Compare commits

..

8 Commits

Author SHA1 Message Date
Conrad Ludgate
9041907019 rewrite with custom json serializer 2025-05-18 13:41:43 +02:00
Conrad Ludgate
53fdcd252f remove locking from extract, use refcell instead 2025-05-17 22:14:26 +02:00
Conrad Ludgate
f5c5b99b58 remove lasso from json logger, use field index for lookup 2025-05-17 22:14:26 +02:00
Conrad Ludgate
ac331090bf refactor json logging state 2025-05-17 22:14:26 +02:00
Conrad Ludgate
176b5a8978 replace indexset with lasso and linear search 2025-05-17 22:14:26 +02:00
Conrad Ludgate
e0da7dd8e9 use faster hasher than siphash 2025-05-17 22:14:26 +02:00
Conrad Ludgate
547fe38abf replace papaya with hashmap+lock. this assumes that spans are rarely accessed in parallel 2025-05-17 22:14:26 +02:00
Conrad Ludgate
c06f9635f5 remove tracing instrument on passthrough 2025-05-17 22:14:26 +02:00
484 changed files with 10749 additions and 36626 deletions

View File

@@ -38,11 +38,6 @@ on:
required: false
default: 1
type: number
rerun-failed:
description: 'rerun failed tests to ignore flaky tests'
required: false
default: true
type: boolean
defaults:
run:
@@ -104,10 +99,11 @@ jobs:
# Set some environment variables used by all the steps.
#
# CARGO_FLAGS is extra options to pass to all "cargo" subcommands.
# CARGO_FLAGS is extra options to pass to "cargo build", "cargo test" etc.
# It also includes --features, if any
#
# CARGO_PROFILE is passed to "cargo build", "cargo test" etc, but not to
# "cargo metadata", because it doesn't accept --release or --debug options.
# 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.
@@ -116,16 +112,16 @@ jobs:
ARCH: ${{ inputs.arch }}
SANITIZERS: ${{ inputs.sanitizers }}
run: |
CARGO_FLAGS="--locked --features testing"
CARGO_FEATURES="--features testing"
if [[ $BUILD_TYPE == "debug" && $ARCH == 'x64' ]]; then
cov_prefix="scripts/coverage --profraw-prefix=$GITHUB_JOB --dir=/tmp/coverage run"
CARGO_PROFILE=""
CARGO_FLAGS="--locked"
elif [[ $BUILD_TYPE == "debug" ]]; then
cov_prefix=""
CARGO_PROFILE=""
CARGO_FLAGS="--locked"
elif [[ $BUILD_TYPE == "release" ]]; then
cov_prefix=""
CARGO_PROFILE="--release"
CARGO_FLAGS="--locked --release"
fi
if [[ $SANITIZERS == 'enabled' ]]; then
make_vars="WITH_SANITIZERS=yes"
@@ -135,8 +131,8 @@ jobs:
{
echo "cov_prefix=${cov_prefix}"
echo "make_vars=${make_vars}"
echo "CARGO_FEATURES=${CARGO_FEATURES}"
echo "CARGO_FLAGS=${CARGO_FLAGS}"
echo "CARGO_PROFILE=${CARGO_PROFILE}"
echo "CARGO_HOME=${GITHUB_WORKSPACE}/.cargo"
} >> $GITHUB_ENV
@@ -188,18 +184,34 @@ jobs:
path: pg_install/v17
key: v1-${{ runner.os }}-${{ runner.arch }}-${{ inputs.build-type }}-pg-${{ steps.pg_v17_rev.outputs.pg_rev }}-bookworm-${{ hashFiles('Makefile', 'build-tools.Dockerfile') }}
- name: Build all
# Note: the Makefile picks up BUILD_TYPE and CARGO_PROFILE from the env variables
run: mold -run make ${make_vars} all -j$(nproc) CARGO_BUILD_FLAGS="$CARGO_FLAGS"
- name: Build postgres v14
if: steps.cache_pg_14.outputs.cache-hit != 'true'
run: mold -run make ${make_vars} postgres-v14 -j$(nproc)
- name: Build postgres v15
if: steps.cache_pg_15.outputs.cache-hit != 'true'
run: mold -run make ${make_vars} postgres-v15 -j$(nproc)
- name: Build postgres v16
if: steps.cache_pg_16.outputs.cache-hit != 'true'
run: mold -run make ${make_vars} postgres-v16 -j$(nproc)
- name: Build postgres v17
if: steps.cache_pg_17.outputs.cache-hit != 'true'
run: mold -run make ${make_vars} postgres-v17 -j$(nproc)
- name: Build neon extensions
run: mold -run make ${make_vars} neon-pg-ext -j$(nproc)
- name: Build walproposer-lib
run: mold -run make ${make_vars} walproposer-lib -j$(nproc)
- name: Build unit tests
if: inputs.sanitizers != 'enabled'
- name: Run cargo build
env:
WITH_TESTS: ${{ inputs.sanitizers != 'enabled' && '--tests' || '' }}
run: |
export ASAN_OPTIONS=detect_leaks=0
${cov_prefix} mold -run cargo build $CARGO_FLAGS $CARGO_PROFILE --tests
${cov_prefix} mold -run cargo build $CARGO_FLAGS $CARGO_FEATURES --bins ${WITH_TESTS}
# Do install *before* running rust tests because they might recompile the
# binaries with different features/flags.
@@ -211,7 +223,7 @@ jobs:
# Install target binaries
mkdir -p /tmp/neon/bin/
binaries=$(
${cov_prefix} cargo metadata $CARGO_FLAGS --format-version=1 --no-deps |
${cov_prefix} cargo metadata $CARGO_FEATURES --format-version=1 --no-deps |
