfigMode;
use pageserver_api::shard::{
ShardCount, ShardIdentity, ShardIndex, ShardNumber, ShardStripeSize, TenantShardId,
ShardCount, ShardIdentity, ShardNumber, ShardStripeSize, TenantShardId,
};
use pageserver_api::upcall_api::ReAttachResponseTenant;
use rand::{distributions::Alphanumeric, Rng};
@@ -46,7 +45,7 @@ use crate::tenant::delete::DeleteTenantFlow;
use crate::tenant::span::debug_assert_current_span_has_tenant_id;
use crate::tenant::storage_layer::inmemory_layer;
use crate::tenant::timeline::ShutdownMode;
use crate::tenant::{AttachedTenantConf, GcError, SpawnMode, Tenant, TenantState};
use crate::tenant::{AttachedTenantConf, SpawnMode, Tenant, TenantState};
use crate::{InitializationOrder, IGNORED_TENANT_FILE_NAME, TEMP_FILE_SUFFIX};
use utils::crashsafe::path_with_suffix_extension;
@@ -55,7 +54,6 @@ use utils::generation::Generation;
use utils::id::{TenantId, TimelineId};
use super::delete::DeleteTenantError;
use super::remote_timeline_client::remote_tenant_path;
use super::secondary::SecondaryTenant;
use super::timeline::detach_ancestor::PreparedTimelineDetach;
use super::TenantSharedResources;
@@ -129,8 +127,6 @@ pub(crate) enum ShardSelector {
First,
/// Pick the shard that holds this key
Page(Key),
/// The shard ID is known: pick the given shard
Known(ShardIndex),
}
/// A convenience for use with the re_attach ControlPlaneClient function: rather
@@ -1371,7 +1367,7 @@ impl TenantManager {
&self,
tenant_shard_id: TenantShardId,
activation_timeout: Duration,
) -> Result<StatusCode, DeleteTenantError> {
) -> Result<(), DeleteTenantError> {
super::span::debug_assert_current_span_has_tenant_id();
// We acquire a SlotGuard during this function to protect against concurrent
// changes while the ::prepare phase of DeleteTenantFlow executes, but then
@@ -1384,79 +1380,18 @@ impl TenantManager {
//
// See https://github.com/neondatabase/neon/issues/5080
// Tenant deletion can happen two ways:
// - Legacy: called on an attached location. The attached Tenant object stays alive in Stopping
// state until deletion is complete.
// - New: called on a pageserver without an attached location. We proceed with deletion from
// remote storage.
//
// See https://github.com/neondatabase/neon/issues/5080 for more context on this transition.
let slot_guard =
tenant_map_acquire_slot(&tenant_shard_id, TenantSlotAcquireMode::MustExist)?;
let slot_guard = tenant_map_acquire_slot(&tenant_shard_id, TenantSlotAcquireMode::Any)?;
match &slot_guard.old_value {
Some(TenantSlot::Attached(tenant)) => {
// Legacy deletion flow: the tenant remains attached, goes to Stopping state, and
// deletion will be resumed across restarts.
