rust/neon
1
0
Fork 0
mirror of https://github.com/neondatabase/neon.git synced 2026-06-04 22:10:39 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #2573 from neondatabase/main

Browse Source 
Release 2022-10-06
This commit is contained in:
Stas Kelvich
2022-10-07 11:07:06 +03:00
committed by GitHub
parent e7939a527a 687ba81366
commit 9316cb1b1f
267 changed files with 18794 additions and 9919 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
.cargo/config.toml
View File

@@ -11,3 +11,6 @@ opt-level = 3
[profile.dev]
# Turn on a small amount of optimization in Development mode.
opt-level = 1
[alias]
build_testing = ["build", "--features", "testing"]

5
.dockerignore
View File

@@ -1,5 +1,6 @@
*
!rust-toolchain.toml
!Cargo.toml
!Cargo.lock
!Makefile
@@ -13,6 +14,8 @@
!pgxn/
!proxy/
!safekeeper/
!vendor/postgres/
!vendor/postgres-v14/
!vendor/postgres-v15/
!workspace_hack/
!neon_local/
!scripts/ninstall.sh

2
.github/ISSUE_TEMPLATE/epic-template.md vendored
View File

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

8
.github/actions/allure-report/action.yml vendored
View File

@@ -11,6 +11,10 @@ inputs:
test_selection:
description: '`test_selector` from run-python-test-set action'
required: false
outputs:
report-url:
description: 'Allure report URL'
value: ${{ steps.generate-report.outputs.report-url }}
runs:
using: "composite"
@@ -182,7 +186,7 @@ runs:
aws s3 cp --only-show-errors ./index.html "s3://${BUCKET}/${REPORT_PREFIX}/latest/index.html"
echo "[Allure Report](${REPORT_URL})" >> ${GITHUB_STEP_SUMMARY}
echo "::set-output name=REPORT_URL::${REPORT_URL}"
echo "::set-output name=report-url::${REPORT_URL}"
- name: Release Allure lock
if: ${{ inputs.action == 'generate' && always() }}
@@ -200,7 +204,7 @@ runs:
- uses: actions/github-script@v6
if: ${{ inputs.action == 'generate' && always() }}
env:
REPORT_URL: ${{ steps.generate-report.outputs.REPORT_URL }}
REPORT_URL: ${{ steps.generate-report.outputs.report-url }}
BUILD_TYPE: ${{ inputs.build_type }}
SHA: ${{ github.event.pull_request.head.sha || github.sha }}
with:

9
.github/actions/download/action.yml vendored
View File

@@ -12,6 +12,9 @@ inputs:
description: "Allow to skip if file doesn't exist, fail otherwise"
default: false
required: false
prefix:
description: "S3 prefix. Default is '${GITHUB_RUN_ID}/${GITHUB_RUN_ATTEMPT}'"
required: false
runs:
using: "composite"
@@ -23,18 +26,18 @@ runs:
TARGET: ${{ inputs.path }}
ARCHIVE: /tmp/downloads/${{ inputs.name }}.tar.zst
SKIP_IF_DOES_NOT_EXIST: ${{ inputs.skip-if-does-not-exist }}
PREFIX: artifacts/${{ inputs.prefix || format('{0}/{1}', github.run_id, github.run_attempt) }}
run: |
BUCKET=neon-github-public-dev
PREFIX=artifacts/${GITHUB_RUN_ID}
FILENAME=$(basename $ARCHIVE)
S3_KEY=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${PREFIX} | jq -r '.Contents[].Key' | grep ${FILENAME} | sort --version-sort | tail -1 || true)
S3_KEY=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${PREFIX%$GITHUB_RUN_ATTEMPT} | jq -r '.Contents[].Key' | grep ${FILENAME} | sort --version-sort | tail -1 || true)
if [ -z "${S3_KEY}" ]; then
if [ "${SKIP_IF_DOES_NOT_EXIST}" = "true" ]; then
echo '::set-output name=SKIPPED::true'
exit 0
else
echo 2>&1 "Neither s3://${BUCKET}/${PREFIX}/${GITHUB_RUN_ATTEMPT}/${FILENAME} nor its version from previous attempts exist"
echo 2>&1 "Neither s3://${BUCKET}/${PREFIX}/${FILENAME} nor its version from previous attempts exist"
exit 1
fi
fi

82
.github/actions/neon-project-create/action.yml vendored Normal file
View File

@@ -0,0 +1,82 @@
name: 'Create Neon Project'
description: 'Create Neon Project using API'
inputs:
api_key:
desctiption: 'Neon API key'
required: true
environment:
desctiption: 'dev (aka captest) or stage'
required: true
region_id:
desctiption: 'Region ID, if not set the project will be created in the default region'
required: false
outputs:
dsn:
description: 'Created Project DSN (for main database)'
value: ${{ steps.create-neon-project.outputs.dsn }}
project_id:
description: 'Created Project ID'
value: ${{ steps.create-neon-project.outputs.project_id }}
runs:
using: "composite"
steps:
- name: Parse Input
id: parse-input
shell: bash -euxo pipefail {0}
run: |
case "${ENVIRONMENT}" in
dev)
API_HOST=console.dev.neon.tech
REGION_ID=${REGION_ID:-eu-west-1}
;;
staging)
API_HOST=console.stage.neon.tech
REGION_ID=${REGION_ID:-us-east-1}
;;
*)
echo 2>&1 "Unknown environment=${ENVIRONMENT}. Allowed 'dev' or 'staging' only"
exit 1
;;
esac
echo "::set-output name=api_host::${API_HOST}"
echo "::set-output name=region_id::${REGION_ID}"
env:
ENVIRONMENT: ${{ inputs.environment }}
REGION_ID: ${{ inputs.region_id }}
- name: Create Neon Project
id: create-neon-project
# A shell without `set -x` to not to expose password/dsn in logs
shell: bash -euo pipefail {0}
run: |
project=$(curl \
"https://${API_HOST}/api/v1/projects" \
--fail \
--header "Accept: application/json" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${API_KEY}" \
--data "{
\"project\": {
\"name\": \"Created by actions/neon-project-create; GITHUB_RUN_ID=${GITHUB_RUN_ID}\",
\"platform_id\": \"aws\",
\"region_id\": \"${REGION_ID}\",
\"settings\": { }
}
}")
# Mask password
echo "::add-mask::$(echo $project | jq --raw-output '.roles[] | select(.name != "web_access") | .password')"
dsn=$(echo $project | jq --raw-output '.roles[] | select(.name != "web_access") | .dsn')/main
echo "::add-mask::${dsn}"
echo "::set-output name=dsn::${dsn}"
project_id=$(echo $project | jq --raw-output '.id')
echo "::set-output name=project_id::${project_id}"
env:
API_KEY: ${{ inputs.api_key }}
API_HOST: ${{ steps.parse-input.outputs.api_host }}
REGION_ID: ${{ steps.parse-input.outputs.region_id }}

54
.github/actions/neon-project-delete/action.yml vendored Normal file
View File

@@ -0,0 +1,54 @@
name: 'Delete Neon Project'
description: 'Delete Neon Project using API'
inputs:
api_key:
desctiption: 'Neon API key'
required: true
environment:
desctiption: 'dev (aka captest) or stage'
required: true
project_id:
desctiption: 'ID of the Project to delete'
required: true
runs:
using: "composite"
steps:
- name: Parse Input
id: parse-input
shell: bash -euxo pipefail {0}
run: |
case "${ENVIRONMENT}" in
dev)
API_HOST=console.dev.neon.tech
;;
staging)
API_HOST=console.stage.neon.tech
;;
*)
echo 2>&1 "Unknown environment=${ENVIRONMENT}. Allowed 'dev' or 'staging' only"
exit 1
;;
esac
echo "::set-output name=api_host::${API_HOST}"
env:
ENVIRONMENT: ${{ inputs.environment }}
- name: Delete Neon Project
shell: bash -euxo pipefail {0}
run: |
# Allow PROJECT_ID to be empty/null for cases when .github/actions/neon-project-create failed
if [ -n "${PROJECT_ID}" ]; then
curl -X "POST" \
"https://${API_HOST}/api/v1/projects/${PROJECT_ID}/delete" \
--fail \
--header "Accept: application/json" \
--header "Content-Type: application/json" \
--header "Authorization: Bearer ${API_KEY}"
fi
env:
API_KEY: ${{ inputs.api_key }}
PROJECT_ID: ${{ inputs.project_id }}
API_HOST: ${{ steps.parse-input.outputs.api_host }}

28
.github/actions/run-python-test-set/action.yml vendored
View File

@@ -5,9 +5,6 @@ inputs:
build_type:
description: 'Type of Rust (neon) and C (postgres) builds. Must be "release" or "debug", or "remote" for the remote cluster'
required: true
rust_toolchain:
description: 'Rust toolchain version to fetch the caches'
required: false
test_selection:
description: 'A python test suite to run'
required: true
@@ -55,7 +52,7 @@ runs:
if: inputs.build_type != 'remote'
uses: ./.github/actions/download
with:
name: neon-${{ runner.os }}-${{ inputs.build_type }}-${{ inputs.rust_toolchain }}-artifact
name: neon-${{ runner.os }}-${{ inputs.build_type }}-artifact
path: /tmp/neon
- name: Checkout
@@ -80,15 +77,16 @@ runs:
env:
NEON_BIN: /tmp/neon/bin
TEST_OUTPUT: /tmp/test_output
# this variable will be embedded in perf test report
# and is needed to distinguish different environments
PLATFORM: github-actions-selfhosted
BUILD_TYPE: ${{ inputs.build_type }}
AWS_ACCESS_KEY_ID: ${{ inputs.real_s3_access_key_id }}
AWS_SECRET_ACCESS_KEY: ${{ inputs.real_s3_secret_access_key }}
shell: bash -euxo pipefail {0}
run: |
# PLATFORM will be embedded in the perf test report
# and it is needed to distinguish different environments
export PLATFORM=${PLATFORM:-github-actions-selfhosted}
export POSTGRES_DISTRIB_DIR=${POSTGRES_DISTRIB_DIR:-/tmp/neon/pg_install}
export DEFAULT_PG_VERSION=${DEFAULT_PG_VERSION:-14}
if [ "${BUILD_TYPE}" = "remote" ]; then
export REMOTE_ENV=1
@@ -115,10 +113,8 @@ runs:
fi
if [[ "${{ inputs.save_perf_report }}" == "true" ]]; then
if [[ "$GITHUB_REF" == "refs/heads/main" ]]; then
mkdir -p "$PERF_REPORT_DIR"
EXTRA_PARAMS="--out-dir $PERF_REPORT_DIR $EXTRA_PARAMS"
fi
mkdir -p "$PERF_REPORT_DIR"
EXTRA_PARAMS="--out-dir $PERF_REPORT_DIR $EXTRA_PARAMS"
fi
if [[ "${{ inputs.build_type }}" == "debug" ]]; then
@@ -131,7 +127,7 @@ runs:
# Wake up the cluster if we use remote neon instance
if [ "${{ inputs.build_type }}" = "remote" ] && [ -n "${BENCHMARK_CONNSTR}" ]; then
${POSTGRES_DISTRIB_DIR}/bin/psql ${BENCHMARK_CONNSTR} -c "SELECT version();"
${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/psql ${BENCHMARK_CONNSTR} -c "SELECT version();"
fi
# Run the tests.
@@ -153,11 +149,9 @@ runs:
-rA $TEST_SELECTION $EXTRA_PARAMS
if [[ "${{ inputs.save_perf_report }}" == "true" ]]; then
if [[ "$GITHUB_REF" == "refs/heads/main" ]]; then
export REPORT_FROM="$PERF_REPORT_DIR"
export REPORT_TO=local
scripts/generate_and_push_perf_report.sh
fi
export REPORT_FROM="$PERF_REPORT_DIR"
export REPORT_TO="$PLATFORM"
scripts/generate_and_push_perf_report.sh
fi
- name: Create Allure report

9
.github/actions/upload/action.yml vendored
View File

@@ -7,6 +7,9 @@ inputs:
path:
description: "A directory or file to upload"
required: true
prefix:
description: "S3 prefix. Default is '${GITHUB_RUN_ID}/${GITHUB_RUN_ATTEMPT}'"
required: false
runs:
using: "composite"
@@ -42,14 +45,14 @@ runs:
env:
SOURCE: ${{ inputs.path }}
ARCHIVE: /tmp/uploads/${{ inputs.name }}.tar.zst
PREFIX: artifacts/${{ inputs.prefix || format('{0}/{1}', github.run_id, github.run_attempt) }}
run: |
BUCKET=neon-github-public-dev
PREFIX=artifacts/${GITHUB_RUN_ID}
FILENAME=$(basename $ARCHIVE)
FILESIZE=$(du -sh ${ARCHIVE} | cut -f1)
time aws s3 mv --only-show-errors ${ARCHIVE} s3://${BUCKET}/${PREFIX}/${GITHUB_RUN_ATTEMPT}/${FILENAME}
time aws s3 mv --only-show-errors ${ARCHIVE} s3://${BUCKET}/${PREFIX}/${FILENAME}
# Ref https://docs.github.com/en/actions/using-workflows/workflow-commands-for-github-actions#adding-a-job-summary
echo "[${FILENAME}](https://${BUCKET}.s3.amazonaws.com/${PREFIX}/${GITHUB_RUN_ATTEMPT}/${FILENAME}) ${FILESIZE}" >> ${GITHUB_STEP_SUMMARY}
echo "[${FILENAME}](https://${BUCKET}.s3.amazonaws.com/${PREFIX}/${FILENAME}) ${FILESIZE}" >> ${GITHUB_STEP_SUMMARY}

4
.github/ansible/deploy.yaml vendored
View File

@@ -58,7 +58,7 @@
creates: "/storage/pageserver/data/tenants"
environment:
NEON_REPO_DIR: "/storage/pageserver/data"
LD_LIBRARY_PATH: "/usr/local/lib"
LD_LIBRARY_PATH: "/usr/local/v14/lib"
become: true
tags:
- pageserver
@@ -132,7 +132,7 @@
creates: "/storage/safekeeper/data/safekeeper.id"
environment:
NEON_REPO_DIR: "/storage/safekeeper/data"
LD_LIBRARY_PATH: "/usr/local/lib"
LD_LIBRARY_PATH: "/usr/local/v14/lib"
become: true
tags:
- safekeeper

6
.github/ansible/get_binaries.sh vendored
View File

@@ -21,10 +21,14 @@ docker pull --quiet neondatabase/neon:${DOCKER_TAG}
ID=$(docker create neondatabase/neon:${DOCKER_TAG})
docker cp ${ID}:/data/postgres_install.tar.gz .
tar -xzf postgres_install.tar.gz -C neon_install
mkdir neon_install/bin/
docker cp ${ID}:/usr/local/bin/pageserver neon_install/bin/
docker cp ${ID}:/usr/local/bin/safekeeper neon_install/bin/
docker cp ${ID}:/usr/local/bin/proxy neon_install/bin/
docker cp ${ID}:/usr/local/bin/postgres neon_install/bin/
docker cp ${ID}:/usr/local/v14/bin/ neon_install/v14/bin/
docker cp ${ID}:/usr/local/v15/bin/ neon_install/v15/bin/
docker cp ${ID}:/usr/local/v14/lib/ neon_install/v14/lib/
docker cp ${ID}:/usr/local/v15/lib/ neon_install/v15/lib/
docker rm -vf ${ID}
# store version to file (for ansible playbooks) and create binaries tarball

5
.github/ansible/staging.hosts vendored
View File

@@ -2,11 +2,16 @@
#zenith-us-stage-ps-1 console_region_id=27
zenith-us-stage-ps-2 console_region_id=27
zenith-us-stage-ps-3 console_region_id=27
zenith-us-stage-ps-4 console_region_id=27
zenith-us-stage-test-ps-1 console_region_id=28
[safekeepers]
zenith-us-stage-sk-4 console_region_id=27
zenith-us-stage-sk-5 console_region_id=27
zenith-us-stage-sk-6 console_region_id=27
zenith-us-stage-test-sk-1 console_region_id=28
zenith-us-stage-test-sk-2 console_region_id=28
zenith-us-stage-test-sk-3 console_region_id=28
[storage:children]
pageservers

2
.github/ansible/systemd/pageserver.service vendored
View File

@@ -5,7 +5,7 @@ After=network.target auditd.service
[Service]
Type=simple
User=pageserver
Environment=RUST_BACKTRACE=1 NEON_REPO_DIR=/storage/pageserver LD_LIBRARY_PATH=/usr/local/lib
Environment=RUST_BACKTRACE=1 NEON_REPO_DIR=/storage/pageserver LD_LIBRARY_PATH=/usr/local/v14/lib
ExecStart=/usr/local/bin/pageserver -c "pg_distrib_dir='/usr/local'" -c "listen_pg_addr='0.0.0.0:6400'" -c "listen_http_addr='0.0.0.0:9898'" -c "broker_endpoints=['{{ etcd_endpoints }}']" -D /storage/pageserver/data
ExecReload=/bin/kill -HUP $MAINPID
KillMode=mixed

2
.github/ansible/systemd/safekeeper.service vendored
View File

@@ -5,7 +5,7 @@ After=network.target auditd.service
[Service]
Type=simple
User=safekeeper
Environment=RUST_BACKTRACE=1 NEON_REPO_DIR=/storage/safekeeper/data LD_LIBRARY_PATH=/usr/local/lib
Environment=RUST_BACKTRACE=1 NEON_REPO_DIR=/storage/safekeeper/data LD_LIBRARY_PATH=/usr/local/v14/lib
ExecStart=/usr/local/bin/safekeeper -l {{ inventory_hostname }}.local:6500 --listen-http {{ inventory_hostname }}.local:7676 -D /storage/safekeeper/data --broker-endpoints={{ etcd_endpoints }} --remote-storage='{bucket_name="{{bucket_name}}", bucket_region="{{bucket_region}}", prefix_in_bucket="{{ env_name }}/wal"}'
ExecReload=/bin/kill -HUP $MAINPID
KillMode=mixed

1
.github/helm-values/neon-stress.proxy.yaml vendored
View File

@@ -1,6 +1,7 @@
fullnameOverride: "neon-stress-proxy"
settings:
authBackend: "link"
authEndpoint: "https://console.dev.neon.tech/authenticate_proxy_request/"
uri: "https://console.dev.neon.tech/psql_session/"

1
.github/helm-values/production.proxy.yaml vendored
View File

@@ -1,4 +1,5 @@
settings:
authBackend: "link"
authEndpoint: "https://console.neon.tech/authenticate_proxy_request/"
uri: "https://console.neon.tech/psql_session/"

1
.github/helm-values/staging.proxy.yaml vendored
View File

@@ -5,6 +5,7 @@ image:
repository: neondatabase/neon
settings:
authBackend: "link"
authEndpoint: "https://console.stage.neon.tech/authenticate_proxy_request/"
uri: "https://console.stage.neon.tech/psql_session/"

153
.github/workflows/benchmarking.yml vendored
View File

@@ -11,9 +11,20 @@ on:
# │ │ ┌───────────── day of the month (1 - 31)
# │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
# │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
- cron: '36 4 * * *' # run once a day, timezone is utc
- cron: '0 3 * * *' # run once a day, timezone is utc
workflow_dispatch: # adds ability to run this manually
inputs:
environment:
description: 'Environment to run remote tests on (dev or staging)'
required: false
region_id:
description: 'Use a particular region. If not set the default region will be used'
required: false
save_perf_report:
type: boolean
description: 'Publish perf report or not. If not set, the report is published only for the main branch'
required: false
defaults:
run:
@@ -35,7 +46,8 @@ jobs:
runs-on: [self-hosted, zenith-benchmarker]
env:
POSTGRES_DISTRIB_DIR: "/usr/pgsql-14"
POSTGRES_DISTRIB_DIR: /tmp/pg_install
DEFAULT_PG_VERSION: 14
steps:
- name: Checkout zenith repo
@@ -60,21 +72,14 @@ jobs:
echo Poetry
poetry --version
echo Pgbench
$POSTGRES_DISTRIB_DIR/bin/pgbench --version
${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/pgbench --version
# FIXME cluster setup is skipped due to various changes in console API
# for now pre created cluster is used. When API gain some stability
# after massive changes dynamic cluster setup will be revived.
# So use pre created cluster. It needs to be started manually, but stop is automatic after 5 minutes of inactivity
- name: Setup cluster
env:
BENCHMARK_CONNSTR: "${{ secrets.BENCHMARK_STAGING_CONNSTR }}"
run: |
set -e
echo "Starting cluster"
# wake up the cluster
$POSTGRES_DISTRIB_DIR/bin/psql $BENCHMARK_CONNSTR -c "SELECT 1"
- name: Create Neon Project
id: create-neon-project
uses: ./.github/actions/neon-project-create
with:
environment: ${{ github.event.inputs.environment || 'staging' }}
api_key: ${{ ( github.event.inputs.environment || 'staging' ) == 'staging' && secrets.NEON_STAGING_API_KEY || secrets.NEON_CAPTEST_API_KEY }}
- name: Run benchmark
# pgbench is installed system wide from official repo
@@ -97,7 +102,7 @@ jobs:
TEST_PG_BENCH_DURATIONS_MATRIX: "300"
TEST_PG_BENCH_SCALES_MATRIX: "10,100"
PLATFORM: "neon-staging"
BENCHMARK_CONNSTR: "${{ secrets.BENCHMARK_STAGING_CONNSTR }}"
BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
REMOTE_ENV: "1" # indicate to test harness that we do not have zenith binaries locally
run: |
# just to be sure that no data was cached on self hosted runner
@@ -115,6 +120,14 @@ jobs:
run: |
REPORT_FROM=$(realpath perf-report-staging) REPORT_TO=staging scripts/generate_and_push_perf_report.sh
- name: Delete Neon Project
if: ${{ always() }}
uses: ./.github/actions/neon-project-delete
with:
environment: staging
project_id: ${{ steps.create-neon-project.outputs.project_id }}
api_key: ${{ secrets.NEON_STAGING_API_KEY }}
- name: Post to a Slack channel
if: ${{ github.event.schedule && failure() }}
uses: slackapi/slack-github-action@v1
@@ -128,14 +141,19 @@ jobs:
env:
TEST_PG_BENCH_DURATIONS_MATRIX: "60m"
TEST_PG_BENCH_SCALES_MATRIX: "10gb"
POSTGRES_DISTRIB_DIR: /usr
POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
DEFAULT_PG_VERSION: 14
TEST_OUTPUT: /tmp/test_output
BUILD_TYPE: remote
SAVE_PERF_REPORT: ${{ github.event.inputs.save_perf_report || ( github.ref == 'refs/heads/main' ) }}
strategy:
fail-fast: false
matrix:
connstr: [ BENCHMARK_CAPTEST_CONNSTR, BENCHMARK_RDS_CONNSTR ]
# neon-captest-new: Run pgbench in a freshly created project
# neon-captest-reuse: Same, but reusing existing project
# neon-captest-prefetch: Same, with prefetching enabled (new project)
platform: [ neon-captest-new, neon-captest-reuse, neon-captest-prefetch, rds-aurora ]
runs-on: dev
container:
@@ -147,26 +165,58 @@ jobs:
steps:
- uses: actions/checkout@v3
- name: Calculate platform
id: calculate-platform
- name: Download Neon artifact
uses: ./.github/actions/download
with:
name: neon-${{ 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: Create Neon Project
if: matrix.platform != 'neon-captest-reuse'
id: create-neon-project
uses: ./.github/actions/neon-project-create
with:
environment: ${{ github.event.inputs.environment || 'dev' }}
api_key: ${{ ( github.event.inputs.environment || 'dev' ) == 'staging' && secrets.NEON_STAGING_API_KEY || secrets.NEON_CAPTEST_API_KEY }}
- name: Set up Connection String
id: set-up-connstr
run: |
case "${PLATFORM}" in
neon-captest-reuse)
CONNSTR=${{ secrets.BENCHMARK_CAPTEST_CONNSTR }}
;;
neon-captest-new | neon-captest-prefetch)
CONNSTR=${{ steps.create-neon-project.outputs.dsn }}
;;
rds-aurora)
CONNSTR=${{ secrets.BENCHMARK_RDS_CONNSTR }}
;;
*)
echo 2>&1 "Unknown PLATFORM=${PLATFORM}. Allowed only 'neon-captest-reuse', 'neon-captest-new', 'neon-captest-prefetch' or 'rds-aurora'"
exit 1
;;
esac
echo "::set-output name=connstr::${CONNSTR}"
psql ${CONNSTR} -c "SELECT version();"
env:
CONNSTR: ${{ matrix.connstr }}
run: |
if [ "${CONNSTR}" = "BENCHMARK_CAPTEST_CONNSTR" ]; then
PLATFORM=neon-captest
elif [ "${CONNSTR}" = "BENCHMARK_RDS_CONNSTR" ]; then
PLATFORM=rds-aurora
else
echo 2>&1 "Unknown CONNSTR=${CONNSTR}. Allowed are BENCHMARK_CAPTEST_CONNSTR, and BENCHMARK_RDS_CONNSTR only"
exit 1
fi
PLATFORM: ${{ matrix.platform }}
echo "::set-output name=PLATFORM::${PLATFORM}"
- name: Install Deps
- name: Set database options
if: matrix.platform == 'neon-captest-prefetch'
run: |
sudo apt -y update
sudo apt install -y postgresql-14
psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE main SET enable_seqscan_prefetch=on"
psql ${BENCHMARK_CONNSTR} -c "ALTER DATABASE main SET seqscan_prefetch_buffers=10"
env:
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
- name: Benchmark init
uses: ./.github/actions/run-python-test-set
@@ -174,11 +224,11 @@ jobs:
build_type: ${{ env.BUILD_TYPE }}
test_selection: performance
run_in_parallel: false
save_perf_report: true
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_init
env:
PLATFORM: ${{ steps.calculate-platform.outputs.PLATFORM }}
BENCHMARK_CONNSTR: ${{ secrets[matrix.connstr] }}
PLATFORM: ${{ matrix.platform }}
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 }}"
@@ -188,39 +238,48 @@ jobs:
build_type: ${{ env.BUILD_TYPE }}
test_selection: performance
run_in_parallel: false
save_perf_report: true
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_simple_update
env:
PLATFORM: ${{ steps.calculate-platform.outputs.PLATFORM }}
BENCHMARK_CONNSTR: ${{ secrets[matrix.connstr] }}
PLATFORM: ${{ matrix.platform }}
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: Benchmark simple-update
- name: Benchmark select-only
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ env.BUILD_TYPE }}
test_selection: performance
run_in_parallel: false
save_perf_report: true
save_perf_report: ${{ env.SAVE_PERF_REPORT }}
extra_params: -m remote_cluster --timeout 21600 -k test_pgbench_remote_select_only
env:
PLATFORM: ${{ steps.calculate-platform.outputs.PLATFORM }}
BENCHMARK_CONNSTR: ${{ secrets[matrix.connstr] }}
PLATFORM: ${{ matrix.platform }}
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: always()
uses: ./.github/actions/allure-report
with:
action: generate
build_type: ${{ env.BUILD_TYPE }}
- name: Delete Neon Project
if: ${{ matrix.platform != 'neon-captest-reuse' && always() }}
uses: ./.github/actions/neon-project-delete
with:
environment: dev
project_id: ${{ steps.create-neon-project.outputs.project_id }}
api_key: ${{ secrets.NEON_CAPTEST_API_KEY }}
- 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: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
slack-message: "Periodic perf testing ${{ matrix.platform }}: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
env:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

203
.github/workflows/build_and_test.yml vendored
View File

@@ -54,7 +54,6 @@ jobs:
fail-fast: false
matrix:
build_type: [ debug, release ]
rust_toolchain: [ 1.58 ]
env:
BUILD_TYPE: ${{ matrix.build_type }}
@@ -77,9 +76,14 @@ jobs:
submodules: true
fetch-depth: 1
- name: Set pg revision for caching
id: pg_ver
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres)
- name: Set pg 14 revision for caching
id: pg_v14_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v14)
shell: bash -euxo pipefail {0}
- name: Set pg 15 revision for caching
id: pg_v15_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v15)
shell: bash -euxo pipefail {0}
# Set some environment variables used by all the steps.
@@ -90,16 +94,18 @@ jobs:
# CARGO_FEATURES is passed to "cargo metadata". It is separate from CARGO_FLAGS,
# because "cargo metadata" doesn't accept --release or --debug options
#
# We run tests with addtional features, that are turned off by default (e.g. in release builds), see
# corresponding Cargo.toml files for their descriptions.
- name: Set env variables
run: |
if [[ $BUILD_TYPE == "debug" ]]; then
cov_prefix="scripts/coverage --profraw-prefix=$GITHUB_JOB --dir=/tmp/coverage run"
CARGO_FEATURES=""
CARGO_FLAGS="--locked"
CARGO_FEATURES="--features testing"
CARGO_FLAGS="--locked --timings $CARGO_FEATURES"
elif [[ $BUILD_TYPE == "release" ]]; then
cov_prefix=""
CARGO_FEATURES="--features profiling"
CARGO_FLAGS="--locked --release $CARGO_FEATURES"
CARGO_FEATURES="--features testing,profiling"
CARGO_FLAGS="--locked --timings --release $CARGO_FEATURES"
fi
echo "cov_prefix=${cov_prefix}" >> $GITHUB_ENV
echo "CARGO_FEATURES=${CARGO_FEATURES}" >> $GITHUB_ENV
@@ -121,19 +127,31 @@ jobs:
target/
# Fall back to older versions of the key, if no cache for current Cargo.lock was found
key: |
v7-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ matrix.rust_toolchain }}-${{ hashFiles('Cargo.lock') }}
v7-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ matrix.rust_toolchain }}-
v9-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ hashFiles('Cargo.lock') }}
v9-${{ runner.os }}-${{ matrix.build_type }}-cargo-
- name: Cache postgres build
id: cache_pg
- name: Cache postgres v14 build
id: cache_pg_14
uses: actions/cache@v3
with:
path: tmp_install/
key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_ver.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
path: pg_install/v14
key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_v14_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
- name: Build postgres
if: steps.cache_pg.outputs.cache-hit != 'true'
run: mold -run make postgres -j$(nproc)
- name: Cache postgres v15 build
id: cache_pg_15
uses: actions/cache@v3
with:
path: pg_install/v15
key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_v15_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
- name: Build postgres v14
if: steps.cache_pg_14.outputs.cache-hit != 'true'
run: mold -run make postgres-v14 -j$(nproc)
shell: bash -euxo pipefail {0}
- name: Build postgres v15
if: steps.cache_pg_15.outputs.cache-hit != 'true'
run: mold -run make postgres-v15 -j$(nproc)
shell: bash -euxo pipefail {0}
- name: Build neon extensions
@@ -142,7 +160,7 @@ jobs:
- name: Run cargo build
run: |
${cov_prefix} mold -run cargo build $CARGO_FLAGS --features failpoints --bins --tests
${cov_prefix} mold -run cargo build $CARGO_FLAGS --bins --tests
shell: bash -euxo pipefail {0}
- name: Run cargo test
@@ -192,15 +210,26 @@ jobs:
shell: bash -euxo pipefail {0}
- name: Install postgres binaries
run: cp -a tmp_install /tmp/neon/pg_install
run: cp -a pg_install /tmp/neon/pg_install
shell: bash -euxo pipefail {0}
- name: Upload Neon artifact
uses: ./.github/actions/upload
with:
name: neon-${{ runner.os }}-${{ matrix.build_type }}-${{ matrix.rust_toolchain }}-artifact
name: neon-${{ runner.os }}-${{ matrix.build_type }}-artifact
path: /tmp/neon
- name: Prepare cargo build timing stats for storing
run: |
mkdir -p "/tmp/neon/cargo-timings/$BUILD_TYPE/"
cp -r ./target/cargo-timings/* "/tmp/neon/cargo-timings/$BUILD_TYPE/"
shell: bash -euxo pipefail {0}
- name: Upload cargo build stats
uses: ./.github/actions/upload
with:
name: neon-${{ runner.os }}-${{ matrix.build_type }}-build-stats
path: /tmp/neon/cargo-timings/
# XXX: keep this after the binaries.list is formed, so the coverage can properly work later
- name: Merge and upload coverage data
if: matrix.build_type == 'debug'
@@ -216,7 +245,6 @@ jobs:
fail-fast: false
matrix:
build_type: [ debug, release ]
rust_toolchain: [ 1.58 ]
steps:
- name: Checkout
uses: actions/checkout@v3
@@ -228,7 +256,6 @@ jobs:
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ matrix.build_type }}
rust_toolchain: ${{ matrix.rust_toolchain }}
test_selection: regress
needs_postgres_source: true
run_with_real_s3: true
@@ -236,10 +263,37 @@ jobs:
real_s3_region: us-west-2
real_s3_access_key_id: "${{ secrets.AWS_ACCESS_KEY_ID_CI_TESTS_S3 }}"
real_s3_secret_access_key: "${{ secrets.AWS_SECRET_ACCESS_KEY_CI_TESTS_S3 }}"
- name: Merge and upload coverage data
if: matrix.build_type == 'debug'
uses: ./.github/actions/save-coverage-data
upload-latest-artifacts:
runs-on: dev
container:
image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
options: --init
needs: [ regress-tests ]
if: github.ref_name == 'main'
steps:
- name: Copy Neon artifact to the latest directory
shell: bash -euxo pipefail {0}
env:
BUCKET: neon-github-public-dev
PREFIX: artifacts/${{ github.run_id }}
run: |
for build_type in debug release; do
FILENAME=neon-${{ runner.os }}-${build_type}-artifact.tar.zst
S3_KEY=$(aws s3api list-objects-v2 --bucket ${BUCKET} --prefix ${PREFIX} | jq -r '.Contents[].Key' | grep ${FILENAME} | sort --version-sort | tail -1 || true)
if [ -z "${S3_KEY}" ]; then
echo 2>&1 "Neither s3://${BUCKET}/${PREFIX}/${FILENAME} nor its version from previous attempts exist"
exit 1
fi
time aws s3 cp --only-show-errors s3://${BUCKET}/${S3_KEY} s3://${BUCKET}/artifacts/latest/${FILENAME}
done
benchmarks:
runs-on: dev
container:
@@ -251,7 +305,6 @@ jobs:
fail-fast: false
matrix:
build_type: [ release ]
rust_toolchain: [ 1.58 ]
steps:
- name: Checkout
uses: actions/checkout@v3
@@ -263,10 +316,9 @@ jobs:
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ matrix.build_type }}
rust_toolchain: ${{ matrix.rust_toolchain }}
test_selection: performance
run_in_parallel: false
save_perf_report: true
save_perf_report: ${{ github.ref == 'refs/heads/main' }}
env:
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
@@ -291,11 +343,26 @@ jobs:
submodules: false
- name: Create Allure report
id: create-allure-report
uses: ./.github/actions/allure-report
with:
action: generate
build_type: ${{ matrix.build_type }}
- name: Store Allure test stat in the DB
if: ${{ steps.create-allure-report.outputs.report-url }}
env:
BUILD_TYPE: ${{ matrix.build_type }}
SHA: ${{ github.event.pull_request.head.sha || github.sha }}
REPORT_URL: ${{ steps.create-allure-report.outputs.report-url }}
TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR }}
shell: bash -euxo pipefail {0}
run: |
curl --fail --output suites.json ${REPORT_URL%/index.html}/data/suites.json
./scripts/pysync
DATABASE_URL="$TEST_RESULT_CONNSTR" poetry run python3 scripts/ingest_regress_test_result.py --revision ${SHA} --reference ${GITHUB_REF} --build-type ${BUILD_TYPE} --ingest suites.json
coverage-report:
runs-on: dev
container:
@@ -306,7 +373,6 @@ jobs:
fail-fast: false
matrix:
build_type: [ debug ]
rust_toolchain: [ 1.58 ]
steps:
- name: Checkout
uses: actions/checkout@v3
@@ -323,12 +389,12 @@ jobs:
!~/.cargo/registry/src
~/.cargo/git/
target/
key: v7-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ matrix.rust_toolchain }}-${{ hashFiles('Cargo.lock') }}
key: v9-${{ runner.os }}-${{ matrix.build_type }}-cargo-${{ hashFiles('Cargo.lock') }}
- name: Get Neon artifact
uses: ./.github/actions/download
with:
name: neon-${{ runner.os }}-${{ matrix.build_type }}-${{ matrix.rust_toolchain }}-artifact
name: neon-${{ runner.os }}-${{ matrix.build_type }}-artifact
path: /tmp/neon
- name: Get coverage artifact
@@ -377,7 +443,7 @@ jobs:
trigger-e2e-tests:
runs-on: dev
container:
image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/base:pinned
options: --init
needs: [ build-neon ]
steps:
@@ -447,7 +513,6 @@ jobs:
compute-node-image:
runs-on: dev
container: gcr.io/kaniko-project/executor:v1.9.0-debug
steps:
- name: Checkout
uses: actions/checkout@v1 # v3 won't work with kaniko
@@ -458,18 +523,57 @@ jobs:
- name: Configure ECR login
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build compute node with extensions
run: /kaniko/executor --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --dockerfile Dockerfile.compute-node --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:$GITHUB_RUN_ID
# compute-node uses postgres 14, which is default now
# cloud repo depends on this image name, thus duplicating it
# remove compute-node when cloud repo is updated
- name: Kaniko build compute node with extensions v14 (compatibility)
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --snapshotMode=redo --context . --dockerfile Dockerfile.compute-node-v14 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:$GITHUB_RUN_ID
compute-node-image-v14:
runs-on: dev
container: gcr.io/kaniko-project/executor:v1.9.0-debug
steps:
- name: Checkout
uses: actions/checkout@v1 # v3 won't work with kaniko
with:
submodules: true
fetch-depth: 0
- name: Configure ECR login
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build compute node with extensions v14
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --context . --dockerfile Dockerfile.compute-node-v14 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:$GITHUB_RUN_ID
compute-node-image-v15:
runs-on: dev
container: gcr.io/kaniko-project/executor:v1.9.0-debug
steps:
- name: Checkout
uses: actions/checkout@v1 # v3 won't work with kaniko
with:
submodules: true
fetch-depth: 0
- name: Configure ECR login
run: echo "{\"credsStore\":\"ecr-login\"}" > /kaniko/.docker/config.json
- name: Kaniko build compute node with extensions v15
run: /kaniko/executor --skip-unused-stages --snapshotMode=redo --cache=true --cache-repo 369495373322.dkr.ecr.eu-central-1.amazonaws.com/cache --context . --dockerfile Dockerfile.compute-node-v15 --destination 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:$GITHUB_RUN_ID
promote-images:
runs-on: dev
needs: [ neon-image, compute-node-image, compute-tools-image ]
needs: [ neon-image, compute-node-image, compute-node-image-v14, compute-node-image-v15, compute-tools-image ]
if: github.event_name != 'workflow_dispatch'
container: amazon/aws-cli
strategy:
fail-fast: false
matrix:
name: [ neon, compute-node, compute-tools ]
# compute-node uses postgres 14, which is default now
# cloud repo depends on this image name, thus duplicating it
# remove compute-node when cloud repo is updated
name: [ neon, compute-node, compute-node-v14, compute-node-v15, compute-tools ]
steps:
- name: Promote image to latest
@@ -501,10 +605,25 @@ jobs:
- name: Pull compute node image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:latest compute-node
- name: Pull compute node v14 image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v14:latest compute-node-v14
- name: Pull compute node v15 image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node-v15:latest compute-node-v15
- name: Pull rust image from ECR
run: crane pull 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned rust
- name: Configure docker login
- name: Push images to production ECR
if: |
(github.ref_name == 'main' || github.ref_name == 'release') &&
github.event_name != 'workflow_dispatch'
run: |
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/neon:$GITHUB_RUN_ID 093970136003.dkr.ecr.us-east-2.amazonaws.com/neon:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-tools:$GITHUB_RUN_ID 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-tools:latest
crane copy 369495373322.dkr.ecr.eu-central-1.amazonaws.com/compute-node:$GITHUB_RUN_ID 093970136003.dkr.ecr.us-east-2.amazonaws.com/compute-node:latest
- 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
@@ -519,10 +638,16 @@ jobs:
- name: Push compute node image to Docker Hub
run: crane push compute-node neondatabase/compute-node:${{needs.tag.outputs.build-tag}}
- name: Push compute node v14 image to Docker Hub
run: crane push compute-node-v14 neondatabase/compute-node-v14:${{needs.tag.outputs.build-tag}}
- name: Push compute node v15 image to Docker Hub
run: crane push compute-node-v15 neondatabase/compute-node-v15:${{needs.tag.outputs.build-tag}}
- name: Push rust image to Docker Hub
run: crane push rust neondatabase/rust:pinned
- name: Add latest tag to images
- name: Add latest tag to images in Docker Hub
if: |
(github.ref_name == 'main' || github.ref_name == 'release') &&
github.event_name != 'workflow_dispatch'
@@ -530,6 +655,8 @@ jobs:
crane tag neondatabase/neon:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-tools:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v14:${{needs.tag.outputs.build-tag}} latest
crane tag neondatabase/compute-node-v15:${{needs.tag.outputs.build-tag}} latest
calculate-deploy-targets:
runs-on: [ self-hosted, Linux, k8s-runner ]
@@ -648,5 +775,5 @@ jobs:
- name: Re-deploy proxy
run: |
DOCKER_TAG=${{needs.tag.outputs.build-tag}}
helm upgrade ${{ matrix.proxy_job }} neondatabase/neon-proxy --namespace default --install -f .github/helm-values/${{ matrix.proxy_config }}.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
helm upgrade ${{ matrix.proxy_job }}-scram neondatabase/neon-proxy --namespace default --install -f .github/helm-values/${{ matrix.proxy_config }}-scram.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
helm upgrade ${{ matrix.proxy_job }} neondatabase/neon-proxy --namespace neon-proxy --install -f .github/helm-values/${{ matrix.proxy_config }}.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s
helm upgrade ${{ matrix.proxy_job }}-scram neondatabase/neon-proxy --namespace neon-proxy --install -f .github/helm-values/${{ matrix.proxy_config }}-scram.yaml --set image.tag=${DOCKER_TAG} --wait --timeout 15m0s