jq -r '.packages[].targets[] | select(.kind | index("bin")) | .name'
)
for bin in $binaries; do
@@ -228,7 +240,7 @@ jobs:
mkdir -p /tmp/neon/test_bin/
test_exe_paths=$(
${cov_prefix} cargo test $CARGO_FLAGS $CARGO_PROFILE --message-format=json --no-run |
${cov_prefix} cargo test $CARGO_FLAGS $CARGO_FEATURES --message-format=json --no-run |
jq -r '.executable | select(. != null)'
)
for bin in $test_exe_paths; do
@@ -262,10 +274,10 @@ jobs:
export LD_LIBRARY_PATH
#nextest does not yet support running doctests
${cov_prefix} cargo test --doc $CARGO_FLAGS $CARGO_PROFILE
${cov_prefix} cargo test --doc $CARGO_FLAGS $CARGO_FEATURES
# run all non-pageserver tests
${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_PROFILE -E '!package(pageserver)'
${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E '!package(pageserver)'
# run pageserver tests
# (When developing new pageserver features gated by config fields, we commonly make the rust
@@ -274,13 +286,13 @@ jobs:
# pageserver tests from non-pageserver tests cuts down the time it takes for this CI step.)
NEON_PAGESERVER_UNIT_TEST_VIRTUAL_FILE_IOENGINE=tokio-epoll-uring \
${cov_prefix} \
cargo nextest run $CARGO_FLAGS $CARGO_PROFILE -E 'package(pageserver)'
cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E 'package(pageserver)'
# Run separate tests for real S3
export ENABLE_REAL_S3_REMOTE_STORAGE=nonempty
export REMOTE_STORAGE_S3_BUCKET=neon-github-ci-tests
export REMOTE_STORAGE_S3_REGION=eu-central-1
${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_PROFILE -E 'package(remote_storage)' -E 'test(test_real_s3)'
${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E 'package(remote_storage)' -E 'test(test_real_s3)'
# Run separate tests for real Azure Blob Storage
# XXX: replace region with `eu-central-1`-like region
@@ -289,17 +301,17 @@ jobs:
export AZURE_STORAGE_ACCESS_KEY="${{ secrets.AZURE_STORAGE_ACCESS_KEY_DEV }}"
export REMOTE_STORAGE_AZURE_CONTAINER="${{ vars.REMOTE_STORAGE_AZURE_CONTAINER }}"
export REMOTE_STORAGE_AZURE_REGION="${{ vars.REMOTE_STORAGE_AZURE_REGION }}"
${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_PROFILE -E 'package(remote_storage)' -E 'test(test_real_azure)'
${cov_prefix} cargo nextest run $CARGO_FLAGS $CARGO_FEATURES -E 'package(remote_storage)' -E 'test(test_real_azure)'
- name: Install postgres binaries
run: |
# Use tar to copy files matching the pattern, preserving the paths in the destionation
tar c \
pg_install/v* \
build/*/src/test/regress/*.so \
build/*/src/test/regress/pg_regress \
build/*/src/test/isolation/isolationtester \
build/*/src/test/isolation/pg_isolation_regress \
pg_install/build/*/src/test/regress/*.so \
pg_install/build/*/src/test/regress/pg_regress \
pg_install/build/*/src/test/isolation/isolationtester \
pg_install/build/*/src/test/isolation/pg_isolation_regress \
| tar x -C /tmp/neon
- name: Upload Neon artifact
@@ -367,7 +379,7 @@ jobs:
- name: Pytest regression tests
continue-on-error: ${{ matrix.lfc_state == 'with-lfc' && inputs.build-type == 'debug' }}
uses: ./.github/actions/run-python-test-set
timeout-minutes: ${{ (inputs.build-type == 'release' && inputs.sanitizers != 'enabled') && 75 || 180 }}
timeout-minutes: ${{ inputs.sanitizers != 'enabled' && 75 || 180 }}
with:
build_type: ${{ inputs.build-type }}
test_selection: regress
@@ -375,14 +387,14 @@ jobs:
run_with_real_s3: true
real_s3_bucket: neon-github-ci-tests
real_s3_region: eu-central-1
rerun_failed: ${{ inputs.rerun-failed }}
rerun_failed: ${{ inputs.test-run-count == 1 }}
pg_version: ${{ matrix.pg_version }}
sanitizers: ${{ inputs.sanitizers }}
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
# `--session-timeout` is equal to (timeout-minutes - 10 minutes) * 60 seconds.
# Attempt to stop tests gracefully to generate test reports
# until they are forcibly stopped by the stricter `timeout-minutes` limit.
extra_params: --session-timeout=${{ (inputs.build-type == 'release' && inputs.sanitizers != 'enabled') && 3000 || 10200 }} --count=${{ inputs.test-run-count }}
extra_params: --session-timeout=${{ inputs.sanitizers != 'enabled' && 3000 || 10200 }} --count=${{ inputs.test-run-count }}
${{ inputs.test-selection != '' && format('-k "{0}"', inputs.test-selection) || '' }}
env:
TEST_RESULT_CONNSTR: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}