let tenant = tenant.clone();
return self
.delete_tenant_attached(slot_guard, tenant, activation_timeout)
.await;
// unwrap is safe because we used MustExist mode when acquiring
let tenant = match slot_guard.get_old_value().as_ref().unwrap() {
TenantSlot::Attached(tenant) => tenant.clone(),
_ => {
// Express "not attached" as equivalent to "not found"
return Err(DeleteTenantError::NotAttached);
}
Some(TenantSlot::Secondary(secondary_tenant)) => {
secondary_tenant.shutdown().await;
let local_tenant_directory = self.conf.tenant_path(&tenant_shard_id);
let tmp_dir = safe_rename_tenant_dir(&local_tenant_directory)
.await
.with_context(|| {
format!("local tenant directory {local_tenant_directory:?} rename")
})?;
spawn_background_purge(tmp_dir);
}
Some(TenantSlot::InProgress(_)) => unreachable!(),
None => {}
};
// Fall through: local state for this tenant is no longer present, proceed with remote delete
let remote_path = remote_tenant_path(&tenant_shard_id);
let keys = match self
.resources
.remote_storage
.list(
Some(&remote_path),
remote_storage::ListingMode::NoDelimiter,
None,
&self.cancel,
)
.await
{
Ok(listing) => listing.keys,
Err(remote_storage::DownloadError::Cancelled) => {
return Err(DeleteTenantError::Cancelled)
}
Err(remote_storage::DownloadError::NotFound) => return Ok(StatusCode::NOT_FOUND),
Err(other) => return Err(DeleteTenantError::Other(anyhow::anyhow!(other))),
};
if keys.is_empty() {
tracing::info!("Remote storage already deleted");
} else {
tracing::info!("Deleting {} keys from remote storage", keys.len());
self.resources
.remote_storage
.delete_objects(&keys, &self.cancel)
.await?;
}
// Callers use 404 as success for deletions, for historical reasons.
Ok(StatusCode::NOT_FOUND)
}
async fn delete_tenant_attached(
&self,
slot_guard: SlotGuard,
tenant: Arc<Tenant>,
activation_timeout: Duration,
) -> Result<StatusCode, DeleteTenantError> {
match tenant.current_state() {
TenantState::Broken { .. } | TenantState::Stopping { .. } => {
// If deletion is already in progress, return success (the semantics of this
@@ -1466,7 +1401,7 @@ impl TenantManager {
// The `delete_progress` lock is held: deletion is already happening
// in the bacckground
slot_guard.revert();
return Ok(StatusCode::ACCEPTED);
return Ok(());
}
}
_ => {
@@ -1499,8 +1434,7 @@ impl TenantManager {
// The Tenant goes back into the map in Stopping state, it will eventually be removed by DeleteTenantFLow
slot_guard.revert();
let () = result?;
Ok(StatusCode::ACCEPTED)
result
}
#[instrument(skip_all, fields(tenant_id=%tenant.get_tenant_shard_id().tenant_id, shard_id=%tenant.get_tenant_shard_id().shard_slug(), new_shard_count=%new_shard_count.literal()))]
@@ -2133,11 +2067,6 @@ impl TenantManager {
return ShardResolveResult::Found(tenant.clone());
}
}
ShardSelector::Known(shard)
if tenant.shard_identity.shard_index() == shard =>
{
return ShardResolveResult::Found(tenant.clone());
}
_ => continue,
}
}
@@ -2897,13 +2826,7 @@ pub(crate) async fn immediate_gc(
}
}
result.map_err(|e| match e {
GcError::TenantCancelled | GcError::TimelineCancelled => ApiError::ShuttingDown,
GcError::TimelineNotFound => {
ApiError::NotFound(anyhow::anyhow!("Timeline not found").into())
}
other => ApiError::InternalServerError(anyhow::anyhow!(other)),
})
result.map_err(ApiError::InternalServerError)
}
#[cfg(test)]

View File

@@ -91,7 +91,8 @@
//!
//! The *actual* remote state lags behind the *desired* remote state while
//! there are in-flight operations.
//! We keep track of the desired remote state in [`UploadQueueInitialized::dirty`].
//! We keep track of the desired remote state in
//! [`UploadQueueInitialized::latest_files`] and [`UploadQueueInitialized::latest_metadata`].
//! It is initialized based on the [`IndexPart`] that was passed during init
//! and updated with every `schedule_*` function call.
//! All this is necessary necessary to compute the future [`IndexPart`]s
@@ -114,7 +115,8 @@
//!
//! # Completion
//!