94
.github/workflows/codestyle.yml vendored
View File

@@ -17,18 +17,21 @@ concurrency:
env:
RUST_BACKTRACE: 1
COPT: '-Werror'
jobs:
check-codestyle-rust:
strategy:
fail-fast: false
matrix:
# If we want to duplicate this job for different
# Rust toolchains (e.g. nightly or 1.37.0), add them here.
rust_toolchain: [1.58]
# XXX: both OSes have rustup
# * https://github.com/actions/runner-images/blob/main/images/macos/macos-12-Readme.md#rust-tools
# * https://github.com/actions/runner-images/blob/main/images/linux/Ubuntu2204-Readme.md#rust-tools
# this is all we need to install our toolchain later via rust-toolchain.toml
# so don't install any toolchain explicitly.
os: [ubuntu-latest, macos-latest]
timeout-minutes: 60
name: run regression test suite
timeout-minutes: 90
name: check codestyle rust and postgres
runs-on: ${{ matrix.os }}
steps:
@@ -38,14 +41,6 @@ jobs:
submodules: true
fetch-depth: 2
- name: Install rust toolchain ${{ matrix.rust_toolchain }}
uses: actions-rs/toolchain@v1
with:
profile: minimal
toolchain: ${{ matrix.rust_toolchain }}
components: rustfmt, clippy
override: true
- name: Check formatting
run: cargo fmt --all -- --check
@@ -59,17 +54,29 @@ jobs:
if: matrix.os == 'macos-latest'
run: brew install flex bison openssl
- name: Set pg revision for caching
id: pg_ver
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres)
- name: Set pg 14 revision for caching
id: pg_v14_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v14)
shell: bash -euxo pipefail {0}
- name: Cache postgres build
id: cache_pg
- name: Set pg 15 revision for caching
id: pg_v15_rev
run: echo ::set-output name=pg_rev::$(git rev-parse HEAD:vendor/postgres-v15)
shell: bash -euxo pipefail {0}
- name: Cache postgres v14 build
id: cache_pg_14
uses: actions/cache@v3
with:
path: |
tmp_install/
key: ${{ runner.os }}-pg-${{ steps.pg_ver.outputs.pg_rev }}
path: pg_install/v14
key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_v14_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
- name: Cache postgres v15 build
id: cache_pg_15
uses: actions/cache@v3
with:
path: pg_install/v15
key: v1-${{ runner.os }}-${{ matrix.build_type }}-pg-${{ steps.pg_v15_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
- name: Set extra env for macOS
if: matrix.os == 'macos-latest'
@@ -77,24 +84,19 @@ jobs:
echo 'LDFLAGS=-L/usr/local/opt/openssl@3/lib' >> $GITHUB_ENV
echo 'CPPFLAGS=-I/usr/local/opt/openssl@3/include' >> $GITHUB_ENV
- name: Build postgres
if: steps.cache_pg.outputs.cache-hit != 'true'
run: make postgres
- name: Build postgres v14
if: steps.cache_pg_14.outputs.cache-hit != 'true'
run: make postgres-v14
shell: bash -euxo pipefail {0}
- name: Build postgres v15
if: steps.cache_pg_15.outputs.cache-hit != 'true'
run: make postgres-v15
shell: bash -euxo pipefail {0}
- name: Build neon extensions
run: make neon-pg-ext
# Plain configure output can contain weird errors like 'error: C compiler cannot create executables'
# and the real cause will be inside config.log
- name: Print configure logs in case of failure
if: failure()
continue-on-error: true
run: |
echo '' && echo '=== config.log ===' && echo ''
cat tmp_install/build/config.log
echo '' && echo '=== configure.log ===' && echo ''
cat tmp_install/build/configure.log
- name: Cache cargo deps
id: cache_cargo
uses: actions/cache@v3
@@ -104,7 +106,7 @@ jobs:
!~/.cargo/registry/src
~/.cargo/git
target
key: v3-${{ runner.os }}-cargo-${{ hashFiles('./Cargo.lock') }}-rust-${{ matrix.rust_toolchain }}
key: v5-${{ runner.os }}-cargo-${{ hashFiles('./Cargo.lock') }}-rust
- name: Run cargo clippy
run: ./run_clippy.sh
@@ -112,6 +114,26 @@ jobs:
- name: Ensure all project builds
run: cargo build --locked --all --all-targets
check-rust-dependencies:
runs-on: dev
container:
image: 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:pinned
options: --init
steps:
- name: Checkout
uses: actions/checkout@v3
with:
submodules: false
fetch-depth: 1
# https://github.com/facebookincubator/cargo-guppy/tree/bec4e0eb29dcd1faac70b1b5360267fc02bf830e/tools/cargo-hakari#2-keep-the-workspace-hack-up-to-date-in-ci
- name: Check every project module is covered by Hakari
run: |
cargo hakari generate --diff # workspace-hack Cargo.toml is up-to-date
cargo hakari manage-deps --dry-run # all workspace crates depend on workspace-hack
shell: bash -euxo pipefail {0}
check-codestyle-python:
runs-on: [ self-hosted, Linux, k8s-runner ]
steps:

20
.github/workflows/pg_clients.yml vendored
View File

@@ -47,17 +47,23 @@ jobs:
shell: bash -euxo pipefail {0}
run: ./scripts/pysync
- name: Create Neon Project
id: create-neon-project
uses: ./.github/actions/neon-project-create
with:
environment: staging
api_key: ${{ secrets.NEON_STAGING_API_KEY }}
- name: Run pytest
env:
REMOTE_ENV: 1
BENCHMARK_CONNSTR: "${{ secrets.BENCHMARK_STAGING_CONNSTR }}"
BENCHMARK_CONNSTR: ${{ steps.create-neon-project.outputs.dsn }}
POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
shell: bash -euxo pipefail {0}
run: |
# Test framework expects we have psql binary;
# but since we don't really need it in this test, let's mock it
mkdir -p "$POSTGRES_DISTRIB_DIR/bin" && touch "$POSTGRES_DISTRIB_DIR/bin/psql";
mkdir -p "$POSTGRES_DISTRIB_DIR/v14/bin" && touch "$POSTGRES_DISTRIB_DIR/v14/bin/psql";
./scripts/pytest \
--junitxml=$TEST_OUTPUT/junit.xml \
--tb=short \
@@ -65,6 +71,14 @@ jobs:
-m "remote_cluster" \
-rA "test_runner/pg_clients"
- name: Delete Neon Project
if: ${{ always() }}
uses: ./.github/actions/neon-project-delete
with:
environment: staging
project_id: ${{ steps.create-neon-project.outputs.project_id }}
api_key: ${{ secrets.NEON_STAGING_API_KEY }}
# We use GitHub's action upload-artifact because `ubuntu-latest` doesn't have configured AWS CLI.
# It will be fixed after switching to gen2 runner
- name: Upload python test logs

5
.gitignore vendored
View File

@@ -1,6 +1,6 @@
/pg_install
/target
/tmp_check
/tmp_install
/tmp_check_cli
__pycache__/
test_output/
@@ -15,3 +15,6 @@ test_output/
*.key
*.crt
*.o
*.so
*.Po

10
.gitmodules vendored
View File

@@ -1,4 +1,8 @@
[submodule "vendor/postgres"]
path = vendor/postgres
url = https://github.com/zenithdb/postgres
[submodule "vendor/postgres-v14"]
path = vendor/postgres-v14
url = https://github.com/neondatabase/postgres.git
branch = main
[submodule "vendor/postgres-v15"]
path = vendor/postgres-v15
url = https://github.com/neondatabase/postgres.git
branch = REL_15_STABLE_neon

404
Cargo.lock generated
View File

@@ -37,6 +37,12 @@ dependencies = [
"memchr",
]
[[package]]
name = "amplify_num"
version = "0.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f27d3d00d3d115395a7a8a4dc045feb7aa82b641e485f7e15f4e67ac16f4f56d"
[[package]]
name = "ansi_term"
version = "0.12.1"
@@ -135,6 +141,15 @@ dependencies = [
"syn",
]
[[package]]
name = "atomic-polyfill"
version = "0.1.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9c041a8d9751a520ee19656232a18971f18946a7900f1520ee4400002244dd89"
dependencies = [
"critical-section",
]
[[package]]
name = "atty"
version = "0.2.14"
@@ -183,9 +198,9 @@ dependencies = [
[[package]]
name = "axum-core"
version = "0.2.7"
version = "0.2.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e4f44a0e6200e9d11a1cdc989e4b358f6e3d354fbf48478f345a17f4e43f8635"
checksum = "d9f0c0a60006f2a293d82d571f635042a72edf927539b7685bd62d361963839b"
dependencies = [
"async-trait",
"bytes",
@@ -193,6 +208,8 @@ dependencies = [
"http",
"http-body",
"mime",
"tower-layer",
"tower-service",
]
[[package]]
@@ -210,6 +227,21 @@ dependencies = [
"rustc-demangle",
]
[[package]]
name = "bare-metal"
version = "0.2.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5deb64efa5bd81e31fcd1938615a6d98c82eafcbcd787162b6f63b91d6bac5b3"
dependencies = [
"rustc_version 0.2.3",
]
[[package]]
name = "bare-metal"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f8fe8f5a8a398345e52358e18ff07cc17a568fbca5c6f73873d3a62056309603"
[[package]]
name = "base64"
version = "0.13.0"
@@ -227,14 +259,14 @@ dependencies = [
[[package]]
name = "bindgen"
version = "0.59.2"
version = "0.60.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2bd2a9a458e8f4304c52c43ebb0cfbd520289f8379a52e329a38afda99bf8eb8"
checksum = "062dddbc1ba4aca46de6338e2bf87771414c335f7b2f2036e8f3e9befebf88e6"
dependencies = [
"bitflags",
"cexpr",
"clang-sys",
"clap 2.34.0",
"clap 3.2.16",
"env_logger",
"lazy_static",
"lazycell",
@@ -248,6 +280,18 @@ dependencies = [
"which",
]
[[package]]
name = "bit_field"
version = "0.10.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "dcb6dd1c2376d2e096796e234a70e17e94cc2d5d54ff8ce42b28cef1d0d359a4"
[[package]]
name = "bitfield"
version = "0.13.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "46afbd2983a5d5a7bd740ccb198caf5b82f45c40c09c0eed36052d91cb92e719"
[[package]]
name = "bitflags"
version = "1.3.2"
@@ -375,13 +419,9 @@ version = "2.34.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a0610544180c38b88101fecf2dd634b174a62eef6946f84dfc6a7127512b381c"
dependencies = [
"ansi_term",
"atty",
"bitflags",
"strsim 0.8.0",
"textwrap 0.11.0",
"unicode-width",
"vec_map",
]
[[package]]
@@ -394,7 +434,7 @@ dependencies = [
"bitflags",
"clap_lex",
"indexmap",
"strsim 0.10.0",
"strsim",
"termcolor",
"textwrap 0.15.0",
]
@@ -457,8 +497,10 @@ dependencies = [
"chrono",
"clap 3.2.16",
"env_logger",
"futures",
"hyper",
"log",
"notify",
"postgres",
"regex",
"serde",
@@ -495,13 +537,16 @@ name = "control_plane"
version = "0.1.0"
dependencies = [
"anyhow",
"clap 3.2.16",
"comfy-table",
"git-version",
"nix",
"once_cell",
"pageserver",
"pageserver_api",
"postgres",
"regex",
"reqwest",
"safekeeper",
"safekeeper_api",
"serde",
"serde_with",
"tar",
@@ -527,6 +572,18 @@ version = "0.8.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5827cebf4670468b8772dd191856768aedcb1b0278a04f989f7766351917b9dc"
[[package]]
name = "cortex-m"
version = "0.7.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "70858629a458fdfd39f9675c4dc309411f2a3f83bede76988d81bf1a0ecee9e0"
dependencies = [
"bare-metal 0.2.5",
"bitfield",
"embedded-hal",
"volatile-register",
]
[[package]]
name = "cpp_demangle"
version = "0.3.5"
@@ -551,7 +608,7 @@ version = "0.6.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3dfea2db42e9927a3845fb268a10a72faed6d416065f77873f05e411457c363e"
dependencies = [
"rustc_version",
"rustc_version 0.4.0",
]
[[package]]
@@ -599,6 +656,18 @@ dependencies = [
"itertools",
]
[[package]]
name = "critical-section"
version = "0.2.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "95da181745b56d4bd339530ec393508910c909c784e8962d15d722bacf0bcbcd"
dependencies = [
"bare-metal 1.0.0",
"cfg-if",
"cortex-m",
"riscv",
]
[[package]]
name = "crossbeam-channel"
version = "0.5.6"
@@ -741,7 +810,7 @@ dependencies = [
"ident_case",
"proc-macro2",
"quote",
"strsim 0.10.0",
"strsim",
"syn",
]
@@ -843,6 +912,16 @@ version = "1.7.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3f107b87b6afc2a64fd13cac55fe06d6c8859f12d4b14cbcdd2c67d0976781be"
[[package]]
name = "embedded-hal"
version = "0.2.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "35949884794ad573cf46071e41c9b60efb0cb311e3ca01f7af807af1debc66ff"
dependencies = [
"nb 0.1.3",
"void",
]
[[package]]
name = "encoding_rs"
version = "0.8.31"
@@ -995,6 +1074,15 @@ dependencies = [
"winapi",
]
[[package]]
name = "fsevent-sys"
version = "4.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "76ee7a02da4d231650c7cea31349b889be2f45ddb3ef3032d2ec8185f6313fd2"
dependencies = [
"libc",
]
[[package]]
name = "futures"
version = "0.3.21"
@@ -1164,6 +1252,15 @@ version = "1.8.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "eabb4a44450da02c90444cf74558da904edde8fb4e9035a9a6a4e15445af0bd7"
[[package]]
name = "hash32"
version = "0.2.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b0c35f58762feb77d74ebe43bdbc3210f09be9fe6742234d573bacc26ed92b67"
dependencies = [
"byteorder",
]
[[package]]
name = "hashbrown"
version = "0.12.3"
@@ -1173,6 +1270,19 @@ dependencies = [
"ahash",
]
[[package]]
name = "heapless"
version = "0.7.16"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "db04bc24a18b9ea980628ecf00e6c0264f3c1426dac36c00cb49b6fbad8b0743"
dependencies = [
"atomic-polyfill",
"hash32",
"rustc_version 0.4.0",
"spin 0.9.4",
"stable_deref_trait",
]
[[package]]
name = "heck"
version = "0.3.3"
@@ -1394,6 +1504,26 @@ dependencies = [
"str_stack",
]
[[package]]
name = "inotify"
version = "0.9.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f8069d3ec154eb856955c1c0fbffefbf5f3c40a104ec912d4797314c1801abff"
dependencies = [
"bitflags",
"inotify-sys",
"libc",
]
[[package]]
name = "inotify-sys"
version = "0.1.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e05c02b5e89bff3b946cedeca278abc628fe811e604f027c45a8aa3cf793d0eb"
dependencies = [
"libc",
]
[[package]]
name = "instant"
version = "0.1.12"
@@ -1453,6 +1583,26 @@ dependencies = [
"simple_asn1",
]
[[package]]
name = "kqueue"
version = "1.0.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4d6112e8f37b59803ac47a42d14f1f3a59bbf72fc6857ffc5be455e28a691f8e"
dependencies = [
"kqueue-sys",
"libc",
]
[[package]]
name = "kqueue-sys"
version = "1.0.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8367585489f01bc55dd27404dcf56b95e6da061a256a666ab23be9ba96a2e587"
dependencies = [
"bitflags",
"libc",
]
[[package]]
name = "kstring"
version = "1.0.6"
@@ -1490,6 +1640,12 @@ dependencies = [
"winapi",
]
[[package]]
name = "libm"
version = "0.2.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "292a948cd991e376cf75541fe5b97a1081d713c618b4f1b9500f8844e49eb565"
[[package]]
name = "lock_api"
version = "0.4.7"
@@ -1649,22 +1805,20 @@ dependencies = [
]
[[package]]
name = "neon_local"
version = "0.1.0"
name = "nb"
version = "0.1.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "801d31da0513b6ec5214e9bf433a77966320625a37860f910be265be6e18d06f"
dependencies = [
"anyhow",
"clap 3.2.16",
"comfy-table",
"control_plane",
"git-version",
"pageserver",
"postgres",
"safekeeper",
"serde_json",
"utils",
"workspace_hack",
"nb 1.0.0",
]
[[package]]
name = "nb"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "546c37ac5d9e56f55e73b677106873d9d9f5190605e41a856503623648488cae"
[[package]]
name = "nix"
version = "0.23.1"
@@ -1694,6 +1848,24 @@ dependencies = [
"minimal-lexical",
]
[[package]]
name = "notify"
version = "5.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ed2c66da08abae1c024c01d635253e402341b4060a12e99b31c7594063bf490a"
dependencies = [
"bitflags",
"crossbeam-channel",
"filetime",
"fsevent-sys",
"inotify",
"kqueue",
"libc",
"mio",
"walkdir",
"winapi",
]
[[package]]
name = "num-bigint"
version = "0.4.3"
@@ -1732,6 +1904,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "578ede34cf02f8924ab9447f50c28075b4d3e5b269972345e7e0372b38c6cdcd"
dependencies = [
"autocfg",
"libm",
]
[[package]]
@@ -1844,7 +2017,10 @@ checksum = "648001efe5d5c0102d8cea768e348da85d90af8ba91f0bea908f157951493cd4"
name = "pageserver"
version = "0.1.0"
dependencies = [
"amplify_num",
"anyhow",
"async-stream",
"async-trait",
"byteorder",
"bytes",
"chrono",
@@ -1866,7 +2042,9 @@ dependencies = [
"itertools",
"metrics",
"nix",
"num-traits",
"once_cell",
"pageserver_api",
"postgres",
"postgres-protocol",
"postgres-types",
@@ -1875,6 +2053,7 @@ dependencies = [
"rand",
"regex",
"remote_storage",
"rstar",
"scopeguard",
"serde",
"serde_json",
@@ -1885,6 +2064,7 @@ dependencies = [
"thiserror",
"tokio",
"tokio-postgres",
"tokio-util",
"toml_edit",
"tracing",
"url",
@@ -1893,6 +2073,17 @@ dependencies = [
"workspace_hack",
]
[[package]]
name = "pageserver_api"
version = "0.1.0"
dependencies = [
"const_format",
"serde",
"serde_with",
"utils",
"workspace_hack",
]
[[package]]
name = "parking_lot"
version = "0.11.2"
@@ -2059,7 +2250,7 @@ dependencies = [
[[package]]
name = "postgres"
version = "0.19.2"
source = "git+https://github.com/zenithdb/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
source = "git+https://github.com/neondatabase/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
dependencies = [
"bytes",
"fallible-iterator",
@@ -2072,7 +2263,7 @@ dependencies = [
[[package]]
name = "postgres-protocol"
version = "0.6.4"
source = "git+https://github.com/zenithdb/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
source = "git+https://github.com/neondatabase/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
dependencies = [
"base64",
"byteorder",
@@ -2090,7 +2281,7 @@ dependencies = [
[[package]]
name = "postgres-types"
version = "0.2.3"
source = "git+https://github.com/zenithdb/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
source = "git+https://github.com/neondatabase/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
dependencies = [
"bytes",
"fallible-iterator",
@@ -2296,6 +2487,7 @@ dependencies = [
"tokio-rustls",
"url",
"utils",
"uuid",
"workspace_hack",
"x509-parser",
]
@@ -2457,6 +2649,7 @@ dependencies = [
"tokio-util",
"toml_edit",
"tracing",
"utils",
"workspace_hack",
]
@@ -2526,12 +2719,33 @@ dependencies = [
"cc",
"libc",
"once_cell",
"spin",
"spin 0.5.2",
"untrusted",
"web-sys",
"winapi",
]
[[package]]
name = "riscv"
version = "0.7.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6907ccdd7a31012b70faf2af85cd9e5ba97657cc3987c4f13f8e4d2c2a088aba"
dependencies = [
"bare-metal 1.0.0",
"bit_field",
"riscv-target",
]
[[package]]
name = "riscv-target"
version = "0.1.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "88aa938cda42a0cf62a20cfe8d139ff1af20c2e681212b5b34adb5a58333f222"
dependencies = [
"lazy_static",
"regex",
]
[[package]]
name = "routerify"
version = "3.0.0"
@@ -2545,6 +2759,17 @@ dependencies = [
"regex",
]
[[package]]
name = "rstar"
version = "0.9.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b40f1bfe5acdab44bc63e6699c28b74f75ec43afb59f3eda01e145aff86a25fa"
dependencies = [
"heapless",
"num-traits",
"smallvec",
]
[[package]]
name = "rstest"
version = "0.12.0"
@@ -2554,7 +2779,7 @@ dependencies = [
"cfg-if",
"proc-macro2",
"quote",
"rustc_version",
"rustc_version 0.4.0",
"syn",
]
@@ -2576,7 +2801,7 @@ dependencies = [
"log",
"rusoto_credential",
"rusoto_signature",
"rustc_version",
"rustc_version 0.4.0",
"serde",
"serde_json",
"tokio",
@@ -2634,7 +2859,7 @@ dependencies = [
"percent-encoding",
"pin-project-lite",
"rusoto_credential",
"rustc_version",
"rustc_version 0.4.0",
"serde",
"sha2 0.9.9",
"tokio",
@@ -2652,13 +2877,22 @@ version = "1.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "08d43f7aa6b08d49f382cde6a7982047c3426db949b1424bc4b7ec9ae12c6ce2"
[[package]]
name = "rustc_version"
version = "0.2.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "138e3e0acb6c9fb258b19b67cb8abd63c00679d2851805ea151465464fe9030a"
dependencies = [
"semver 0.9.0",
]
[[package]]
name = "rustc_version"
version = "0.4.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bfa0f585226d2e68097d4f95d113b15b83a82e819ab25717ec0590d9584ef366"
dependencies = [
"semver",
"semver 1.0.13",
]
[[package]]
@@ -2732,16 +2966,19 @@ dependencies = [
"hyper",
"metrics",
"once_cell",
"parking_lot 0.12.1",
"postgres",
"postgres-protocol",
"postgres_ffi",
"regex",
"remote_storage",
"safekeeper_api",
"serde",
"serde_json",
"serde_with",
"signal-hook",
"tempfile",
"thiserror",
"tokio",
"tokio-postgres",
"toml_edit",
@@ -2751,6 +2988,17 @@ dependencies = [
"workspace_hack",
]
[[package]]
name = "safekeeper_api"
version = "0.1.0"
dependencies = [
"const_format",
"serde",
"serde_with",
"utils",
"workspace_hack",
]
[[package]]
name = "same-file"
version = "1.0.6"
@@ -2809,12 +3057,27 @@ dependencies = [
"libc",
]
[[package]]
name = "semver"
version = "0.9.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "1d7eb9ef2c18661902cc47e535f9bc51b78acd254da71d375c2f6720d9a40403"
dependencies = [
"semver-parser",
]
[[package]]
name = "semver"
version = "1.0.13"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "93f6841e709003d68bb2deee8c343572bf446003ec20a583e76f7b15cebf3711"
[[package]]
name = "semver-parser"
version = "0.7.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "388a1df253eca08550bef6c72392cfe7c30914bf41df5269b68cbd6ff8f570a3"
[[package]]
name = "serde"
version = "1.0.142"
@@ -3008,6 +3271,15 @@ version = "0.5.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6e63cff320ae2c57904679ba7cb63280a3dc4613885beafb148ee7bf9aa9042d"
[[package]]
name = "spin"
version = "0.9.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "7f6002a767bff9e83f8eeecf883ecb8011875a21ae8da43bffb817a57e78cc09"
dependencies = [
"lock_api",
]
[[package]]
name = "stable_deref_trait"
version = "1.2.0"
@@ -3030,12 +3302,6 @@ dependencies = [
"unicode-normalization",
]
[[package]]
name = "strsim"
version = "0.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8ea5119cdb4c55b55d432abb513a0429384878c15dde60cc77b1c99de1a95a6a"
[[package]]
name = "strsim"
version = "0.10.0"
@@ -3306,7 +3572,7 @@ dependencies = [
[[package]]
name = "tokio-postgres"
version = "0.7.6"
source = "git+https://github.com/zenithdb/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
source = "git+https://github.com/neondatabase/rust-postgres.git?rev=d052ee8b86fff9897c77b0fe89ea9daba0e1fa38#d052ee8b86fff9897c77b0fe89ea9daba0e1fa38"
dependencies = [
"async-trait",
"byteorder",
@@ -3495,9 +3761,9 @@ checksum = "b6bc1c9ce2b5135ac7f93c72918fc37feb872bdc6a5533a8b85eb4b86bfdae52"
[[package]]
name = "tracing"
version = "0.1.34"
version = "0.1.36"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5d0ecdcb44a79f0fe9844f0c4f33a342cbcbb5117de8001e6ba0dc2351327d09"
checksum = "2fce9567bd60a67d08a16488756721ba392f24f29006402881e43b19aac64307"
dependencies = [
"cfg-if",
"log",
@@ -3519,11 +3785,11 @@ dependencies = [
[[package]]
name = "tracing-core"
version = "0.1.26"
version = "0.1.29"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f54c8ca710e81886d498c2fd3331b56c93aa248d49de2222ad2742247c60072f"
checksum = "5aeea4303076558a00714b823f9ad67d58a3bbda1df83d8827d21193156e22f7"
dependencies = [
"lazy_static",
"once_cell",
"valuable",
]
@@ -3640,6 +3906,7 @@ name = "utils"
version = "0.1.0"
dependencies = [
"anyhow",
"async-trait",
"bincode",
"byteorder",
"bytes",
@@ -3667,6 +3934,7 @@ dependencies = [
"tempfile",
"thiserror",
"tokio",
"tokio-rustls",
"tracing",
"tracing-subscriber",
"workspace_hack",
@@ -3677,6 +3945,10 @@ name = "uuid"
version = "0.8.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "bc5cf98d8186244414c848017f0e2676b3fcb46807f6668a97dfe67359a3c4b7"
dependencies = [
"getrandom",
"serde",
]
[[package]]
name = "valuable"
@@ -3684,24 +3956,39 @@ version = "0.1.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "830b7e5d4d90034032940e4ace0d9a9a057e7a45cd94e6c007832e39edb82f6d"
[[package]]
name = "vcell"
version = "0.1.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "77439c1b53d2303b20d9459b1ade71a83c716e3f9c34f3228c00e6f185d6c002"
[[package]]
name = "vcpkg"
version = "0.2.15"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "accd4ea62f7bb7a82fe23066fb0957d48ef677f6eeb8215f372f52e48bb32426"
[[package]]
name = "vec_map"
version = "0.8.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f1bddf1187be692e79c5ffeab891132dfb0f236ed36a43c7ed39f1165ee20191"
[[package]]
name = "version_check"
version = "0.9.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "49874b5167b65d7193b8aba1567f5c7d93d001cafc34600cee003eda787e483f"
[[package]]
name = "void"
version = "1.0.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6a02e4885ed3bc0f2de90ea6dd45ebcbb66dacffe03547fadbb0eeae2770887d"
[[package]]
name = "volatile-register"
version = "0.2.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9ee8f19f9d74293faf70901bc20ad067dc1ad390d2cbf1e3f75f721ffee908b6"
dependencies = [
"vcell",
]
[[package]]
name = "wal_craft"
version = "0.1.0"
@@ -3714,6 +4001,7 @@ dependencies = [
"postgres",
"postgres_ffi",
"tempfile",
"workspace_hack",
]
[[package]]
@@ -3947,16 +4235,10 @@ dependencies = [
"bstr",
"bytes",
"chrono",
"clap 2.34.0",
"crossbeam-utils",
"either",
"fail",
"futures-channel",
"futures-task",
"futures-util",
"generic-array",
"hashbrown",
"hex",
"hyper",
"indexmap",
"itoa 0.4.8",
"libc",
@@ -3973,12 +4255,14 @@ dependencies = [
"regex-syntax",
"scopeguard",
"serde",
"stable_deref_trait",
"syn",
"time 0.3.12",
"tokio",
"tokio-util",
"tracing",
"tracing-core",
"uuid",
]
[[package]]

56
Cargo.toml
View File

@@ -6,7 +6,6 @@ members = [
"proxy",
"safekeeper",
"workspace_hack",
"neon_local",
"libs/*",
]
@@ -15,7 +14,60 @@ members = [
# Besides, debug info should not affect the performance.
debug = true
[profile.release-line-debug]
inherits = "release"
debug = 1 # true = 2 = all symbols, 1 = line only
[profile.release-line-debug-lto]
inherits = "release"
debug = 1 # true = 2 = all symbols, 1 = line only
lto = true
[profile.release-line-debug-size]
inherits = "release"
debug = 1 # true = 2 = all symbols, 1 = line only
opt-level = "s"
[profile.release-line-debug-zize]
inherits = "release"
debug = 1 # true = 2 = all symbols, 1 = line only
opt-level = "z"
[profile.release-line-debug-size-lto]
inherits = "release"
debug = 1 # true = 2 = all symbols, 1 = line only
opt-level = "s"
lto = true
[profile.release-line-debug-zize-lto]
inherits = "release"
debug = 1 # true = 2 = all symbols, 1 = line only
opt-level = "z"
lto = true
[profile.release-no-debug]
inherits = "release"
debug = false # true = 2 = all symbols, 1 = line only
[profile.release-no-debug-size]
inherits = "release"
debug = false # true = 2 = all symbols, 1 = line only
opt-level = "s"
[profile.release-no-debug-zize]
inherits = "release"
debug = false # true = 2 = all symbols, 1 = line only
opt-level = "z"
[profile.release-no-debug-size-lto]
inherits = "release"
debug = false # true = 2 = all symbols, 1 = line only
opt-level = "s"
lto = true
[profile.release-no-debug-zize-lto]
inherits = "release"
debug = false # true = 2 = all symbols, 1 = line only
opt-level = "z"
lto = true
# This is only needed for proxy's tests.
# TODO: we should probably fork `tokio-postgres-rustls` instead.
[patch.crates-io]
tokio-postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }

34
Dockerfile
View File

@@ -10,17 +10,19 @@ ARG TAG=pinned
FROM $REPOSITORY/$IMAGE:$TAG AS pg-build
WORKDIR /home/nonroot
COPY --chown=nonroot vendor/postgres vendor/postgres
COPY --chown=nonroot vendor/postgres-v14 vendor/postgres-v14
COPY --chown=nonroot vendor/postgres-v15 vendor/postgres-v15
COPY --chown=nonroot pgxn pgxn
COPY --chown=nonroot Makefile Makefile
COPY --chown=nonroot scripts/ninstall.sh scripts/ninstall.sh
ENV BUILD_TYPE release
RUN set -e \
&& mold -run make -j $(nproc) -s neon-pg-ext \
&& rm -rf tmp_install/build \
&& tar -C tmp_install -czf /home/nonroot/postgres_install.tar.gz .
&& rm -rf pg_install/build \
&& tar -C pg_install -czf /home/nonroot/postgres_install.tar.gz .
# Build zenith binaries
# Build neon binaries
FROM $REPOSITORY/$IMAGE:$TAG AS build
WORKDIR /home/nonroot
ARG GIT_VERSION=local
@@ -35,13 +37,14 @@ ARG CACHEPOT_BUCKET=neon-github-dev
#ARG AWS_ACCESS_KEY_ID
#ARG AWS_SECRET_ACCESS_KEY
COPY --from=pg-build /home/nonroot/tmp_install/include/postgresql/server tmp_install/include/postgresql/server
COPY --from=pg-build /home/nonroot/pg_install/v14/include/postgresql/server pg_install/v14/include/postgresql/server
COPY --from=pg-build /home/nonroot/pg_install/v15/include/postgresql/server pg_install/v15/include/postgresql/server
COPY . .
# Show build caching stats to check if it was used in the end.
# Has to be the part of the same RUN since cachepot daemon is killed in the end of this RUN, losing the compilation stats.
RUN set -e \
&& mold -run cargo build --locked --release \
&& mold -run cargo build --bin pageserver --bin safekeeper --bin proxy --locked --release \
&& cachepot -s
# Build final image
@@ -57,28 +60,29 @@ RUN set -e \
openssl \
ca-certificates \
&& rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* \
&& useradd -d /data zenith \
&& chown -R zenith:zenith /data
&& useradd -d /data neon \
&& chown -R neon:neon /data
COPY --from=build --chown=zenith:zenith /home/nonroot/target/release/pageserver /usr/local/bin
COPY --from=build --chown=zenith:zenith /home/nonroot/target/release/safekeeper /usr/local/bin
COPY --from=build --chown=zenith:zenith /home/nonroot/target/release/proxy /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/pageserver /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/safekeeper /usr/local/bin
COPY --from=build --chown=neon:neon /home/nonroot/target/release/proxy /usr/local/bin
COPY --from=pg-build /home/nonroot/tmp_install/ /usr/local/
COPY --from=pg-build /home/nonroot/pg_install/v14 /usr/local/v14/
COPY --from=pg-build /home/nonroot/pg_install/v15 /usr/local/v15/
COPY --from=pg-build /home/nonroot/postgres_install.tar.gz /data/
# By default, pageserver uses `.neon/` working directory in WORKDIR, so create one and fill it with the dummy config.
# Now, when `docker run ... pageserver` is run, it can start without errors, yet will have some default dummy values.
RUN mkdir -p /data/.neon/ && chown -R zenith:zenith /data/.neon/ \
RUN mkdir -p /data/.neon/ && chown -R neon:neon /data/.neon/ \
&& /usr/local/bin/pageserver -D /data/.neon/ --init \
-c "id=1234" \
-c "broker_endpoints=['http://etcd:2379']" \
-c "pg_distrib_dir='/usr/local'" \
-c "pg_distrib_dir='/usr/local/'" \
-c "listen_pg_addr='0.0.0.0:6400'" \
-c "listen_http_addr='0.0.0.0:9898'"
VOLUME ["/data"]
USER zenith
USER neon
EXPOSE 6400
EXPOSE 9898
CMD ["/bin/bash"]

114
Dockerfile.compute-node
View File

@@ -1,114 +0,0 @@
ARG TAG=pinned
# apparently, ARGs don't get replaced in RUN commands in kaniko
# ARG POSTGIS_VERSION=3.3.0
# ARG PLV8_VERSION=3.1.4
FROM debian:bullseye-slim AS build-deps
RUN apt update && \
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev zlib1g-dev libxml2-dev \
libcurl4-openssl-dev libossp-uuid-dev
# Build Postgres from the neon postgres repository.
FROM build-deps AS pg-build
COPY vendor/postgres postgres
RUN cd postgres && \
./configure CFLAGS='-O2 -g3' --enable-debug --with-uuid=ossp && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C contrib/ install && \
# Install headers
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/include install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/interfaces/libpq install
# Build PostGIS from the upstream PostGIS mirror. PostGIS compiles against neon postgres sources without changes.
# Perhaps we could even use the upstream binaries, compiled against vanilla Postgres, but it would require some
# investigation to check that it works, and also keeps working in the future. So for now, we compile our own binaries.
FROM build-deps AS postgis-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y gdal-bin libgdal-dev libprotobuf-c-dev protobuf-c-compiler xsltproc wget
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.0.tar.gz && \
tar xvzf postgis-3.3.0.tar.gz && \
cd postgis-3.3.0 && \
./autogen.sh && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
./configure && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
cd extensions/postgis && \
make clean && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_raster.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_tiger_geocoder.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_topology.control
# Build plv8
FROM build-deps AS plv8-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y git curl wget make ninja-build build-essential libncurses5 python3-dev pkg-config libc++-dev libc++abi-dev libglib2.0-dev
# https://github.com/plv8/plv8/issues/475
# Debian bullseye provides binutils 2.35 when >= 2.38 is necessary
RUN echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing binutils
RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.4.tar.gz && \
tar xvzf v3.1.4.tar.gz && \
cd plv8-3.1.4 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
rm -rf /plv8-* && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/plv8.control
# compile neon extensions
FROM build-deps AS neon-pg-ext-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
COPY pgxn/ pgxn/
RUN make -j $(getconf _NPROCESSORS_ONLN) \
PG_CONFIG=/usr/local/pgsql/bin/pg_config \
-C pgxn/neon \
-s install
# Compile and run the Neon-specific `compute_ctl` binary
FROM 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:$TAG AS compute-tools
USER nonroot
# Copy entire project to get Cargo.* files with proper dependencies for the whole project
COPY --chown=nonroot . .
RUN cd compute_tools && cargo build --locked --release
# Put it all together into the final image
FROM debian:bullseye-slim
# Add user postgres
RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
echo "postgres:test_console_pass" | chpasswd && \
mkdir /var/db/postgres/compute && mkdir /var/db/postgres/specs && \
chown -R postgres:postgres /var/db/postgres && \
chmod 0750 /var/db/postgres/compute && \
echo '/usr/local/lib' >> /etc/ld.so.conf && /sbin/ldconfig
# TODO: Check if we can make the extension setup more modular versus a linear build
# currently plv8-build copies the output /usr/local/pgsql from postgis-build, etc#
COPY --from=neon-pg-ext-build --chown=postgres /usr/local/pgsql /usr/local
COPY --from=compute-tools --chown=postgres /home/nonroot/target/release/compute_ctl /usr/local/bin/compute_ctl
RUN apt update && \
apt install -y libreadline-dev libossp-uuid-dev gdal-bin libgdal-dev libprotobuf-c-dev && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# Debian bullseye provides GLIBC 2.31 when 2.34 is necessary as we compiled plv8 with that version
RUN echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing binutils && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
# "temporary" symlink for old control-plane
RUN ln -s /usr/local/bin/compute_ctl /usr/local/bin/zenith_ctl
USER postgres
ENTRYPOINT ["/usr/local/bin/compute_ctl"]