View File

@@ -110,7 +110,7 @@ jobs:
build-walproposer-lib:
if: |
contains(inputs.pg_versions, 'v17') || inputs.rebuild_everything ||
inputs.pg_versions != '[]' || inputs.rebuild_everything ||
contains(github.event.pull_request.labels.*.name, 'run-extra-build-macos') ||
contains(github.event.pull_request.labels.*.name, 'run-extra-build-*') ||
github.ref_name == 'main'
@@ -144,7 +144,7 @@ jobs:
id: cache_walproposer_lib
uses: actions/cache@5a3ec84eff668545956fd18022155c47e93e2684 # v4.2.3
with:
path: build/walproposer-lib
path: pg_install/build/walproposer-lib
key: v1-${{ runner.os }}-${{ runner.arch }}-${{ env.BUILD_TYPE }}-walproposer_lib-v17-${{ steps.pg_rev.outputs.pg_rev }}-${{ hashFiles('Makefile') }}
- name: Checkout submodule vendor/postgres-v17
@@ -169,11 +169,11 @@ jobs:
run:
make walproposer-lib -j$(sysctl -n hw.ncpu)
- name: Upload "build/walproposer-lib" artifact
- name: Upload "pg_install/build/walproposer-lib" artifact
uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
with:
name: build--walproposer-lib
path: build/walproposer-lib
name: pg_install--build--walproposer-lib
path: pg_install/build/walproposer-lib
# The artifact is supposed to be used by the next job in the same workflow,
# so theres no need to store it for too long.
retention-days: 1
@@ -226,11 +226,11 @@ jobs:
name: pg_install--v17
path: pg_install/v17
- name: Download "build/walproposer-lib" artifact
- name: Download "pg_install/build/walproposer-lib" artifact
uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
with:
name: build--walproposer-lib
path: build/walproposer-lib
name: pg_install--build--walproposer-lib
path: pg_install/build/walproposer-lib
# `actions/download-artifact` doesn't preserve permissions:
# https://github.com/actions/download-artifact?tab=readme-ov-file#permission-loss