//! Once an operation has completed, we update [`UploadQueueInitialized::clean`] immediately,
//! Once an operation has completed, we update
//! [`UploadQueueInitialized::projected_remote_consistent_lsn`] immediately,
//! and submit a request through the DeletionQueue to update
//! [`UploadQueueInitialized::visible_remote_consistent_lsn`] after it has
//! validated that our generation is not stale. It is this visible value
@@ -195,7 +197,6 @@ pub(crate) use upload::upload_initdb_dir;
use utils::backoff::{
self, exponential_backoff, DEFAULT_BASE_BACKOFF_SECONDS, DEFAULT_MAX_BACKOFF_SECONDS,
};
use utils::pausable_failpoint;
use std::collections::{HashMap, VecDeque};
use std::sync::atomic::{AtomicU32, Ordering};
@@ -414,7 +415,6 @@ impl RemoteTimelineClient {
Ok(())
}
/// Returns `None` if nothing is yet uplodaded, `Some(disk_consistent_lsn)` otherwise.
pub fn remote_consistent_lsn_projected(&self) -> Option<Lsn> {
match &mut *self.upload_queue.lock().unwrap() {
UploadQueue::Uninitialized => None,
@@ -441,11 +441,13 @@ impl RemoteTimelineClient {
/// Returns true if this timeline was previously detached at this Lsn and the remote timeline
/// client is currently initialized.
pub(crate) fn is_previous_ancestor_lsn(&self, lsn: Lsn) -> bool {
// technically this is a dirty read, but given how timeline detach ancestor is implemented
// via tenant restart, the lineage has always been uploaded.
self.upload_queue
.lock()
.unwrap()
.initialized_mut()
.map(|uq| uq.clean.0.lineage.is_previous_ancestor_lsn(lsn))
.map(|uq| uq.latest_lineage.is_previous_ancestor_lsn(lsn))
.unwrap_or(false)
}
@@ -454,6 +456,7 @@ impl RemoteTimelineClient {
current_remote_index_part
.layer_metadata
.values()
// If we don't have the file size for the layer, don't account for it in the metric.
.map(|ilmd| ilmd.file_size)
.sum()
} else {
@@ -581,9 +584,9 @@ impl RemoteTimelineClient {
// As documented in the struct definition, it's ok for latest_metadata to be
// ahead of what's _actually_ on the remote during index upload.
upload_queue.dirty.metadata = metadata.clone();
upload_queue.latest_metadata = metadata.clone();
self.schedule_index_upload(upload_queue)?;
self.schedule_index_upload(upload_queue);
Ok(())
}
@@ -602,9 +605,9 @@ impl RemoteTimelineClient {
let mut guard = self.upload_queue.lock().unwrap();
let upload_queue = guard.initialized_mut()?;
upload_queue.dirty.metadata.apply(update);
upload_queue.latest_metadata.apply(update);
self.schedule_index_upload(upload_queue)?;
self.schedule_index_upload(upload_queue);
Ok(())
}
@@ -616,8 +619,8 @@ impl RemoteTimelineClient {
) -> anyhow::Result<()> {
let mut guard = self.upload_queue.lock().unwrap();
let upload_queue = guard.initialized_mut()?;
upload_queue.dirty.last_aux_file_policy = last_aux_file_policy;
self.schedule_index_upload(upload_queue)?;
upload_queue.last_aux_file_policy = last_aux_file_policy;
self.schedule_index_upload(upload_queue);
Ok(())
}
///
@@ -635,44 +638,30 @@ impl RemoteTimelineClient {
let upload_queue = guard.initialized_mut()?;
if upload_queue.latest_files_changes_since_metadata_upload_scheduled > 0 {
self.schedule_index_upload(upload_queue)?;
self.schedule_index_upload(upload_queue);
}
Ok(())
}
/// Launch an index-file upload operation in the background (internal function)
fn schedule_index_upload(
self: &Arc<Self>,
upload_queue: &mut UploadQueueInitialized,
) -> anyhow::Result<()> {
let disk_consistent_lsn = upload_queue.dirty.metadata.disk_consistent_lsn();
// fix up the duplicated field
upload_queue.dirty.disk_consistent_lsn = disk_consistent_lsn;
// make sure it serializes before doing it in perform_upload_task so that it doesn't
// look like a retryable error
let void = std::io::sink();
serde_json::to_writer(void, &upload_queue.dirty).context("serialize index_part.json")?;
let index_part = &upload_queue.dirty;
fn schedule_index_upload(self: &Arc<Self>, upload_queue: &mut UploadQueueInitialized) {