200
Dockerfile.compute-node-v14 Normal file
View File

@@ -0,0 +1,200 @@
ARG TAG=pinned
# apparently, ARGs don't get replaced in RUN commands in kaniko
# ARG POSTGIS_VERSION=3.3.0
# ARG PLV8_VERSION=3.1.4
# ARG PG_VERSION=v14
#
# Layer "build-deps"
#
FROM debian:bullseye-slim AS build-deps
RUN echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update
RUN apt update && \
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev zlib1g-dev libxml2-dev \
libcurl4-openssl-dev libossp-uuid-dev wget pkg-config libglib2.0-dev
#
# Layer "pg-build"
# Build Postgres from the neon postgres repository.
#
FROM build-deps AS pg-build
COPY vendor/postgres-v14 postgres
RUN cd postgres && \
./configure CFLAGS='-O2 -g3' --enable-debug --with-uuid=ossp && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C contrib/ install && \
# Install headers
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/include install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/interfaces/libpq install
#
# Layer "postgis-build"
# Build PostGIS from the upstream PostGIS mirror.
#
# PostGIS compiles against neon postgres sources without changes. Perhaps we
# could even use the upstream binaries, compiled against vanilla Postgres, but
# it would require some investigation to check that it works, and also keeps
# working in the future. So for now, we compile our own binaries.
FROM build-deps AS postgis-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y gdal-bin libgdal-dev libprotobuf-c-dev protobuf-c-compiler xsltproc
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.0.tar.gz && \
tar xvzf postgis-3.3.0.tar.gz && \
cd postgis-3.3.0 && \
./autogen.sh && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
./configure && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
cd extensions/postgis && \
make clean && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_raster.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_tiger_geocoder.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_topology.control
#
# Layer "plv8-build"
# Build plv8
#
FROM build-deps AS plv8-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y ninja-build python3-dev libc++-dev libc++abi-dev libncurses5
# https://github.com/plv8/plv8/issues/475
# Debian bullseye provides binutils 2.35 when >= 2.38 is necessary
RUN apt update && \
apt install -y --no-install-recommends -t testing binutils
RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.4.tar.gz && \
tar xvzf v3.1.4.tar.gz && \
cd plv8-3.1.4 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
rm -rf /plv8-* && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/plv8.control
#
# Layer "h3-pg-build"
# Build h3_pg
#
FROM build-deps AS h3-pg-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
# packaged cmake is too old
RUN apt update && \
apt install -y --no-install-recommends -t testing cmake
RUN wget https://github.com/uber/h3/archive/refs/tags/v4.0.1.tar.gz -O h3.tgz && \
tar xvzf h3.tgz && \
cd h3-4.0.1 && \
mkdir build && \
cd build && \
cmake .. -DCMAKE_BUILD_TYPE=Release && \
make -j $(getconf _NPROCESSORS_ONLN) && \
DESTDIR=/h3 make install && \
cp -R /h3/usr / && \
rm -rf build
RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.0.1.tar.gz -O h3-pg.tgz && \
tar xvzf h3-pg.tgz && \
cd h3-pg-4.0.1 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3.control
#
# Layer "neon-pg-ext-build"
# compile neon extensions
#
FROM build-deps AS neon-pg-ext-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
# plv8 still sometimes crashes during the creation
# COPY --from=plv8-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /h3/usr /
COPY pgxn/ pgxn/
RUN make -j $(getconf _NPROCESSORS_ONLN) \
PG_CONFIG=/usr/local/pgsql/bin/pg_config \
-C pgxn/neon \
-s install
# Compile and run the Neon-specific `compute_ctl` binary
FROM 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:$TAG AS compute-tools
USER nonroot
# Copy entire project to get Cargo.* files with proper dependencies for the whole project
COPY --chown=nonroot . .
RUN cd compute_tools && cargo build --locked --profile release-line-debug-size-lto
#
# Clean up postgres folder before inclusion
#
FROM neon-pg-ext-build AS postgres-cleanup-layer
COPY --from=neon-pg-ext-build /usr/local/pgsql /usr/local/pgsql
# Remove binaries from /bin/ that we won't use (or would manually copy & install otherwise)
RUN cd /usr/local/pgsql/bin && rm ecpg raster2pgsql shp2pgsql pgtopo_export pgtopo_import pgsql2shp
# Remove headers that we won't need anymore - we've completed installation of all extensions
RUN rm -r /usr/local/pgsql/include
# Remove now-useless PGXS src infrastructure
RUN rm -r /usr/local/pgsql/lib/pgxs/src
# Remove static postgresql libraries - all compilation is finished, so we
# can now remove these files - they must be included in other binaries by now
# if they were to be used by other libraries.
RUN rm /usr/local/pgsql/lib/lib*.a
#
# Final layer
# Put it all together into the final image
#
FROM debian:bullseye-slim
# Add user postgres
RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
echo "postgres:test_console_pass" | chpasswd && \
mkdir /var/db/postgres/compute && mkdir /var/db/postgres/specs && \
chown -R postgres:postgres /var/db/postgres && \
chmod 0750 /var/db/postgres/compute && \
echo '/usr/local/lib' >> /etc/ld.so.conf && /sbin/ldconfig
COPY --from=postgres-cleanup-layer --chown=postgres /usr/local/pgsql /usr/local
COPY --from=compute-tools --chown=postgres /home/nonroot/target/release-line-debug-size-lto/compute_ctl /usr/local/bin/compute_ctl
# Install:
# libreadline8 for psql
# libossp-uuid16 for extension ossp-uuid
# libgeos, libgdal, libproj and libprotobuf-c1 for PostGIS
# GLIBC 2.34 for plv8.
# Debian bullseye provides GLIBC 2.31, so we install the library from testing
#
# Lastly, link compute_ctl into zenith_ctl while we're at it,
# so that we don't need to put this in another layer.
RUN apt update && \
apt install --no-install-recommends -y \
libreadline8 \
libossp-uuid16 \
libgeos-c1v5 \
libgdal28 \
libproj19 \
libprotobuf-c1 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
echo "Installing GLIBC 2.34" && \
echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing libc6 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
ln /usr/local/bin/compute_ctl /usr/local/bin/zenith_ctl
USER postgres
ENTRYPOINT ["/usr/local/bin/compute_ctl"]

205
Dockerfile.compute-node-v15 Normal file
View File

@@ -0,0 +1,205 @@
#
# This file is identical to the Dockerfile.compute-node-v14 file
# except for the version of Postgres that is built.
#
ARG TAG=pinned
# apparently, ARGs don't get replaced in RUN commands in kaniko
# ARG POSTGIS_VERSION=3.3.1
# ARG PLV8_VERSION=3.1.4
# ARG PG_VERSION=v15
#
# Layer "build-deps"
#
FROM debian:bullseye-slim AS build-deps
RUN echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update
RUN apt update && \
apt install -y git autoconf automake libtool build-essential bison flex libreadline-dev zlib1g-dev libxml2-dev \
libcurl4-openssl-dev libossp-uuid-dev wget pkg-config libglib2.0-dev
#
# Layer "pg-build"
# Build Postgres from the neon postgres repository.
#
FROM build-deps AS pg-build
COPY vendor/postgres-v15 postgres
RUN cd postgres && \
./configure CFLAGS='-O2 -g3' --enable-debug --with-uuid=ossp && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C contrib/ install && \
# Install headers
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/include install && \
make MAKELEVEL=0 -j $(getconf _NPROCESSORS_ONLN) -s -C src/interfaces/libpq install
#
# Layer "postgis-build"
# Build PostGIS from the upstream PostGIS mirror.
#
# PostGIS compiles against neon postgres sources without changes. Perhaps we
# could even use the upstream binaries, compiled against vanilla Postgres, but
# it would require some investigation to check that it works, and also keeps
# working in the future. So for now, we compile our own binaries.
FROM build-deps AS postgis-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y gdal-bin libgdal-dev libprotobuf-c-dev protobuf-c-compiler xsltproc
RUN wget https://download.osgeo.org/postgis/source/postgis-3.3.1.tar.gz && \
tar xvzf postgis-3.3.1.tar.gz && \
cd postgis-3.3.1 && \
./autogen.sh && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
./configure && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
cd extensions/postgis && \
make clean && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_raster.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_tiger_geocoder.control && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/postgis_topology.control
#
# Layer "plv8-build"
# Build plv8
#
FROM build-deps AS plv8-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
RUN apt update && \
apt install -y ninja-build python3-dev libc++-dev libc++abi-dev libncurses5
# https://github.com/plv8/plv8/issues/475
# Debian bullseye provides binutils 2.35 when >= 2.38 is necessary
RUN apt update && \
apt install -y --no-install-recommends -t testing binutils
RUN wget https://github.com/plv8/plv8/archive/refs/tags/v3.1.4.tar.gz && \
tar xvzf v3.1.4.tar.gz && \
cd plv8-3.1.4 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
rm -rf /plv8-* && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/plv8.control
#
# Layer "h3-pg-build"
# Build h3_pg
#
FROM build-deps AS h3-pg-build
COPY --from=pg-build /usr/local/pgsql/ /usr/local/pgsql/
# packaged cmake is too old
RUN apt update && \
apt install -y --no-install-recommends -t testing cmake
RUN wget https://github.com/uber/h3/archive/refs/tags/v4.0.1.tar.gz -O h3.tgz && \
tar xvzf h3.tgz && \
cd h3-4.0.1 && \
mkdir build && \
cd build && \
cmake .. -DCMAKE_BUILD_TYPE=Release && \
make -j $(getconf _NPROCESSORS_ONLN) && \
DESTDIR=/h3 make install && \
cp -R /h3/usr / && \
rm -rf build
RUN wget https://github.com/zachasme/h3-pg/archive/refs/tags/v4.0.1.tar.gz -O h3-pg.tgz && \
tar xvzf h3-pg.tgz && \
cd h3-pg-4.0.1 && \
export PATH="/usr/local/pgsql/bin:$PATH" && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/h3.control
#
# Layer "neon-pg-ext-build"
# compile neon extensions
#
FROM build-deps AS neon-pg-ext-build
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
# plv8 still sometimes crashes during the creation
# COPY --from=plv8-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=h3-pg-build /h3/usr /
COPY pgxn/ pgxn/
RUN make -j $(getconf _NPROCESSORS_ONLN) \
PG_CONFIG=/usr/local/pgsql/bin/pg_config \
-C pgxn/neon \
-s install
# Compile and run the Neon-specific `compute_ctl` binary
FROM 369495373322.dkr.ecr.eu-central-1.amazonaws.com/rust:$TAG AS compute-tools
USER nonroot
# Copy entire project to get Cargo.* files with proper dependencies for the whole project
COPY --chown=nonroot . .
RUN cd compute_tools && cargo build --locked --profile release-line-debug-size-lto
#
# Clean up postgres folder before inclusion
#
FROM neon-pg-ext-build AS postgres-cleanup-layer
COPY --from=neon-pg-ext-build /usr/local/pgsql /usr/local/pgsql
# Remove binaries from /bin/ that we won't use (or would manually copy & install otherwise)
RUN cd /usr/local/pgsql/bin && rm ecpg raster2pgsql shp2pgsql pgtopo_export pgtopo_import pgsql2shp
# Remove headers that we won't need anymore - we've completed installation of all extensions
RUN rm -r /usr/local/pgsql/include
# Remove now-useless PGXS src infrastructure
RUN rm -r /usr/local/pgsql/lib/pgxs/src
# Remove static postgresql libraries - all compilation is finished, so we
# can now remove these files - they must be included in other binaries by now
# if they were to be used by other libraries.
RUN rm /usr/local/pgsql/lib/lib*.a
#
# Final layer
# Put it all together into the final image
#
FROM debian:bullseye-slim
# Add user postgres
RUN mkdir /var/db && useradd -m -d /var/db/postgres postgres && \
echo "postgres:test_console_pass" | chpasswd && \
mkdir /var/db/postgres/compute && mkdir /var/db/postgres/specs && \
chown -R postgres:postgres /var/db/postgres && \
chmod 0750 /var/db/postgres/compute && \
echo '/usr/local/lib' >> /etc/ld.so.conf && /sbin/ldconfig
COPY --from=postgres-cleanup-layer --chown=postgres /usr/local/pgsql /usr/local
COPY --from=compute-tools --chown=postgres /home/nonroot/target/release-line-debug-size-lto/compute_ctl /usr/local/bin/compute_ctl
# Install:
# libreadline8 for psql
# libossp-uuid16 for extension ossp-uuid
# libgeos, libgdal, libproj and libprotobuf-c1 for PostGIS
# GLIBC 2.34 for plv8.
# Debian bullseye provides GLIBC 2.31, so we install the library from testing
#
# Lastly, link compute_ctl into zenith_ctl while we're at it,
# so that we don't need to put this in another layer.
RUN apt update && \
apt install --no-install-recommends -y \
libreadline8 \
libossp-uuid16 \
libgeos-c1v5 \
libgdal28 \
libproj19 \
libprotobuf-c1 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
echo "Installing GLIBC 2.34" && \
echo "deb http://ftp.debian.org/debian testing main" >> /etc/apt/sources.list && \
echo "APT::Default-Release \"stable\";" > /etc/apt/apt.conf.d/default-release && \
apt update && \
apt install -y --no-install-recommends -t testing libc6 && \
rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* && \
ln /usr/local/bin/compute_ctl /usr/local/bin/zenith_ctl
USER postgres
ENTRYPOINT ["/usr/local/bin/compute_ctl"]

7
Dockerfile.compute-node.legacy
View File

@@ -22,7 +22,7 @@ FROM $REPOSITORY/$IMAGE:$TAG AS compute-deps
#
# Image with Postgres build deps
#
FROM debian:buster-slim AS build-deps
FROM debian:bullseye-slim AS build-deps
RUN apt-get update && apt-get -yq install automake libtool build-essential bison flex libreadline-dev zlib1g-dev libxml2-dev \
libcurl4-openssl-dev libossp-uuid-dev
@@ -37,7 +37,8 @@ RUN adduser postgres
RUN mkdir /pg && chown postgres:postgres /pg
# Copy source files
COPY ./vendor/postgres /pg/
# version 14 is default for now
COPY ./vendor/postgres-v14 /pg/
COPY ./pgxn /pg/
# Build and install Postgres locally
@@ -58,7 +59,7 @@ WORKDIR /pg
#
# Final compute node image to be exported
#
FROM debian:buster-slim
FROM debian:bullseye-slim
# libreadline-dev is required to run psql
RUN apt-get update && apt-get -yq install libreadline-dev libossp-uuid-dev

179
Makefile
View File

@@ -1,15 +1,7 @@
ROOT_PROJECT_DIR := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
# Where to install Postgres, default is ./tmp_install, maybe useful for package managers
POSTGRES_INSTALL_DIR ?= $(ROOT_PROJECT_DIR)/tmp_install
# Seccomp BPF is only available for Linux
UNAME_S := $(shell uname -s)
ifeq ($(UNAME_S),Linux)
SECCOMP = --with-libseccomp
else
SECCOMP =
endif
# Where to install Postgres, default is ./pg_install, maybe useful for package managers
POSTGRES_INSTALL_DIR ?= $(ROOT_PROJECT_DIR)/pg_install/
#
# We differentiate between release / debug build types using the BUILD_TYPE
@@ -28,6 +20,12 @@ else
$(error Bad build type '$(BUILD_TYPE)', see Makefile for options)
endif
# Seccomp BPF is only available for Linux
UNAME_S := $(shell uname -s)
ifeq ($(UNAME_S),Linux)
PG_CONFIGURE_OPTS += --with-libseccomp
endif
# macOS with brew-installed openssl requires explicit paths
# It can be configured with OPENSSL_PREFIX variable
UNAME_S := $(shell uname -s)
@@ -36,6 +34,12 @@ ifeq ($(UNAME_S),Darwin)
PG_CONFIGURE_OPTS += --with-includes=$(OPENSSL_PREFIX)/include --with-libraries=$(OPENSSL_PREFIX)/lib
endif
# Use -C option so that when PostgreSQL "make install" installs the
# headers, the mtime of the headers are not changed when there have
# been no changes to the files. Changing the mtime triggers an
# unnecessary rebuild of 'postgres_ffi'.
PG_CONFIGURE_OPTS += INSTALL='$(ROOT_PROJECT_DIR)/scripts/ninstall.sh -C'
# Choose whether we should be silent or verbose
CARGO_BUILD_FLAGS += --$(if $(filter s,$(MAKEFLAGS)),quiet,verbose)
# Fix for a corner case when make doesn't pass a jobserver
@@ -48,75 +52,136 @@ CARGO_CMD_PREFIX += $(if $(filter n,$(MAKEFLAGS)),,+)
CARGO_CMD_PREFIX += CARGO_TERM_PROGRESS_WHEN=never CI=1
#
# Top level Makefile to build Zenith and PostgreSQL
# Top level Makefile to build Neon and PostgreSQL
#
.PHONY: all
all: zenith postgres neon-pg-ext
all: neon postgres neon-pg-ext
### Zenith Rust bits
### Neon Rust bits
#
# The 'postgres_ffi' depends on the Postgres headers.
.PHONY: zenith
zenith: postgres-headers
+@echo "Compiling Zenith"
.PHONY: neon
neon: postgres-v14-headers postgres-v15-headers
+@echo "Compiling Neon"
$(CARGO_CMD_PREFIX) cargo build $(CARGO_BUILD_FLAGS)
### PostgreSQL parts
$(POSTGRES_INSTALL_DIR)/build/config.status:
+@echo "Configuring postgres build"
mkdir -p $(POSTGRES_INSTALL_DIR)/build
(cd $(POSTGRES_INSTALL_DIR)/build && \
$(ROOT_PROJECT_DIR)/vendor/postgres/configure CFLAGS='$(PG_CFLAGS)' \
# The rules are duplicated for Postgres v14 and 15. We may want to refactor
# to avoid the duplication in the future, but it's tolerable for now.
#
$(POSTGRES_INSTALL_DIR)/build/v14/config.status:
+@echo "Configuring Postgres v14 build"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/v14
(cd $(POSTGRES_INSTALL_DIR)/build/v14 && \
$(ROOT_PROJECT_DIR)/vendor/postgres-v14/configure CFLAGS='$(PG_CFLAGS)' \
$(PG_CONFIGURE_OPTS) \
$(SECCOMP) \
--prefix=$(abspath $(POSTGRES_INSTALL_DIR)) > configure.log)
--prefix=$(abspath $(POSTGRES_INSTALL_DIR))/v14 > configure.log)
# nicer alias for running 'configure'
.PHONY: postgres-configure
postgres-configure: $(POSTGRES_INSTALL_DIR)/build/config.status
$(POSTGRES_INSTALL_DIR)/build/v15/config.status:
+@echo "Configuring Postgres v15 build"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/v15
(cd $(POSTGRES_INSTALL_DIR)/build/v15 && \
$(ROOT_PROJECT_DIR)/vendor/postgres-v15/configure CFLAGS='$(PG_CFLAGS)' \
$(PG_CONFIGURE_OPTS) \
--prefix=$(abspath $(POSTGRES_INSTALL_DIR))/v15 > configure.log)
# Install the PostgreSQL header files into $(POSTGRES_INSTALL_DIR)/include
.PHONY: postgres-headers
postgres-headers: postgres-configure
+@echo "Installing PostgreSQL headers"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/src/include MAKELEVEL=0 install
# nicer alias to run 'configure'
.PHONY: postgres-v14-configure
postgres-v14-configure: $(POSTGRES_INSTALL_DIR)/build/v14/config.status
# Compile and install PostgreSQL and contrib/neon
.PHONY: postgres
postgres: postgres-configure \
postgres-headers # to prevent `make install` conflicts with zenith's `postgres-headers`
+@echo "Compiling PostgreSQL"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build MAKELEVEL=0 install
+@echo "Compiling libpq"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/src/interfaces/libpq install
+@echo "Compiling pg_buffercache"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/contrib/pg_buffercache install
+@echo "Compiling pageinspect"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/contrib/pageinspect install
.PHONY: postgres-v15-configure
postgres-v15-configure: $(POSTGRES_INSTALL_DIR)/build/v15/config.status
.PHONY: postgres-clean
postgres-clean:
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build MAKELEVEL=0 clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/contrib/pg_buffercache clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/contrib/pageinspect clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/src/interfaces/libpq clean
# Install the PostgreSQL header files into $(POSTGRES_INSTALL_DIR)/<version>/include
.PHONY: postgres-v14-headers
postgres-v14-headers: postgres-v14-configure
+@echo "Installing PostgreSQL v14 headers"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/src/include MAKELEVEL=0 install
neon-pg-ext: postgres
+@echo "Compiling neon"
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/bin/pg_config \
-C $(ROOT_PROJECT_DIR)/pgxn/neon install
+@echo "Compiling neon_test_utils"
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/bin/pg_config \
-C $(ROOT_PROJECT_DIR)/pgxn/neon_test_utils install
.PHONY: postgres-v15-headers
postgres-v15-headers: postgres-v15-configure
+@echo "Installing PostgreSQL v15 headers"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/src/include MAKELEVEL=0 install
# Compile and install PostgreSQL
.PHONY: postgres-v14
postgres-v14: postgres-v14-configure \
postgres-v14-headers # to prevent `make install` conflicts with neon's `postgres-headers`
+@echo "Compiling PostgreSQL v14"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14 MAKELEVEL=0 install
+@echo "Compiling libpq v14"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/src/interfaces/libpq install
+@echo "Compiling pg_buffercache v14"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/contrib/pg_buffercache install
+@echo "Compiling pageinspect v14"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/contrib/pageinspect install
.PHONY: postgres-v15
postgres-v15: postgres-v15-configure \
postgres-v15-headers # to prevent `make install` conflicts with neon's `postgres-headers`
+@echo "Compiling PostgreSQL v15"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15 MAKELEVEL=0 install
+@echo "Compiling libpq v15"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/src/interfaces/libpq install
+@echo "Compiling pg_buffercache v15"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/contrib/pg_buffercache install
+@echo "Compiling pageinspect v15"
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/contrib/pageinspect install
# shorthand to build all Postgres versions
postgres: postgres-v14 postgres-v15
.PHONY: postgres-v14-clean
postgres-v14-clean:
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14 MAKELEVEL=0 clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/contrib/pg_buffercache clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/contrib/pageinspect clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v14/src/interfaces/libpq clean
.PHONY: postgres-v15-clean
postgres-v15-clean:
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15 MAKELEVEL=0 clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/contrib/pg_buffercache clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/contrib/pageinspect clean
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/v15/src/interfaces/libpq clean
neon-pg-ext-v14: postgres-v14
+@echo "Compiling neon v14"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-v14
(cd $(POSTGRES_INSTALL_DIR)/build/neon-v14 && \
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v14/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile install)
+@echo "Compiling neon_test_utils" v14
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-v14
(cd $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-v14 && \
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v14/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_test_utils/Makefile install)
neon-pg-ext-v15: postgres-v15
+@echo "Compiling neon v15"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-v15
(cd $(POSTGRES_INSTALL_DIR)/build/neon-v15 && \
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v15/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile install)
+@echo "Compiling neon_test_utils" v15
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-v15
(cd $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-v15 && \
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v15/bin/pg_config CFLAGS='$(PG_CFLAGS) $(COPT)' \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_test_utils/Makefile install)
.PHONY: neon-pg-ext-clean
$(MAKE) -C $(ROOT_PROJECT_DIR)/pgxn/neon clean
$(MAKE) -C $(ROOT_PROJECT_DIR)/pgxn/neon_test_utils clean
neon-pg-ext: neon-pg-ext-v14 neon-pg-ext-v15
postgres-headers: postgres-v14-headers postgres-v15-headers
postgres-clean: postgres-v14-clean postgres-v15-clean
# This doesn't remove the effects of 'configure'.
.PHONY: clean
clean:
cd $(POSTGRES_INSTALL_DIR)/build && $(MAKE) clean
cd $(POSTGRES_INSTALL_DIR)/build/v14 && $(MAKE) clean
cd $(POSTGRES_INSTALL_DIR)/build/v15 && $(MAKE) clean
$(CARGO_CMD_PREFIX) cargo clean
cd pgxn/neon && $(MAKE) clean
cd pgxn/neon_test_utils && $(MAKE) clean

4
NOTICE
View File

@@ -1,5 +1,5 @@
Neon
Copyright 2022 Neon Inc.
The PostgreSQL submodule in vendor/postgres is licensed under the
PostgreSQL license. See vendor/postgres/COPYRIGHT.
The PostgreSQL submodules in vendor/postgres-v14 and vendor/postgres-v15 are licensed under the
PostgreSQL license. See vendor/postgres-v14/COPYRIGHT and vendor/postgres-v15/COPYRIGHT.