View File

@@ -58,7 +58,6 @@ jobs:
test-cfg: ${{ inputs.pg-versions }}
test-selection: ${{ inputs.test-selection }}
test-run-count: ${{ fromJson(inputs.run-count) }}
rerun-failed: false
secrets: inherit
create-test-report:

View File

@@ -199,28 +199,6 @@ jobs:
build-tools-image: ${{ needs.build-build-tools-image.outputs.image }}-bookworm
secrets: inherit
validate-compute-manifest:
runs-on: ubuntu-22.04
needs: [ meta, check-permissions ]
# We do need to run this in `.*-rc-pr` because of hotfixes.
if: ${{ contains(fromJSON('["pr", "push-main", "storage-rc-pr", "proxy-rc-pr", "compute-rc-pr"]'), needs.meta.outputs.run-kind) }}
steps:
- name: Harden the runner (Audit all outbound calls)
uses: step-security/harden-runner@4d991eb9b905ef189e4c376166672c3f2f230481 # v2.11.0
with:
egress-policy: audit
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- name: Set up Node.js
uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
with:
node-version: '24'
- name: Validate manifest against schema
run: |
make -C compute manifest-schema-validation
build-and-test-locally:
needs: [ meta, build-build-tools-image ]
# We do need to run this in `.*-rc-pr` because of hotfixes.
@@ -336,8 +314,7 @@ jobs:
test_selection: performance
run_in_parallel: false
save_perf_report: ${{ github.ref_name == 'main' }}
# test_pageserver_max_throughput_getpage_at_latest_lsn is run in separate workflow periodic_pagebench.yml because it needs snapshots
extra_params: --splits 5 --group ${{ matrix.pytest_split_group }} --ignore=test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py
extra_params: --splits 5 --group ${{ matrix.pytest_split_group }}
benchmark_durations: ${{ needs.get-benchmarks-durations.outputs.json }}
pg_version: v16
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
@@ -670,7 +647,7 @@ jobs:
ghcr.io/neondatabase/neon:${{ needs.meta.outputs.build-tag }}-bookworm-arm64
compute-node-image-arch:
needs: [ check-permissions, meta ]
needs: [ check-permissions, build-build-tools-image, meta ]
if: ${{ contains(fromJSON('["push-main", "pr", "compute-rc-pr"]'), needs.meta.outputs.run-kind) }}
permissions:
id-token: write # aws-actions/configure-aws-credentials
@@ -743,6 +720,7 @@ jobs:
GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
PG_VERSION=${{ matrix.version.pg }}
BUILD_TAG=${{ needs.meta.outputs.release-tag || needs.meta.outputs.build-tag }}
TAG=${{ needs.build-build-tools-image.outputs.image-tag }}-${{ matrix.version.debian }}
DEBIAN_VERSION=${{ matrix.version.debian }}
provenance: false
push: true
@@ -762,6 +740,7 @@ jobs:
GIT_VERSION=${{ github.event.pull_request.head.sha || github.sha }}
PG_VERSION=${{ matrix.version.pg }}
BUILD_TAG=${{ needs.meta.outputs.release-tag || needs.meta.outputs.build-tag }}
TAG=${{ needs.build-build-tools-image.outputs.image-tag }}-${{ matrix.version.debian }}
DEBIAN_VERSION=${{ matrix.version.debian }}
provenance: false
push: true