let disk_consistent_lsn = upload_queue.latest_metadata.disk_consistent_lsn();
info!(
"scheduling metadata upload up to consistent LSN {disk_consistent_lsn} with {} files ({} changed)",
index_part.layer_metadata.len(),
upload_queue.latest_files.len(),
upload_queue.latest_files_changes_since_metadata_upload_scheduled,
);
let op = UploadOp::UploadMetadata {
uploaded: Box::new(index_part.clone()),
};
let index_part = IndexPart::from(&*upload_queue);
let op = UploadOp::UploadMetadata(Box::new(index_part), disk_consistent_lsn);
self.metric_begin(&op);
upload_queue.queued_operations.push_back(op);
upload_queue.latest_files_changes_since_metadata_upload_scheduled = 0;
// Launch the task immediately, if possible
self.launch_queued_tasks(upload_queue);
Ok(())
}
pub(crate) async fn schedule_reparenting_and_wait(
@@ -685,16 +674,16 @@ impl RemoteTimelineClient {
let mut guard = self.upload_queue.lock().unwrap();
let upload_queue = guard.initialized_mut()?;
let Some(prev) = upload_queue.dirty.metadata.ancestor_timeline() else {
let Some(prev) = upload_queue.latest_metadata.ancestor_timeline() else {
return Err(anyhow::anyhow!(
"cannot reparent without a current ancestor"
));
};
upload_queue.dirty.metadata.reparent(new_parent);
upload_queue.dirty.lineage.record_previous_ancestor(&prev);
upload_queue.latest_metadata.reparent(new_parent);
upload_queue.latest_lineage.record_previous_ancestor(&prev);
self.schedule_index_upload(upload_queue)?;
self.schedule_index_upload(upload_queue);
self.schedule_barrier0(upload_queue)
};
@@ -715,17 +704,16 @@ impl RemoteTimelineClient {
let mut guard = self.upload_queue.lock().unwrap();
let upload_queue = guard.initialized_mut()?;
upload_queue.dirty.metadata.detach_from_ancestor(&adopted);
upload_queue.dirty.lineage.record_detaching(&adopted);
upload_queue.latest_metadata.detach_from_ancestor(&adopted);
upload_queue.latest_lineage.record_detaching(&adopted);
for layer in layers {
upload_queue
.dirty
.layer_metadata
.latest_files
.insert(layer.layer_desc().layer_name(), layer.metadata());
}
self.schedule_index_upload(upload_queue)?;
self.schedule_index_upload(upload_queue);
let barrier = self.schedule_barrier0(upload_queue);
self.launch_queued_tasks(upload_queue);
@@ -757,8 +745,7 @@ impl RemoteTimelineClient {
let metadata = layer.metadata();
upload_queue
.dirty
.layer_metadata
.latest_files
.insert(layer.layer_desc().layer_name(), metadata.clone());
upload_queue.latest_files_changes_since_metadata_upload_scheduled += 1;
@@ -788,8 +775,8 @@ impl RemoteTimelineClient {
let mut guard = self.upload_queue.lock().unwrap();
let upload_queue = guard.initialized_mut()?;
let with_metadata = self
.schedule_unlinking_of_layers_from_index_part0(upload_queue, names.iter().cloned())?;
let with_metadata =
self.schedule_unlinking_of_layers_from_index_part0(upload_queue, names.iter().cloned());
self.schedule_deletion_of_unlinked0(upload_queue, with_metadata);
@@ -813,7 +800,7 @@ impl RemoteTimelineClient {
let names = gc_layers.iter().map(|x| x.layer_desc().layer_name());
self.schedule_unlinking_of_layers_from_index_part0(upload_queue, names)?;
self.schedule_unlinking_of_layers_from_index_part0(upload_queue, names);
self.launch_queued_tasks(upload_queue);
@@ -826,7 +813,7 @@ impl RemoteTimelineClient {
self: &Arc<Self>,
upload_queue: &mut UploadQueueInitialized,
names: I,
) -> anyhow::Result<Vec<(LayerName, LayerFileMetadata)>>
) -> Vec<(LayerName, LayerFileMetadata)>
where
I: IntoIterator<Item = LayerName>,
{
@@ -836,7 +823,7 @@ impl RemoteTimelineClient {
let with_metadata: Vec<_> = names
.into_iter()
.filter_map(|name| {
let meta = upload_queue.dirty.layer_metadata.remove(&name);
let meta = upload_queue.latest_files.remove(&name);
if let Some(meta) = meta {
upload_queue.latest_files_changes_since_metadata_upload_scheduled += 1;
@@ -868,10 +855,10 @@ impl RemoteTimelineClient {
// index_part update, because that needs to be uploaded before we can actually delete the
// files.