45
README.md
View File

@@ -25,6 +25,7 @@ Pageserver consists of:
- WAL receiver - service that receives WAL from WAL service and stores it in the repository.
- Page service - service that communicates with compute nodes and responds with pages from the repository.
- WAL redo - service that builds pages from base images and WAL records on Page service request
## Running local installation
@@ -68,6 +69,17 @@ brew install libpq
brew link --force libpq
```
#### Rustc version
The project uses [rust toolchain file](./rust-toolchain.toml) to define the version it's built with in CI for testing and local builds.
This file is automatically picked up by [`rustup`](https://rust-lang.github.io/rustup/overrides.html#the-toolchain-file) that installs (if absent) and uses the toolchain version pinned in the file.
rustup users who want to build with another toolchain can use [`rustup override`](https://rust-lang.github.io/rustup/overrides.html#directory-overrides) command to set a specific toolchain for the project's directory.
non-rustup users most probably are not getting the same toolchain automatically from the file, so are responsible to manually verify their toolchain matches the version in the file.
Newer rustc versions most probably will work fine, yet older ones might not be supported due to some new features used by the project or the crates.
#### Building on Linux
1. Build neon and patched postgres
@@ -77,9 +89,9 @@ brew link --force libpq
git clone --recursive https://github.com/neondatabase/neon.git
cd neon
# The preferred and default is to make a debug build. This will create a
# demonstrably slower build than a release build. If you want to use a release
# build, utilize "BUILD_TYPE=release make -j`nproc`"
# The preferred and default is to make a debug build. This will create a
# demonstrably slower build than a release build. For a release build,
# use "BUILD_TYPE=release make -j`nproc`"
make -j`nproc`
```
@@ -93,15 +105,15 @@ make -j`nproc`
git clone --recursive https://github.com/neondatabase/neon.git
cd neon
# The preferred and default is to make a debug build. This will create a
# demonstrably slower build than a release build. If you want to use a release
# build, utilize "BUILD_TYPE=release make -j`sysctl -n hw.logicalcpu`"
# The preferred and default is to make a debug build. This will create a
# demonstrably slower build than a release build. For a release build,
# use "BUILD_TYPE=release make -j`sysctl -n hw.logicalcpu`"
make -j`sysctl -n hw.logicalcpu`
```
#### Dependency installation notes
To run the `psql` client, install the `postgresql-client` package or modify `PATH` and `LD_LIBRARY_PATH` to include `tmp_install/bin` and `tmp_install/lib`, respectively.
To run the `psql` client, install the `postgresql-client` package or modify `PATH` and `LD_LIBRARY_PATH` to include `pg_install/bin` and `pg_install/lib`, respectively.
To run the integration tests or Python scripts (not required to use the code), install
Python (3.9 or higher), and install python3 packages using `./scripts/pysync` (requires [poetry](https://python-poetry.org/)) in the project directory.
@@ -113,16 +125,18 @@ Python (3.9 or higher), and install python3 packages using `./scripts/pysync` (r
# Create repository in .neon with proper paths to binaries and data
# Later that would be responsibility of a package install script
> ./target/debug/neon_local init
initializing tenantid 9ef87a5bf0d92544f6fafeeb3239695c
created initial timeline de200bd42b49cc1814412c7e592dd6e9 timeline.lsn 0/16B5A50
initial timeline de200bd42b49cc1814412c7e592dd6e9 created
pageserver init succeeded
Starting pageserver at '127.0.0.1:64000' in '.neon'
Pageserver started
Successfully initialized timeline 7dd0907914ac399ff3be45fb252bfdb7
Stopping pageserver gracefully...done!
# start pageserver and safekeeper
> ./target/debug/neon_local start
Starting etcd broker using /usr/bin/etcd
Starting pageserver at '127.0.0.1:64000' in '.neon'
Pageserver started
initializing for sk 1 for 7676
Starting safekeeper at '127.0.0.1:5454' in '.neon/safekeepers/sk1'
Safekeeper started
@@ -208,7 +222,12 @@ Ensure your dependencies are installed as described [here](https://github.com/ne
```sh
git clone --recursive https://github.com/neondatabase/neon.git
make # builds also postgres and installs it to ./tmp_install
# either:
CARGO_BUILD_FLAGS="--features=testing" make
# or:
make debug
./scripts/pytest
```

6
compute_tools/Cargo.toml
View File

@@ -8,14 +8,16 @@ anyhow = "1.0"
chrono = "0.4"
clap = "3.0"
env_logger = "0.9"
futures = "0.3.13"
hyper = { version = "0.14", features = ["full"] }
log = { version = "0.4", features = ["std", "serde"] }
postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
notify = "5.0.0"
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
regex = "1"
serde = { version = "1.0", features = ["derive"] }
serde_json = "1"
tar = "0.4"
tokio = { version = "1.17", features = ["macros", "rt", "rt-multi-thread"] }
tokio-postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
tokio-postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
url = "2.2.2"
workspace_hack = { version = "0.1", path = "../workspace_hack" }

16
compute_tools/src/compute.rs
View File

@@ -178,7 +178,6 @@ impl ComputeNode {
.args(&["--sync-safekeepers"])
.env("PGDATA", &self.pgdata) // we cannot use -D in this mode
.stdout(Stdio::piped())
.stderr(Stdio::piped())
.spawn()
.expect("postgres --sync-safekeepers failed to start");
@@ -191,10 +190,10 @@ impl ComputeNode {
if !sync_output.status.success() {
anyhow::bail!(
"postgres --sync-safekeepers exited with non-zero status: {}. stdout: {}, stderr: {}",
"postgres --sync-safekeepers exited with non-zero status: {}. stdout: {}",
sync_output.status,
String::from_utf8(sync_output.stdout).expect("postgres --sync-safekeepers exited, and stdout is not utf-8"),
String::from_utf8(sync_output.stderr).expect("postgres --sync-safekeepers exited, and stderr is not utf-8"),
String::from_utf8(sync_output.stdout)
.expect("postgres --sync-safekeepers exited, and stdout is not utf-8"),
);
}
@@ -258,14 +257,7 @@ impl ComputeNode {
.spawn()
.expect("cannot start postgres process");
// Try default Postgres port if it is not provided
let port = self
.spec
.cluster
.settings
.find("port")
.unwrap_or_else(|| "5432".to_string());
wait_for_postgres(&mut pg, &port, pgdata_path)?;
wait_for_postgres(&mut pg, pgdata_path)?;
// If connection fails,
// it may be the old node with `zenith_admin` superuser.

105
compute_tools/src/pg_helpers.rs
View File

@@ -1,18 +1,19 @@
use std::fmt::Write;
use std::fs;
use std::fs::File;
use std::io::{BufRead, BufReader};
use std::net::{SocketAddr, TcpStream};
use std::os::unix::fs::PermissionsExt;
use std::path::Path;
use std::process::Child;
use std::str::FromStr;
use std::{fs, thread, time};
use std::time::{Duration, Instant};
use anyhow::{bail, Result};
use postgres::{Client, Transaction};
use serde::Deserialize;
const POSTGRES_WAIT_TIMEOUT: u64 = 60 * 1000; // milliseconds
use notify::{RecursiveMode, Watcher};
const POSTGRES_WAIT_TIMEOUT: Duration = Duration::from_millis(60 * 1000); // milliseconds
/// Rust representation of Postgres role info with only those fields
/// that matter for us.
@@ -230,52 +231,112 @@ pub fn get_existing_dbs(client: &mut Client) -> Result<Vec<Database>> {
Ok(postgres_dbs)
}
/// Wait for Postgres to become ready to accept connections:
/// - state should be `ready` in the `pgdata/postmaster.pid`
/// - and we should be able to connect to 127.0.0.1:5432
pub fn wait_for_postgres(pg: &mut Child, port: &str, pgdata: &Path) -> Result<()> {
/// Wait for Postgres to become ready to accept connections. It's ready to
/// accept connections when the state-field in `pgdata/postmaster.pid` says
/// 'ready'.
pub fn wait_for_postgres(pg: &mut Child, pgdata: &Path) -> Result<()> {
let pid_path = pgdata.join("postmaster.pid");
let mut slept: u64 = 0; // ms
let pause = time::Duration::from_millis(100);
let timeout = time::Duration::from_millis(10);
let addr = SocketAddr::from_str(&format!("127.0.0.1:{}", port)).unwrap();
// PostgreSQL writes line "ready" to the postmaster.pid file, when it has
// completed initialization and is ready to accept connections. We want to
// react quickly and perform the rest of our initialization as soon as
// PostgreSQL starts accepting connections. Use 'notify' to be notified
// whenever the PID file is changed, and whenever it changes, read it to
// check if it's now "ready".
//
// You cannot actually watch a file before it exists, so we first watch the
// data directory, and once the postmaster.pid file appears, we switch to
// watch the file instead. We also wake up every 100 ms to poll, just in
// case we miss some events for some reason. Not strictly necessary, but
// better safe than sorry.
let (tx, rx) = std::sync::mpsc::channel();
let (mut watcher, rx): (Box<dyn Watcher>, _) = match notify::recommended_watcher(move |res| {
let _ = tx.send(res);
}) {
Ok(watcher) => (Box::new(watcher), rx),
Err(e) => {
match e.kind {
notify::ErrorKind::Io(os) if os.raw_os_error() == Some(38) => {
// docker on m1 macs does not support recommended_watcher
// but return "Function not implemented (os error 38)"
// see https://github.com/notify-rs/notify/issues/423
let (tx, rx) = std::sync::mpsc::channel();
loop {
// Sleep POSTGRES_WAIT_TIMEOUT at max (a bit longer actually if consider a TCP timeout,
// but postgres starts listening almost immediately, even if it is not really
// ready to accept connections).
if slept >= POSTGRES_WAIT_TIMEOUT {
bail!("timed out while waiting for Postgres to start");
// let's poll it faster than what we check the results for (100ms)
let config =
notify::Config::default().with_poll_interval(Duration::from_millis(50));
let watcher = notify::PollWatcher::new(
move |res| {
let _ = tx.send(res);
},
config,
)?;
(Box::new(watcher), rx)
}
_ => return Err(e.into()),
}
}
};
watcher.watch(pgdata, RecursiveMode::NonRecursive)?;
let started_at = Instant::now();
let mut postmaster_pid_seen = false;
loop {
if let Ok(Some(status)) = pg.try_wait() {
// Postgres exited, that is not what we expected, bail out earlier.
let code = status.code().unwrap_or(-1);
bail!("Postgres exited unexpectedly with code {}", code);
}
let res = rx.recv_timeout(Duration::from_millis(100));
log::debug!("woken up by notify: {res:?}");
// If there are multiple events in the channel already, we only need to be
// check once. Swallow the extra events before we go ahead to check the
// pid file.
while let Ok(res) = rx.try_recv() {
log::debug!("swallowing extra event: {res:?}");
}
// Check that we can open pid file first.
if let Ok(file) = File::open(&pid_path) {
if !postmaster_pid_seen {
log::debug!("postmaster.pid appeared");
watcher
.unwatch(pgdata)
.expect("Failed to remove pgdata dir watch");
watcher
.watch(&pid_path, RecursiveMode::NonRecursive)
.expect("Failed to add postmaster.pid file watch");
postmaster_pid_seen = true;
}
let file = BufReader::new(file);
let last_line = file.lines().last();
// Pid file could be there and we could read it, but it could be empty, for example.
if let Some(Ok(line)) = last_line {
let status = line.trim();
let can_connect = TcpStream::connect_timeout(&addr, timeout).is_ok();
log::debug!("last line of postmaster.pid: {status:?}");
// Now Postgres is ready to accept connections
if status == "ready" && can_connect {
if status == "ready" {
break;
}
}
}
thread::sleep(pause);
slept += 100;
// Give up after POSTGRES_WAIT_TIMEOUT.
let duration = started_at.elapsed();
if duration >= POSTGRES_WAIT_TIMEOUT {
bail!("timed out while waiting for Postgres to start");
}
}
log::info!("PostgreSQL is now running, continuing to configure it");
Ok(())
}

11
control_plane/Cargo.toml
View File

@@ -4,8 +4,11 @@ version = "0.1.0"
edition = "2021"
[dependencies]
clap = "3.0"
comfy-table = "5.0.1"
git-version = "0.3.5"
tar = "0.4.38"
postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
serde = { version = "1.0", features = ["derive"] }
serde_with = "1.12.0"
toml = "0.5"
@@ -16,7 +19,9 @@ thiserror = "1"
nix = "0.23"
reqwest = { version = "0.11", default-features = false, features = ["blocking", "json", "rustls-tls"] }
pageserver = { path = "../pageserver" }
safekeeper = { path = "../safekeeper" }
# Note: Do not directly depend on pageserver or safekeeper; use pageserver_api or safekeeper_api
# instead, so that recompile times are better.
pageserver_api = { path = "../libs/pageserver_api" }
safekeeper_api = { path = "../libs/safekeeper_api" }
utils = { path = "../libs/utils" }
workspace_hack = { version = "0.1", path = "../workspace_hack" }

2
control_plane/simple.conf
View File

@@ -1,4 +1,4 @@
# Minimal zenith environment with one safekeeper. This is equivalent to the built-in
# Minimal neon environment with one safekeeper. This is equivalent to the built-in
# defaults that you get with no --config
[pageserver]
listen_pg_addr = '127.0.0.1:64000'

152
neon_local/src/main.rs → control_plane/src/bin/neon_local.rs
View File

@@ -1,3 +1,10 @@
//!
//! `neon_local` is an executable that can be used to create a local
//! Neon environment, for testing purposes. The local environment is
//! quite different from the cloud environment with Kubernetes, but it
//! easier to work with locally. The python tests in `test_runner`
//! rely on `neon_local` to set up the environment for each test.
//!
use anyhow::{anyhow, bail, Context, Result};
use clap::{App, AppSettings, Arg, ArgMatches};
use control_plane::compute::ComputeControlPlane;
@@ -5,12 +12,12 @@ use control_plane::local_env::{EtcdBroker, LocalEnv};
use control_plane::safekeeper::SafekeeperNode;
use control_plane::storage::PageServerNode;
use control_plane::{etcd, local_env};
use pageserver::config::defaults::{
use pageserver_api::models::TimelineInfo;
use pageserver_api::{
DEFAULT_HTTP_LISTEN_ADDR as DEFAULT_PAGESERVER_HTTP_ADDR,
DEFAULT_PG_LISTEN_ADDR as DEFAULT_PAGESERVER_PG_ADDR,
};
use pageserver::http::models::TimelineInfo;
use safekeeper::defaults::{
use safekeeper_api::{
DEFAULT_HTTP_LISTEN_PORT as DEFAULT_SAFEKEEPER_HTTP_PORT,
DEFAULT_PG_LISTEN_PORT as DEFAULT_SAFEKEEPER_PG_PORT,
};
@@ -20,10 +27,10 @@ use std::process::exit;
use std::str::FromStr;
use utils::{
auth::{Claims, Scope},
id::{NodeId, TenantId, TenantTimelineId, TimelineId},
lsn::Lsn,
postgres_backend::AuthType,
project_git_version,
zid::{NodeId, ZTenantId, ZTenantTimelineId, ZTimelineId},
};
// Default id of a safekeeper node, if not specified on the command line.
@@ -32,6 +39,8 @@ const DEFAULT_PAGESERVER_ID: NodeId = NodeId(1);
const DEFAULT_BRANCH_NAME: &str = "main";
project_git_version!(GIT_VERSION);
const DEFAULT_PG_VERSION: &str = "14";
fn default_conf(etcd_binary_path: &Path) -> String {
format!(
r#"
@@ -65,7 +74,7 @@ struct TimelineTreeEl {
/// Name, recovered from neon config mappings
pub name: Option<String>,
/// Holds all direct children of this timeline referenced using `timeline_id`.
pub children: BTreeSet<ZTimelineId>,
pub children: BTreeSet<TimelineId>,
}
// Main entry point for the 'neon_local' CLI utility
@@ -98,6 +107,13 @@ fn main() -> Result<()> {
.takes_value(true)
.required(false);
let pg_version_arg = Arg::new("pg-version")
.long("pg-version")
.help("Postgres version to use for the initial tenant")
.required(false)
.takes_value(true)
.default_value(DEFAULT_PG_VERSION);
let port_arg = Arg::new("port")
.long("port")
.required(false)
@@ -139,6 +155,7 @@ fn main() -> Result<()> {
.required(false)
.value_name("config"),
)
.arg(pg_version_arg.clone())
)
.subcommand(
App::new("timeline")
@@ -157,7 +174,9 @@ fn main() -> Result<()> {
.subcommand(App::new("create")
.about("Create a new blank timeline")
.arg(tenant_id_arg.clone())
.arg(branch_name_arg.clone()))
.arg(branch_name_arg.clone())
.arg(pg_version_arg.clone())
)
.subcommand(App::new("import")
.about("Import timeline from basebackup directory")
.arg(tenant_id_arg.clone())
@@ -171,7 +190,9 @@ fn main() -> Result<()> {
.arg(Arg::new("wal-tarfile").long("wal-tarfile").takes_value(true)
.help("Wal to add after base"))
.arg(Arg::new("end-lsn").long("end-lsn").takes_value(true)
.help("Lsn the basebackup ends at")))
.help("Lsn the basebackup ends at"))
.arg(pg_version_arg.clone())
)
).subcommand(
App::new("tenant")
.setting(AppSettings::ArgRequiredElseHelp)
@@ -181,6 +202,7 @@ fn main() -> Result<()> {
.arg(tenant_id_arg.clone())
.arg(timeline_id_arg.clone().help("Use a specific timeline id when creating a tenant and its initial timeline"))
.arg(Arg::new("config").short('c').takes_value(true).multiple_occurrences(true).required(false))
.arg(pg_version_arg.clone())
)
.subcommand(App::new("config")
.arg(tenant_id_arg.clone())
@@ -232,8 +254,9 @@ fn main() -> Result<()> {
Arg::new("config-only")
.help("Don't do basebackup, create compute node with only config files")
.long("config-only")
.required(false)
))
.required(false))
.arg(pg_version_arg.clone())
)
.subcommand(App::new("start")
.about("Start a postgres compute node.\n This command actually creates new node from scratch, but preserves existing config files")
.arg(pg_node_arg.clone())
@@ -241,7 +264,9 @@ fn main() -> Result<()> {
.arg(branch_name_arg.clone())
.arg(timeline_id_arg.clone())
.arg(lsn_arg.clone())
.arg(port_arg.clone()))
.arg(port_arg.clone())
.arg(pg_version_arg.clone())
)
.subcommand(
App::new("stop")
.arg(pg_node_arg.clone())
@@ -314,7 +339,7 @@ fn main() -> Result<()> {
///
fn print_timelines_tree(
timelines: Vec<TimelineInfo>,
mut timeline_name_mappings: HashMap<ZTenantTimelineId, String>,
mut timeline_name_mappings: HashMap<TenantTimelineId, String>,
) -> Result<()> {
let mut timelines_hash = timelines
.iter()
@@ -325,7 +350,7 @@ fn print_timelines_tree(
info: t.clone(),
children: BTreeSet::new(),
name: timeline_name_mappings
.remove(&ZTenantTimelineId::new(t.tenant_id, t.timeline_id)),
.remove(&TenantTimelineId::new(t.tenant_id, t.timeline_id)),
},
)
})
@@ -367,7 +392,7 @@ fn print_timeline(
nesting_level: usize,
is_last: &[bool],
timeline: &TimelineTreeEl,
timelines: &HashMap<ZTimelineId, TimelineTreeEl>,
timelines: &HashMap<TimelineId, TimelineTreeEl>,
) -> Result<()> {
let local_remote = match (timeline.info.local.as_ref(), timeline.info.remote.as_ref()) {
(None, None) => unreachable!("in this case no info for a timeline is found"),
@@ -445,8 +470,8 @@ fn print_timeline(
/// Connects to the pageserver to query this information.
fn get_timeline_infos(
env: &local_env::LocalEnv,
tenant_id: &ZTenantId,
) -> Result<HashMap<ZTimelineId, TimelineInfo>> {
tenant_id: &TenantId,
) -> Result<HashMap<TimelineId, TimelineInfo>> {
Ok(PageServerNode::from_env(env)
.timeline_list(tenant_id)?
.into_iter()
@@ -455,7 +480,7 @@ fn get_timeline_infos(
}
// Helper function to parse --tenant_id option, or get the default from config file
fn get_tenant_id(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow::Result<ZTenantId> {
fn get_tenant_id(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow::Result<TenantId> {
if let Some(tenant_id_from_arguments) = parse_tenant_id(sub_match).transpose() {
tenant_id_from_arguments
} else if let Some(default_id) = env.default_tenant_id {
@@ -465,18 +490,18 @@ fn get_tenant_id(sub_match: &ArgMatches, env: &local_env::LocalEnv) -> anyhow::R
}
}
fn parse_tenant_id(sub_match: &ArgMatches) -> anyhow::Result<Option<ZTenantId>> {
fn parse_tenant_id(sub_match: &ArgMatches) -> anyhow::Result<Option<TenantId>> {
sub_match
.value_of("tenant-id")
.map(ZTenantId::from_str)
.map(TenantId::from_str)
.transpose()
.context("Failed to parse tenant id from the argument string")
}
fn parse_timeline_id(sub_match: &ArgMatches) -> anyhow::Result<Option<ZTimelineId>> {
fn parse_timeline_id(sub_match: &ArgMatches) -> anyhow::Result<Option<TimelineId>> {
sub_match
.value_of("timeline-id")
.map(ZTimelineId::from_str)
.map(TimelineId::from_str)
.transpose()
.context("Failed to parse timeline id from the argument string")
}
@@ -494,12 +519,19 @@ fn handle_init(init_match: &ArgMatches) -> anyhow::Result<LocalEnv> {
default_conf(&EtcdBroker::locate_etcd()?)
};
let pg_version = init_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
let mut env =
LocalEnv::parse_config(&toml_file).context("Failed to create neon configuration")?;
env.init().context("Failed to initialize neon repository")?;
// default_tenantid was generated by the `env.init()` call above
let initial_tenant_id = env.default_tenant_id.unwrap();
env.init(pg_version)
.context("Failed to initialize neon repository")?;
let initial_tenant_id = env
.default_tenant_id
.expect("default_tenant_id should be generated by the `env.init()` call above");
// Initialize pageserver, create initial tenant and timeline.
let pageserver = PageServerNode::from_env(&env);
@@ -508,6 +540,7 @@ fn handle_init(init_match: &ArgMatches) -> anyhow::Result<LocalEnv> {
Some(initial_tenant_id),
initial_timeline_id_arg,
&pageserver_config_overrides(init_match),
pg_version,
)
.unwrap_or_else(|e| {
eprintln!("pageserver init failed: {e}");
@@ -536,13 +569,7 @@ fn handle_tenant(tenant_match: &ArgMatches, env: &mut local_env::LocalEnv) -> an
match tenant_match.subcommand() {
Some(("list", _)) => {
for t in pageserver.tenant_list()? {
println!(
"{} {}",
t.id,
t.state
.map(|s| s.to_string())
.unwrap_or_else(|| String::from(""))
);
println!("{} {:?}", t.id, t.state);
}
}
Some(("create", create_match)) => {
@@ -556,8 +583,19 @@ fn handle_tenant(tenant_match: &ArgMatches, env: &mut local_env::LocalEnv) -> an
// Create an initial timeline for the new tenant
let new_timeline_id = parse_timeline_id(create_match)?;
let timeline_info =
pageserver.timeline_create(new_tenant_id, new_timeline_id, None, None)?;
let pg_version = create_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
let timeline_info = pageserver.timeline_create(
new_tenant_id,
new_timeline_id,
None,
None,
Some(pg_version),
)?;
let new_timeline_id = timeline_info.timeline_id;
let last_record_lsn = timeline_info
.local
@@ -606,7 +644,15 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
let new_branch_name = create_match
.value_of("branch-name")
.ok_or_else(|| anyhow!("No branch name provided"))?;
let timeline_info = pageserver.timeline_create(tenant_id, None, None, None)?;
let pg_version = create_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
let timeline_info =
pageserver.timeline_create(tenant_id, None, None, None, Some(pg_version))?;
let new_timeline_id = timeline_info.timeline_id;
let last_record_lsn = timeline_info
@@ -649,12 +695,19 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
// TODO validate both or none are provided
let pg_wal = end_lsn.zip(wal_tarfile);
let pg_version = import_match
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
let mut cplane = ComputeControlPlane::load(env.clone())?;
println!("Importing timeline into pageserver ...");
pageserver.timeline_import(tenant_id, timeline_id, base, pg_wal)?;
pageserver.timeline_import(tenant_id, timeline_id, base, pg_wal, pg_version)?;
println!("Creating node for imported timeline ...");
env.register_branch_mapping(name.to_string(), tenant_id, timeline_id)?;
cplane.new_node(tenant_id, name, timeline_id, None, None)?;
cplane.new_node(tenant_id, name, timeline_id, None, None, pg_version)?;
println!("Done");
}
Some(("branch", branch_match)) => {
@@ -681,6 +734,7 @@ fn handle_timeline(timeline_match: &ArgMatches, env: &mut local_env::LocalEnv) -
None,
start_lsn,
Some(ancestor_timeline_id),
None,
)?;
let new_timeline_id = timeline_info.timeline_id;
@@ -758,7 +812,7 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
};
let branch_name = timeline_name_mappings
.get(&ZTenantTimelineId::new(tenant_id, node.timeline_id))
.get(&TenantTimelineId::new(tenant_id, node.timeline_id))
.map(|name| name.as_str())
.unwrap_or("?");
@@ -796,7 +850,14 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
Some(p) => Some(p.parse()?),
None => None,
};
cplane.new_node(tenant_id, &node_name, timeline_id, lsn, port)?;
let pg_version = sub_args
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
cplane.new_node(tenant_id, &node_name, timeline_id, lsn, port, pg_version)?;
}
"start" => {
let port: Option<u16> = match sub_args.value_of("port") {
@@ -809,7 +870,7 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
let node = cplane.nodes.get(&(tenant_id, node_name.to_owned()));
let auth_token = if matches!(env.pageserver.auth_type, AuthType::ZenithJWT) {
let auth_token = if matches!(env.pageserver.auth_type, AuthType::NeonJWT) {
let claims = Claims::new(Some(tenant_id), Scope::Tenant);
Some(env.generate_auth_token(&claims)?)
@@ -834,16 +895,23 @@ fn handle_pg(pg_match: &ArgMatches, env: &local_env::LocalEnv) -> Result<()> {
.map(Lsn::from_str)
.transpose()
.context("Failed to parse Lsn from the request")?;
let pg_version = sub_args
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?;
// when used with custom port this results in non obvious behaviour
// port is remembered from first start command, i e
// start --port X
// stop
// start <-- will also use port X even without explicit port argument
println!(
"Starting new postgres {} on timeline {} ...",
node_name, timeline_id
"Starting new postgres (v{}) {} on timeline {} ...",
pg_version, node_name, timeline_id
);
let node = cplane.new_node(tenant_id, node_name, timeline_id, lsn, port)?;
let node =
cplane.new_node(tenant_id, node_name, timeline_id, lsn, port, pg_version)?;
node.start(&auth_token)?;
}
}

69
control_plane/src/compute.rs
View File

@@ -13,12 +13,12 @@ use std::time::Duration;
use anyhow::{Context, Result};
use utils::{
connstring::connection_host_port,
id::{TenantId, TimelineId},
lsn::Lsn,
postgres_backend::AuthType,
zid::{ZTenantId, ZTimelineId},
};
use crate::local_env::LocalEnv;
use crate::local_env::{LocalEnv, DEFAULT_PG_VERSION};
use crate::postgresql_conf::PostgresConf;
use crate::storage::PageServerNode;
@@ -28,7 +28,7 @@ use crate::storage::PageServerNode;
pub struct ComputeControlPlane {
base_port: u16,
pageserver: Arc<PageServerNode>,
pub nodes: BTreeMap<(ZTenantId, String), Arc<PostgresNode>>,
pub nodes: BTreeMap<(TenantId, String), Arc<PostgresNode>>,
env: LocalEnv,
}
@@ -76,11 +76,12 @@ impl ComputeControlPlane {
pub fn new_node(
&mut self,
tenant_id: ZTenantId,
tenant_id: TenantId,
name: &str,
timeline_id: ZTimelineId,
timeline_id: TimelineId,
lsn: Option<Lsn>,
port: Option<u16>,
pg_version: u32,
) -> Result<Arc<PostgresNode>> {
let port = port.unwrap_or_else(|| self.get_port());
let node = Arc::new(PostgresNode {
@@ -93,6 +94,7 @@ impl ComputeControlPlane {
lsn,
tenant_id,
uses_wal_proposer: false,
pg_version,
});
node.create_pgdata()?;
@@ -114,10 +116,11 @@ pub struct PostgresNode {
pub env: LocalEnv,
pageserver: Arc<PageServerNode>,
is_test: bool,
pub timeline_id: ZTimelineId,
pub timeline_id: TimelineId,
pub lsn: Option<Lsn>, // if it's a read-only node. None for primary
pub tenant_id: ZTenantId,
pub tenant_id: TenantId,
uses_wal_proposer: bool,
pg_version: u32,
}
impl PostgresNode {
@@ -148,10 +151,18 @@ impl PostgresNode {
// Read a few options from the config file
let context = format!("in config file {}", cfg_path_str);
let port: u16 = conf.parse_field("port", &context)?;
let timeline_id: ZTimelineId = conf.parse_field("neon.timeline_id", &context)?;
let tenant_id: ZTenantId = conf.parse_field("neon.tenant_id", &context)?;
let timeline_id: TimelineId = conf.parse_field("neon.timeline_id", &context)?;
let tenant_id: TenantId = conf.parse_field("neon.tenant_id", &context)?;
let uses_wal_proposer = conf.get("neon.safekeepers").is_some();
// Read postgres version from PG_VERSION file to determine which postgres version binary to use.
// If it doesn't exist, assume broken data directory and use default pg version.
let pg_version_path = entry.path().join("PG_VERSION");
let pg_version_str =
fs::read_to_string(pg_version_path).unwrap_or_else(|_| DEFAULT_PG_VERSION.to_string());
let pg_version = u32::from_str(&pg_version_str)?;
// parse recovery_target_lsn, if any
let recovery_target_lsn: Option<Lsn> =
conf.parse_field_optional("recovery_target_lsn", &context)?;
@@ -167,17 +178,24 @@ impl PostgresNode {
lsn: recovery_target_lsn,
tenant_id,
uses_wal_proposer,
pg_version,
})
}
fn sync_safekeepers(&self, auth_token: &Option<String>) -> Result<Lsn> {
let pg_path = self.env.pg_bin_dir().join("postgres");
fn sync_safekeepers(&self, auth_token: &Option<String>, pg_version: u32) -> Result<Lsn> {
let pg_path = self.env.pg_bin_dir(pg_version).join("postgres");
let mut cmd = Command::new(&pg_path);
cmd.arg("--sync-safekeepers")
.env_clear()
.env("LD_LIBRARY_PATH", self.env.pg_lib_dir().to_str().unwrap())
.env("DYLD_LIBRARY_PATH", self.env.pg_lib_dir().to_str().unwrap())
.env(
"LD_LIBRARY_PATH",
self.env.pg_lib_dir(pg_version).to_str().unwrap(),
)
.env(
"DYLD_LIBRARY_PATH",
self.env.pg_lib_dir(pg_version).to_str().unwrap(),
)
.env("PGDATA", self.pgdata().to_str().unwrap())
.stdout(Stdio::piped())
// Comment this to avoid capturing stderr (useful if command hangs)
@@ -259,8 +277,8 @@ impl PostgresNode {
})
}
// Connect to a page server, get base backup, and untar it to initialize a
// new data directory
// Write postgresql.conf with default configuration
// and PG_VERSION file to the data directory of a new node.
fn setup_pg_conf(&self, auth_type: AuthType) -> Result<()> {
let mut conf = PostgresConf::new();
conf.append("max_wal_senders", "10");
@@ -292,7 +310,7 @@ impl PostgresNode {
// variable during compute pg startup. It is done this way because
// otherwise user will be able to retrieve the value using SHOW
// command or pg_settings
let password = if let AuthType::ZenithJWT = auth_type {
let password = if let AuthType::NeonJWT = auth_type {
"$ZENITH_AUTH_TOKEN"
} else {
""
@@ -301,7 +319,7 @@ impl PostgresNode {
// Also note that not all parameters are supported here. Because in compute we substitute $ZENITH_AUTH_TOKEN
// We parse this string and build it back with token from env var, and for simplicity rebuild
// uses only needed variables namely host, port, user, password.
format!("postgresql://no_user:{}@{}:{}", password, host, port)
format!("postgresql://no_user:{password}@{host}:{port}")
};
conf.append("shared_preload_libraries", "neon");
conf.append_line("");
@@ -357,6 +375,9 @@ impl PostgresNode {
let mut file = File::create(self.pgdata().join("postgresql.conf"))?;
file.write_all(conf.to_string().as_bytes())?;
let mut file = File::create(self.pgdata().join("PG_VERSION"))?;
file.write_all(self.pg_version.to_string().as_bytes())?;
Ok(())
}
@@ -368,7 +389,7 @@ impl PostgresNode {
// latest data from the pageserver. That is a bit clumsy but whole bootstrap
// procedure evolves quite actively right now, so let's think about it again
// when things would be more stable (TODO).
let lsn = self.sync_safekeepers(auth_token)?;
let lsn = self.sync_safekeepers(auth_token, self.pg_version)?;
if lsn == Lsn(0) {
None
} else {
@@ -401,7 +422,7 @@ impl PostgresNode {
}
fn pg_ctl(&self, args: &[&str], auth_token: &Option<String>) -> Result<()> {
let pg_ctl_path = self.env.pg_bin_dir().join("pg_ctl");
let pg_ctl_path = self.env.pg_bin_dir(self.pg_version).join("pg_ctl");
let mut cmd = Command::new(pg_ctl_path);
cmd.args(
[
@@ -417,8 +438,14 @@ impl PostgresNode {
.concat(),
)
.env_clear()
.env("LD_LIBRARY_PATH", self.env.pg_lib_dir().to_str().unwrap())
.env("DYLD_LIBRARY_PATH", self.env.pg_lib_dir().to_str().unwrap());
.env(
"LD_LIBRARY_PATH",
self.env.pg_lib_dir(self.pg_version).to_str().unwrap(),
)
.env(
"DYLD_LIBRARY_PATH",
self.env.pg_lib_dir(self.pg_version).to_str().unwrap(),
);
if let Some(token) = auth_token {
cmd.env("ZENITH_AUTH_TOKEN", token);
}

99
control_plane/src/local_env.rs
View File

@@ -14,12 +14,14 @@ use std::path::{Path, PathBuf};
use std::process::{Command, Stdio};
use utils::{
auth::{encode_from_key_file, Claims, Scope},
id::{NodeId, TenantId, TenantTimelineId, TimelineId},
postgres_backend::AuthType,
zid::{NodeId, ZTenantId, ZTenantTimelineId, ZTimelineId},
};
use crate::safekeeper::SafekeeperNode;
pub const DEFAULT_PG_VERSION: u32 = 14;
//
// This data structures represents neon_local CLI config
//
@@ -48,13 +50,13 @@ pub struct LocalEnv {
// Path to pageserver binary.
#[serde(default)]
pub zenith_distrib_dir: PathBuf,
pub neon_distrib_dir: PathBuf,
// Default tenant ID to use with the 'zenith' command line utility, when
// --tenantid is not explicitly specified.
// Default tenant ID to use with the 'neon_local' command line utility, when
// --tenant_id is not explicitly specified.
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub default_tenant_id: Option<ZTenantId>,
pub default_tenant_id: Option<TenantId>,
// used to issue tokens during e.g pg start
#[serde(default)]
@@ -69,11 +71,11 @@ pub struct LocalEnv {
/// Keep human-readable aliases in memory (and persist them to config), to hide ZId hex strings from the user.
#[serde(default)]
// A `HashMap<String, HashMap<ZTenantId, ZTimelineId>>` would be more appropriate here,
// A `HashMap<String, HashMap<TenantId, TimelineId>>` would be more appropriate here,
// but deserialization into a generic toml object as `toml::Value::try_from` fails with an error.
// https://toml.io/en/v1.0.0 does not contain a concept of "a table inside another table".
#[serde_as(as = "HashMap<_, Vec<(DisplayFromStr, DisplayFromStr)>>")]
branch_name_mappings: HashMap<String, Vec<(ZTenantId, ZTimelineId)>>,
branch_name_mappings: HashMap<String, Vec<(TenantId, TimelineId)>>,
}
/// Etcd broker config for cluster internal communication.
@@ -195,29 +197,50 @@ impl Default for SafekeeperConf {
}
impl LocalEnv {
// postgres installation paths
pub fn pg_bin_dir(&self) -> PathBuf {
self.pg_distrib_dir.join("bin")
pub fn pg_distrib_dir_raw(&self) -> PathBuf {
self.pg_distrib_dir.clone()
}
pub fn pg_lib_dir(&self) -> PathBuf {
self.pg_distrib_dir.join("lib")
pub fn pg_distrib_dir(&self, pg_version: u32) -> PathBuf {
let path = self.pg_distrib_dir.clone();
match pg_version {
14 => path.join(format!("v{pg_version}")),
15 => path.join(format!("v{pg_version}")),
_ => panic!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pg_bin_dir(&self, pg_version: u32) -> PathBuf {
match pg_version {
14 => self.pg_distrib_dir(pg_version).join("bin"),
15 => self.pg_distrib_dir(pg_version).join("bin"),
_ => panic!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pg_lib_dir(&self, pg_version: u32) -> PathBuf {
match pg_version {
14 => self.pg_distrib_dir(pg_version).join("lib"),
15 => self.pg_distrib_dir(pg_version).join("lib"),
_ => panic!("Unsupported postgres version: {}", pg_version),
}
}
pub fn pageserver_bin(&self) -> anyhow::Result<PathBuf> {
Ok(self.zenith_distrib_dir.join("pageserver"))
Ok(self.neon_distrib_dir.join("pageserver"))
}
pub fn safekeeper_bin(&self) -> anyhow::Result<PathBuf> {
Ok(self.zenith_distrib_dir.join("safekeeper"))
Ok(self.neon_distrib_dir.join("safekeeper"))
}
pub fn pg_data_dirs_path(&self) -> PathBuf {
self.base_data_dir.join("pgdatadirs").join("tenants")
}
pub fn pg_data_dir(&self, tenantid: &ZTenantId, branch_name: &str) -> PathBuf {
pub fn pg_data_dir(&self, tenant_id: &TenantId, branch_name: &str) -> PathBuf {
self.pg_data_dirs_path()
.join(tenantid.to_string())
.join(tenant_id.to_string())
.join(branch_name)
}
@@ -233,8 +256,8 @@ impl LocalEnv {
pub fn register_branch_mapping(
&mut self,
branch_name: String,
tenant_id: ZTenantId,
timeline_id: ZTimelineId,
tenant_id: TenantId,
timeline_id: TimelineId,
) -> anyhow::Result<()> {
let existing_values = self
.branch_name_mappings
@@ -260,22 +283,22 @@ impl LocalEnv {
pub fn get_branch_timeline_id(
&self,
branch_name: &str,
tenant_id: ZTenantId,
) -> Option<ZTimelineId> {
tenant_id: TenantId,
) -> Option<TimelineId> {
self.branch_name_mappings
.get(branch_name)?
.iter()
.find(|(mapped_tenant_id, _)| mapped_tenant_id == &tenant_id)
.map(|&(_, timeline_id)| timeline_id)
.map(ZTimelineId::from)
.map(TimelineId::from)
}
pub fn timeline_name_mappings(&self) -> HashMap<ZTenantTimelineId, String> {
pub fn timeline_name_mappings(&self) -> HashMap<TenantTimelineId, String> {
self.branch_name_mappings
.iter()
.flat_map(|(name, tenant_timelines)| {
tenant_timelines.iter().map(|&(tenant_id, timeline_id)| {
(ZTenantTimelineId::new(tenant_id, timeline_id), name.clone())
(TenantTimelineId::new(tenant_id, timeline_id), name.clone())
})
})
.collect()
@@ -289,24 +312,26 @@ impl LocalEnv {
let mut env: LocalEnv = toml::from_str(toml)?;
// Find postgres binaries.
// Follow POSTGRES_DISTRIB_DIR if set, otherwise look in "tmp_install".
// Follow POSTGRES_DISTRIB_DIR if set, otherwise look in "pg_install".
// Note that later in the code we assume, that distrib dirs follow the same pattern
// for all postgres versions.
if env.pg_distrib_dir == Path::new("") {
if let Some(postgres_bin) = env::var_os("POSTGRES_DISTRIB_DIR") {
env.pg_distrib_dir = postgres_bin.into();
} else {
let cwd = env::current_dir()?;
env.pg_distrib_dir = cwd.join("tmp_install")
env.pg_distrib_dir = cwd.join("pg_install")
}
}
// Find zenith binaries.
if env.zenith_distrib_dir == Path::new("") {
env.zenith_distrib_dir = env::current_exe()?.parent().unwrap().to_owned();
// Find neon binaries.
if env.neon_distrib_dir == Path::new("") {
env.neon_distrib_dir = env::current_exe()?.parent().unwrap().to_owned();
}
// If no initial tenant ID was given, generate it.
if env.default_tenant_id.is_none() {
env.default_tenant_id = Some(ZTenantId::generate());
env.default_tenant_id = Some(TenantId::generate());
}
env.base_data_dir = base_path();
@@ -320,12 +345,12 @@ impl LocalEnv {
if !repopath.exists() {
bail!(
"Zenith config is not found in {}. You need to run 'neon_local init' first",
"Neon config is not found in {}. You need to run 'neon_local init' first",
repopath.to_str().unwrap()
);
}
// TODO: check that it looks like a zenith repository
// TODO: check that it looks like a neon repository
// load and parse file
let config = fs::read_to_string(repopath.join("config"))?;
@@ -384,7 +409,7 @@ impl LocalEnv {
//
// Initialize a new Neon repository
//
pub fn init(&mut self) -> anyhow::Result<()> {
pub fn init(&mut self, pg_version: u32) -> anyhow::Result<()> {
// check if config already exists
let base_path = &self.base_data_dir;
ensure!(
@@ -397,17 +422,17 @@ impl LocalEnv {
"directory '{}' already exists. Perhaps already initialized?",
base_path.display()
);
if !self.pg_distrib_dir.join("bin/postgres").exists() {
if !self.pg_bin_dir(pg_version).join("postgres").exists() {
bail!(
"Can't find postgres binary at {}",
self.pg_distrib_dir.display()
self.pg_bin_dir(pg_version).display()
);
}
for binary in ["pageserver", "safekeeper"] {
if !self.zenith_distrib_dir.join(binary).exists() {
if !self.neon_distrib_dir.join(binary).exists() {
bail!(
"Can't find binary '{binary}' in zenith distrib dir '{}'",
self.zenith_distrib_dir.display()
"Can't find binary '{binary}' in neon distrib dir '{}'",
self.neon_distrib_dir.display()
);
}
}

2
control_plane/src/postgresql_conf.rs
View File

@@ -2,7 +2,7 @@
/// Module for parsing postgresql.conf file.
///
/// NOTE: This doesn't implement the full, correct postgresql.conf syntax. Just
/// enough to extract a few settings we need in Zenith, assuming you don't do
/// enough to extract a few settings we need in Neon, assuming you don't do
/// funny stuff like include-directives or funny escaping.
use anyhow::{bail, Context, Result};
use once_cell::sync::Lazy;

12
control_plane/src/safekeeper.rs
View File

@@ -12,12 +12,12 @@ use nix::unistd::Pid;
use postgres::Config;
use reqwest::blocking::{Client, RequestBuilder, Response};
use reqwest::{IntoUrl, Method};
use safekeeper::http::models::TimelineCreateRequest;
use safekeeper_api::models::TimelineCreateRequest;
use thiserror::Error;
use utils::{
connstring::connection_address,
http::error::HttpErrorBody,
zid::{NodeId, ZTenantId, ZTimelineId},
id::{NodeId, TenantId, TimelineId},
};
use crate::local_env::{LocalEnv, SafekeeperConf};
@@ -46,7 +46,7 @@ impl ResponseErrorMessageExt for Response {
return Ok(self);
}
// reqwest do not export it's error construction utility functions, so lets craft the message ourselves
// reqwest does not export its error construction utility functions, so let's craft the message ourselves
let url = self.url().to_owned();
Err(SafekeeperHttpError::Response(
match self.json::<HttpErrorBody>() {
@@ -269,7 +269,7 @@ impl SafekeeperNode {
fn http_request<U: IntoUrl>(&self, method: Method, url: U) -> RequestBuilder {
// TODO: authentication
//if self.env.auth_type == AuthType::ZenithJWT {
//if self.env.auth_type == AuthType::NeonJWT {
// builder = builder.bearer_auth(&self.env.safekeeper_auth_token)
//}
self.http_client.request(method, url)
@@ -284,8 +284,8 @@ impl SafekeeperNode {
pub fn timeline_create(
&self,
tenant_id: ZTenantId,
timeline_id: ZTimelineId,
tenant_id: TenantId,
timeline_id: TimelineId,
peer_ids: Vec<NodeId>,
) -> Result<()> {
Ok(self

147
control_plane/src/storage.rs
View File

@@ -11,7 +11,7 @@ use anyhow::{bail, Context};
use nix::errno::Errno;
use nix::sys::signal::{kill, Signal};
use nix::unistd::Pid;
use pageserver::http::models::{
use pageserver_api::models::{
TenantConfigRequest, TenantCreateRequest, TenantInfo, TimelineCreateRequest, TimelineInfo,
};
use postgres::{Config, NoTls};
@@ -21,9 +21,9 @@ use thiserror::Error;
use utils::{
connstring::connection_address,
http::error::HttpErrorBody,
id::{TenantId, TimelineId},
lsn::Lsn,
postgres_backend::AuthType,
zid::{ZTenantId, ZTimelineId},
};
use crate::local_env::LocalEnv;
@@ -57,7 +57,7 @@ impl ResponseErrorMessageExt for Response {
return Ok(self);
}
// reqwest do not export it's error construction utility functions, so lets craft the message ourselves
// reqwest does not export its error construction utility functions, so let's craft the message ourselves
let url = self.url().to_owned();
Err(PageserverHttpError::Response(
match self.json::<HttpErrorBody>() {
@@ -83,7 +83,7 @@ pub struct PageServerNode {
impl PageServerNode {
pub fn from_env(env: &LocalEnv) -> PageServerNode {
let password = if env.pageserver.auth_type == AuthType::ZenithJWT {
let password = if env.pageserver.auth_type == AuthType::NeonJWT {
&env.pageserver.auth_token
} else {
""
@@ -109,14 +109,18 @@ impl PageServerNode {
pub fn initialize(
&self,
create_tenant: Option<ZTenantId>,
initial_timeline_id: Option<ZTimelineId>,
create_tenant: Option<TenantId>,
initial_timeline_id: Option<TimelineId>,
config_overrides: &[&str],
) -> anyhow::Result<ZTimelineId> {
pg_version: u32,
) -> anyhow::Result<TimelineId> {
let id = format!("id={}", self.env.pageserver.id);
// FIXME: the paths should be shell-escaped to handle paths with spaces, quotas etc.
let pg_distrib_dir_param =
format!("pg_distrib_dir='{}'", self.env.pg_distrib_dir.display());
let pg_distrib_dir_param = format!(
"pg_distrib_dir='{}'",
self.env.pg_distrib_dir_raw().display()
);
let authg_type_param = format!("auth_type='{}'", self.env.pageserver.auth_type);
let listen_http_addr_param = format!(
"listen_http_addr='{}'",
@@ -159,7 +163,7 @@ impl PageServerNode {
self.start_node(&init_config_overrides, &self.env.base_data_dir, true)?;
let init_result = self
.try_init_timeline(create_tenant, initial_timeline_id)
.try_init_timeline(create_tenant, initial_timeline_id, pg_version)
.context("Failed to create initial tenant and timeline for pageserver");
match &init_result {
Ok(initial_timeline_id) => {
@@ -173,12 +177,18 @@ impl PageServerNode {
fn try_init_timeline(
&self,
new_tenant_id: Option<ZTenantId>,
new_timeline_id: Option<ZTimelineId>,
) -> anyhow::Result<ZTimelineId> {
new_tenant_id: Option<TenantId>,
new_timeline_id: Option<TimelineId>,
pg_version: u32,
) -> anyhow::Result<TimelineId> {
let initial_tenant_id = self.tenant_create(new_tenant_id, HashMap::new())?;
let initial_timeline_info =
self.timeline_create(initial_tenant_id, new_timeline_id, None, None)?;
let initial_timeline_info = self.timeline_create(
initial_tenant_id,
new_timeline_id,
None,
None,
Some(pg_version),
)?;
Ok(initial_timeline_info.timeline_id)
}
@@ -345,7 +355,7 @@ impl PageServerNode {
fn http_request<U: IntoUrl>(&self, method: Method, url: U) -> RequestBuilder {
let mut builder = self.http_client.request(method, url);
if self.env.pageserver.auth_type == AuthType::ZenithJWT {
if self.env.pageserver.auth_type == AuthType::NeonJWT {
builder = builder.bearer_auth(&self.env.pageserver.auth_token)
}
builder
@@ -368,46 +378,53 @@ impl PageServerNode {
pub fn tenant_create(
&self,
new_tenant_id: Option<ZTenantId>,
new_tenant_id: Option<TenantId>,
settings: HashMap<&str, &str>,
) -> anyhow::Result<ZTenantId> {
) -> anyhow::Result<TenantId> {
let mut settings = settings.clone();
let request = TenantCreateRequest {
new_tenant_id,
checkpoint_distance: settings
.remove("checkpoint_distance")
.map(|x| x.parse::<u64>())
.transpose()?,
checkpoint_timeout: settings.remove("checkpoint_timeout").map(|x| x.to_string()),
compaction_target_size: settings
.remove("compaction_target_size")
.map(|x| x.parse::<u64>())
.transpose()?,
compaction_period: settings.remove("compaction_period").map(|x| x.to_string()),
compaction_threshold: settings
.remove("compaction_threshold")
.map(|x| x.parse::<usize>())
.transpose()?,
gc_horizon: settings
.remove("gc_horizon")
.map(|x| x.parse::<u64>())
.transpose()?,
gc_period: settings.remove("gc_period").map(|x| x.to_string()),
image_creation_threshold: settings
.remove("image_creation_threshold")
.map(|x| x.parse::<usize>())
.transpose()?,
pitr_interval: settings.remove("pitr_interval").map(|x| x.to_string()),
walreceiver_connect_timeout: settings
.remove("walreceiver_connect_timeout")
.map(|x| x.to_string()),
lagging_wal_timeout: settings
.remove("lagging_wal_timeout")
.map(|x| x.to_string()),
max_lsn_wal_lag: settings
.remove("max_lsn_wal_lag")
.map(|x| x.parse::<NonZeroU64>())
.transpose()
.context("Failed to parse 'max_lsn_wal_lag' as non zero integer")?,
};
if !settings.is_empty() {
bail!("Unrecognized tenant settings: {settings:?}")
}
self.http_request(Method::POST, format!("{}/tenant", self.http_base_url))
.json(&TenantCreateRequest {
new_tenant_id,
checkpoint_distance: settings
.get("checkpoint_distance")
.map(|x| x.parse::<u64>())
.transpose()?,
checkpoint_timeout: settings.get("checkpoint_timeout").map(|x| x.to_string()),
compaction_target_size: settings
.get("compaction_target_size")
.map(|x| x.parse::<u64>())
.transpose()?,
compaction_period: settings.get("compaction_period").map(|x| x.to_string()),
compaction_threshold: settings
.get("compaction_threshold")
.map(|x| x.parse::<usize>())
.transpose()?,
gc_horizon: settings
.get("gc_horizon")
.map(|x| x.parse::<u64>())
.transpose()?,
gc_period: settings.get("gc_period").map(|x| x.to_string()),
image_creation_threshold: settings
.get("image_creation_threshold")
.map(|x| x.parse::<usize>())
.transpose()?,
pitr_interval: settings.get("pitr_interval").map(|x| x.to_string()),
walreceiver_connect_timeout: settings
.get("walreceiver_connect_timeout")
.map(|x| x.to_string()),
lagging_wal_timeout: settings.get("lagging_wal_timeout").map(|x| x.to_string()),
max_lsn_wal_lag: settings
.get("max_lsn_wal_lag")
.map(|x| x.parse::<NonZeroU64>())
.transpose()
.context("Failed to parse 'max_lsn_wal_lag' as non zero integer")?,
})
.json(&request)
.send()?
.error_from_body()?
.json::<Option<String>>()
@@ -422,7 +439,7 @@ impl PageServerNode {
})
}
pub fn tenant_config(&self, tenant_id: ZTenantId, settings: HashMap<&str, &str>) -> Result<()> {
pub fn tenant_config(&self, tenant_id: TenantId, settings: HashMap<&str, &str>) -> Result<()> {
self.http_request(Method::PUT, format!("{}/tenant/config", self.http_base_url))
.json(&TenantConfigRequest {
tenant_id,
@@ -471,7 +488,7 @@ impl PageServerNode {
Ok(())
}
pub fn timeline_list(&self, tenant_id: &ZTenantId) -> anyhow::Result<Vec<TimelineInfo>> {
pub fn timeline_list(&self, tenant_id: &TenantId) -> anyhow::Result<Vec<TimelineInfo>> {
let timeline_infos: Vec<TimelineInfo> = self
.http_request(
Method::GET,
@@ -486,10 +503,11 @@ impl PageServerNode {
pub fn timeline_create(
&self,
tenant_id: ZTenantId,
new_timeline_id: Option<ZTimelineId>,
tenant_id: TenantId,
new_timeline_id: Option<TimelineId>,
ancestor_start_lsn: Option<Lsn>,
ancestor_timeline_id: Option<ZTimelineId>,
ancestor_timeline_id: Option<TimelineId>,
pg_version: Option<u32>,
) -> anyhow::Result<TimelineInfo> {
self.http_request(
Method::POST,
@@ -499,6 +517,7 @@ impl PageServerNode {
new_timeline_id,
ancestor_start_lsn,
ancestor_timeline_id,
pg_version,
})
.send()?
.error_from_body()?
@@ -524,10 +543,11 @@ impl PageServerNode {
/// * `pg_wal` - if there's any wal to import: (end lsn, path to `pg_wal.tar`)
pub fn timeline_import(
&self,
tenant_id: ZTenantId,
timeline_id: ZTimelineId,
tenant_id: TenantId,
timeline_id: TimelineId,
base: (Lsn, PathBuf),
pg_wal: Option<(Lsn, PathBuf)>,
pg_version: u32,
) -> anyhow::Result<()> {
let mut client = self.pg_connection_config.connect(NoTls).unwrap();
@@ -546,8 +566,9 @@ impl PageServerNode {
};
// Import base
let import_cmd =
format!("import basebackup {tenant_id} {timeline_id} {start_lsn} {end_lsn}");
let import_cmd = format!(
"import basebackup {tenant_id} {timeline_id} {start_lsn} {end_lsn} {pg_version}"
);
let mut writer = client.copy_in(&import_cmd)?;
io::copy(&mut base_reader, &mut writer)?;
writer.finish()?;

1
docs/SUMMARY.md
View File

@@ -79,4 +79,5 @@
- [014-storage-lsm](rfcs/014-storage-lsm.md)
- [015-storage-messaging](rfcs/015-storage-messaging.md)
- [016-connection-routing](rfcs/016-connection-routing.md)
- [017-timeline-data-management](rfcs/017-timeline-data-management.md)
- [cluster-size-limits](rfcs/cluster-size-limits.md)

4
docs/authentication.md
View File

@@ -2,14 +2,14 @@
### Overview
Current state of authentication includes usage of JWT tokens in communication between compute and pageserver and between CLI and pageserver. JWT token is signed using RSA keys. CLI generates a key pair during call to `zenith init`. Using following openssl commands:
Current state of authentication includes usage of JWT tokens in communication between compute and pageserver and between CLI and pageserver. JWT token is signed using RSA keys. CLI generates a key pair during call to `neon_local init`. Using following openssl commands:
```bash
openssl genrsa -out private_key.pem 2048
openssl rsa -in private_key.pem -pubout -outform PEM -out public_key.pem
```
CLI also generates signed token and saves it in the config for later access to pageserver. Now authentication is optional. Pageserver has two variables in config: `auth_validation_public_key_path` and `auth_type`, so when auth type present and set to `ZenithJWT` pageserver will require authentication for connections. Actual JWT is passed in password field of connection string. There is a caveat for psql, it silently truncates passwords to 100 symbols, so to correctly pass JWT via psql you have to either use PGPASSWORD environment variable, or store password in psql config file.
CLI also generates signed token and saves it in the config for later access to pageserver. Now authentication is optional. Pageserver has two variables in config: `auth_validation_public_key_path` and `auth_type`, so when auth type present and set to `NeonJWT` pageserver will require authentication for connections. Actual JWT is passed in password field of connection string. There is a caveat for psql, it silently truncates passwords to 100 symbols, so to correctly pass JWT via psql you have to either use PGPASSWORD environment variable, or store password in psql config file.
Currently there is no authentication between compute and safekeepers, because this communication layer is under heavy refactoring. After this refactoring support for authentication will be added there too. Now safekeeper supports "hardcoded" token passed via environment variable to be able to use callmemaybe command in pageserver.

25
docs/core_changes.md
View File

@@ -148,31 +148,6 @@ relcache? (I think we do cache nblocks in relcache already, check why that's not
Neon)
## Misc change in vacuumlazy.c
```
index 8aab6e324e..c684c4fbee 100644
--- a/src/backend/access/heap/vacuumlazy.c
+++ b/src/backend/access/heap/vacuumlazy.c
@@ -1487,7 +1487,10 @@ lazy_scan_heap(LVRelState *vacrel, VacuumParams *params, bool aggressive)
else if (all_visible_according_to_vm && !PageIsAllVisible(page)
&& VM_ALL_VISIBLE(vacrel->rel, blkno, &vmbuffer))
{
- elog(WARNING, "page is not marked all-visible but visibility map bit is set in relation \"%s\" page %u",
+ /* ZENITH-XXX: all visible hint is not wal-logged
+ * FIXME: Replay visibilitymap changes in pageserver
+ */
+ elog(DEBUG1, "page is not marked all-visible but visibility map bit is set in relation \"%s\" page %u",
vacrel->relname, blkno);
visibilitymap_clear(vacrel->rel, blkno, vmbuffer,
VISIBILITYMAP_VALID_BITS);
```
Is this still needed? If that WARNING happens, it looks like potential corruption that we should
fix!
## Use buffer manager when extending VM or FSM
```

18
docs/multitenancy.md
View File

@@ -2,26 +2,26 @@
### Overview
Zenith supports multitenancy. One pageserver can serve multiple tenants at once. Tenants can be managed via zenith CLI. During page server setup tenant can be created using ```zenith init --create-tenant``` Also tenants can be added into the system on the fly without pageserver restart. This can be done using the following cli command: ```zenith tenant create``` Tenants use random identifiers which can be represented as a 32 symbols hexadecimal string. So zenith tenant create accepts desired tenant id as an optional argument. The concept of timelines/branches is working independently per tenant.
Neon supports multitenancy. One pageserver can serve multiple tenants at once. Tenants can be managed via neon_local CLI. During page server setup tenant can be created using ```neon_local init --create-tenant``` Also tenants can be added into the system on the fly without pageserver restart. This can be done using the following cli command: ```neon_local tenant create``` Tenants use random identifiers which can be represented as a 32 symbols hexadecimal string. So neon_local tenant create accepts desired tenant id as an optional argument. The concept of timelines/branches is working independently per tenant.
### Tenants in other commands
By default during `zenith init` new tenant is created on the pageserver. Newly created tenant's id is saved to cli config, so other commands can use it automatically if no direct argument `--tenantid=<tenantid>` is provided. So generally tenantid more frequently appears in internal pageserver interface. Its commands take tenantid argument to distinguish to which tenant operation should be applied. CLI support creation of new tenants.
By default during `neon_local init` new tenant is created on the pageserver. Newly created tenant's id is saved to cli config, so other commands can use it automatically if no direct argument `--tenant_id=<tenant_id>` is provided. So generally tenant_id more frequently appears in internal pageserver interface. Its commands take tenant_id argument to distinguish to which tenant operation should be applied. CLI support creation of new tenants.
Examples for cli:
```sh
zenith tenant list
neon_local tenant list
zenith tenant create // generates new id
neon_local tenant create // generates new id
zenith tenant create ee6016ec31116c1b7c33dfdfca38892f
neon_local tenant create ee6016ec31116c1b7c33dfdfca38892f
zenith pg create main // default tenant from zenith init
neon_local pg create main // default tenant from neon init
zenith pg create main --tenantid=ee6016ec31116c1b7c33dfdfca38892f
neon_local pg create main --tenant_id=ee6016ec31116c1b7c33dfdfca38892f
zenith branch --tenantid=ee6016ec31116c1b7c33dfdfca38892f
neon_local branch --tenant_id=ee6016ec31116c1b7c33dfdfca38892f
```
### Data layout
@@ -56,4 +56,4 @@ Tenant id is passed to postgres via GUC the same way as the timeline. Tenant id
### Safety
For now particular tenant can only appear on a particular pageserver. Set of safekeepers are also pinned to particular (tenantid, timeline) pair so there can only be one writer for particular (tenantid, timeline).
For now particular tenant can only appear on a particular pageserver. Set of safekeepers are also pinned to particular (tenant_id, timeline_id) pair so there can only be one writer for particular (tenant_id, timeline_id).