View File

@@ -1,151 +0,0 @@
name: Build and Test Fully
on:
schedule:
# * is a special character in YAML so you have to quote this string
# ┌───────────── minute (0 - 59)
# │ ┌───────────── hour (0 - 23)
# │ │ ┌───────────── day of the month (1 - 31)
# │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
# │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
- cron: '0 3 * * *' # run once a day, timezone is utc
workflow_dispatch:
defaults:
run:
shell: bash -euxo pipefail {0}
concurrency:
# Allow only one workflow per any non-`main` branch.
group: ${{ github.workflow }}-${{ github.ref_name }}-${{ github.ref_name == 'main' && github.sha || 'anysha' }}
cancel-in-progress: true
env:
RUST_BACKTRACE: 1
COPT: '-Werror'
jobs:
tag:
runs-on: [ self-hosted, small ]
container: ${{ vars.NEON_DEV_AWS_ACCOUNT_ID }}.dkr.ecr.${{ vars.AWS_ECR_REGION }}.amazonaws.com/base:pinned
outputs:
build-tag: ${{steps.build-tag.outputs.tag}}
steps:
# Need `fetch-depth: 0` to count the number of commits in the branch
- name: Harden the runner (Audit all outbound calls)
uses: step-security/harden-runner@4d991eb9b905ef189e4c376166672c3f2f230481 # v2.11.0
with:
egress-policy: audit
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
with:
fetch-depth: 0
- name: Get build tag
run: |
echo run:$GITHUB_RUN_ID
echo ref:$GITHUB_REF_NAME
echo rev:$(git rev-list --count HEAD)
if [[ "$GITHUB_REF_NAME" == "main" ]]; then
echo "tag=$(git rev-list --count HEAD)" >> $GITHUB_OUTPUT
elif [[ "$GITHUB_REF_NAME" == "release" ]]; then
echo "tag=release-$(git rev-list --count HEAD)" >> $GITHUB_OUTPUT
elif [[ "$GITHUB_REF_NAME" == "release-proxy" ]]; then
echo "tag=release-proxy-$(git rev-list --count HEAD)" >> $GITHUB_OUTPUT
elif [[ "$GITHUB_REF_NAME" == "release-compute" ]]; then
echo "tag=release-compute-$(git rev-list --count HEAD)" >> $GITHUB_OUTPUT
else
echo "GITHUB_REF_NAME (value '$GITHUB_REF_NAME') is not set to either 'main' or 'release', 'release-proxy', 'release-compute'"
echo "tag=$GITHUB_RUN_ID" >> $GITHUB_OUTPUT
fi
shell: bash
id: build-tag
build-build-tools-image:
uses: ./.github/workflows/build-build-tools-image.yml
secrets: inherit
build-and-test-locally:
needs: [ tag, build-build-tools-image ]
strategy:
fail-fast: false
matrix:
arch: [ x64, arm64 ]
build-type: [ debug, release ]
uses: ./.github/workflows/_build-and-test-locally.yml
with:
arch: ${{ matrix.arch }}
build-tools-image: ${{ needs.build-build-tools-image.outputs.image }}-bookworm
build-tag: ${{ needs.tag.outputs.build-tag }}
build-type: ${{ matrix.build-type }}
rerun-failed: false
test-cfg: '[{"pg_version":"v14", "lfc_state": "with-lfc"},
{"pg_version":"v15", "lfc_state": "with-lfc"},
{"pg_version":"v16", "lfc_state": "with-lfc"},
{"pg_version":"v17", "lfc_state": "with-lfc"},
{"pg_version":"v14", "lfc_state": "without-lfc"},
{"pg_version":"v15", "lfc_state": "without-lfc"},
{"pg_version":"v16", "lfc_state": "without-lfc"},
{"pg_version":"v17", "lfc_state": "withouts-lfc"}]'
secrets: inherit
create-test-report:
needs: [ build-and-test-locally, build-build-tools-image ]
if: ${{ !cancelled() }}
permissions:
id-token: write # aws-actions/configure-aws-credentials
statuses: write
contents: write
pull-requests: write
outputs:
report-url: ${{ steps.create-allure-report.outputs.report-url }}
runs-on: [ self-hosted, small ]
container:
image: ${{ needs.build-build-tools-image.outputs.image }}-bookworm
credentials:
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
options: --init
steps:
- name: Harden the runner (Audit all outbound calls)
uses: step-security/harden-runner@4d991eb9b905ef189e4c376166672c3f2f230481 # v2.11.0
with:
egress-policy: audit
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- name: Create Allure report
if: ${{ !cancelled() }}
id: create-allure-report
uses: ./.github/actions/allure-report-generate
with:
store-test-results-into-db: true
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
env:
REGRESS_TEST_RESULT_CONNSTR_NEW: ${{ secrets.REGRESS_TEST_RESULT_CONNSTR_NEW }}
- uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
if: ${{ !cancelled() }}
with:
# Retry script for 5XX server errors: https://github.com/actions/github-script#retries
retries: 5
script: |
const report = {
reportUrl: "${{ steps.create-allure-report.outputs.report-url }}",
reportJsonUrl: "${{ steps.create-allure-report.outputs.report-json-url }}",
}
const coverage = {}
const script = require("./scripts/comment-test-report.js")
await script({
github,
context,
fetch,
report,
coverage,
})