if upload_queue.latest_files_changes_since_metadata_upload_scheduled > 0 {
self.schedule_index_upload(upload_queue)?;
self.schedule_index_upload(upload_queue);
}
Ok(with_metadata)
with_metadata
}
/// Schedules deletion for layer files which have previously been unlinked from the
@@ -962,7 +949,7 @@ impl RemoteTimelineClient {
let names = compacted_from.iter().map(|x| x.layer_desc().layer_name());
self.schedule_unlinking_of_layers_from_index_part0(upload_queue, names)?;
self.schedule_unlinking_of_layers_from_index_part0(upload_queue, names);
self.launch_queued_tasks(upload_queue);
Ok(())
@@ -1097,7 +1084,7 @@ impl RemoteTimelineClient {
let deleted_at = Utc::now().naive_utc();
stopped.deleted_at = SetDeletedFlagProgress::InProgress(deleted_at);
let mut index_part = stopped.upload_queue_for_deletion.dirty.clone();
let mut index_part = IndexPart::from(&stopped.upload_queue_for_deletion);
index_part.deleted_at = Some(deleted_at);
index_part
};
@@ -1205,7 +1192,7 @@ impl RemoteTimelineClient {
&self.storage_impl,
uploaded.local_path(),
&remote_path,
uploaded.metadata().file_size,
uploaded.metadata().file_size(),
cancel,
)
.await
@@ -1308,8 +1295,7 @@ impl RemoteTimelineClient {
stopped
.upload_queue_for_deletion
.dirty
.layer_metadata
.latest_files
.drain()
.map(|(file_name, meta)| {
remote_layer_path(
@@ -1446,7 +1432,7 @@ impl RemoteTimelineClient {
// Can always be scheduled.
true
}
UploadOp::UploadMetadata { .. } => {
UploadOp::UploadMetadata(_, _) => {
// These can only be performed after all the preceding operations
// have finished.