2
docs/pageserver-services.md
View File

@@ -109,7 +109,7 @@ Repository
The repository stores all the page versions, or WAL records needed to
reconstruct them. Each tenant has a separate Repository, which is
stored in the .neon/tenants/<tenantid> directory.
stored in the .neon/tenants/<tenant_id> directory.
Repository is an abstract trait, defined in `repository.rs`. It is
implemented by the LayeredRepository object in

10
docs/pageserver-storage.md
View File

@@ -123,7 +123,7 @@ The files are called "layer files". Each layer file covers a range of keys, and
a range of LSNs (or a single LSN, in case of image layers). You can think of it
as a rectangle in the two-dimensional key-LSN space. The layer files for each
timeline are stored in the timeline's subdirectory under
`.neon/tenants/<tenantid>/timelines`.
`.neon/tenants/<tenant_id>/timelines`.
There are two kind of layer files: images, and delta layers. An image file
contains a snapshot of all keys at a particular LSN, whereas a delta file
@@ -351,7 +351,7 @@ branch.
Note: It doesn't make any difference if the child branch is created
when the end of the main branch was at LSN 250, or later when the tip of
the main branch had already moved on. The latter case, creating a
branch at a historic LSN, is how we support PITR in Zenith.
branch at a historic LSN, is how we support PITR in Neon.
# Garbage collection
@@ -396,9 +396,9 @@ table:
main/orders_200_300 DELETE
main/orders_300 STILL NEEDED BY orders_300_400
main/orders_300_400 KEEP, NEWER THAN GC HORIZON
main/orders_400 ..
main/orders_400_500 ..
main/orders_500 ..
main/orders_400 ..
main/orders_400_500 ..
main/orders_500 ..
main/customers_100 DELETE
main/customers_100_200 DELETE
main/customers_200 KEEP, NO NEWER VERSION

4
docs/pageserver-tenant-migration.md
View File

@@ -9,7 +9,7 @@ This feature allows to migrate a timeline from one pageserver to another by util
Pageserver implements two new http handlers: timeline attach and timeline detach.
Timeline migration is performed in a following way:
1. Timeline attach is called on a target pageserver. This asks pageserver to download latest checkpoint uploaded to s3.
2. For now it is necessary to manually initialize replication stream via callmemaybe call so target pageserver initializes replication from safekeeper (it is desired to avoid this and initialize replication directly in attach handler, but this requires some refactoring (probably [#997](https://github.com/zenithdb/zenith/issues/997)/[#1049](https://github.com/zenithdb/zenith/issues/1049))
2. For now it is necessary to manually initialize replication stream via callmemaybe call so target pageserver initializes replication from safekeeper (it is desired to avoid this and initialize replication directly in attach handler, but this requires some refactoring (probably [#997](https://github.com/neondatabase/neon/issues/997)/[#1049](https://github.com/neondatabase/neon/issues/1049))
3. Replication state can be tracked via timeline detail pageserver call.
4. Compute node should be restarted with new pageserver connection string. Issue with multiple compute nodes for one timeline is handled on the safekeeper consensus level. So this is not a problem here.Currently responsibility for rescheduling the compute with updated config lies on external coordinator (console).
5. Timeline is detached from old pageserver. On disk data is removed.
@@ -18,5 +18,5 @@ Timeline migration is performed in a following way:
### Implementation details
Now safekeeper needs to track which pageserver it is replicating to. This introduces complications into replication code:
* We need to distinguish different pageservers (now this is done by connection string which is imperfect and is covered here: https://github.com/zenithdb/zenith/issues/1105). Callmemaybe subscription management also needs to track that (this is already implemented).
* We need to distinguish different pageservers (now this is done by connection string which is imperfect and is covered here: https://github.com/neondatabase/neon/issues/1105). Callmemaybe subscription management also needs to track that (this is already implemented).
* We need to track which pageserver is the primary. This is needed to avoid reconnections to non primary pageservers. Because we shouldn't reconnect to them when they decide to stop their walreceiver. I e this can appear when there is a load on the compute and we are trying to detach timeline from old pageserver. In this case callmemaybe will try to reconnect to it because replication termination condition is not met (page server with active compute could never catch up to the latest lsn, so there is always some wal tail)

47
docs/pageserver-thread-mgmt.md
View File

@@ -1,26 +1,39 @@
## Thread management
Each thread in the system is tracked by the `thread_mgr` module. It
maintains a registry of threads, and which tenant or timeline they are
operating on. This is used for safe shutdown of a tenant, or the whole
system.
The pageserver uses Tokio for handling concurrency. Everything runs in
Tokio tasks, although some parts are written in blocking style and use
spawn_blocking().
Each Tokio task is tracked by the `task_mgr` module. It maintains a
registry of tasks, and which tenant or timeline they are operating
on.
### Handling shutdown
When a tenant or timeline is deleted, we need to shut down all threads
operating on it, before deleting the data on disk. A thread registered
in the thread registry can check if it has been requested to shut down,
by calling `is_shutdown_requested()`. For async operations, there's also
a `shudown_watcher()` async task that can be used to wake up on shutdown.
When a tenant or timeline is deleted, we need to shut down all tasks
operating on it, before deleting the data on disk. There's a function,
`shutdown_tasks`, to request all tasks of a particular tenant or
timeline to shutdown. It will also wait for them to finish.
A task registered in the task registry can check if it has been
requested to shut down, by calling `is_shutdown_requested()`. There's
also a `shudown_watcher()` Future that can be used with `tokio::select!`
or similar, to wake up on shutdown.
### Sync vs async
The primary programming model in the page server is synchronous,
blocking code. However, there are some places where async code is
used. Be very careful when mixing sync and async code.
Async is primarily used to wait for incoming data on network
connections. For example, all WAL receivers have a shared thread pool,
with one async Task for each connection. Once a piece of WAL has been
received from the network, the thread calls the blocking functions in
We use async to wait for incoming data on network connections, and to
perform other long-running operations. For example, each WAL receiver
connection is handled by a tokio Task. Once a piece of WAL has been
received from the network, the task calls the blocking functions in
the Repository to process the WAL.
The core storage code in `layered_repository/` is synchronous, with
blocking locks and I/O calls. The current model is that we consider
disk I/Os to be short enough that we perform them while running in a
Tokio task. If that becomes a problem, we should use `spawn_blocking`
before entering the synchronous parts of the code, or switch to using
tokio I/O functions.
Be very careful when mixing sync and async code!

2
docs/rfcs/013-term-history.md
View File

@@ -70,7 +70,7 @@ two options.
...start sending WAL conservatively since the horizon (1.1), and truncate
obsolete part of WAL only when recovery is finished, i.e. epochStartLsn (4) is
reached, i.e. 2.3 transferred -- that's what https://github.com/zenithdb/zenith/pull/505 proposes.
reached, i.e. 2.3 transferred -- that's what https://github.com/neondatabase/neon/pull/505 proposes.
Then the following is possible:

413
docs/rfcs/017-timeline-data-management.md Normal file
View File

@@ -0,0 +1,413 @@
# Name
Tenant and timeline data management in pageserver
## Summary
This RFC attempts to describe timeline-related data management as it's done now in pageserver, highlight current complexities caused by this and propose a set of changes to mitigate them.
The main goal is to prepare for future [on-demand layer downloads](https://github.com/neondatabase/neon/issues/2029), yet timeline data is one of the core primitive of pageserver, so a number of other RFCs are affected either.
Due to that, this document won't have a single implementation, rather requiring a set of code changes to achieve the final state.
RFC considers the repository at the `main` branch, commit [`28243d68e60ffc7e69f158522f589f7d2e09186d`](https://github.com/neondatabase/neon/tree/28243d68e60ffc7e69f158522f589f7d2e09186d) on the time of writing.
## Motivation
In recent discussions, it became more clear that timeline-related code becomes harder to change: it consists of multiple disjoint modules, each requiring a synchronization to access.
The lower the code is, the complex the sync gets since many concurrent processes are involved and require orchestration to keep the data consistent.
As the number of modules and isolated data grows per timeline, more questions and corner cases arise:
- https://github.com/neondatabase/neon/issues/1559
right now it's not straightened out what to do when the synchronization task fails for too many times: every separate module's data has to be treated differently.
- https://github.com/neondatabase/neon/issues/1751
GC and compaction file activities are not well known outside their tasks code, causing race bugs
- https://github.com/neondatabase/neon/issues/2003
Even the tenant management gets affected: we have to alter its state based on timeline state, yet the data for making the decision is separated and the synchronisation logic has bugs
- more issues were brought in discussions, but apparently they were too specific to the code to mention them in the issues.
For instance, `tenant_mgr` itself is a static object that we can not mock anyhow, which reduces our capabilities to test the data synchronization logic.
In fact, we have zero Rust tests that cover the case of synchronizing more than one module's data.
On demand layer downloads would require us to dynamically manage the layer files, which we almost not doing at all on the module level, resulting in the most of their APIs dealing with timelines, rather than the layer files.
The disjoint data that would require data synchronization with possibly a chain of lock acquisitions, some async and some sync, and it would be hard to unit test it with the current code state.
Neither this helps to easy start the on-demand download epic, nor it's easy to add more timeline-related code on top, whatever the task is.
We have to develop a vision on a number of topics before progressing safely:
- timeline and tenant data structure and how should we access it
- sync and async worlds and in what way that should evolve
- unit tests for the complex logic
This RFC aims to provide a general overview of the existing situation and propose ways to improve it.
The changes proposed are quite big and no single PR is expected to do the adjustments, they should gradually be done during the on-demand download work later.
## What is a timeline and its data
First, we need to define what data we want to manage per timeline.
Currently, the data every timeline operates is:
- a set of layer files, on the FS
Never updated files, created after pageserver's checkpoints and compaction runs, can be removed from the local FS due to compaction, gc or timeline deletion.
- a set of layer files, on the remote storage
Identically named and placed in tenant subdirectories files on the remote storage (S3), copied by a special background sync thread
- a `metadata` file, on the FS
Updated after every checkpoint with the never `disk_consistent_lsn` and `latest_gc_cutoff_lsn` values. Used to quickly restore timeline's basic metadata on pageserver restart.
Also contains data about the ancestor, if the timeline was branched off another timeline.
- an `index_part.json` file, on the remote storage
Contains `metadata` file contents and a list of layer files, available in the current S3 "directory" for the timeline.
Used to avoid potentially slow and expensive `S3 list` command, updated by the remotes storage sync thread after every operation with the remote layer files.
- LayerMap and PageCache, in memory
Dynamic, used to store and retrieve the page data to users.
- timeline info, in memory
LSNs, walreceiver data, `RemoteTimelineIndex` and other data to share via HTTP API and internal processes.
- metrics data, in memory
Data to push or provide to Prometheus, Opentelemetry, etc.
Besides the data, every timeline currently needs an etcd connection to receive WAL events and connect to safekeepers.
Timeline could be an ancestor to another one, forming a dependency tree, which is implicit right now: every time relations are looked up in place, based on the corresponding `TimelineMetadata` struct contents.
Yet, there's knowledge on a tenant as a group of timelines, belonging to a single user which is used in GC and compaction tasks, run on every tenant.
`tenant_mgr` manages tenant creation and its task startup, along with the remote storage sync for timeline layers.
Last file being managed per-tenant is the tenant config file, created and updated on the local FS to hold tenant-specific configuration between restarts.
It's not yet anyhow synchronized with the remote storage, so only exists on the local FS.
### How the data is stored
We have multiple places where timeline data is stored:
- `tenant_mgr` [holds](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/tenant_mgr.rs#L43) a static `static ref TENANTS: RwLock<HashMap<ZTenantId, Tenant>>` with the `Tenant` having the `local_timelines: HashMap<ZTimelineId, Arc<DatadirTimelineImpl>>` inside
- same `Tenant` above has actually two references to timelines: another via its `repo: Arc<RepositoryImpl>` with `pub type RepositoryImpl = LayeredRepository;` that [holds](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/layered_repository.rs#L178) `Mutex<HashMap<ZTimelineId, LayeredTimelineEntry>>`
- `RemoteTimelineIndex` [contains](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/storage_sync/index.rs#L84) the metadata about timelines on the remote storage (S3) for sync reasons and possible HTTP API queries
- `walreceiver` [stores](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/walreceiver.rs#L60) the metadata for possible HTTP API queries and its [internal state](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/walreceiver/connection_manager.rs#L245) with a reference to the timeline, its current connections and etcd subscription (if any)
- `PageCache` contains timeline-related data, and is created globally for the whole pageserver
- implicitly, we also have files on local FS, that contain timeline state. We operate on those files and for some operations (GC, compaction) yet we don't anyhow synchronize the access to the files per se: there are more high-level locks, ensuring only one of a group of operations is running at a time.
On practice though, `LayerMap` and layer files are tightly coupled together: current low-level code requires a timeline to be loaded into the memory to work with it, and the code removes the layer files after removing the entry from the `LayerMap` first.
Based on this, a high-level pageserver's module diagram with data and entities could be:
![timeline tenant state diagram](./images/017-timeline-data-management/timeline_tenant_state.svg)
A few comments on the diagram:
- the diagram does not show all the data and replaces a few newtypes and type aliases (for example, completely ignores "unloaded" timelines due to reasons described below)
It aims to show main data and means of synchronizing it.
- modules tend to isolate their data inside and provide access to it via API
Due to multitenancy, that results in a common pattern for storing both tenant and timeline data: `RwLock` or `Mutex` around the `HashMap<Id, Data>`, gc and compaction tasks also use the same lock pattern to ensure no concurrent runs are happening.
- part of the modules is asynchronous, while the other is not, that complicates the data access
Currently, anything that's not related to tasks (walreceiver, storage sync, GC, compaction) is blocking.
Async tasks that try to access the data in the sync world, have to call `std::sync::Mutex::lock` method, which blocks the thread the callee async task runs on, also blocking other async tasks running in the same thread. Methods of `std::sync::RwLock` have the same issues, forcing async tasks either to block or spawn another, "blocking" task on a separate thread.
Sync tasks that try to access the data in the async world, cannot use `.await` hence have to have some `Runtime` doing those calls for them. [`tokio::sync::Mutex`](https://docs.rs/tokio/1.19.2/tokio/sync/struct.Mutex.html#method.blocking_lock) and [`tokio::sync::RwLock`](https://docs.rs/tokio/1.19.2/tokio/sync/struct.RwLock.html#method.blocking_read) provide an API to simplify such calls. Similarly, both `std::sync` and `tokio::sync` have channels that are able to communicate into one direction without blocking and requiring `.await` calls, hence can be used to connect both worlds without locking.
Some modules are in transition, started as async "blocking" tasks and being fully synchronous in their entire code below the start. Current idea is to transfer them to the async further, but it's not yet done.
- locks are used in two different ways:
- `RwLock<HashMap<..>>` ones to hold the shared data and ensure its atomic updates
- `Mutex<()>` for synchronizing the tasks, used to implicitly order the data access
The "shared data" locks of the first kind are mainly accessed briefly to either look up or alter the data, yet there are a few notable exceptions, such as
`latest_gc_cutoff_lsn: RwLock<Lsn>` that is explicitly held in a few places to prevent GC thread from progressing. Those are covered later in the data access diagrams.
- some synchronizations are not yet implemented
E.g. asynchronous storage sync module does not synchronize with almost synchronous GC and compaction tasks when the layer files are uploaded to the remote storage.
That occasionally results in the files being deleted before the storage upload task is run for this layer, but due to the incremental nature of the layer files, we can handle such situations without issues.
- `LayeredRepository` covers lots of responsibilities: GC and compaction task synchronisation, timeline access (`local_timelines` in `Tenant` is not used directly before the timeline from the repository is accessed), layer flushing to FS, layer sync to remote storage scheduling, etc.
### How is this data accessed?
There are multiple ways the data is accessed, from different sources:
1. [HTTP requests](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/http/routes.rs)
High-level CRUD API for managing tenants, timelines and getting data about them.
Current API list (modified for readability):
```rust
.get("/v1/status", status_handler) // pageserver status
.get("/v1/tenant", tenant_list_handler)
.post("/v1/tenant", tenant_create_handler) // can create "empty" timelines or branch off the existing ones
.get("/v1/tenant/:tenant_id", tenant_status) // the only tenant public metadata
.put("/v1/tenant/config", tenant_config_handler) // tenant config data and local file manager
.get("/v1/tenant/:tenant_id/timeline", timeline_list_handler)
.post("/v1/tenant/:tenant_id/timeline", timeline_create_handler)
.post("/v1/tenant/:tenant_id/attach", tenant_attach_handler) // download entire tenant from the remote storage and load its timelines memory
.post("/v1/tenant/:tenant_id/detach", tenant_detach_handler) // delete all tenant timelines from memory, remote corresponding storage and local FS files
.get("/v1/tenant/:tenant_id/timeline/:timeline_id", timeline_detail_handler)
.delete("/v1/tenant/:tenant_id/timeline/:timeline_id", timeline_delete_handler)
.get("/v1/tenant/:tenant_id/timeline/:timeline_id/wal_receiver", wal_receiver_get_handler) // get walreceiver stats metadata
```
Overall, neither HTTP operation goes below `LayeredRepository` level and does not interact with layers: instead, they manage tenant and timeline entities, their configuration and metadata.
`GET` data is small (relative to layer files contents), updated via brief `.write()/.lock()` calls and read via copying/cloning the data to release the lock soon.
It does not mean that the operations themselves are short, e.g. `tenant_attach_handler` downloads multiple files from the remote storage which might take time, yet the final data is inserted in memory via one brief write under the lock.
Non-`GET` operations mostly follow the same rule, with two differences:
- `tenant_detach_handler` has to wait for its background tasks to stop before shutting down, which requires more work with locks
- `timeline_create_handler` currently requires GC to be paused before branching the timeline, which requires orchestrating too.
This is the only HTTP operation, able to load the timeline into memory: rest of the operations are reading the metadata or, as in `tenant_attach_handler`, schedule a deferred task to download timeline and load it into memory.
"Timeline data synchronization" section below describes both complex cases in more details.
2. [libpq requests](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/page_service.rs)
Is the main interface of pageserver, intended to handle libpq (and similar) requests.
Operates on `LayeredTimeline` and, lower, `LayerMap` modules; all timelines accessed during the operation are loaded into memory immediately (if not loaded already), operations bail on timeline load errors.
- `pagestream`
Page requests: `get_rel_exists`, `get_rel_size`, `get_page_at_lsn`, `get_db_size`
Main API points, intended to be used by `compute` to show the data to the user. All require requests to be made at certain Lsn, if this Lsn is not available in the memory, request processing is paused until that happens or bails after a timeout.
- `basebackup` and `fullbackup`
Options to generate postgres-compatible backup archives.
- `import basebackup`
- `import wal`
Import the `pg_wal` section of the basebackup archive.
- `get_last_record_rlsn`, `get_lsn_by_timestamp`
"Metadata" retrieval methods, that still requires internal knowledge about layers.
- `set`, `fallpoints`, `show`
Utility methods to support various edge cases or help with debugging/testing.
- `do_gc`, `compact`, `checkpoint`
Manual triggers for corresponding tenant tasks (GC, compaction) and inmemory layer flushing on disk (checkpointing), with upload task scheduling as a follow-up.
Apart from loading into memory, every timeline layer has to be accessed using specific set of locking primitives, especially if a write operations happens: otherwise, GC or compaction might spoil the data. User API is implicitly affected by this synchronization during branching, when a GC has to be orchestrated properly before the new timeline could be branched off the existing one.
See "Timeline data synchronization" section for the united synchronization diagram on the topic.
3. internal access
Entities within pageserver that update files on local FS and remote storage, metadata in memory; has to use internal data for those operations.
Places that access internal, lower data are also required to have the corresponding timeline successfully loaded into memory and accessed with corresponding synchronization.
If ancestors' data is accessed via its child branch, it means more than one timeline has to be loaded into memory entirely and more locking primitives usage involved.
Right now, all ancestors are resolved in-place: every place that has to check timeline's ancestor has to lock the timelines map, check if one is loaded into the memory, load it there or bail if it's not present, and get the information required and so on.
- periodic GC and compaction tasks
Alter metadata (GC info), in-memory data (layer relations, page caches, etc.) and layer files on disk.
Same as its libpq counterparts, needs full synchronization with the low level layer management code.
- storage sync task
Alters metadata (`RemoteTimelineIndex`), layer files on remote storage (upload, delete) and local FS (download) and in-memory data (registers downloaded timelines in the repository).
Currently, does not know anything about layer files contents, rather focusing on the file structure and metadata file updates: due to the fact that the layer files cannot be updated (only created or deleted), storage sync is able to back up the files to the remote storage without further low-level synchronizations: only when the timeline is downloaded, a load operation is needed to run, possibly pausing GC and compaction tasks.
- walreceiver and walingest task
Per timeline, subscribes for etcd events from safekeeper and eventually spawns a walreceiver connection task to receive WAL from a safekeeper node.
Fills memory with data, eventually triggering a checkpoint task that creates a new layer file in the local FS and schedules a remote storage sync upload task.
During WAL receiving, also updates a separate in-memory data structure with the walreceiver stats, used later via HTTP API.
Layer updates require low-level set of sync primitives used to preserve the data consistency.
- checkpoint (layer freeze) task
Periodic, short-lived tasks to generate a new layer file in the FS. Requires low level synchronization in the end, when the layer is being registered after creating and has additional mode to ensure only one concurrent compaction happens at a time.
### Timeline data synchronization
Here's a high-level timeline data access diagram, considering the synchronization locks, based on the state diagram above.
For brevity, diagrams do not show `RwLock<HashMap<..>>` data accesses, considering them almost instant to happen.
`RwLock<LayerMap>` is close to be an exception to the previous rule, since it's taken in a multiple places to ensure all layers are inserted correctly.
Yet the only long operation in the current code is a `.write()` lock on the map during its creation, while all other lock usages tend to be short in the current code.
Note though, that due to current "working with loaded timeline only", prevailing amount of the locks taken on the struct are `.write()` locks, not the `.read()` ones.
To simplify the diagrams, these accesses are now considered "fast" data access, not the synchronization attempts.
`write_lock` synchronization diagram:
![timeline data access synchronization(1)](./images/017-timeline-data-management/timeline_data_access_sync_1.svg)
Comments:
- `write_lock: Mutex<()>` ensures that all timeline data being written into **in-memory layers** is done without races, one concurrent write at a time
- `layer_flush_lock: Mutex<()>` and layer flushing seems to be slightly bloated with various ways to create a layer on disk and write it in memory
The lock itself seem to repeat `write_lock` purpose when it touches in-memory layers, and also to limit the on-disk layer creations.
Yet the latter is not really done consistently, since remote storage sync manages to download and register the new layers without touching the locks
- `freeze_inmem_layer(true)` that touches both `write_lock` and `layer_flush_lock` seems not very aligned with the rest of the locks to those primitives; it also now restricts the layer creation concurrency even more, yet there are various `freeze_inmem_layer(false)` that are ignoring those restrictions at the same time
![timeline data access synchronization(2)](./images/017-timeline-data-management/timeline_data_access_sync_2.svg)
Comments:
- `partitioning: Mutex<(KeyPartitioning, Lsn)>` lock is a data sync lock that's not used to synchronize the tasks (all other such kinds were considered "almost instant" and omitted on the diagram), yet is very similar to what `write_lock` and `layer_flush_lock` do: it ensures the timeline in-memory data is up-to-date with the layer files state on disk, which is what `LayerMap` is for.
- there are multiple locks that do similar task management operations:
- `gc_cs: Mutex<()>` and `latest_gc_cutoff_lsn: RwLock<Lsn>` ensures that branching and gc are not run concurrently
- `layer_removal_cs: Mutex<()>` lock ensure gc, compaction and timeline deletion via HTTP API do not run concurrently
- `file_lock: RwLock<()>` is used as a semaphore, to ensure "all" gc and compaction tasks are shut down and do not start
Yet that lock does take only gc and compaction from internal loops: libpq call is not cancelled and waited upon.
Those operations do not seem to belong to a timeline. Moreover, some of those could be eliminated entirely due to duplication of their tasks.
## Proposed implementation
### How to structure timeline data access better
- adjust tenant state handling
Current [`TenantState`](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/tenant_mgr.rs#L108) [changes](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/tenant_mgr.rs#L317) mainly indicates whether GC and compaction tasks are running or not; another state, `Broken` shows only in case any timeline does not load during startup.
We could start both GC and compaction tasks at the time the tenant is created and adjust the tasks to throttle/sleep on timeline absence and wake up when the first one is added.
The latter becomes more important on download on demand, since we won't have the entire timeline in reach to verify its correctness. Moreover, if any network connection happens, the timeline could fail temporarily and entire tenant should be marked as broken due to that.
Since nothing verifies the `TenantState` via HTTP API currently, it makes sense to remove the whole state entirely and don't write the code to synchronize its changes.
Instead, we could indicate internal issues for every timeline and have a better API to "stop" timeline processing without deleting its data, making our API less restrictive.
- remove the "unloaded" status for the timeline
Current approach to timeline management [assumes](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/layered_repository.rs#L486-L493)
```rust
#[derive(Clone)]
enum LayeredTimelineEntry {
Loaded(Arc<LayeredTimeline>),
Unloaded {
id: ZTimelineId,
metadata: TimelineMetadata,
},
}
```
supposes that timelines have to be in `Unloaded` state.
The difference between both variants is whether its layer map was loaded from disk and kept in memory (Loaded) or not (Unloaded).
The idea behind such separation was to lazy load timelines in memory with all their layers only after its first access and potentially unload them later.
Yet now there's no public API methods, that deal with unloaded timelines' layers: all of them either bail when such timeline is worked on, or load it into memory and continue working.
Moreover, every timeline in the local FS is loaded on pageserver startup now, so only two places where `Unloaded` variant is used are branching and timeline attach, with both loading the timeline into memory before the end of the operation.
Even if that loading into memory bails for some reason, next GC or compaction task periodic run would load such timeline into memory.
There are a few timeline methods that return timeline metadata without loading its layers, but such metadata also comes from the `metadata` FS file, not the layer files (so no page info could be retrieved without loading the entire layer map first).
With the layer on-demand download, it's not feasible anymore to wait for the entire layer map to be loaded into the memory, since it might not even be available on the local FS when requested: `LayerMap` needs to be changed to contain metadata to retrieve the missing layers and handle partially present on the local FS timeline state.
To accommodate to that and move away from the redundant status, a timeline should always be "loaded" with its metadata read from the disk and its layer map prepared to be downloaded when requested, per layer.
Layers in the layer map, on the other hand, could be in various state: loaded, unloaded, downloading, downloading failed, etc. and their state has to be handled instead, if we want to support on-demand download in the future.
This way, tenants and timelines could always try to serve requests and do their internal tasks periodically, trying to recover.
- scale down the remote storage sync to per layer file, not per timeline as now
Due to the reasons from the previous bullet, current remote storage model needs its timeline download approach to be changed.
Right now, a timeline is marked as "ready" only after all its layers on the remote storage are downloaded on the local storage.
With the on-demand download approach, only remote storage timeline metadata should be downloaded from S3, leaving the rest of the layers ready for download if/when it's requested.
Note: while the remote storage sync should operate per layer, it should stay global for all tenants, to better manage S3 limits and sync queue priorities.
Yet the only place using remote storage should be the layer map.
- encapsulate `tenant_mgr` logic into a regular Rust struct, unite with part of the `Repository` and anything else needed to manage the timeline data in a single place and to test it independently
[`Repository`](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/repository.rs#L187) trait gets closer to `tenant_mgr` in terms of functionality: there are two background task-related functions, that are run on all timelines of a tenant: `gc_iteration` (it does allow running on a single timeline, but GC task runs it on all timelines) and `compaction_iteration` that are related to service tasks, not the data storage; and the metadata management functions, also not really related to the timeline contents.
`tenant_mgr` proxies some of the `Repository` calls, yet both service tasks use `tenant_mgr` to access the data they need, creating a circular dependency between their APIs.
To avoid excessive synchronization between components, taking multiple locks for that and static state, we can organize the data access and updates in one place.
One potential benefit Rust gets from this is the ability to track and manage timeline resources, if all the related data is located in one place.
- move `RemoteStorage` usage from `LayeredRepository` into `LayerMap`, as the rest of the layer-based entities (layer files, etc.)
Layer == file in our model, since pageserver always either tries to load the LayerMap from disk for the timeline not in memory, or assumes the file contents matches its memory.
`LayeredRepository` is one of the most loaded objects currently and not everything from it deserves unification with the `tenant_mgr`.
In particular, layer files need to be better prepared for future download on demand functionality, where every layer could be dynamically loaded and unloaded from memory and local FS.
Current amount of locks and sync-async separation would make it hard to implement truly dynamic (un)loading; moreover, we would need retries with backoffs, since the unloaded layer files are most probably not available on the local FS either and network is not always reliable.
One of the solutions to the issue is already being developed for the remote storage sync: [SyncQueue](https://github.com/neondatabase/neon/blob/28243d68e60ffc7e69f158522f589f7d2e09186d/pageserver/src/storage_sync.rs#L463)
The queue is able to batch CRUD layer operations (both for local and remote FS contexts) and reorder them to increase the sync speed.
Similar approach could be generalized for all layer modifications, including in-memory ones such as GC or compaction: this way, we could manage all layer modifications and reads in one place with lesser locks and tests that are closer to unit tests.
- change the approach to locking synchronization
A number of locks in the timeline seem to be used to coordinate gc, compaction tasks and related processes.
It should be done in a task manager or other place, external to the timeline.
Timeline contents still needs to be synchronized, considering the task work, so fields like `latest_gc_cutoff_lsn: RwLock<Lsn>` are expected to stay for that purpose, but general amount of locks should be reduced.
### Putting it all together
If the proposal bullets applied to the diagrams above, the state could be represented as:
![timeline timeline tenant state](./images/017-timeline-data-management/proposed_timeline_tenant_state.svg)
The reorders aim to put all tasks into separated modules, with strictly defined interfaces and as less knowledge about other components, as possible.
This way, all timeline data is now in the `data_storage`, including the GC, walreceiver, `RemoteTimelineIndex`, `LayerMap`, etc. with some API to get the data in the way,
more convenient for the data sync system inside.
So far, it seems that a few maps with `Arc<RwLock<SeparateData>>` with actual data operations added inside each `SeparateData` struct, if needed.
`page_cache` is proposed to placed into the same `data_storage` since it contains tenant timelines' data: this way, all metadata and data is in the same struct, simplifying things with Rust's borrow checker and allowing us to share internals between data modules and later might simplify timeline in-memory size tracking.
`task_manager` is related to data storage and manages all tenant and timeline tasks, manages shared resources (runtimes, thread pools, etcd connection, etc.) and synchronizes tasks.
All locks such as `gc_cs` belong to this module tree, as primitives inherently related to the task synchronization.
Tasks have to access timelines and their metadata, but should do that through `data_storage` API and similar.
`task_manager` should (re)start, stop and track all tasks that are run in it, selecting an appropriate runtime depending on a task kind (we have async/sync task separation, CPU and IO bound tasks separation, ...)
Some locks such as `layer_removal_cs` one are not needed, if the only component that starts the tasks ensures they don't run concurrently.
`LayeredTimeline` is still split into two parts, more high-level with whatever primitives needed to sync its state, and the actual state storage with `LayerMap` and other low level entities.
Only `LayerMap` knows what storage it's layer files are taken from (inmem, local FS, etc.), and it's responsible for synchronizing the layers when needed, as also reacting to sync events, successful or not.
Last but not least, `tenant config file` has to be backed into a remote storage, as tenant-specific information for all timelines.
Tenant and timelines have volatile information that's now partially mixed with constant information (e.g. fields in `metadata` file), that model should be better split and handled, in case we want to properly support its backups and synchronization.
![proposed timeline data access synchronization(1)](./images/017-timeline-data-management/proposed_timeline_data_access_sync_1.svg)
There's still a need to keep inmemory layer buffer synchronized during layer freezing, yet that could happen on a layer level, not on a timeline level, as `write_lock` used to be, so we could lower the sync primitives one layer deeper, preparing us for download on demand feature, where multiple layers could be concurrently streamed and written from various data sources.
Flushing the frozen layer requires creating a new layer on disk and further remote storage upload, so `LayerMap` has to get those flushed bytes and queue them later: no need to block in the timeline itself for anything again, rather locking on the layer level, if needed.
![proposed timeline data access synchronization(2)](./images/017-timeline-data-management/proposed_timeline_data_access_sync_2.svg)
Lock diagrams legend:
![lock diagrams legend](./images/017-timeline-data-management/lock_legend.svg)
After the frozen layers are flushed, something has to ensure that the layer structure is intact, so a repartitioning lock is needed still, and could also guard the layer map structure changes, since both are needed either way.
This locking belongs to the `LowLevelLayeredTimeline` from the proposed data structure diagram, as the place with all such data being held.
Similarly, branching is still required to be done after certain Lsn in our current model, but this needs only one lock to synchronize and that could be the `gc_cs: Mutex<()>` lock.
It raises the question of where this lock has to be placed, it's the only place that requires pausing a GC task during external, HTTP request handling.
The right place for the lock seems to be the `task_manager` that could manage GC in more fine-grained way to accommodate the incoming branching request.
There's no explicit lock sync between GC, compaction or other mutually exclusive tasks: it is a job of the `task_manager` to ensure those are not run concurrently.

8
docs/rfcs/cluster-size-limits.md
View File

@@ -15,7 +15,7 @@ The stateless compute node that performs validation is separate from the storage
Limit the maximum size of a PostgreSQL instance to limit free tier users (and other tiers in the future).
First of all, this is needed to control our free tier production costs.
Another reason to limit resources is risk management — we haven't (fully) tested and optimized zenith for big clusters,
Another reason to limit resources is risk management — we haven't (fully) tested and optimized neon for big clusters,
so we don't want to give users access to the functionality that we don't think is ready.
## Components
@@ -43,20 +43,20 @@ Then this size should be reported to compute node.
`current_timeline_size` value is included in the walreceiver's custom feedback message: `ReplicationFeedback.`
(PR about protocol changes https://github.com/zenithdb/zenith/pull/1037).
(PR about protocol changes https://github.com/neondatabase/neon/pull/1037).
This message is received by the safekeeper and propagated to compute node as a part of `AppendResponse`.
Finally, when compute node receives the `current_timeline_size` from safekeeper (or from pageserver directly), it updates the global variable.
And then every zenith_extend() operation checks if limit is reached `(current_timeline_size > neon.max_cluster_size)` and throws `ERRCODE_DISK_FULL` error if so.
And then every neon_extend() operation checks if limit is reached `(current_timeline_size > neon.max_cluster_size)` and throws `ERRCODE_DISK_FULL` error if so.
(see Postgres error codes [https://www.postgresql.org/docs/devel/errcodes-appendix.html](https://www.postgresql.org/docs/devel/errcodes-appendix.html))
TODO:
We can allow autovacuum processes to bypass this check, simply checking `IsAutoVacuumWorkerProcess()`.
It would be nice to allow manual VACUUM and VACUUM FULL to bypass the check, but it's uneasy to distinguish these operations at the low level.
See issues https://github.com/neondatabase/neon/issues/1245
https://github.com/zenithdb/zenith/issues/1445
https://github.com/neondatabase/neon/issues/1445
TODO:
We should warn users if the limit is soon to be reached.