View File

@@ -79,7 +79,6 @@ jobs:
build-tools-image: ${{ needs.build-build-tools-image.outputs.image }}-bookworm
build-tag: ${{ needs.tag.outputs.build-tag }}
build-type: ${{ matrix.build-type }}
rerun-failed: false
test-cfg: '[{"pg_version":"v17"}]'
sanitizers: enabled
secrets: inherit

View File

@@ -33,19 +33,11 @@ jobs:
fail-fast: false # allow other variants to continue even if one fails
matrix:
include:
# test only read-only custom scripts in new branch without database maintenance
- target: new_branch
custom_scripts: select_any_webhook_with_skew.sql@300 select_recent_webhook.sql@397 select_prefetch_webhook.sql@3
test_maintenance: false
# test all custom scripts in new branch with database maintenance
- target: new_branch
custom_scripts: insert_webhooks.sql@200 select_any_webhook_with_skew.sql@300 select_recent_webhook.sql@397 select_prefetch_webhook.sql@3 IUD_one_transaction.sql@100
test_maintenance: true
# test all custom scripts in reuse branch with database maintenance
- target: reuse_branch
custom_scripts: insert_webhooks.sql@200 select_any_webhook_with_skew.sql@300 select_recent_webhook.sql@397 select_prefetch_webhook.sql@3 IUD_one_transaction.sql@100
test_maintenance: true
max-parallel: 1 # we want to run each benchmark sequentially to not have noisy neighbors on shared storage (PS, SK)
max-parallel: 1 # we want to run each stripe size sequentially to be able to compare the results
permissions:
contents: write
statuses: write
@@ -153,7 +145,6 @@ jobs:
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
- name: Benchmark database maintenance
if: ${{ matrix.test_maintenance == 'true' }}
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ env.BUILD_TYPE }}