upload_queue.inprogress_tasks.is_empty()
@@ -1488,7 +1474,7 @@ impl RemoteTimelineClient {
UploadOp::UploadLayer(_, _) => {
upload_queue.num_inprogress_layer_uploads += 1;
}
UploadOp::UploadMetadata { .. } => {
UploadOp::UploadMetadata(_, _) => {
upload_queue.num_inprogress_metadata_uploads += 1;
}
UploadOp::Delete(_) => {
@@ -1587,7 +1573,7 @@ impl RemoteTimelineClient {
&self.storage_impl,
local_path,
&remote_path,
layer_metadata.file_size,
layer_metadata.file_size(),
&self.cancel,
)
.measure_remote_op(
@@ -1597,13 +1583,22 @@ impl RemoteTimelineClient {
)
.await
}
UploadOp::UploadMetadata { ref uploaded } => {
UploadOp::UploadMetadata(ref index_part, _lsn) => {
let mention_having_future_layers = if cfg!(feature = "testing") {
index_part
.layer_metadata
.keys()
.any(|x| x.is_in_future(*_lsn))
} else {
false
};
let res = upload::upload_index_part(
&self.storage_impl,
&self.tenant_shard_id,
&self.timeline_id,
self.generation,
uploaded,
index_part,
&self.cancel,
)
.measure_remote_op(
@@ -1613,21 +1608,10 @@ impl RemoteTimelineClient {
)
.await;
if res.is_ok() {
self.update_remote_physical_size_gauge(Some(uploaded));
let mention_having_future_layers = if cfg!(feature = "testing") {
uploaded
.layer_metadata
.keys()
.any(|x| x.is_in_future(uploaded.metadata.disk_consistent_lsn()))
} else {
false
};
self.update_remote_physical_size_gauge(Some(index_part));
if mention_having_future_layers {
// find rationale near crate::tenant::timeline::init::cleanup_future_layer
tracing::info!(
disk_consistent_lsn = %uploaded.metadata.disk_consistent_lsn(),
"uploaded an index_part.json with future layers -- this is ok! if shutdown now, expect future layer cleanup"
);
tracing::info!(disk_consistent_lsn=%_lsn, "uploaded an index_part.json with future layers -- this is ok! if shutdown now, expect future layer cleanup");
}
}
res
@@ -1728,23 +1712,11 @@ impl RemoteTimelineClient {
upload_queue.num_inprogress_layer_uploads -= 1;
None
}
UploadOp::UploadMetadata { ref uploaded } => {
UploadOp::UploadMetadata(_, lsn) => {
upload_queue.num_inprogress_metadata_uploads -= 1;
// XXX monotonicity check?
// the task id is reused as a monotonicity check for storing the "clean"
// IndexPart.
let last_updater = upload_queue.clean.1;
let is_later = last_updater.is_some_and(|task_id| task_id < task.task_id);
let monotone = is_later || last_updater.is_none();
assert!(monotone, "no two index uploads should be completing at the same time, prev={last_updater:?}, task.task_id={}", task.task_id);
// not taking ownership is wasteful
upload_queue.clean.0.clone_from(uploaded);
upload_queue.clean.1 = Some(task.task_id);
let lsn = upload_queue.clean.0.metadata.disk_consistent_lsn();
upload_queue.projected_remote_consistent_lsn = Some(lsn);
if self.generation.is_none() {
// Legacy mode: skip validating generation
upload_queue.visible_remote_consistent_lsn.store(lsn);
@@ -1796,9 +1768,9 @@ impl RemoteTimelineClient {
UploadOp::UploadLayer(_, m) => (
RemoteOpFileKind::Layer,
RemoteOpKind::Upload,
RemoteTimelineClientMetricsCallTrackSize::Bytes(m.file_size),
RemoteTimelineClientMetricsCallTrackSize::Bytes(m.file_size()),
),
UploadOp::UploadMetadata { .. } => (
UploadOp::UploadMetadata(_, _) => (
RemoteOpFileKind::Index,
RemoteOpKind::Upload,
DontTrackSize {
@@ -1874,9 +1846,11 @@ impl RemoteTimelineClient {
// Deletion is not really perf sensitive so there shouldnt be any problems with cloning a fraction of it.
let upload_queue_for_deletion = UploadQueueInitialized {
task_counter: 0,
dirty: initialized.dirty.clone(),
clean: initialized.clean.clone(),
latest_files: initialized.latest_files.clone(),
latest_files_changes_since_metadata_upload_scheduled: 0,
latest_metadata: initialized.latest_metadata.clone(),
latest_lineage: initialized.latest_lineage.clone(),
projected_remote_consistent_lsn: None,
visible_remote_consistent_lsn: initialized
.visible_remote_consistent_lsn
.clone(),
@@ -1889,6 +1863,7 @@ impl RemoteTimelineClient {
dangling_files: HashMap::default(),
shutting_down: false,
shutdown_ready: Arc::new(tokio::sync::Semaphore::new(0)),
last_aux_file_policy: initialized.last_aux_file_policy,
};
let upload_queue = std::mem::replace(

Some files were not shown because too many files have changed in this diff Show More