4
docs/rfcs/images/017-timeline-data-management/lock_legend.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 15 KiB

4
docs/rfcs/images/017-timeline-data-management/proposed_timeline_data_access_sync_1.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 62 KiB

4
docs/rfcs/images/017-timeline-data-management/proposed_timeline_data_access_sync_2.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 33 KiB

4
docs/rfcs/images/017-timeline-data-management/proposed_timeline_tenant_state.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 34 KiB

4
docs/rfcs/images/017-timeline-data-management/timeline_data_access_sync_1.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 70 KiB

4
docs/rfcs/images/017-timeline-data-management/timeline_data_access_sync_2.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 49 KiB

4
docs/rfcs/images/017-timeline-data-management/timeline_tenant_state.svg Normal file
View File

File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 33 KiB

4
docs/settings.md
View File

@@ -155,9 +155,11 @@ for other files and for sockets for incoming connections.
#### pg_distrib_dir
A directory with Postgres installation to use during pageserver activities.
Since pageserver supports several postgres versions, `pg_distrib_dir` contains
a subdirectory for each version with naming convention `v{PG_MAJOR_VERSION}/`.
Inside that dir, a `bin/postgres` binary should be present.
The default distrib dir is `./tmp_install/`.
The default distrib dir is `./pg_install/`.
#### workdir (-D)

68
docs/sourcetree.md
View File

@@ -10,7 +10,7 @@ Intended to be used in integration tests and in CLI tools for local installation
`/docs`:
Documentation of the Zenith features and concepts.
Documentation of the Neon features and concepts.
Now it is mostly dev documentation.
`/monitoring`:
@@ -19,7 +19,7 @@ TODO
`/pageserver`:
Zenith storage service.
Neon storage service.
The pageserver has a few different duties:
- Store and manage the data.
@@ -40,21 +40,21 @@ and create new databases and accounts (control plane API in our case).
Integration tests, written in Python using the `pytest` framework.
`/vendor/postgres`:
`/vendor/postgres-v14`:
PostgreSQL source tree, with the modifications needed for Neon.
`/vendor/postgres/contrib/neon`:
`/pgxn/neon`:
PostgreSQL extension that implements storage manager API and network communications with remote page server.
`/vendor/postgres/contrib/neon_test_utils`:
`/pgxn/neon_test_utils`:
PostgreSQL extension that contains functions needed for testing and debugging.
`/safekeeper`:
The zenith WAL service that receives WAL from a primary compute nodes and streams it to the pageserver.
The neon WAL service that receives WAL from a primary compute nodes and streams it to the pageserver.
It acts as a holding area and redistribution center for recently generated WAL.
For more detailed info, see [walservice.md](./walservice.md)
@@ -64,11 +64,6 @@ The workspace_hack crate exists only to pin down some dependencies.
We use [cargo-hakari](https://crates.io/crates/cargo-hakari) for automation.
`/zenith`
Main entry point for the 'zenith' CLI utility.
TODO: Doesn't it belong to control_plane?
`/libs`:
Unites granular neon helper crates under the hood.
@@ -101,7 +96,7 @@ A single virtual environment with all dependencies is described in the single `P
sudo apt install python3.9
```
- Install `poetry`
- Exact version of `poetry` is not important, see installation instructions available at poetry's [website](https://python-poetry.org/docs/#installation)`.
- Exact version of `poetry` is not important, see installation instructions available at poetry's [website](https://python-poetry.org/docs/#installation).
- Install dependencies via `./scripts/pysync`.
- Note that CI uses specific Python version (look for `PYTHON_VERSION` [here](https://github.com/neondatabase/docker-images/blob/main/rust/Dockerfile))
so if you have different version some linting tools can yield different result locally vs in the CI.
@@ -134,3 +129,52 @@ Also consider:
To add new package or change an existing one you can use `poetry add` or `poetry update` or edit `pyproject.toml` manually. Do not forget to run `poetry lock` in the latter case.
More details are available in poetry's [documentation](https://python-poetry.org/docs/).
## Configuring IDEs
Neon consists of three projects in different languages which use different project models.
* A bunch of Rust crates, all available from the root `Cargo.toml`.
* Integration tests in Python in the `test_runner` directory. Some stand-alone Python scripts exist as well.
* Postgres and our Postgres extensions in C built with Makefiles under `vendor/postgres` and `pgxn`.
### CLion
You can use CLion with the [Rust plugin](https://plugins.jetbrains.com/plugin/8182-rust) to develop Neon. It should pick up Rust and Python projects whenever you open Neon's repository as a project. We have not tried setting up a debugger, though.
C code requires some extra care, as it's built via Make, not CMake. Some of our developers have successfully used [compilation database](https://www.jetbrains.com/help/clion/compilation-database.html#compdb_generate) for CLion. It is a JSON file which lists all C source files and corresponding compilation keys. CLion can use it instead of `CMakeLists.txt`. To set up a project with a compilation database:
1. Clone the Neon repository and install all dependencies, including Python. Do not open it with CLion just yet.
2. Run the following commands in the repository's root:
```bash
# Install a `compiledb` tool which can parse make's output and generate the compilation database.
poetry add -D compiledb
# Clean the build tree so we can rebuild from scratch.
# Unfortunately, our and Postgres Makefiles do not work well with either --dry-run or --assume-new,
# so we don't know a way to generate the compilation database without recompiling everything,
# see https://github.com/neondatabase/neon/issues/2378#issuecomment-1241421325
make distclean
# Rebuild the Postgres parts from scratch and save the compilation commands to the compilation database.
# You can alter the -j parameter to your liking.
# Note that we only build for a specific version of Postgres. The extension code is shared, but headers are
# different, so we set up CLion to only use a specific version of the headers.
make -j$(nproc) --print-directory postgres-v15 neon-pg-ext-v15 | poetry run compiledb --verbose --no-build
# Uninstall the tool
poetry remove -D compiledb
# Make sure the compile_commands.json file is not committed.
echo /compile_commands.json >>.git/info/exclude
```
3. Open CLion, click "Open File or Project" and choose the generated `compile_commands.json` file to be opened "as a project". You cannot add a compilation database into an existing CLion project, you have to create a new one. _Do not_ open the directory as a project, open the file.
4. The newly created project should start indexing Postgres source code in C, as well as the C standard library. You may have to [configure the C compiler for the compilation database](https://www.jetbrains.com/help/clion/compilation-database.html#compdb_toolchain).
5. Open the `Cargo.toml` file in an editor in the same project. CLion should pick up the hint and start indexing Rust code.
6. Now you have a CLion project which knows about C files, Rust files. It should pick up Python files automatically as well.
7. Set up correct code indentation in CLion's settings: Editor > Code Style > C/C++, choose the "Project" scheme on the top, and tick the "Use tab character" on the "Tabs and Indents" tab. Ensure that "Tab size" is 4.
You can also enable Cargo Clippy diagnostics and enable Rustfmt instead of built-in code formatter.
Whenever you change layout of C files, you may need to regenerate the compilation database. No need to re-create the CLion project, changes should be picked up automatically.
Known issues (fixes and suggestions are welcome):
* Test results may be hard to read in CLion, both for unit tests in Rust and integration tests in Python. Use command line to run them instead.
* CLion does not support non-local Python interpreters, unlike PyCharm. E.g. if you use WSL, CLion does not see `poetry` and installed dependencies. Python support is limited.
* Cargo Clippy diagnostics in CLion may take a lot of resources.
* `poetry add -D` updates some packages and changes `poetry.lock` drastically even when followed by `poetry remove -D`. Feel free to `git checkout poetry.lock` and `./scripts/pysync` to revert these changes.

26
libs/etcd_broker/src/subscription_key.rs
View File

@@ -11,7 +11,7 @@ use std::{fmt::Display, str::FromStr};
use once_cell::sync::Lazy;
use regex::{Captures, Regex};
use utils::zid::{NodeId, ZTenantId, ZTenantTimelineId};
use utils::id::{NodeId, TenantId, TenantTimelineId};
/// The subscription kind to the timeline updates from safekeeper.
#[derive(Debug, Clone, PartialEq, Eq, Hash)]
@@ -30,13 +30,13 @@ pub enum SubscriptionKind {
/// Get every update in etcd.
All,
/// Get etcd updates for any timeiline of a certain tenant, affected by any operation from any node kind.
TenantTimelines(ZTenantId),
TenantTimelines(TenantId),
/// Get etcd updates for a certain timeline of a tenant, affected by any operation from any node kind.
Timeline(ZTenantTimelineId),
Timeline(TenantTimelineId),
/// Get etcd timeline updates, specific to a certain node kind.
Node(ZTenantTimelineId, NodeKind),
Node(TenantTimelineId, NodeKind),
/// Get etcd timeline updates for a certain operation on specific nodes.
Operation(ZTenantTimelineId, NodeKind, OperationKind),
Operation(TenantTimelineId, NodeKind, OperationKind),
}
/// All kinds of nodes, able to write into etcd.
@@ -67,7 +67,7 @@ static SUBSCRIPTION_FULL_KEY_REGEX: Lazy<Regex> = Lazy::new(|| {
/// No other etcd keys are considered during system's work.
#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
pub struct SubscriptionFullKey {
pub id: ZTenantTimelineId,
pub id: TenantTimelineId,
pub node_kind: NodeKind,
pub operation: OperationKind,
pub node_id: NodeId,
@@ -83,7 +83,7 @@ impl SubscriptionKey {
}
/// Subscribes to a given timeline info updates from safekeepers.
pub fn sk_timeline_info(cluster_prefix: String, timeline: ZTenantTimelineId) -> Self {
pub fn sk_timeline_info(cluster_prefix: String, timeline: TenantTimelineId) -> Self {
Self {
cluster_prefix,
kind: SubscriptionKind::Operation(
@@ -97,7 +97,7 @@ impl SubscriptionKey {
/// Subscribes to all timeine updates during specific operations, running on the corresponding nodes.
pub fn operation(
cluster_prefix: String,
timeline: ZTenantTimelineId,
timeline: TenantTimelineId,
node_kind: NodeKind,
operation: OperationKind,
) -> Self {
@@ -175,7 +175,7 @@ impl FromStr for SubscriptionFullKey {
};
Ok(Self {
id: ZTenantTimelineId::new(
id: TenantTimelineId::new(
parse_capture(&key_captures, 1)?,
parse_capture(&key_captures, 2)?,
),
@@ -247,7 +247,7 @@ impl FromStr for SkOperationKind {
#[cfg(test)]
mod tests {
use utils::zid::ZTimelineId;
use utils::id::TimelineId;
use super::*;
@@ -256,9 +256,9 @@ mod tests {
let prefix = "neon";
let node_kind = NodeKind::Safekeeper;
let operation_kind = OperationKind::Safekeeper(SkOperationKind::WalBackup);
let tenant_id = ZTenantId::generate();
let timeline_id = ZTimelineId::generate();
let id = ZTenantTimelineId::new(tenant_id, timeline_id);
let tenant_id = TenantId::generate();
let timeline_id = TimelineId::generate();
let id = TenantTimelineId::new(tenant_id, timeline_id);
let node_id = NodeId(1);
let timeline_subscription_keys = [

19
libs/metrics/src/lib.rs
View File

@@ -3,7 +3,7 @@
//! Otherwise, we might not see all metrics registered via
//! a default registry.
use once_cell::sync::Lazy;
use prometheus::core::{AtomicU64, GenericGauge, GenericGaugeVec};
use prometheus::core::{AtomicU64, Collector, GenericGauge, GenericGaugeVec};
pub use prometheus::opts;
pub use prometheus::register;
pub use prometheus::{core, default_registry, proto};
@@ -17,6 +17,7 @@ pub use prometheus::{register_int_counter_vec, IntCounterVec};
pub use prometheus::{register_int_gauge, IntGauge};
pub use prometheus::{register_int_gauge_vec, IntGaugeVec};
pub use prometheus::{Encoder, TextEncoder};
use prometheus::{Registry, Result};
mod wrappers;
pub use wrappers::{CountedReader, CountedWriter};
@@ -32,13 +33,27 @@ macro_rules! register_uint_gauge_vec {
}};
}
/// Special internal registry, to collect metrics independently from the default registry.
/// Was introduced to fix deadlock with lazy registration of metrics in the default registry.
static INTERNAL_REGISTRY: Lazy<Registry> = Lazy::new(Registry::new);
/// Register a collector in the internal registry. MUST be called before the first call to `gather()`.
/// Otherwise, we can have a deadlock in the `gather()` call, trying to register a new collector
/// while holding the lock.
pub fn register_internal(c: Box<dyn Collector>) -> Result<()> {
INTERNAL_REGISTRY.register(c)
}
/// Gathers all Prometheus metrics and records the I/O stats just before that.
///
/// Metrics gathering is a relatively simple and standalone operation, so
/// it might be fine to do it this way to keep things simple.
pub fn gather() -> Vec<prometheus::proto::MetricFamily> {
update_rusage_metrics();
prometheus::gather()
let mut mfs = prometheus::gather();
let mut internal_mfs = INTERNAL_REGISTRY.gather();
mfs.append(&mut internal_mfs);
mfs
}
static DISK_IO_BYTES: Lazy<IntGaugeVec> = Lazy::new(|| {

12
libs/pageserver_api/Cargo.toml Normal file
View File

@@ -0,0 +1,12 @@
[package]
name = "pageserver_api"
version = "0.1.0"
edition = "2021"
[dependencies]
serde = { version = "1.0", features = ["derive"] }
serde_with = "1.12.0"
const_format = "0.2.21"
utils = { path = "../utils" }
workspace_hack = { version = "0.1", path = "../../workspace_hack" }

9
libs/pageserver_api/src/lib.rs Normal file
View File

@@ -0,0 +1,9 @@
use const_format::formatcp;
/// Public API types
pub mod models;
pub const DEFAULT_PG_LISTEN_PORT: u16 = 64000;
pub const DEFAULT_PG_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_PG_LISTEN_PORT}");
pub const DEFAULT_HTTP_LISTEN_PORT: u16 = 9898;
pub const DEFAULT_HTTP_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_HTTP_LISTEN_PORT}");

61
pageserver/src/http/models.rs → libs/pageserver_api/src/models.rs
View File

@@ -3,26 +3,35 @@ use std::num::NonZeroU64;
use serde::{Deserialize, Serialize};
use serde_with::{serde_as, DisplayFromStr};
use utils::{
id::{NodeId, TenantId, TimelineId},
lsn::Lsn,
zid::{NodeId, ZTenantId, ZTimelineId},
};
// These enums are used in the API response fields.
use crate::repository::LocalTimelineState;
use crate::tenant_mgr::TenantState;
/// A state of a tenant in pageserver's memory.
#[derive(Debug, Clone, Copy, PartialEq, Eq, serde::Serialize, serde::Deserialize)]
pub enum TenantState {
/// Tenant is fully operational, its background jobs might be running or not.
Active { background_jobs_running: bool },
/// A tenant is recognized by pageserver, but not yet ready to operate:
/// e.g. not present locally and being downloaded or being read into memory from the file system.
Paused,
/// A tenant is recognized by the pageserver, but no longer used for any operations, as failed to get activated.
Broken,
}
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TimelineCreateRequest {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub new_timeline_id: Option<ZTimelineId>,
pub new_timeline_id: Option<TimelineId>,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<ZTimelineId>,
pub ancestor_timeline_id: Option<TimelineId>,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_start_lsn: Option<Lsn>,
pub pg_version: Option<u32>,
}
#[serde_as]
@@ -30,7 +39,7 @@ pub struct TimelineCreateRequest {
pub struct TenantCreateRequest {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub new_tenant_id: Option<ZTenantId>,
pub new_tenant_id: Option<TenantId>,
pub checkpoint_distance: Option<u64>,
pub checkpoint_timeout: Option<String>,
pub compaction_target_size: Option<u64>,
@@ -48,7 +57,7 @@ pub struct TenantCreateRequest {
#[serde_as]
#[derive(Serialize, Deserialize)]
#[serde(transparent)]
pub struct TenantCreateResponse(#[serde_as(as = "DisplayFromStr")] pub ZTenantId);
pub struct TenantCreateResponse(#[serde_as(as = "DisplayFromStr")] pub TenantId);
#[derive(Serialize)]
pub struct StatusResponse {
@@ -56,7 +65,7 @@ pub struct StatusResponse {
}
impl TenantCreateRequest {
pub fn new(new_tenant_id: Option<ZTenantId>) -> TenantCreateRequest {
pub fn new(new_tenant_id: Option<TenantId>) -> TenantCreateRequest {
TenantCreateRequest {
new_tenant_id,
..Default::default()
@@ -67,7 +76,7 @@ impl TenantCreateRequest {
#[serde_as]
#[derive(Serialize, Deserialize)]
pub struct TenantConfigRequest {
pub tenant_id: ZTenantId,
pub tenant_id: TenantId,
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub checkpoint_distance: Option<u64>,
@@ -85,7 +94,7 @@ pub struct TenantConfigRequest {
}
impl TenantConfigRequest {
pub fn new(tenant_id: ZTenantId) -> TenantConfigRequest {
pub fn new(tenant_id: TenantId) -> TenantConfigRequest {
TenantConfigRequest {
tenant_id,
checkpoint_distance: None,
@@ -108,8 +117,8 @@ impl TenantConfigRequest {
#[derive(Serialize, Deserialize, Clone)]
pub struct TenantInfo {
#[serde_as(as = "DisplayFromStr")]
pub id: ZTenantId,
pub state: Option<TenantState>,
pub id: TenantId,
pub state: TenantState,
pub current_physical_size: Option<u64>, // physical size is only included in `tenant_status` endpoint
pub has_in_progress_downloads: Option<bool>,
}
@@ -118,7 +127,7 @@ pub struct TenantInfo {
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct LocalTimelineInfo {
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_timeline_id: Option<ZTimelineId>,
pub ancestor_timeline_id: Option<TimelineId>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub ancestor_lsn: Option<Lsn>,
#[serde_as(as = "DisplayFromStr")]
@@ -133,13 +142,13 @@ pub struct LocalTimelineInfo {
pub current_physical_size: Option<u64>, // is None when timeline is Unloaded
pub current_logical_size_non_incremental: Option<u64>,
pub current_physical_size_non_incremental: Option<u64>,
pub timeline_state: LocalTimelineState,
pub wal_source_connstr: Option<String>,
#[serde_as(as = "Option<DisplayFromStr>")]
pub last_received_msg_lsn: Option<Lsn>,
/// the timestamp (in microseconds) of the last received message
pub last_received_msg_ts: Option<u128>,
pub pg_version: u32,
}
#[serde_as]
@@ -157,9 +166,27 @@ pub struct RemoteTimelineInfo {
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct TimelineInfo {
#[serde_as(as = "DisplayFromStr")]
pub tenant_id: ZTenantId,
pub tenant_id: TenantId,
#[serde_as(as = "DisplayFromStr")]
pub timeline_id: ZTimelineId,
pub timeline_id: TimelineId,
pub local: Option<LocalTimelineInfo>,
pub remote: Option<RemoteTimelineInfo>,
}
pub type ConfigureFailpointsRequest = Vec<FailpointConfig>;
/// Information for configuring a single fail point
#[derive(Debug, Serialize, Deserialize)]
pub struct FailpointConfig {
/// Name of the fail point
pub name: String,
/// List of actions to take, using the format described in `fail::cfg`
///
/// We also support `actions = "exit"` to cause the fail point to immediately exit.
pub actions: String,
}
#[derive(Debug, Serialize, Deserialize)]
pub struct TimelineGcRequest {
pub gc_horizon: Option<u64>,
}

5
libs/postgres_ffi/Cargo.toml
View File

@@ -21,8 +21,9 @@ workspace_hack = { version = "0.1", path = "../../workspace_hack" }
[dev-dependencies]
env_logger = "0.9"
postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
wal_craft = { path = "wal_craft" }
[build-dependencies]
bindgen = "0.59.1"
anyhow = "1.0"
bindgen = "0.60.1"

8
libs/postgres_ffi/README.md
View File

@@ -9,9 +9,11 @@ should be auto-generated too, but that's a TODO.
The PostgreSQL on-disk file format is not portable across different
CPU architectures and operating systems. It is also subject to change
in each major PostgreSQL version. Currently, this module is based on
PostgreSQL v14, but in the future we will probably need a separate
copy for each PostgreSQL version.
in each major PostgreSQL version. Currently, this module supports
PostgreSQL v14 and v15: bindings and code that depends on them are version-specific.
This code is organized in modules: `postgres_ffi::v14` and `postgres_ffi::v15`
Version independend code is explicitly exported into shared `postgres_ffi`.
TODO: Currently, there is also some code that deals with WAL records
in pageserver/src/waldecoder.rs. That should be moved into this

174
libs/postgres_ffi/build.rs
View File

@@ -4,6 +4,7 @@ use std::env;
use std::path::PathBuf;
use std::process::Command;
use anyhow::{anyhow, Context};
use bindgen::callbacks::ParseCallbacks;
#[derive(Debug)]
@@ -42,93 +43,112 @@ impl ParseCallbacks for PostgresFfiCallbacks {
}
}
fn main() {
fn main() -> anyhow::Result<()> {
// Tell cargo to invalidate the built crate whenever the wrapper changes
println!("cargo:rerun-if-changed=bindgen_deps.h");
// Finding the location of C headers for the Postgres server:
// - if POSTGRES_INSTALL_DIR is set look into it, otherwise look into `<project_root>/tmp_install`
// - if there's a `bin/pg_config` file use it for getting include server, otherwise use `<project_root>/tmp_install/include/postgresql/server`
let mut pg_install_dir = if let Some(postgres_install_dir) = env::var_os("POSTGRES_INSTALL_DIR")
{
// - if POSTGRES_INSTALL_DIR is set look into it, otherwise look into `<project_root>/pg_install`
// - if there's a `bin/pg_config` file use it for getting include server, otherwise use `<project_root>/pg_install/{PG_MAJORVERSION}/include/postgresql/server`
let pg_install_dir = if let Some(postgres_install_dir) = env::var_os("POSTGRES_INSTALL_DIR") {
postgres_install_dir.into()
} else {
PathBuf::from("tmp_install")
PathBuf::from("pg_install")
};
if pg_install_dir.is_relative() {
let cwd = env::current_dir().unwrap();
pg_install_dir = cwd.join("..").join("..").join(pg_install_dir);
}
let pg_config_bin = pg_install_dir.join("bin").join("pg_config");
let inc_server_path: String = if pg_config_bin.exists() {
let output = Command::new(pg_config_bin)
.arg("--includedir-server")
.output()
.expect("failed to execute `pg_config --includedir-server`");
if !output.status.success() {
panic!("`pg_config --includedir-server` failed")
for pg_version in &["v14", "v15"] {
let mut pg_install_dir_versioned = pg_install_dir.join(pg_version);
if pg_install_dir_versioned.is_relative() {
let cwd = env::current_dir().context("Failed to get current_dir")?;
pg_install_dir_versioned = cwd.join("..").join("..").join(pg_install_dir_versioned);
}
String::from_utf8(output.stdout).unwrap().trim_end().into()
} else {
pg_install_dir
.join("include")
.join("postgresql")
.join("server")
.into_os_string()
.into_string()
.unwrap()
};
let pg_config_bin = pg_install_dir_versioned
.join(pg_version)
.join("bin")
.join("pg_config");
let inc_server_path: String = if pg_config_bin.exists() {
let output = Command::new(pg_config_bin)
.arg("--includedir-server")
.output()
.context("failed to execute `pg_config --includedir-server`")?;
// The bindgen::Builder is the main entry point
// to bindgen, and lets you build up options for
// the resulting bindings.
let bindings = bindgen::Builder::default()
//
// All the needed PostgreSQL headers are included from 'bindgen_deps.h'
//
.header("bindgen_deps.h")
//
// Tell cargo to invalidate the built crate whenever any of the
// included header files changed.
//
.parse_callbacks(Box::new(PostgresFfiCallbacks))
//
// These are the types and constants that we want to generate bindings for
//
.allowlist_type("BlockNumber")
.allowlist_type("OffsetNumber")
.allowlist_type("MultiXactId")
.allowlist_type("MultiXactOffset")
.allowlist_type("MultiXactStatus")
.allowlist_type("ControlFileData")
.allowlist_type("CheckPoint")
.allowlist_type("FullTransactionId")
.allowlist_type("XLogRecord")
.allowlist_type("XLogPageHeaderData")
.allowlist_type("XLogLongPageHeaderData")
.allowlist_var("XLOG_PAGE_MAGIC")
.allowlist_var("PG_CONTROL_FILE_SIZE")
.allowlist_var("PG_CONTROLFILEDATA_OFFSETOF_CRC")
.allowlist_type("PageHeaderData")
.allowlist_type("DBState")
// Because structs are used for serialization, tell bindgen to emit
// explicit padding fields.
.explicit_padding(true)
//
.clang_arg(format!("-I{inc_server_path}"))
//
// Finish the builder and generate the bindings.
//
.generate()
.expect("Unable to generate bindings");
if !output.status.success() {
panic!("`pg_config --includedir-server` failed")
}
// Write the bindings to the $OUT_DIR/bindings.rs file.
let out_path = PathBuf::from(env::var("OUT_DIR").unwrap());
bindings
.write_to_file(out_path.join("bindings.rs"))
.expect("Couldn't write bindings!");
String::from_utf8(output.stdout)
.context("pg_config output is not UTF-8")?
.trim_end()
.into()
} else {
let server_path = pg_install_dir_versioned
.join("include")
.join("postgresql")
.join("server")
.into_os_string();
server_path
.into_string()
.map_err(|s| anyhow!("Bad postgres server path {s:?}"))?
};
// The bindgen::Builder is the main entry point
// to bindgen, and lets you build up options for
// the resulting bindings.
let bindings = bindgen::Builder::default()
//
// All the needed PostgreSQL headers are included from 'bindgen_deps.h'
//
.header("bindgen_deps.h")
//
// Tell cargo to invalidate the built crate whenever any of the
// included header files changed.
//
.parse_callbacks(Box::new(PostgresFfiCallbacks))
//
// These are the types and constants that we want to generate bindings for
//
.allowlist_type("BlockNumber")
.allowlist_type("OffsetNumber")
.allowlist_type("XLogRecPtr")
.allowlist_type("XLogSegNo")
.allowlist_type("TimeLineID")
.allowlist_type("TimestampTz")
.allowlist_type("MultiXactId")
.allowlist_type("MultiXactOffset")
.allowlist_type("MultiXactStatus")
.allowlist_type("ControlFileData")
.allowlist_type("CheckPoint")
.allowlist_type("FullTransactionId")
.allowlist_type("XLogRecord")
.allowlist_type("XLogPageHeaderData")
.allowlist_type("XLogLongPageHeaderData")
.allowlist_var("XLOG_PAGE_MAGIC")
.allowlist_var("PG_CONTROL_FILE_SIZE")
.allowlist_var("PG_CONTROLFILEDATA_OFFSETOF_CRC")
.allowlist_type("PageHeaderData")
.allowlist_type("DBState")
// Because structs are used for serialization, tell bindgen to emit
// explicit padding fields.
.explicit_padding(true)
//
.clang_arg(format!("-I{inc_server_path}"))
//
// Finish the builder and generate the bindings.
//
.generate()
.context("Unable to generate bindings")?;
// Write the bindings to the $OUT_DIR/bindings_$pg_version.rs file.
let out_path: PathBuf = env::var("OUT_DIR")
.context("Couldn't read OUT_DIR environment variable var")?
.into();
let filename = format!("bindings_{pg_version}.rs");
bindings
.write_to_file(out_path.join(filename))
.context("Couldn't write bindings")?;
}
Ok(())
}

156
libs/postgres_ffi/src/lib.rs
View File

@@ -3,31 +3,41 @@
#![allow(non_snake_case)]
// bindgen creates some unsafe code with no doc comments.
#![allow(clippy::missing_safety_doc)]
// suppress warnings on rust 1.53 due to bindgen unit tests.
// https://github.com/rust-lang/rust-bindgen/issues/1651
#![allow(deref_nullptr)]
// noted at 1.63 that in many cases there's a u32 -> u32 transmutes in bindgen code.
#![allow(clippy::useless_transmute)]
// modules included with the postgres_ffi macro depend on the types of the specific version's
// types, and trigger a too eager lint.
#![allow(clippy::duplicate_mod)]
use bytes::Bytes;
use utils::bin_ser::SerializeError;
use utils::lsn::Lsn;
macro_rules! postgres_ffi {
($version:ident) => {
#[path = "."]
pub mod $version {
// fixme: does this have to be 'pub'?
pub mod bindings {
// bindgen generates bindings for a lot of stuff we don't need
#![allow(dead_code)]
use serde::{Deserialize, Serialize};
include!(concat!(env!("OUT_DIR"), "/bindings.rs"));
include!(concat!(
env!("OUT_DIR"),
"/bindings_",
stringify!($version),
".rs"
));
include!(concat!("pg_constants_", stringify!($version), ".rs"));
}
pub mod controlfile_utils;
pub mod nonrelfile_utils;
pub mod pg_constants;
pub mod relfile_utils;
pub mod waldecoder;
pub mod waldecoder_handler;
pub mod xlog_utils;
pub const PG_MAJORVERSION: &str = stringify!($version);
// Re-export some symbols from bindings
pub use bindings::DBState_DB_SHUTDOWNED;
pub use bindings::{CheckPoint, ControlFileData, XLogRecord};
@@ -36,20 +46,75 @@ macro_rules! postgres_ffi {
}
postgres_ffi!(v14);
postgres_ffi!(v15);
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::{uint32, uint64, Oid};
pub use v14::bindings::{BlockNumber, OffsetNumber};
pub use v14::bindings::{MultiXactId, TransactionId};
pub use v14::bindings::{TimeLineID, TimestampTz, XLogRecPtr, XLogSegNo};
// Likewise for these, although the assumption that these don't change is a little more iffy.
pub use v14::bindings::{MultiXactOffset, MultiXactStatus};
pub use v14::bindings::{PageHeaderData, XLogRecord};
pub use v14::xlog_utils::{XLOG_SIZE_OF_XLOG_RECORD, XLOG_SIZE_OF_XLOG_SHORT_PHD};
pub use v14::bindings::{CheckPoint, ControlFileData};
// from pg_config.h. These can be changed with configure options --with-blocksize=BLOCKSIZE and
// --with-segsize=SEGSIZE, but assume the defaults for now.
pub const BLCKSZ: u16 = 8192;
pub const RELSEG_SIZE: u32 = 1024 * 1024 * 1024 / (BLCKSZ as u32);
pub const XLOG_BLCKSZ: usize = 8192;
pub const WAL_SEGMENT_SIZE: usize = 16 * 1024 * 1024;
pub const MAX_SEND_SIZE: usize = XLOG_BLCKSZ * 16;
// Export some version independent functions that are used outside of this mod
pub use v14::xlog_utils::encode_logical_message;
pub use v14::xlog_utils::get_current_timestamp;
pub use v14::xlog_utils::to_pg_timestamp;
pub use v14::xlog_utils::XLogFileName;
pub use v14::bindings::DBState_DB_SHUTDOWNED;
pub fn bkpimage_is_compressed(bimg_info: u8, version: u32) -> anyhow::Result<bool> {
match version {
14 => Ok(bimg_info & v14::bindings::BKPIMAGE_IS_COMPRESSED != 0),
15 => Ok(bimg_info & v15::bindings::BKPIMAGE_COMPRESS_PGLZ != 0
|| bimg_info & v15::bindings::BKPIMAGE_COMPRESS_LZ4 != 0
|| bimg_info & v15::bindings::BKPIMAGE_COMPRESS_ZSTD != 0),
_ => anyhow::bail!("Unknown version {}", version),
}
}
pub fn generate_wal_segment(
segno: u64,
system_id: u64,
pg_version: u32,
) -> Result<Bytes, SerializeError> {
match pg_version {
14 => v14::xlog_utils::generate_wal_segment(segno, system_id),
15 => v15::xlog_utils::generate_wal_segment(segno, system_id),
_ => Err(SerializeError::BadInput),
}
}
pub fn generate_pg_control(
pg_control_bytes: &[u8],
checkpoint_bytes: &[u8],
lsn: Lsn,
pg_version: u32,
) -> anyhow::Result<(Bytes, u64)> {
match pg_version {
14 => v14::xlog_utils::generate_pg_control(pg_control_bytes, checkpoint_bytes, lsn),
15 => v15::xlog_utils::generate_pg_control(pg_control_bytes, checkpoint_bytes, lsn),
_ => anyhow::bail!("Unknown version {}", pg_version),
}
}
// PG timeline is always 1, changing it doesn't have any useful meaning in Neon.
//
@@ -62,7 +127,7 @@ pub const PG_TLI: u32 = 1;
// See TransactionIdIsNormal in transam.h
pub const fn transaction_id_is_normal(id: TransactionId) -> bool {
id > v14::pg_constants::FIRST_NORMAL_TRANSACTION_ID
id > pg_constants::FIRST_NORMAL_TRANSACTION_ID
}
// See TransactionIdPrecedes in transam.c
@@ -97,3 +162,76 @@ pub fn page_set_lsn(pg: &mut [u8], lsn: Lsn) {
pg[0..4].copy_from_slice(&((lsn.0 >> 32) as u32).to_le_bytes());
pg[4..8].copy_from_slice(&(lsn.0 as u32).to_le_bytes());
}
pub mod waldecoder {
use crate::{v14, v15};
use bytes::{Buf, Bytes, BytesMut};
use std::num::NonZeroU32;
use thiserror::Error;
use utils::lsn::Lsn;
pub enum State {
WaitingForRecord,
ReassemblingRecord {
recordbuf: BytesMut,
contlen: NonZeroU32,
},
SkippingEverything {
skip_until_lsn: Lsn,
},
}
pub struct WalStreamDecoder {
pub lsn: Lsn,
pub pg_version: u32,
pub inputbuf: BytesMut,
pub state: State,
}
#[derive(Error, Debug, Clone)]
#[error("{msg} at {lsn}")]
pub struct WalDecodeError {
pub msg: String,
pub lsn: Lsn,
}
impl WalStreamDecoder {
pub fn new(lsn: Lsn, pg_version: u32) -> WalStreamDecoder {
WalStreamDecoder {
lsn,
pg_version,
inputbuf: BytesMut::new(),
state: State::WaitingForRecord,
}
}
// The latest LSN position fed to the decoder.
pub fn available(&self) -> Lsn {
self.lsn + self.inputbuf.remaining() as u64
}
pub fn feed_bytes(&mut self, buf: &[u8]) {
self.inputbuf.extend_from_slice(buf);
}
pub fn poll_decode(&mut self) -> Result<Option<(Lsn, Bytes)>, WalDecodeError> {
match self.pg_version {
// This is a trick to support both versions simultaneously.
// See WalStreamDecoderHandler comments.
14 => {
use self::v14::waldecoder_handler::WalStreamDecoderHandler;
self.poll_decode_internal()
}
15 => {
use self::v15::waldecoder_handler::WalStreamDecoderHandler;
self.poll_decode_internal()
}
_ => Err(WalDecodeError {
msg: format!("Unknown version {}", self.pg_version),
lsn: self.lsn,
}),
}
}
}
}

2
libs/postgres_ffi/src/nonrelfile_utils.rs
View File

@@ -1,7 +1,7 @@
//!
//! Common utilities for dealing with PostgreSQL non-relation files.
//!
use super::pg_constants;
use crate::pg_constants;
use crate::transaction_id_precedes;
use bytes::BytesMut;
use log::*;

26
libs/postgres_ffi/src/pg_constants.rs
View File

@@ -1,14 +1,16 @@
//!
//! Misc constants, copied from PostgreSQL headers.
//!
//! Only place version-independent constants here.
//!
//! TODO: These probably should be auto-generated using bindgen,
//! rather than copied by hand. Although on the other hand, it's nice
//! to have them all here in one place, and have the ability to add
//! comments on them.
//!
use super::bindings::PageHeaderData;
use crate::BLCKSZ;
use crate::{PageHeaderData, XLogRecord};
//
// From pg_tablespace_d.h
@@ -16,14 +18,6 @@ use crate::BLCKSZ;
pub const DEFAULTTABLESPACE_OID: u32 = 1663;
pub const GLOBALTABLESPACE_OID: u32 = 1664;
//
// Fork numbers, from relpath.h
//
pub const MAIN_FORKNUM: u8 = 0;
pub const FSM_FORKNUM: u8 = 1;
pub const VISIBILITYMAP_FORKNUM: u8 = 2;
pub const INIT_FORKNUM: u8 = 3;
// From storage_xlog.h
pub const XLOG_SMGR_CREATE: u8 = 0x10;
pub const XLOG_SMGR_TRUNCATE: u8 = 0x20;
@@ -114,7 +108,6 @@ pub const XLOG_NEXTOID: u8 = 0x30;
pub const XLOG_SWITCH: u8 = 0x40;
pub const XLOG_FPI_FOR_HINT: u8 = 0xA0;
pub const XLOG_FPI: u8 = 0xB0;
pub const DB_SHUTDOWNED: u32 = 1;
// From multixact.h
pub const FIRST_MULTIXACT_ID: u32 = 1;
@@ -169,14 +162,10 @@ pub const RM_HEAP_ID: u8 = 10;
pub const XLR_INFO_MASK: u8 = 0x0F;
pub const XLR_RMGR_INFO_MASK: u8 = 0xF0;
// from dbcommands_xlog.h
pub const XLOG_DBASE_CREATE: u8 = 0x00;
pub const XLOG_DBASE_DROP: u8 = 0x10;
pub const XLOG_TBLSPC_CREATE: u8 = 0x00;
pub const XLOG_TBLSPC_DROP: u8 = 0x10;
pub const SIZEOF_XLOGRECORD: u32 = 24;
pub const SIZEOF_XLOGRECORD: u32 = std::mem::size_of::<XLogRecord>() as u32;
//
// from xlogrecord.h
@@ -197,8 +186,6 @@ pub const BKPBLOCK_SAME_REL: u8 = 0x80; /* RelFileNode omitted, same as previous
/* Information stored in bimg_info */
pub const BKPIMAGE_HAS_HOLE: u8 = 0x01; /* page image has "hole" */
pub const BKPIMAGE_IS_COMPRESSED: u8 = 0x02; /* page image is compressed */
pub const BKPIMAGE_APPLY: u8 = 0x04; /* page image should be restored during replay */
/* From transam.h */
pub const FIRST_NORMAL_TRANSACTION_ID: u32 = 3;
@@ -206,15 +193,10 @@ pub const INVALID_TRANSACTION_ID: u32 = 0;
pub const FIRST_BOOTSTRAP_OBJECT_ID: u32 = 12000;
pub const FIRST_NORMAL_OBJECT_ID: u32 = 16384;
/* FIXME: pageserver should request wal_seg_size from compute node */
pub const WAL_SEGMENT_SIZE: usize = 16 * 1024 * 1024;
pub const XLOG_CHECKPOINT_SHUTDOWN: u8 = 0x00;
pub const XLOG_CHECKPOINT_ONLINE: u8 = 0x10;
pub const XLP_LONG_HEADER: u16 = 0x0002;
pub const PG_MAJORVERSION: &str = "14";
// List of subdirectories inside pgdata.
// Copied from src/bin/initdb/initdb.c
pub const PGDATA_SUBDIRS: [&str; 22] = [

5
libs/postgres_ffi/src/pg_constants_v14.rs Normal file
View File

@@ -0,0 +1,5 @@
pub const XLOG_DBASE_CREATE: u8 = 0x00;
pub const XLOG_DBASE_DROP: u8 = 0x10;
pub const BKPIMAGE_IS_COMPRESSED: u8 = 0x02; /* page image is compressed */
pub const BKPIMAGE_APPLY: u8 = 0x04; /* page image should be restored during replay */

10
libs/postgres_ffi/src/pg_constants_v15.rs Normal file
View File

@@ -0,0 +1,10 @@
pub const XACT_XINFO_HAS_DROPPED_STATS: u32 = 1u32 << 8;
pub const XLOG_DBASE_CREATE_FILE_COPY: u8 = 0x00;
pub const XLOG_DBASE_CREATE_WAL_LOG: u8 = 0x00;
pub const XLOG_DBASE_DROP: u8 = 0x20;
pub const BKPIMAGE_APPLY: u8 = 0x02; /* page image should be restored during replay */
pub const BKPIMAGE_COMPRESS_PGLZ: u8 = 0x04; /* page image is compressed */
pub const BKPIMAGE_COMPRESS_LZ4: u8 = 0x08; /* page image is compressed */
pub const BKPIMAGE_COMPRESS_ZSTD: u8 = 0x10; /* page image is compressed */

25
libs/postgres_ffi/src/relfile_utils.rs
View File

@@ -1,10 +1,17 @@
//!
//! Common utilities for dealing with PostgreSQL relation files.
//!
use super::pg_constants;
use once_cell::sync::OnceCell;
use regex::Regex;
//
// Fork numbers, from relpath.h
//
pub const MAIN_FORKNUM: u8 = 0;
pub const FSM_FORKNUM: u8 = 1;
pub const VISIBILITYMAP_FORKNUM: u8 = 2;
pub const INIT_FORKNUM: u8 = 3;
#[derive(Debug, Clone, thiserror::Error, PartialEq, Eq)]
pub enum FilePathError {
#[error("invalid relation fork name")]
@@ -23,10 +30,10 @@ impl From<core::num::ParseIntError> for FilePathError {
pub fn forkname_to_number(forkname: Option<&str>) -> Result<u8, FilePathError> {
match forkname {
// "main" is not in filenames, it's implicit if the fork name is not present
None => Ok(pg_constants::MAIN_FORKNUM),
Some("fsm") => Ok(pg_constants::FSM_FORKNUM),
Some("vm") => Ok(pg_constants::VISIBILITYMAP_FORKNUM),
Some("init") => Ok(pg_constants::INIT_FORKNUM),
None => Ok(MAIN_FORKNUM),
Some("fsm") => Ok(FSM_FORKNUM),
Some("vm") => Ok(VISIBILITYMAP_FORKNUM),
Some("init") => Ok(INIT_FORKNUM),
Some(_) => Err(FilePathError::InvalidForkName),
}
}
@@ -34,10 +41,10 @@ pub fn forkname_to_number(forkname: Option<&str>) -> Result<u8, FilePathError> {
/// Convert Postgres fork number to the right suffix of the relation data file.
pub fn forknumber_to_name(forknum: u8) -> Option<&'static str> {
match forknum {
pg_constants::MAIN_FORKNUM => None,
pg_constants::FSM_FORKNUM => Some("fsm"),
pg_constants::VISIBILITYMAP_FORKNUM => Some("vm"),
pg_constants::INIT_FORKNUM => Some("init"),
MAIN_FORKNUM => None,
FSM_FORKNUM => Some("fsm"),
VISIBILITYMAP_FORKNUM => Some("vm"),
INIT_FORKNUM => Some("init"),
_ => Some("UNKNOWN FORKNUM"),
}
}

66
libs/postgres_ffi/src/waldecoder.rs → libs/postgres_ffi/src/waldecoder_handler.rs
View File

@@ -8,63 +8,35 @@
//! to look deeper into the WAL records to also understand which blocks they modify, the code
//! for that is in pageserver/src/walrecord.rs
//!
use super::super::waldecoder::{State, WalDecodeError, WalStreamDecoder};
use super::bindings::{XLogLongPageHeaderData, XLogPageHeaderData, XLogRecord, XLOG_PAGE_MAGIC};
use super::pg_constants;
use super::xlog_utils::*;
use crate::WAL_SEGMENT_SIZE;
use bytes::{Buf, BufMut, Bytes, BytesMut};
use crc32c::*;
use log::*;
use std::cmp::min;
use std::num::NonZeroU32;
use thiserror::Error;
use utils::lsn::Lsn;
enum State {
WaitingForRecord,
ReassemblingRecord {
recordbuf: BytesMut,
contlen: NonZeroU32,
},
SkippingEverything {
skip_until_lsn: Lsn,
},
}
pub struct WalStreamDecoder {
lsn: Lsn,
inputbuf: BytesMut,
state: State,
}
#[derive(Error, Debug, Clone)]
#[error("{msg} at {lsn}")]
pub struct WalDecodeError {
msg: String,
lsn: Lsn,
pub trait WalStreamDecoderHandler {
fn validate_page_header(&self, hdr: &XLogPageHeaderData) -> Result<(), WalDecodeError>;
fn poll_decode_internal(&mut self) -> Result<Option<(Lsn, Bytes)>, WalDecodeError>;
fn complete_record(&mut self, recordbuf: Bytes) -> Result<(Lsn, Bytes), WalDecodeError>;
}
//
// WalRecordStream is a Stream that returns a stream of WAL records
// FIXME: This isn't a proper rust stream
// This is a trick to support several postgres versions simultaneously.
//
impl WalStreamDecoder {
pub fn new(lsn: Lsn) -> WalStreamDecoder {
WalStreamDecoder {
lsn,
inputbuf: BytesMut::new(),
state: State::WaitingForRecord,
}
}
// The latest LSN position fed to the decoder.
pub fn available(&self) -> Lsn {
self.lsn + self.inputbuf.remaining() as u64
}
pub fn feed_bytes(&mut self, buf: &[u8]) {
self.inputbuf.extend_from_slice(buf);
}
// Page decoding code depends on postgres bindings, so it is compiled for each version.
// Thus WalStreamDecoder implements several WalStreamDecoderHandler traits.
// WalStreamDecoder poll_decode() method dispatches to the right handler based on the postgres version.
// Other methods are internal and are not dispatched.
//
// It is similar to having several impl blocks for the same struct,
// but the impls here are in different modules, so need to use a trait.
//
impl WalStreamDecoderHandler for WalStreamDecoder {
fn validate_page_header(&self, hdr: &XLogPageHeaderData) -> Result<(), WalDecodeError> {
let validate_impl = || {
if hdr.xlp_magic != XLOG_PAGE_MAGIC as u16 {
@@ -125,7 +97,7 @@ impl WalStreamDecoder {
/// Ok(None): there is not enough data in the input buffer. Feed more by calling the `feed_bytes` function
/// Err(WalDecodeError): an error occurred while decoding, meaning the input was invalid.
///
pub fn poll_decode(&mut self) -> Result<Option<(Lsn, Bytes)>, WalDecodeError> {
fn poll_decode_internal(&mut self) -> Result<Option<(Lsn, Bytes)>, WalDecodeError> {
// Run state machine that validates page headers, and reassembles records
// that cross page boundaries.
loop {
@@ -133,7 +105,7 @@ impl WalStreamDecoder {
// However, we may have to skip some page headers if we're processing the XLOG_SWITCH record or skipping padding for whatever reason.
match self.state {
State::WaitingForRecord | State::ReassemblingRecord { .. } => {
if self.lsn.segment_offset(pg_constants::WAL_SEGMENT_SIZE) == 0 {
if self.lsn.segment_offset(WAL_SEGMENT_SIZE) == 0 {
// parse long header
if self.inputbuf.remaining() < XLOG_SIZE_OF_XLOG_LONG_PHD {
@@ -265,7 +237,7 @@ impl WalStreamDecoder {
// to the next WAL segment.
let next_lsn = if xlogrec.is_xlog_switch_record() {
trace!("saw xlog switch record at {}", self.lsn);
self.lsn + self.lsn.calc_padding(pg_constants::WAL_SEGMENT_SIZE as u64)
self.lsn + self.lsn.calc_padding(WAL_SEGMENT_SIZE as u64)
} else {
// Pad to an 8-byte boundary
self.lsn.align()

81
libs/postgres_ffi/src/xlog_utils.rs
View File

@@ -9,15 +9,16 @@
use crc32c::crc32c_append;
use super::super::waldecoder::WalStreamDecoder;
use super::bindings::{
CheckPoint, FullTransactionId, XLogLongPageHeaderData, XLogPageHeaderData, XLogRecord,
XLOG_PAGE_MAGIC,
CheckPoint, ControlFileData, DBState_DB_SHUTDOWNED, FullTransactionId, TimeLineID, TimestampTz,
XLogLongPageHeaderData, XLogPageHeaderData, XLogRecPtr, XLogRecord, XLogSegNo, XLOG_PAGE_MAGIC,
};
use super::pg_constants;
use super::pg_constants::WAL_SEGMENT_SIZE;
use crate::v14::waldecoder::WalStreamDecoder;
use super::PG_MAJORVERSION;
use crate::pg_constants;
use crate::PG_TLI;
use crate::{uint32, uint64, Oid};
use crate::{WAL_SEGMENT_SIZE, XLOG_BLCKSZ};
use bytes::BytesMut;
use bytes::{Buf, Bytes};
@@ -37,11 +38,9 @@ use utils::bin_ser::SerializeError;
use utils::lsn::Lsn;
pub const XLOG_FNAME_LEN: usize = 24;
pub const XLOG_BLCKSZ: usize = 8192;
pub const XLP_FIRST_IS_CONTRECORD: u16 = 0x0001;
pub const XLP_REM_LEN_OFFS: usize = 2 + 2 + 4 + 8;
pub const XLOG_RECORD_CRC_OFFS: usize = 4 + 4 + 8 + 1 + 1 + 2;
pub const MAX_SEND_SIZE: usize = XLOG_BLCKSZ * 16;
pub const XLOG_SIZE_OF_XLOG_SHORT_PHD: usize = std::mem::size_of::<XLogPageHeaderData>();
pub const XLOG_SIZE_OF_XLOG_LONG_PHD: usize = std::mem::size_of::<XLogLongPageHeaderData>();
@@ -49,11 +48,6 @@ pub const XLOG_SIZE_OF_XLOG_RECORD: usize = std::mem::size_of::<XLogRecord>();
#[allow(clippy::identity_op)]
pub const SIZE_OF_XLOG_RECORD_DATA_HEADER_SHORT: usize = 1 * 2;
pub type XLogRecPtr = u64;
pub type TimeLineID = u32;
pub type TimestampTz = i64;
pub type XLogSegNo = u64;
/// Interval of checkpointing metadata file. We should store metadata file to enforce
/// predicate that checkpoint.nextXid is larger than any XID in WAL.
/// But flushing checkpoint file for each transaction seems to be too expensive,
@@ -63,12 +57,10 @@ pub type XLogSegNo = u64;
/// in order to let CLOG_TRUNCATE mechanism correctly extend CLOG.
const XID_CHECKPOINT_INTERVAL: u32 = 1024;
#[allow(non_snake_case)]
pub fn XLogSegmentsPerXLogId(wal_segsz_bytes: usize) -> XLogSegNo {
(0x100000000u64 / wal_segsz_bytes as u64) as XLogSegNo
}
#[allow(non_snake_case)]
pub fn XLogSegNoOffsetToRecPtr(
segno: XLogSegNo,
offset: u32,
@@ -77,7 +69,6 @@ pub fn XLogSegNoOffsetToRecPtr(
segno * (wal_segsz_bytes as u64) + (offset as u64)
}
#[allow(non_snake_case)]
pub fn XLogFileName(tli: TimeLineID, logSegNo: XLogSegNo, wal_segsz_bytes: usize) -> String {
format!(
"{:>08X}{:>08X}{:>08X}",
@@ -87,7 +78,6 @@ pub fn XLogFileName(tli: TimeLineID, logSegNo: XLogSegNo, wal_segsz_bytes: usize
)
}
#[allow(non_snake_case)]
pub fn XLogFromFileName(fname: &str, wal_seg_size: usize) -> (XLogSegNo, TimeLineID) {
let tli = u32::from_str_radix(&fname[0..8], 16).unwrap();
let log = u32::from_str_radix(&fname[8..16], 16).unwrap() as XLogSegNo;
@@ -95,12 +85,10 @@ pub fn XLogFromFileName(fname: &str, wal_seg_size: usize) -> (XLogSegNo, TimeLin
(log * XLogSegmentsPerXLogId(wal_seg_size) + seg, tli)
}
#[allow(non_snake_case)]
pub fn IsXLogFileName(fname: &str) -> bool {
return fname.len() == XLOG_FNAME_LEN && fname.chars().all(|c| c.is_ascii_hexdigit());
}
#[allow(non_snake_case)]
pub fn IsPartialXLogFileName(fname: &str) -> bool {
fname.ends_with(".partial") && IsXLogFileName(&fname[0..fname.len() - 8])
}
@@ -120,6 +108,30 @@ pub fn normalize_lsn(lsn: Lsn, seg_sz: usize) -> Lsn {
}
}
pub fn generate_pg_control(
pg_control_bytes: &[u8],
checkpoint_bytes: &[u8],
lsn: Lsn,
) -> anyhow::Result<(Bytes, u64)> {
let mut pg_control = ControlFileData::decode(pg_control_bytes)?;
let mut checkpoint = CheckPoint::decode(checkpoint_bytes)?;
// Generate new pg_control needed for bootstrap
checkpoint.redo = normalize_lsn(lsn, WAL_SEGMENT_SIZE).0;
//reset some fields we don't want to preserve
//TODO Check this.
//We may need to determine the value from twophase data.
checkpoint.oldestActiveXid = 0;
//save new values in pg_control
pg_control.checkPoint = 0;
pg_control.checkPointCopy = checkpoint;
pg_control.state = DBState_DB_SHUTDOWNED;
Ok((pg_control.encode(), pg_control.system_identifier))
}
pub fn get_current_timestamp() -> TimestampTz {
to_pg_timestamp(SystemTime::now())
}
@@ -151,7 +163,10 @@ pub fn find_end_of_wal(
let mut result = start_lsn;
let mut curr_lsn = start_lsn;
let mut buf = [0u8; XLOG_BLCKSZ];
let mut decoder = WalStreamDecoder::new(start_lsn);
let pg_version = PG_MAJORVERSION[1..3].parse::<u32>().unwrap();
debug!("find_end_of_wal PG_VERSION: {}", pg_version);
let mut decoder = WalStreamDecoder::new(start_lsn, pg_version);
// loop over segments
loop {
@@ -161,7 +176,7 @@ pub fn find_end_of_wal(
match open_wal_segment(&seg_file_path)? {
None => {
// no more segments
info!(
debug!(
"find_end_of_wal reached end at {:?}, segment {:?} doesn't exist",
result, seg_file_path
);
@@ -184,7 +199,7 @@ pub fn find_end_of_wal(
match decoder.poll_decode() {
Ok(Some(record)) => result = record.0,
Err(e) => {
info!(
debug!(
"find_end_of_wal reached end at {:?}, decode error: {:?}",
result, e
);
@@ -318,9 +333,9 @@ impl CheckPoint {
// We need this segment to start compute node.
//
pub fn generate_wal_segment(segno: u64, system_id: u64) -> Result<Bytes, SerializeError> {
let mut seg_buf = BytesMut::with_capacity(pg_constants::WAL_SEGMENT_SIZE as usize);
let mut seg_buf = BytesMut::with_capacity(WAL_SEGMENT_SIZE as usize);
let pageaddr = XLogSegNoOffsetToRecPtr(segno, 0, pg_constants::WAL_SEGMENT_SIZE);
let pageaddr = XLogSegNoOffsetToRecPtr(segno, 0, WAL_SEGMENT_SIZE);
let hdr = XLogLongPageHeaderData {
std: {
XLogPageHeaderData {
@@ -333,7 +348,7 @@ pub fn generate_wal_segment(segno: u64, system_id: u64) -> Result<Bytes, Seriali
}
},
xlp_sysid: system_id,
xlp_seg_size: pg_constants::WAL_SEGMENT_SIZE as u32,
xlp_seg_size: WAL_SEGMENT_SIZE as u32,
xlp_xlog_blcksz: XLOG_BLCKSZ as u32,
};
@@ -341,7 +356,7 @@ pub fn generate_wal_segment(segno: u64, system_id: u64) -> Result<Bytes, Seriali
seg_buf.extend_from_slice(&hdr_bytes);
//zero out the rest of the file
seg_buf.resize(pg_constants::WAL_SEGMENT_SIZE, 0);
seg_buf.resize(WAL_SEGMENT_SIZE, 0);
Ok(seg_buf.freeze())
}
@@ -426,6 +441,7 @@ pub fn encode_logical_message(prefix: &str, message: &str) -> Vec<u8> {
#[cfg(test)]
mod tests {
use super::super::PG_MAJORVERSION;
use super::*;
use regex::Regex;
use std::cmp::min;
@@ -434,23 +450,26 @@ mod tests {
use utils::const_assert;
fn init_logging() {
let _ = env_logger::Builder::from_env(
env_logger::Env::default()
.default_filter_or("wal_craft=info,postgres_ffi::xlog_utils=trace"),
)
let _ = env_logger::Builder::from_env(env_logger::Env::default().default_filter_or(
format!("wal_craft=info,postgres_ffi::{PG_MAJORVERSION}::xlog_utils=trace"),
))
.is_test(true)
.try_init();
}
fn test_end_of_wal<C: wal_craft::Crafter>(test_name: &str) {
use wal_craft::*;
let pg_version = PG_MAJORVERSION[1..3].parse::<u32>().unwrap();
// Craft some WAL
let top_path = PathBuf::from(env!("CARGO_MANIFEST_DIR"))
.join("..")
.join("..");
let cfg = Conf {
pg_distrib_dir: top_path.join("tmp_install"),
datadir: top_path.join(format!("test_output/{}", test_name)),
pg_version,
pg_distrib_dir: top_path.join("pg_install"),
datadir: top_path.join(format!("test_output/{}-{PG_MAJORVERSION}", test_name)),
};
if cfg.datadir.exists() {
fs::remove_dir_all(&cfg.datadir).unwrap();

3
libs/postgres_ffi/wal_craft/Cargo.toml
View File

@@ -11,6 +11,7 @@ clap = "3.0"
env_logger = "0.9"
log = "0.4"
once_cell = "1.13.0"
postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres_ffi = { path = "../" }
tempfile = "3.2"
workspace_hack = { version = "0.1", path = "../../../workspace_hack" }

15
libs/postgres_ffi/wal_craft/src/bin/wal_craft.rs
View File

@@ -37,9 +37,16 @@ fn main() -> Result<()> {
Arg::new("pg-distrib-dir")
.long("pg-distrib-dir")
.takes_value(true)
.help("Directory with Postgres distribution (bin and lib directories, e.g. tmp_install)")
.help("Directory with Postgres distributions (bin and lib directories, e.g. pg_install containing subpath `v14/bin/postgresql`)")
.default_value("/usr/local")
)
.arg(
Arg::new("pg-version")
.long("pg-version")
.help("Postgres version to use for the initial tenant")
.required(true)
.takes_value(true)
)
)
.subcommand(
App::new("in-existing")
@@ -82,8 +89,14 @@ fn main() -> Result<()> {
}
Ok(())
}
Some(("with-initdb", arg_matches)) => {
let cfg = Conf {
pg_version: arg_matches
.value_of("pg-version")
.unwrap()
.parse::<u32>()
.context("Failed to parse postgres version from the argument string")?,
pg_distrib_dir: arg_matches.value_of("pg-distrib-dir").unwrap().into(),
datadir: arg_matches.value_of("datadir").unwrap().into(),
};

21
libs/postgres_ffi/wal_craft/src/lib.rs
View File

@@ -4,10 +4,8 @@ use log::*;
use once_cell::sync::Lazy;
use postgres::types::PgLsn;
use postgres::Client;
use postgres_ffi::v14::pg_constants::WAL_SEGMENT_SIZE;
use postgres_ffi::v14::xlog_utils::{
XLOG_BLCKSZ, XLOG_SIZE_OF_XLOG_RECORD, XLOG_SIZE_OF_XLOG_SHORT_PHD,
};
use postgres_ffi::{WAL_SEGMENT_SIZE, XLOG_BLCKSZ};
use postgres_ffi::{XLOG_SIZE_OF_XLOG_RECORD, XLOG_SIZE_OF_XLOG_SHORT_PHD};
use std::cmp::Ordering;
use std::fs;
use std::path::{Path, PathBuf};
@@ -17,6 +15,7 @@ use tempfile::{tempdir, TempDir};
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct Conf {
pub pg_version: u32,
pub pg_distrib_dir: PathBuf,
pub datadir: PathBuf,
}
@@ -38,12 +37,22 @@ pub static REQUIRED_POSTGRES_CONFIG: Lazy<Vec<&'static str>> = Lazy::new(|| {
});
impl Conf {
pub fn pg_distrib_dir(&self) -> PathBuf {
let path = self.pg_distrib_dir.clone();
match self.pg_version {
14 => path.join(format!("v{}", self.pg_version)),
15 => path.join(format!("v{}", self.pg_version)),
_ => panic!("Unsupported postgres version: {}", self.pg_version),
}
}
fn pg_bin_dir(&self) -> PathBuf {
self.pg_distrib_dir.join("bin")
self.pg_distrib_dir().join("bin")
}
fn pg_lib_dir(&self) -> PathBuf {
self.pg_distrib_dir.join("lib")
self.pg_distrib_dir().join("lib")
}
pub fn wal_dir(&self) -> PathBuf {

1
libs/remote_storage/Cargo.toml
View File

@@ -7,6 +7,7 @@ edition = "2021"
anyhow = { version = "1.0", features = ["backtrace"] }
async-trait = "0.1"
metrics = { version = "0.1", path = "../metrics" }
utils = { version = "0.1", path = "../utils" }
once_cell = "1.13.0"
rusoto_core = "0.48"
rusoto_s3 = "0.48"

221
libs/remote_storage/src/lib.rs
View File

@@ -9,13 +9,13 @@ mod local_fs;
mod s3_bucket;
use std::{
borrow::Cow,
collections::HashMap,
ffi::OsStr,
fmt::Debug,
fmt::{Debug, Display},
num::{NonZeroU32, NonZeroUsize},
ops::Deref,
path::{Path, PathBuf},
pin::Pin,
sync::Arc,
};
use anyhow::{bail, Context};
@@ -24,10 +24,7 @@ use tokio::io;
use toml_edit::Item;
use tracing::info;
pub use self::{
local_fs::LocalFs,
s3_bucket::{S3Bucket, S3ObjectKey},
};
pub use self::{local_fs::LocalFs, s3_bucket::S3Bucket};
/// How many different timelines can be processed simultaneously when synchronizing layers with the remote storage.
/// During regular work, pageserver produces one layer file per timeline checkpoint, with bursts of concurrency
@@ -42,22 +39,62 @@ pub const DEFAULT_REMOTE_STORAGE_MAX_SYNC_ERRORS: u32 = 10;
/// https://aws.amazon.com/premiumsupport/knowledge-center/s3-request-limit-avoid-throttling/
pub const DEFAULT_REMOTE_STORAGE_S3_CONCURRENCY_LIMIT: usize = 100;
const REMOTE_STORAGE_PREFIX_SEPARATOR: char = '/';
#[derive(Clone, PartialEq, Eq)]
pub struct RemoteObjectId(String);
///
/// A key that refers to an object in remote storage. It works much like a Path,
/// but it's a separate datatype so that you don't accidentally mix local paths
/// and remote keys.
///
impl RemoteObjectId {
// Needed to retrieve last component for RemoteObjectId.
// In other words a file name
/// Turn a/b/c or a/b/c/ into c
pub fn object_name(&self) -> Option<&str> {
// corner case, char::to_string is not const, thats why this is more verbose than it needs to be
// see https://github.com/rust-lang/rust/issues/88674
if self.0.len() == 1 && self.0.chars().next().unwrap() == REMOTE_STORAGE_PREFIX_SEPARATOR {
return None;
}
if self.0.ends_with(REMOTE_STORAGE_PREFIX_SEPARATOR) {
self.0.rsplit(REMOTE_STORAGE_PREFIX_SEPARATOR).nth(1)
} else {
self.0
.rsplit_once(REMOTE_STORAGE_PREFIX_SEPARATOR)
.map(|(_, last)| last)
}
}
}
impl Debug for RemoteObjectId {
fn fmt(&self, fmt: &mut std::fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
Debug::fmt(&self.0, fmt)
}
}
impl Display for RemoteObjectId {
fn fmt(&self, fmt: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
Display::fmt(&self.0, fmt)
}
}
/// Storage (potentially remote) API to manage its state.
/// This storage tries to be unaware of any layered repository context,
/// providing basic CRUD operations for storage files.
#[async_trait::async_trait]
pub trait RemoteStorage: Send + Sync {
/// A way to uniquely reference a file in the remote storage.
type RemoteObjectId;
pub trait RemoteStorage: Send + Sync + 'static {
/// Attempts to derive the storage path out of the local path, if the latter is correct.
fn remote_object_id(&self, local_path: &Path) -> anyhow::Result<Self::RemoteObjectId>;
fn remote_object_id(&self, local_path: &Path) -> anyhow::Result<RemoteObjectId>;
/// Gets the download path of the given storage file.
fn local_path(&self, remote_object_id: &Self::RemoteObjectId) -> anyhow::Result<PathBuf>;
fn local_path(&self, remote_object_id: &RemoteObjectId) -> anyhow::Result<PathBuf>;
/// Lists all items the storage has right now.
async fn list(&self) -> anyhow::Result<Vec<Self::RemoteObjectId>>;
async fn list(&self) -> anyhow::Result<Vec<RemoteObjectId>>;
/// Lists all top level subdirectories for a given prefix
/// Note: here we assume that if the prefix is passed it was obtained via remote_object_id
@@ -65,34 +102,39 @@ pub trait RemoteStorage: Send + Sync {
/// so this method doesnt need to.
async fn list_prefixes(
&self,
prefix: Option<&Self::RemoteObjectId>,
) -> anyhow::Result<Vec<Self::RemoteObjectId>>;
prefix: Option<&RemoteObjectId>,
) -> anyhow::Result<Vec<RemoteObjectId>>;
/// Streams the local file contents into remote into the remote storage entry.
async fn upload(
&self,
from: impl io::AsyncRead + Unpin + Send + Sync + 'static,
from: Box<(dyn io::AsyncRead + Unpin + Send + Sync + 'static)>,
// S3 PUT request requires the content length to be specified,
// otherwise it starts to fail with the concurrent connection count increasing.
from_size_bytes: usize,
to: &Self::RemoteObjectId,
to: &RemoteObjectId,
metadata: Option<StorageMetadata>,
) -> anyhow::Result<()>;
/// Streams the remote storage entry contents into the buffered writer given, returns the filled writer.
/// Returns the metadata, if any was stored with the file previously.
async fn download(&self, from: &Self::RemoteObjectId) -> Result<Download, DownloadError>;
async fn download(&self, from: &RemoteObjectId) -> Result<Download, DownloadError>;
/// Streams a given byte range of the remote storage entry contents into the buffered writer given, returns the filled writer.
/// Returns the metadata, if any was stored with the file previously.
async fn download_byte_range(
&self,
from: &Self::RemoteObjectId,
from: &RemoteObjectId,
start_inclusive: u64,
end_exclusive: Option<u64>,
) -> Result<Download, DownloadError>;
async fn delete(&self, path: &Self::RemoteObjectId) -> anyhow::Result<()>;
async fn delete(&self, path: &RemoteObjectId) -> anyhow::Result<()>;
/// Downcast to LocalFs implementation. For tests.
fn as_local(&self) -> Option<&LocalFs> {
None
}
}
pub struct Download {
@@ -135,33 +177,91 @@ impl std::error::Error for DownloadError {}
/// Every storage, currently supported.
/// Serves as a simple way to pass around the [`RemoteStorage`] without dealing with generics.
pub enum GenericRemoteStorage {
Local(LocalFs),
S3(S3Bucket),
#[derive(Clone)]
pub struct GenericRemoteStorage(Arc<dyn RemoteStorage>);
impl Deref for GenericRemoteStorage {
type Target = dyn RemoteStorage;
fn deref(&self) -> &Self::Target {
self.0.as_ref()
}
}
impl GenericRemoteStorage {
pub fn new(
pub fn new(storage: impl RemoteStorage) -> Self {
Self(Arc::new(storage))
}
pub fn from_config(
working_directory: PathBuf,
storage_config: &RemoteStorageConfig,
) -> anyhow::Result<Self> {
match &storage_config.storage {
) -> anyhow::Result<GenericRemoteStorage> {
Ok(match &storage_config.storage {
RemoteStorageKind::LocalFs(root) => {
info!("Using fs root '{}' as a remote storage", root.display());
LocalFs::new(root.clone(), working_directory).map(GenericRemoteStorage::Local)
GenericRemoteStorage::new(LocalFs::new(root.clone(), working_directory)?)
}
RemoteStorageKind::AwsS3(s3_config) => {
info!("Using s3 bucket '{}' in region '{}' as a remote storage, prefix in bucket: '{:?}', bucket endpoint: '{:?}'",
s3_config.bucket_name, s3_config.bucket_region, s3_config.prefix_in_bucket, s3_config.endpoint);
S3Bucket::new(s3_config, working_directory).map(GenericRemoteStorage::S3)
s3_config.bucket_name, s3_config.bucket_region, s3_config.prefix_in_bucket, s3_config.endpoint);
GenericRemoteStorage::new(S3Bucket::new(s3_config, working_directory)?)
}
}
})
}
pub fn as_local(&self) -> Option<&LocalFs> {
match self {
Self::Local(local_fs) => Some(local_fs),
_ => None,
/// Takes storage object contents and its size and uploads to remote storage,
/// mapping `from_path` to the corresponding remote object id in the storage.
///
/// The storage object does not have to be present on the `from_path`,
/// this path is used for the remote object id conversion only.
pub async fn upload_storage_object(
&self,
from: Box<dyn tokio::io::AsyncRead + Unpin + Send + Sync + 'static>,
from_size_bytes: usize,
from_path: &Path,
) -> anyhow::Result<()> {
let target_storage_path = self.remote_object_id(from_path).with_context(|| {
format!(
"Failed to get the storage path for source local path '{}'",
from_path.display()
)
})?;
self.upload(from, from_size_bytes, &target_storage_path, None)
.await
.with_context(|| {
format!(
"Failed to upload from '{}' to storage path '{:?}'",
from_path.display(),
target_storage_path
)
})
}
/// Downloads the storage object into the `to_path` provided.
/// `byte_range` could be specified to dowload only a part of the file, if needed.
pub async fn download_storage_object(
&self,
byte_range: Option<(u64, Option<u64>)>,
to_path: &Path,
) -> Result<Download, DownloadError> {
let remote_object_path = self
.remote_object_id(to_path)
.with_context(|| {
format!(
"Failed to get the storage path for target local path '{}'",
to_path.display()
)
})
.map_err(DownloadError::BadInput)?;
match byte_range {
Some((start, end)) => {
self.download_byte_range(&remote_object_path, start, end)
.await
}
None => self.download(&remote_object_path).await,
}
}
}
@@ -242,20 +342,6 @@ impl Debug for S3Config {
}
}
pub fn path_with_suffix_extension(original_path: impl AsRef<Path>, suffix: &str) -> PathBuf {
let new_extension = match original_path
.as_ref()
.extension()
.map(OsStr::to_string_lossy)
{
Some(extension) => Cow::Owned(format!("{extension}.{suffix}")),
None => Cow::Borrowed(suffix),
};
original_path
.as_ref()
.with_extension(new_extension.as_ref())
}
impl RemoteStorageConfig {
pub fn from_toml(toml: &toml_edit::Item) -> anyhow::Result<RemoteStorageConfig> {
let local_path = toml.get("local_path");
@@ -345,26 +431,21 @@ mod tests {
use super::*;
#[test]
fn test_path_with_suffix_extension() {
let p = PathBuf::from("/foo/bar");
assert_eq!(
&path_with_suffix_extension(&p, "temp").to_string_lossy(),
"/foo/bar.temp"
);
let p = PathBuf::from("/foo/bar");
assert_eq!(
&path_with_suffix_extension(&p, "temp.temp").to_string_lossy(),
"/foo/bar.temp.temp"
);
let p = PathBuf::from("/foo/bar.baz");
assert_eq!(
&path_with_suffix_extension(&p, "temp.temp").to_string_lossy(),
"/foo/bar.baz.temp.temp"
);
let p = PathBuf::from("/foo/bar.baz");
assert_eq!(
&path_with_suffix_extension(&p, ".temp").to_string_lossy(),
"/foo/bar.baz..temp"
);
fn object_name() {
let k = RemoteObjectId("a/b/c".to_owned());
assert_eq!(k.object_name(), Some("c"));
let k = RemoteObjectId("a/b/c/".to_owned());
assert_eq!(k.object_name(), Some("c"));
let k = RemoteObjectId("a/".to_owned());
assert_eq!(k.object_name(), Some("a"));
// XXX is it impossible to have an empty key?
let k = RemoteObjectId("".to_owned());
assert_eq!(k.object_name(), None);
let k = RemoteObjectId("/".to_owned());
assert_eq!(k.object_name(), None);
}
}

137
libs/remote_storage/src/local_fs.rs
View File

@@ -16,11 +16,23 @@ use tokio::{
io::{self, AsyncReadExt, AsyncSeekExt, AsyncWriteExt},
};
use tracing::*;
use utils::crashsafe_dir::path_with_suffix_extension;
use crate::{path_with_suffix_extension, Download, DownloadError};
use crate::{Download, DownloadError, RemoteObjectId};
use super::{strip_path_prefix, RemoteStorage, StorageMetadata};
const LOCAL_FS_TEMP_FILE_SUFFIX: &str = "___temp";
/// Convert a Path in the remote storage into a RemoteObjectId
fn remote_object_id_from_path(path: &Path) -> anyhow::Result<RemoteObjectId> {
Ok(RemoteObjectId(
path.to_str()
.ok_or_else(|| anyhow::anyhow!("unexpected characters found in path"))?
.to_string(),
))
}
pub struct LocalFs {
working_directory: PathBuf,
storage_root: PathBuf,
@@ -43,11 +55,17 @@ impl LocalFs {
})
}
fn resolve_in_storage(&self, path: &Path) -> anyhow::Result<PathBuf> {
///
/// Get the absolute path in the local filesystem to given remote object.
///
/// This is public so that it can be used in tests. Should not be used elsewhere.
///
pub fn resolve_in_storage(&self, remote_object_id: &RemoteObjectId) -> anyhow::Result<PathBuf> {
let path = PathBuf::from(&remote_object_id.0);
if path.is_relative() {
Ok(self.storage_root.join(path))
} else if path.starts_with(&self.storage_root) {
Ok(path.to_path_buf())
Ok(path)
} else {
bail!(
"Path '{}' does not belong to the current storage",
@@ -85,38 +103,42 @@ impl LocalFs {
#[async_trait::async_trait]
impl RemoteStorage for LocalFs {
type RemoteObjectId = PathBuf;
fn remote_object_id(&self, local_path: &Path) -> anyhow::Result<Self::RemoteObjectId> {
Ok(self.storage_root.join(
/// Convert a "local" path into a "remote path"
fn remote_object_id(&self, local_path: &Path) -> anyhow::Result<RemoteObjectId> {
let path = self.storage_root.join(
strip_path_prefix(&self.working_directory, local_path)
.context("local path does not belong to this storage")?,
))
);
remote_object_id_from_path(&path)
}
fn local_path(&self, storage_path: &Self::RemoteObjectId) -> anyhow::Result<PathBuf> {
let relative_path = strip_path_prefix(&self.storage_root, storage_path)
fn local_path(&self, remote_object_id: &RemoteObjectId) -> anyhow::Result<PathBuf> {
let storage_path = PathBuf::from(&remote_object_id.0);
let relative_path = strip_path_prefix(&self.storage_root, &storage_path)
.context("local path does not belong to this storage")?;
Ok(self.working_directory.join(relative_path))
}
async fn list(&self) -> anyhow::Result<Vec<Self::RemoteObjectId>> {
async fn list(&self) -> anyhow::Result<Vec<RemoteObjectId>> {
get_all_files(&self.storage_root, true).await
}
async fn list_prefixes(
&self,
prefix: Option<&Self::RemoteObjectId>,
) -> anyhow::Result<Vec<Self::RemoteObjectId>> {
let path = prefix.unwrap_or(&self.storage_root);
prefix: Option<&RemoteObjectId>,
) -> anyhow::Result<Vec<RemoteObjectId>> {
let path = match prefix {
Some(prefix) => Path::new(&prefix.0),
None => &self.storage_root,
};
get_all_files(path, false).await
}
async fn upload(
&self,
from: impl io::AsyncRead + Unpin + Send + Sync + 'static,
from: Box<(dyn io::AsyncRead + Unpin + Send + Sync + 'static)>,
from_size_bytes: usize,
to: &Self::RemoteObjectId,
to: &RemoteObjectId,
metadata: Option<StorageMetadata>,
) -> anyhow::Result<()> {
let target_file_path = self.resolve_in_storage(to)?;
@@ -124,7 +146,8 @@ impl RemoteStorage for LocalFs {
// We need this dance with sort of durable rename (without fsyncs)
// to prevent partial uploads. This was really hit when pageserver shutdown
// cancelled the upload and partial file was left on the fs
let temp_file_path = path_with_suffix_extension(&target_file_path, "temp");
let temp_file_path =
path_with_suffix_extension(&target_file_path, LOCAL_FS_TEMP_FILE_SUFFIX);
let mut destination = io::BufWriter::new(
fs::OpenOptions::new()
.write(true)
@@ -197,7 +220,7 @@ impl RemoteStorage for LocalFs {
Ok(())
}
async fn download(&self, from: &Self::RemoteObjectId) -> Result<Download, DownloadError> {
async fn download(&self, from: &RemoteObjectId) -> Result<Download, DownloadError> {
let file_path = self
.resolve_in_storage(from)
.map_err(DownloadError::BadInput)?;
@@ -231,7 +254,7 @@ impl RemoteStorage for LocalFs {
async fn download_byte_range(
&self,
from: &Self::RemoteObjectId,
from: &RemoteObjectId,
start_inclusive: u64,
end_exclusive: Option<u64>,
) -> Result<Download, DownloadError> {
@@ -285,7 +308,7 @@ impl RemoteStorage for LocalFs {
}
}
async fn delete(&self, path: &Self::RemoteObjectId) -> anyhow::Result<()> {
async fn delete(&self, path: &RemoteObjectId) -> anyhow::Result<()> {
let file_path = self.resolve_in_storage(path)?;
if file_path.exists() && file_path.is_file() {
Ok(fs::remove_file(file_path).await?)
@@ -296,6 +319,10 @@ impl RemoteStorage for LocalFs {
)
}
}
fn as_local(&self) -> Option<&LocalFs> {
Some(self)
}
}
fn storage_metadata_path(original_path: &Path) -> PathBuf {
@@ -305,7 +332,7 @@ fn storage_metadata_path(original_path: &Path) -> PathBuf {
fn get_all_files<'a, P>(
directory_path: P,
recursive: bool,
) -> Pin<Box<dyn Future<Output = anyhow::Result<Vec<PathBuf>>> + Send + Sync + 'a>>
) -> Pin<Box<dyn Future<Output = anyhow::Result<Vec<RemoteObjectId>>> + Send + Sync + 'a>>
where
P: AsRef<Path> + Send + Sync + 'a,
{
@@ -322,12 +349,12 @@ where
debug!("{:?} us a symlink, skipping", entry_path)
} else if file_type.is_dir() {
if recursive {
paths.extend(get_all_files(entry_path, true).await?.into_iter())
paths.extend(get_all_files(&entry_path, true).await?.into_iter())
} else {
paths.push(dir_entry.path())
paths.push(remote_object_id_from_path(&dir_entry.path())?)
}
} else {
paths.push(dir_entry.path());
paths.push(remote_object_id_from_path(&dir_entry.path())?);
}
}
Ok(paths)
@@ -389,9 +416,15 @@ mod pure_tests {
.join("file_name");
let expected_path = storage_root.join(local_path.strip_prefix(&workdir)?);
let actual_path = PathBuf::from(
storage
.remote_object_id(&local_path)
.expect("Matching path should map to storage path normally")
.0,
);
assert_eq!(
expected_path,
storage.remote_object_id(&local_path).expect("Matching path should map to storage path normally"),
actual_path,
"File paths from workdir should be stored in local fs storage with the same path they have relative to the workdir"
);
@@ -452,7 +485,9 @@ mod pure_tests {
assert_eq!(
local_path,
storage
.local_path(&storage_root.join(local_path.strip_prefix(&workdir)?))
.local_path(&remote_object_id_from_path(
&storage_root.join(local_path.strip_prefix(&workdir)?)
)?)
.expect("For a valid input, valid local path should be parsed"),
"Should be able to parse metadata out of the correctly named remote delta file"
);
@@ -476,8 +511,7 @@ mod pure_tests {
#[test]
fn local_path_negatives() -> anyhow::Result<()> {
#[track_caller]
#[allow(clippy::ptr_arg)] // have to use &PathBuf due to `storage.local_path` parameter requirements
fn local_path_error(storage: &LocalFs, storage_path: &PathBuf) -> String {
fn local_path_error(storage: &LocalFs, storage_path: &RemoteObjectId) -> String {
match storage.local_path(storage_path) {
Ok(wrong_path) => panic!(
"Expected local path input {:?} to cause an error, but got file path: {:?}",
@@ -494,7 +528,8 @@ mod pure_tests {
};
let totally_wrong_path = "wrong_wrong_wrong";
let error_message = local_path_error(&storage, &PathBuf::from(totally_wrong_path));
let error_message =
local_path_error(&storage, &RemoteObjectId(totally_wrong_path.to_string()));
assert!(error_message.contains(totally_wrong_path));
Ok(())
@@ -537,7 +572,7 @@ mod fs_tests {
storage: &LocalFs,
#[allow(clippy::ptr_arg)]
// have to use &PathBuf due to `storage.local_path` parameter requirements
remote_storage_path: &PathBuf,
remote_storage_path: &RemoteObjectId,
expected_metadata: Option<&StorageMetadata>,
) -> anyhow::Result<String> {
let mut download = storage
@@ -568,12 +603,20 @@ mod fs_tests {
"whatever_contents",
)
.await?;
let target_path = PathBuf::from("/").join("somewhere").join("else");
match storage.upload(file, size, &target_path, None).await {
let target_path = "/somewhere/else";
match storage
.upload(
Box::new(file),
size,
&RemoteObjectId(target_path.to_string()),
None,
)
.await
{
Ok(()) => panic!("Should not allow storing files with wrong target path"),
Err(e) => {
let message = format!("{:?}", e);
assert!(message.contains(&target_path.display().to_string()));
assert!(message.contains(target_path));
assert!(message.contains("does not belong to the current storage"));
}
}
@@ -606,20 +649,20 @@ mod fs_tests {
// Check that you get an error if the size parameter doesn't match the actual
// size of the stream.
storage
.upload(content.clone(), 0, &id, None)
.upload(Box::new(content.clone()), 0, &id, None)
.await
.expect_err("upload with zero size succeeded");
storage
.upload(content.clone(), 4, &id, None)
.upload(Box::new(content.clone()), 4, &id, None)
.await
.expect_err("upload with too short size succeeded");
storage
.upload(content.clone(), 6, &id, None)
.upload(Box::new(content.clone()), 6, &id, None)
.await
.expect_err("upload with too large size succeeded");
// Correct size is 5, this should succeed.
storage.upload(content, 5, &id, None).await?;
storage.upload(Box::new(content), 5, &id, None).await?;
Ok(())
}
@@ -643,8 +686,8 @@ mod fs_tests {
"We should upload and download the same contents"
);
let non_existing_path = PathBuf::from("somewhere").join("else");
match storage.download(&non_existing_path).await {
let non_existing_path = "somewhere/else";
match storage.download(&RemoteObjectId(non_existing_path.to_string())).await {
Err(DownloadError::NotFound) => {} // Should get NotFound for non existing keys
other => panic!("Should get a NotFound error when downloading non-existing storage files, but got: {other:?}"),
}
@@ -783,7 +826,7 @@ mod fs_tests {
Err(e) => {
let error_string = e.to_string();
assert!(error_string.contains("does not exist"));
assert!(error_string.contains(&upload_target.display().to_string()));
assert!(error_string.contains(&upload_target.0));
}
}
Ok(())
@@ -844,15 +887,19 @@ mod fs_tests {
storage: &LocalFs,
name: &str,
metadata: Option<StorageMetadata>,
) -> anyhow::Result<PathBuf> {
) -> anyhow::Result<RemoteObjectId> {
let timeline_path = workdir.join("timelines").join("some_timeline");
let relative_timeline_path = timeline_path.strip_prefix(&workdir)?;
let storage_path = storage.storage_root.join(relative_timeline_path).join(name);
let remote_object_id = RemoteObjectId(storage_path.to_str().unwrap().to_string());
let from_path = storage.working_directory.join(name);
let (file, size) = create_file_for_upload(&from_path, &dummy_contents(name)).await?;
storage.upload(file, size, &storage_path, metadata).await?;
Ok(storage_path)
storage
.upload(Box::new(file), size, &remote_object_id, metadata)
.await?;
remote_object_id_from_path(&storage_path)
}
async fn create_file_for_upload(
@@ -877,9 +924,9 @@ mod fs_tests {
format!("contents for {name}")
}
async fn list_files_sorted(storage: &LocalFs) -> anyhow::Result<Vec<PathBuf>> {
async fn list_files_sorted(storage: &LocalFs) -> anyhow::Result<Vec<RemoteObjectId>> {
let mut files = storage.list().await?;
files.sort();
files.sort_by(|a, b| a.0.cmp(&b.0));
Ok(files)
}
}

164
libs/remote_storage/src/s3_bucket.rs
View File

@@ -19,7 +19,10 @@ use tokio::{io, sync::Semaphore};
use tokio_util::io::ReaderStream;
use tracing::debug;
use crate::{strip_path_prefix, Download, DownloadError, RemoteStorage, S3Config};
use crate::{
strip_path_prefix, Download, DownloadError, RemoteObjectId, RemoteStorage, S3Config,
REMOTE_STORAGE_PREFIX_SEPARATOR,
};
use super::StorageMetadata;
@@ -88,50 +91,26 @@ pub(super) mod metrics {
}
}
const S3_PREFIX_SEPARATOR: char = '/';
fn download_destination(
id: &RemoteObjectId,
workdir: &Path,
prefix_to_strip: Option<&str>,
) -> PathBuf {
let path_without_prefix = match prefix_to_strip {
Some(prefix) => id.0.strip_prefix(prefix).unwrap_or_else(|| {
panic!(
"Could not strip prefix '{}' from S3 object key '{}'",
prefix, id.0
)
}),
None => &id.0,
};
#[derive(Debug, Eq, PartialEq, PartialOrd, Ord, Hash)]
pub struct S3ObjectKey(String);
impl S3ObjectKey {
/// Turn a/b/c or a/b/c/ into c
pub fn object_name(&self) -> Option<&str> {
// corner case, char::to_string is not const, thats why this is more verbose than it needs to be
// see https://github.com/rust-lang/rust/issues/88674
if self.0.len() == 1 && self.0.chars().next().unwrap() == S3_PREFIX_SEPARATOR {
return None;
}
if self.0.ends_with(S3_PREFIX_SEPARATOR) {
self.0.rsplit(S3_PREFIX_SEPARATOR).nth(1)
} else {
self.0
.rsplit_once(S3_PREFIX_SEPARATOR)
.map(|(_, last)| last)
}
}
fn key(&self) -> &str {
&self.0
}
fn download_destination(&self, workdir: &Path, prefix_to_strip: Option<&str>) -> PathBuf {
let path_without_prefix = match prefix_to_strip {
Some(prefix) => self.0.strip_prefix(prefix).unwrap_or_else(|| {
panic!(
"Could not strip prefix '{}' from S3 object key '{}'",
prefix, self.0
)
}),
None => &self.0,
};
workdir.join(
path_without_prefix
.split(S3_PREFIX_SEPARATOR)
.collect::<PathBuf>(),
)
}
workdir.join(
path_without_prefix
.split(REMOTE_STORAGE_PREFIX_SEPARATOR)
.collect::<PathBuf>(),
)
}
/// AWS S3 storage.
@@ -193,12 +172,12 @@ impl S3Bucket {
let prefix_in_bucket = aws_config.prefix_in_bucket.as_deref().map(|prefix| {
let mut prefix = prefix;
while prefix.starts_with(S3_PREFIX_SEPARATOR) {
while prefix.starts_with(REMOTE_STORAGE_PREFIX_SEPARATOR) {
prefix = &prefix[1..]
}
let mut prefix = prefix.to_string();
while prefix.ends_with(S3_PREFIX_SEPARATOR) {
while prefix.ends_with(REMOTE_STORAGE_PREFIX_SEPARATOR) {
prefix.pop();
}
prefix
@@ -249,23 +228,25 @@ impl S3Bucket {
#[async_trait::async_trait]
impl RemoteStorage for S3Bucket {
type RemoteObjectId = S3ObjectKey;
fn remote_object_id(&self, local_path: &Path) -> anyhow::Result<Self::RemoteObjectId> {
fn remote_object_id(&self, local_path: &Path) -> anyhow::Result<RemoteObjectId> {
let relative_path = strip_path_prefix(&self.workdir, local_path)?;
let mut key = self.prefix_in_bucket.clone().unwrap_or_default();
for segment in relative_path {
key.push(S3_PREFIX_SEPARATOR);
key.push(REMOTE_STORAGE_PREFIX_SEPARATOR);
key.push_str(&segment.to_string_lossy());
}
Ok(S3ObjectKey(key))
Ok(RemoteObjectId(key))
}
fn local_path(&self, storage_path: &Self::RemoteObjectId) -> anyhow::Result<PathBuf> {
Ok(storage_path.download_destination(&self.workdir, self.prefix_in_bucket.as_deref()))
fn local_path(&self, storage_path: &RemoteObjectId) -> anyhow::Result<PathBuf> {
Ok(download_destination(
storage_path,
&self.workdir,
self.prefix_in_bucket.as_deref(),
))
}
async fn list(&self) -> anyhow::Result<Vec<Self::RemoteObjectId>> {
async fn list(&self) -> anyhow::Result<Vec<RemoteObjectId>> {
let mut document_keys = Vec::new();
let mut continuation_token = None;
@@ -296,7 +277,7 @@ impl RemoteStorage for S3Bucket {
.contents
.unwrap_or_default()
.into_iter()
.filter_map(|o| Some(S3ObjectKey(o.key?))),
.filter_map(|o| Some(RemoteObjectId(o.key?))),
);
match fetch_response.continuation_token {
@@ -312,8 +293,8 @@ impl RemoteStorage for S3Bucket {
/// Note: it wont include empty "directories"
async fn list_prefixes(
&self,
prefix: Option<&Self::RemoteObjectId>,
) -> anyhow::Result<Vec<Self::RemoteObjectId>> {
prefix: Option<&RemoteObjectId>,
) -> anyhow::Result<Vec<RemoteObjectId>> {
// get the passed prefix or if it is not set use prefix_in_bucket value
let list_prefix = prefix
.map(|p| p.0.clone())
@@ -321,8 +302,8 @@ impl RemoteStorage for S3Bucket {
.map(|mut p| {
// required to end with a separator
// otherwise request will return only the entry of a prefix
if !p.ends_with(S3_PREFIX_SEPARATOR) {
p.push(S3_PREFIX_SEPARATOR);
if !p.ends_with(REMOTE_STORAGE_PREFIX_SEPARATOR) {
p.push(REMOTE_STORAGE_PREFIX_SEPARATOR);
}
p
});
@@ -345,7 +326,7 @@ impl RemoteStorage for S3Bucket {
bucket: self.bucket_name.clone(),
prefix: list_prefix.clone(),
continuation_token,
delimiter: Some(S3_PREFIX_SEPARATOR.to_string()),
delimiter: Some(REMOTE_STORAGE_PREFIX_SEPARATOR.to_string()),
..ListObjectsV2Request::default()
})
.await
@@ -359,7 +340,7 @@ impl RemoteStorage for S3Bucket {
.common_prefixes
.unwrap_or_default()
.into_iter()
.filter_map(|o| Some(S3ObjectKey(o.prefix?))),
.filter_map(|o| Some(RemoteObjectId(o.prefix?))),
);
match fetch_response.continuation_token {
@@ -373,9 +354,9 @@ impl RemoteStorage for S3Bucket {
async fn upload(
&self,
from: impl io::AsyncRead + Unpin + Send + Sync + 'static,
from: Box<(dyn io::AsyncRead + Unpin + Send + Sync + 'static)>,
from_size_bytes: usize,
to: &Self::RemoteObjectId,
to: &RemoteObjectId,
metadata: Option<StorageMetadata>,
) -> anyhow::Result<()> {
let _guard = self
@@ -392,7 +373,7 @@ impl RemoteStorage for S3Bucket {
from_size_bytes,
)),
bucket: self.bucket_name.clone(),
key: to.key().to_owned(),
key: to.0.to_owned(),
metadata: metadata.map(|m| m.0),
..PutObjectRequest::default()
})
@@ -404,10 +385,10 @@ impl RemoteStorage for S3Bucket {
Ok(())
}
async fn download(&self, from: &Self::RemoteObjectId) -> Result<Download, DownloadError> {
async fn download(&self, from: &RemoteObjectId) -> Result<Download, DownloadError> {
self.download_object(GetObjectRequest {
bucket: self.bucket_name.clone(),
key: from.key().to_owned(),
key: from.0.to_owned(),
..GetObjectRequest::default()
})
.await
@@ -415,7 +396,7 @@ impl RemoteStorage for S3Bucket {
async fn download_byte_range(
&self,
from: &Self::RemoteObjectId,
from: &RemoteObjectId,
start_inclusive: u64,
end_exclusive: Option<u64>,
) -> Result<Download, DownloadError> {
@@ -429,14 +410,14 @@ impl RemoteStorage for S3Bucket {
self.download_object(GetObjectRequest {
bucket: self.bucket_name.clone(),
key: from.key().to_owned(),
key: from.0.to_owned(),
range,
..GetObjectRequest::default()
})
.await
}
async fn delete(&self, path: &Self::RemoteObjectId) -> anyhow::Result<()> {
async fn delete(&self, remote_object_id: &RemoteObjectId) -> anyhow::Result<()> {
let _guard = self
.concurrency_limiter
.acquire()
@@ -448,7 +429,7 @@ impl RemoteStorage for S3Bucket {
self.client
.delete_object(DeleteObjectRequest {
bucket: self.bucket_name.clone(),
key: path.key().to_owned(),
key: remote_object_id.0.to_owned(),
..DeleteObjectRequest::default()
})
.await
@@ -467,43 +448,24 @@ mod tests {
use super::*;
#[test]
fn object_name() {
let k = S3ObjectKey("a/b/c".to_owned());
assert_eq!(k.object_name(), Some("c"));
let k = S3ObjectKey("a/b/c/".to_owned());
assert_eq!(k.object_name(), Some("c"));
let k = S3ObjectKey("a/".to_owned());
assert_eq!(k.object_name(), Some("a"));
// XXX is it impossible to have an empty key?
let k = S3ObjectKey("".to_owned());
assert_eq!(k.object_name(), None);
let k = S3ObjectKey("/".to_owned());
assert_eq!(k.object_name(), None);
}
#[test]
fn download_destination() -> anyhow::Result<()> {
fn test_download_destination() -> anyhow::Result<()> {
let workdir = tempdir()?.path().to_owned();
let local_path = workdir.join("one").join("two").join("test_name");
let relative_path = local_path.strip_prefix(&workdir)?;
let key = S3ObjectKey(format!(
let key = RemoteObjectId(format!(
"{}{}",
S3_PREFIX_SEPARATOR,
REMOTE_STORAGE_PREFIX_SEPARATOR,
relative_path
.iter()
.map(|segment| segment.to_str().unwrap())
.collect::<Vec<_>>()
.join(&S3_PREFIX_SEPARATOR.to_string()),
.join(&REMOTE_STORAGE_PREFIX_SEPARATOR.to_string()),
));
assert_eq!(
local_path,
key.download_destination(&workdir, None),
download_destination(&key, &workdir, None),
"Download destination should consist of s3 path joined with the workdir prefix"
);
@@ -520,8 +482,8 @@ mod tests {
let storage = dummy_storage(workdir);
let expected_key = S3ObjectKey(format!(
"{}{S3_PREFIX_SEPARATOR}{segment_1}{S3_PREFIX_SEPARATOR}{segment_2}",
let expected_key = RemoteObjectId(format!(
"{}{REMOTE_STORAGE_PREFIX_SEPARATOR}{segment_1}{REMOTE_STORAGE_PREFIX_SEPARATOR}{segment_2}",
storage.prefix_in_bucket.as_deref().unwrap_or_default(),
));
@@ -592,7 +554,7 @@ mod tests {
storage.prefix_in_bucket.as_deref(),
);
assert_eq!(
s3_key.download_destination(&workdir, storage.prefix_in_bucket.as_deref()),
download_destination(&s3_key, &workdir, storage.prefix_in_bucket.as_deref()),
storage
.local_path(&s3_key)
.expect("For a valid input, valid S3 info should be parsed"),
@@ -604,7 +566,7 @@ mod tests {
storage.prefix_in_bucket.as_deref(),
);
assert_eq!(
s3_key.download_destination(&workdir, storage.prefix_in_bucket.as_deref()),
download_destination(&s3_key, &workdir, storage.prefix_in_bucket.as_deref()),
storage
.local_path(&s3_key)
.expect("For a valid input, valid S3 info should be parsed"),
@@ -645,11 +607,11 @@ mod tests {
}
}
fn create_s3_key(relative_file_path: &Path, prefix: Option<&str>) -> S3ObjectKey {
S3ObjectKey(relative_file_path.iter().fold(
fn create_s3_key(relative_file_path: &Path, prefix: Option<&str>) -> RemoteObjectId {
RemoteObjectId(relative_file_path.iter().fold(
prefix.unwrap_or_default().to_string(),
|mut path_string, segment| {
path_string.push(S3_PREFIX_SEPARATOR);
path_string.push(REMOTE_STORAGE_PREFIX_SEPARATOR);
path_string.push_str(segment.to_str().unwrap());
path_string
},

12
libs/safekeeper_api/Cargo.toml Normal file
View File

@@ -0,0 +1,12 @@
[package]
name = "safekeeper_api"
version = "0.1.0"
edition = "2021"
[dependencies]
serde = { version = "1.0", features = ["derive"] }
serde_with = "1.12.0"
const_format = "0.2.21"
utils = { path = "../utils" }
workspace_hack = { version = "0.1", path = "../../workspace_hack" }

10
libs/safekeeper_api/src/lib.rs Normal file
View File

@@ -0,0 +1,10 @@
use const_format::formatcp;
/// Public API types
pub mod models;
pub const DEFAULT_PG_LISTEN_PORT: u16 = 5454;
pub const DEFAULT_PG_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_PG_LISTEN_PORT}");
pub const DEFAULT_HTTP_LISTEN_PORT: u16 = 7676;
pub const DEFAULT_HTTP_LISTEN_ADDR: &str = formatcp!("127.0.0.1:{DEFAULT_HTTP_LISTEN_PORT}");

4
safekeeper/src/http/models.rs → libs/safekeeper_api/src/models.rs
View File

@@ -1,8 +1,8 @@
use serde::{Deserialize, Serialize};
use utils::zid::{NodeId, ZTimelineId};
use utils::id::{NodeId, TimelineId};
#[derive(Serialize, Deserialize)]
pub struct TimelineCreateRequest {
pub timeline_id: ZTimelineId,
pub timeline_id: TimelineId,
pub peer_ids: Vec<NodeId>,
}

6
libs/utils/Cargo.toml
View File

@@ -4,18 +4,20 @@ version = "0.1.0"
edition = "2021"
[dependencies]
async-trait = "0.1"
anyhow = "1.0"
bincode = "1.3"
bytes = "1.0.1"
hyper = { version = "0.14.7", features = ["full"] }
pin-project-lite = "0.2.7"
postgres = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres-protocol = { git = "https://github.com/zenithdb/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
postgres-protocol = { git = "https://github.com/neondatabase/rust-postgres.git", rev="d052ee8b86fff9897c77b0fe89ea9daba0e1fa38" }
routerify = "3"
serde = { version = "1.0", features = ["derive"] }
serde_json = "1"
thiserror = "1.0"
tokio = { version = "1.17", features = ["macros"]}
tokio-rustls = "0.23"
tracing = "0.1"
tracing-subscriber = { version = "0.3", features = ["env-filter"] }
nix = "0.23.0"

14
libs/utils/benches/benchmarks.rs
View File

@@ -1,22 +1,22 @@
#![allow(unused)]
use criterion::{criterion_group, criterion_main, Criterion};
use utils::zid;
use utils::id;
pub fn bench_zid_stringify(c: &mut Criterion) {
pub fn bench_id_stringify(c: &mut Criterion) {
// Can only use public methods.
let ztl = zid::ZTenantTimelineId::generate();
let ttid = id::TenantTimelineId::generate();
c.bench_function("zid.to_string", |b| {
c.bench_function("id.to_string", |b| {
b.iter(|| {
// FIXME measurement overhead?
//for _ in 0..1000 {
// ztl.tenant_id.to_string();
// ttid.tenant_id.to_string();
//}
ztl.tenant_id.to_string();
ttid.tenant_id.to_string();
})
});
}
criterion_group!(benches, bench_zid_stringify);
criterion_group!(benches, bench_id_stringify);
criterion_main!(benches);

14
libs/utils/src/auth.rs
View File

@@ -14,7 +14,7 @@ use jsonwebtoken::{
use serde::{Deserialize, Serialize};
use serde_with::{serde_as, DisplayFromStr};
use crate::zid::ZTenantId;
use crate::id::TenantId;
const JWT_ALGORITHM: Algorithm = Algorithm::RS256;
@@ -30,23 +30,23 @@ pub enum Scope {
pub struct Claims {
#[serde(default)]
#[serde_as(as = "Option<DisplayFromStr>")]
pub tenant_id: Option<ZTenantId>,
pub tenant_id: Option<TenantId>,
pub scope: Scope,
}
impl Claims {
pub fn new(tenant_id: Option<ZTenantId>, scope: Scope) -> Self {
pub fn new(tenant_id: Option<TenantId>, scope: Scope) -> Self {
Self { tenant_id, scope }
}
}
pub fn check_permission(claims: &Claims, tenantid: Option<ZTenantId>) -> Result<()> {
match (&claims.scope, tenantid) {
pub fn check_permission(claims: &Claims, tenant_id: Option<TenantId>) -> Result<()> {
match (&claims.scope, tenant_id) {
(Scope::Tenant, None) => {
bail!("Attempt to access management api with tenant scope. Permission denied")
}
(Scope::Tenant, Some(tenantid)) => {
if claims.tenant_id.unwrap() != tenantid {
(Scope::Tenant, Some(tenant_id)) => {
if claims.tenant_id.unwrap() != tenant_id {
bail!("Tenant id mismatch. Permission denied")
}
Ok(())

49
libs/utils/src/crashsafe_dir.rs
View File

@@ -1,7 +1,9 @@
use std::{
borrow::Cow,
ffi::OsStr,
fs::{self, File},
io,
path::Path,
path::{Path, PathBuf},
};
/// Similar to [`std::fs::create_dir`], except we fsync the
@@ -74,6 +76,22 @@ pub fn create_dir_all(path: impl AsRef<Path>) -> io::Result<()> {
Ok(())
}
/// Adds a suffix to the file(directory) name, either appending the suffix to the end of its extension,
/// or if there's no extension, creates one and puts a suffix there.
pub fn path_with_suffix_extension(original_path: impl AsRef<Path>, suffix: &str) -> PathBuf {
let new_extension = match original_path
.as_ref()
.extension()
.map(OsStr::to_string_lossy)
{
Some(extension) => Cow::Owned(format!("{extension}.{suffix}")),
None => Cow::Borrowed(suffix),
};
original_path
.as_ref()
.with_extension(new_extension.as_ref())
}
#[cfg(test)]
mod tests {
use tempfile::tempdir;
@@ -122,4 +140,33 @@ mod tests {
let invalid_dir_path = file_path.join("folder");
create_dir_all(&invalid_dir_path).unwrap_err();
}
#[test]
fn test_path_with_suffix_extension() {
let p = PathBuf::from("/foo/bar");
assert_eq!(
&path_with_suffix_extension(&p, "temp").to_string_lossy(),
"/foo/bar.temp"
);
let p = PathBuf::from("/foo/bar");
assert_eq!(
&path_with_suffix_extension(&p, "temp.temp").to_string_lossy(),
"/foo/bar.temp.temp"
);
let p = PathBuf::from("/foo/bar.baz");
assert_eq!(
&path_with_suffix_extension(&p, "temp.temp").to_string_lossy(),
"/foo/bar.baz.temp.temp"
);
let p = PathBuf::from("/foo/bar.baz");
assert_eq!(
&path_with_suffix_extension(&p, ".temp").to_string_lossy(),
"/foo/bar.baz..temp"
);
let p = PathBuf::from("/foo/bar/dir/");
assert_eq!(
&path_with_suffix_extension(&p, ".temp").to_string_lossy(),
"/foo/bar/dir..temp"
);
}
}

15
libs/utils/src/http/endpoint.rs
View File

@@ -1,6 +1,6 @@
use crate::auth::{self, Claims, JwtAuth};
use crate::http::error;
use crate::zid::ZTenantId;
use crate::id::TenantId;
use anyhow::anyhow;
use hyper::header::AUTHORIZATION;
use hyper::{header::CONTENT_TYPE, Body, Request, Response, Server};
@@ -9,6 +9,7 @@ use once_cell::sync::Lazy;
use routerify::ext::RequestExt;
use routerify::RequestInfo;
use routerify::{Middleware, Router, RouterBuilder, RouterService};
use tokio::task::JoinError;
use tracing::info;
use std::future::Future;
@@ -35,7 +36,13 @@ async fn prometheus_metrics_handler(_req: Request<Body>) -> Result<Response<Body
let mut buffer = vec![];
let encoder = TextEncoder::new();
let metrics = metrics::gather();
let metrics = tokio::task::spawn_blocking(move || {
// Currently we take a lot of mutexes while collecting metrics, so it's
// better to spawn a blocking task to avoid blocking the event loop.
metrics::gather()
})
.await
.map_err(|e: JoinError| ApiError::InternalServerError(e.into()))?;
encoder.encode(&metrics, &mut buffer).unwrap();
let response = Response::builder()
@@ -137,9 +144,9 @@ pub fn auth_middleware<B: hyper::body::HttpBody + Send + Sync + 'static>(
})
}
pub fn check_permission(req: &Request<Body>, tenantid: Option<ZTenantId>) -> Result<(), ApiError> {
pub fn check_permission(req: &Request<Body>, tenant_id: Option<TenantId>) -> Result<(), ApiError> {
match req.context::<Claims>() {
Some(claims) => Ok(auth::check_permission(&claims, tenantid)
Some(claims) => Ok(auth::check_permission(&claims, tenant_id)
.map_err(|err| ApiError::Forbidden(err.to_string()))?),
None => Ok(()), // claims is None because auth is disabled
}

17
libs/utils/src/http/error.rs
View File

@@ -1,12 +1,11 @@
use anyhow::anyhow;
use hyper::{header, Body, Response, StatusCode};
use serde::{Deserialize, Serialize};
use thiserror::Error;
#[derive(Debug, Error)]
pub enum ApiError {
#[error("Bad request: {0}")]
BadRequest(String),
#[error("Bad request: {0:#?}")]
BadRequest(anyhow::Error),
#[error("Forbidden: {0}")]
Forbidden(String),
@@ -15,24 +14,20 @@ pub enum ApiError {
Unauthorized(String),
#[error("NotFound: {0}")]
NotFound(String),
NotFound(anyhow::Error),
#[error("Conflict: {0}")]
Conflict(String),
#[error(transparent)]
InternalServerError(#[from] anyhow::Error),
InternalServerError(anyhow::Error),
}
impl ApiError {
pub fn from_err<E: Into<anyhow::Error>>(err: E) -> Self {
Self::InternalServerError(anyhow!(err))
}
pub fn into_response(self) -> Response<Body> {
match self {
ApiError::BadRequest(_) => HttpErrorBody::response_from_msg_and_status(
self.to_string(),
ApiError::BadRequest(err) => HttpErrorBody::response_from_msg_and_status(
format!("{err:#?}"), // use debug printing so that we give the cause
StatusCode::BAD_REQUEST,
),
ApiError::Forbidden(_) => {

13
libs/utils/src/http/json.rs
View File

@@ -1,3 +1,4 @@
use anyhow::Context;
use bytes::Buf;
use hyper::{header, Body, Request, Response, StatusCode};
use serde::{Deserialize, Serialize};
@@ -9,20 +10,24 @@ pub async fn json_request<T: for<'de> Deserialize<'de>>(
) -> Result<T, ApiError> {
let whole_body = hyper::body::aggregate(request.body_mut())
.await
.map_err(ApiError::from_err)?;
.context("Failed to read request body")
.map_err(ApiError::BadRequest)?;
serde_json::from_reader(whole_body.reader())
.map_err(|err| ApiError::BadRequest(format!("Failed to parse json request {}", err)))
.context("Failed to parse json request")
.map_err(ApiError::BadRequest)
}
pub fn json_response<T: Serialize>(
status: StatusCode,
data: T,
) -> Result<Response<Body>, ApiError> {
let json = serde_json::to_string(&data).map_err(ApiError::from_err)?;
let json = serde_json::to_string(&data)
.context("Failed to serialize JSON response")
.map_err(ApiError::InternalServerError)?;
let response = Response::builder()
.status(status)
.header(header::CONTENT_TYPE, "application/json")
.body(Body::from(json))
.map_err(ApiError::from_err)?;
.map_err(|e| ApiError::InternalServerError(e.into()))?;
Ok(response)
}

2
libs/utils/src/http/mod.rs
View File

@@ -3,6 +3,6 @@ pub mod error;
pub mod json;
pub mod request;
/// Current fast way to apply simple http routing in various Zenith binaries.
/// Current fast way to apply simple http routing in various Neon binaries.
/// Re-exported for sake of uniform approach, that could be later replaced with better alternatives, if needed.
pub use routerify::{ext::RequestExt, RouterBuilder, RouterService};

13
libs/utils/src/http/request.rs
View File

@@ -1,6 +1,7 @@
use std::str::FromStr;
use super::error::ApiError;
use anyhow::anyhow;
use hyper::{body::HttpBody, Body, Request};
use routerify::ext::RequestExt;
@@ -10,9 +11,8 @@ pub fn get_request_param<'a>(
) -> Result<&'a str, ApiError> {
match request.param(param_name) {
Some(arg) => Ok(arg),
None => Err(ApiError::BadRequest(format!(
"no {} specified in path param",
param_name
None => Err(ApiError::BadRequest(anyhow!(
"no {param_name} specified in path param",
))),
}
}
@@ -23,16 +23,15 @@ pub fn parse_request_param<T: FromStr>(
) -> Result<T, ApiError> {
match get_request_param(request, param_name)?.parse() {
Ok(v) => Ok(v),
Err(_) => Err(ApiError::BadRequest(format!(
"failed to parse {}",
param_name
Err(_) => Err(ApiError::BadRequest(anyhow!(
"failed to parse {param_name}",
))),
}
}
pub async fn ensure_no_body(request: &mut Request<Body>) -> Result<(), ApiError> {
match request.body_mut().data().await {
Some(_) => Err(ApiError::BadRequest("Unexpected request body".into())),
Some(_) => Err(ApiError::BadRequest(anyhow!("Unexpected request body"))),
None => Ok(()),
}
}

88
libs/utils/src/zid.rs → libs/utils/src/id.rs
View File

@@ -4,7 +4,7 @@ use hex::FromHex;
use rand::Rng;
use serde::{Deserialize, Serialize};
/// Zenith ID is a 128-bit random ID.
/// Neon ID is a 128-bit random ID.
/// Used to represent various identifiers. Provides handy utility methods and impls.
///
/// NOTE: It (de)serializes as an array of hex bytes, so the string representation would look
@@ -13,13 +13,13 @@ use serde::{Deserialize, Serialize};
/// Use `#[serde_as(as = "DisplayFromStr")]` to (de)serialize it as hex string instead: `ad50847381e248feaac9876cc71ae418`.
/// Check the `serde_with::serde_as` documentation for options for more complex types.
#[derive(Clone, Copy, PartialEq, Eq, Hash, Serialize, Deserialize, PartialOrd, Ord)]
struct ZId([u8; 16]);
struct Id([u8; 16]);
impl ZId {
pub fn get_from_buf(buf: &mut dyn bytes::Buf) -> ZId {
impl Id {
pub fn get_from_buf(buf: &mut dyn bytes::Buf) -> Id {
let mut arr = [0u8; 16];
buf.copy_to_slice(&mut arr);
ZId::from(arr)
Id::from(arr)
}
pub fn as_arr(&self) -> [u8; 16] {
@@ -29,7 +29,7 @@ impl ZId {
pub fn generate() -> Self {
let mut tli_buf = [0u8; 16];
rand::thread_rng().fill(&mut tli_buf);
ZId::from(tli_buf)
Id::from(tli_buf)
}
fn hex_encode(&self) -> String {
@@ -44,54 +44,54 @@ impl ZId {
}
}
impl FromStr for ZId {
impl FromStr for Id {
type Err = hex::FromHexError;
fn from_str(s: &str) -> Result<ZId, Self::Err> {
fn from_str(s: &str) -> Result<Id, Self::Err> {
Self::from_hex(s)
}
}
// this is needed for pretty serialization and deserialization of ZId's using serde integration with hex crate
impl FromHex for ZId {
// this is needed for pretty serialization and deserialization of Id's using serde integration with hex crate
impl FromHex for Id {
type Error = hex::FromHexError;
fn from_hex<T: AsRef<[u8]>>(hex: T) -> Result<Self, Self::Error> {
let mut buf: [u8; 16] = [0u8; 16];
hex::decode_to_slice(hex, &mut buf)?;
Ok(ZId(buf))
Ok(Id(buf))
}
}
impl AsRef<[u8]> for ZId {
impl AsRef<[u8]> for Id {
fn as_ref(&self) -> &[u8] {
&self.0
}
}
impl From<[u8; 16]> for ZId {
impl From<[u8; 16]> for Id {
fn from(b: [u8; 16]) -> Self {
ZId(b)
Id(b)
}
}
impl fmt::Display for ZId {
impl fmt::Display for Id {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(&self.hex_encode())
}
}
impl fmt::Debug for ZId {
impl fmt::Debug for Id {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.write_str(&self.hex_encode())
}
}
macro_rules! zid_newtype {
macro_rules! id_newtype {
($t:ident) => {
impl $t {
pub fn get_from_buf(buf: &mut dyn bytes::Buf) -> $t {
$t(ZId::get_from_buf(buf))
$t(Id::get_from_buf(buf))
}
pub fn as_arr(&self) -> [u8; 16] {
@@ -99,11 +99,11 @@ macro_rules! zid_newtype {
}
pub fn generate() -> Self {
$t(ZId::generate())
$t(Id::generate())
}
pub const fn from_array(b: [u8; 16]) -> Self {
$t(ZId(b))
$t(Id(b))
}
}
@@ -111,14 +111,14 @@ macro_rules! zid_newtype {
type Err = hex::FromHexError;
fn from_str(s: &str) -> Result<$t, Self::Err> {
let value = ZId::from_str(s)?;
let value = Id::from_str(s)?;
Ok($t(value))
}
}
impl From<[u8; 16]> for $t {
fn from(b: [u8; 16]) -> Self {
$t(ZId::from(b))
$t(Id::from(b))
}
}
@@ -126,7 +126,7 @@ macro_rules! zid_newtype {
type Error = hex::FromHexError;
fn from_hex<T: AsRef<[u8]>>(hex: T) -> Result<Self, Self::Error> {
Ok($t(ZId::from_hex(hex)?))
Ok($t(Id::from_hex(hex)?))
}
}
@@ -150,7 +150,7 @@ macro_rules! zid_newtype {
};
}
/// Zenith timeline IDs are different from PostgreSQL timeline
/// Neon timeline IDs are different from PostgreSQL timeline
/// IDs. They serve a similar purpose though: they differentiate
/// between different "histories" of the same cluster. However,
/// PostgreSQL timeline IDs are a bit cumbersome, because they are only
@@ -158,7 +158,7 @@ macro_rules! zid_newtype {
/// timeline history. Those limitations mean that we cannot generate a
/// new PostgreSQL timeline ID by just generating a random number. And
/// that in turn is problematic for the "pull/push" workflow, where you
/// have a local copy of a zenith repository, and you periodically sync
/// have a local copy of a Neon repository, and you periodically sync
/// the local changes with a remote server. When you work "detached"
/// from the remote server, you cannot create a PostgreSQL timeline ID
/// that's guaranteed to be different from all existing timelines in
@@ -168,55 +168,55 @@ macro_rules! zid_newtype {
/// branches? If they pick the same one, and later try to push the
/// branches to the same remote server, they will get mixed up.
///
/// To avoid those issues, Zenith has its own concept of timelines that
/// To avoid those issues, Neon has its own concept of timelines that
/// is separate from PostgreSQL timelines, and doesn't have those
/// limitations. A zenith timeline is identified by a 128-bit ID, which
/// limitations. A Neon timeline is identified by a 128-bit ID, which
/// is usually printed out as a hex string.
///
/// NOTE: It (de)serializes as an array of hex bytes, so the string representation would look
/// like `[173,80,132,115,129,226,72,254,170,201,135,108,199,26,228,24]`.
/// See [`ZId`] for alternative ways to serialize it.
/// See [`Id`] for alternative ways to serialize it.
#[derive(Clone, Copy, PartialEq, Eq, Hash, Ord, PartialOrd, Serialize, Deserialize)]
pub struct ZTimelineId(ZId);
pub struct TimelineId(Id);
zid_newtype!(ZTimelineId);
id_newtype!(TimelineId);
/// Zenith Tenant Id represents identifiar of a particular tenant.
/// Neon Tenant Id represents identifiar of a particular tenant.
/// Is used for distinguishing requests and data belonging to different users.
///
/// NOTE: It (de)serializes as an array of hex bytes, so the string representation would look
/// like `[173,80,132,115,129,226,72,254,170,201,135,108,199,26,228,24]`.
/// See [`ZId`] for alternative ways to serialize it.
/// See [`Id`] for alternative ways to serialize it.
#[derive(Clone, Copy, PartialEq, Eq, Hash, Serialize, Deserialize, PartialOrd, Ord)]
pub struct ZTenantId(ZId);
pub struct TenantId(Id);
zid_newtype!(ZTenantId);
id_newtype!(TenantId);
// A pair uniquely identifying Zenith instance.
// A pair uniquely identifying Neon instance.
#[derive(Debug, Clone, Copy, PartialOrd, Ord, PartialEq, Eq, Hash, Serialize, Deserialize)]
pub struct ZTenantTimelineId {
pub tenant_id: ZTenantId,
pub timeline_id: ZTimelineId,
pub struct TenantTimelineId {
pub tenant_id: TenantId,
pub timeline_id: TimelineId,
}
impl ZTenantTimelineId {
pub fn new(tenant_id: ZTenantId, timeline_id: ZTimelineId) -> Self {
ZTenantTimelineId {
impl TenantTimelineId {
pub fn new(tenant_id: TenantId, timeline_id: TimelineId) -> Self {
TenantTimelineId {
tenant_id,
timeline_id,
}
}
pub fn generate() -> Self {
Self::new(ZTenantId::generate(), ZTimelineId::generate())
Self::new(TenantId::generate(), TimelineId::generate())
}
pub fn empty() -> Self {
Self::new(ZTenantId::from([0u8; 16]), ZTimelineId::from([0u8; 16]))
Self::new(TenantId::from([0u8; 16]), TimelineId::from([0u8; 16]))
}
}
impl fmt::Display for ZTenantTimelineId {
impl fmt::Display for TenantTimelineId {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
write!(f, "{}/{}", self.tenant_id, self.timeline_id)
}

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