View File

@@ -1,175 +0,0 @@
name: large oltp growth
# workflow to grow the reuse branch of large oltp benchmark continuously (about 16 GB per run)
on:
# uncomment to run on push for debugging your PR
# push:
# branches: [ bodobolero/increase_large_oltp_workload ]
schedule:
# * is a special character in YAML so you have to quote this string
# ┌───────────── minute (0 - 59)
# │ ┌───────────── hour (0 - 23)
# │ │ ┌───────────── day of the month (1 - 31)
# │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
# │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
- cron: '0 6 * * *' # 06:00 UTC
- cron: '0 8 * * *' # 08:00 UTC
- cron: '0 10 * * *' # 10:00 UTC
- cron: '0 12 * * *' # 12:00 UTC
- cron: '0 14 * * *' # 14:00 UTC
- cron: '0 16 * * *' # 16:00 UTC
workflow_dispatch: # adds ability to run this manually
defaults:
run:
shell: bash -euxo pipefail {0}
concurrency:
# Allow only one workflow globally because we need dedicated resources which only exist once
group: large-oltp-growth
cancel-in-progress: true
permissions:
contents: read
jobs:
oltp:
strategy:
fail-fast: false # allow other variants to continue even if one fails
matrix:
include:
# for now only grow the reuse branch, not the other branches.
- target: reuse_branch
custom_scripts:
- grow_action_blocks.sql
- grow_action_kwargs.sql
- grow_device_fingerprint_event.sql
- grow_edges.sql
- grow_hotel_rate_mapping.sql
- grow_ocr_pipeline_results_version.sql
- grow_priceline_raw_response.sql
- grow_relabled_transactions.sql
- grow_state_values.sql
- grow_values.sql
- grow_vertices.sql
- update_accounting_coding_body_tracking_category_selection.sql
- update_action_blocks.sql
- update_action_kwargs.sql
- update_denormalized_approval_workflow.sql
- update_device_fingerprint_event.sql
- update_edges.sql
- update_heron_transaction_enriched_log.sql
- update_heron_transaction_enrichment_requests.sql
- update_hotel_rate_mapping.sql
- update_incoming_webhooks.sql
- update_manual_transaction.sql
- update_ml_receipt_matching_log.sql
- update_ocr_pipeine_results_version.sql
- update_orc_pipeline_step_results.sql
- update_orc_pipeline_step_results_version.sql
- update_priceline_raw_response.sql
- update_quickbooks_transactions.sql
- update_raw_finicity_transaction.sql
- update_relabeled_transactions.sql
- update_state_values.sql
- update_stripe_authorization_event_log.sql
- update_transaction.sql
- update_values.sql
- update_vertices.sql
max-parallel: 1 # we want to run each growth workload sequentially (for now there is just one)
permissions:
contents: write
statuses: write
id-token: write # aws-actions/configure-aws-credentials
env:
TEST_PG_BENCH_DURATIONS_MATRIX: "1h"
TEST_PGBENCH_CUSTOM_SCRIPTS: ${{ join(matrix.custom_scripts, ' ') }}
POSTGRES_DISTRIB_DIR: /tmp/neon/pg_install
PG_VERSION: 16 # pre-determined by pre-determined project
TEST_OUTPUT: /tmp/test_output
BUILD_TYPE: remote
PLATFORM: ${{ matrix.target }}
runs-on: [ self-hosted, us-east-2, x64 ]
container:
image: ghcr.io/neondatabase/build-tools:pinned-bookworm
credentials:
username: ${{ github.actor }}
password: ${{ secrets.GITHUB_TOKEN }}
options: --init
steps:
- name: Harden the runner (Audit all outbound calls)
uses: step-security/harden-runner@4d991eb9b905ef189e4c376166672c3f2f230481 # v2.11.0
with:
egress-policy: audit
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
- name: Configure AWS credentials # necessary to download artefacts
uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2
with:
aws-region: eu-central-1
role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
role-duration-seconds: 18000 # 5 hours is currently max associated with IAM role
- name: Download Neon artifact
uses: ./.github/actions/download
with:
name: neon-${{ runner.os }}-${{ runner.arch }}-release-artifact
path: /tmp/neon/
prefix: latest
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
- name: Set up Connection String
id: set-up-connstr
run: |
case "${{ matrix.target }}" in
reuse_branch)
CONNSTR=${{ secrets.BENCHMARK_LARGE_OLTP_REUSE_CONNSTR }}
;;
*)
echo >&2 "Unknown target=${{ matrix.target }}"
exit 1
;;
esac
CONNSTR_WITHOUT_POOLER="${CONNSTR//-pooler/}"
echo "connstr=${CONNSTR}" >> $GITHUB_OUTPUT
echo "connstr_without_pooler=${CONNSTR_WITHOUT_POOLER}" >> $GITHUB_OUTPUT
- name: pgbench with custom-scripts
uses: ./.github/actions/run-python-test-set
with:
build_type: ${{ env.BUILD_TYPE }}
test_selection: performance
run_in_parallel: false
save_perf_report: true
extra_params: -m remote_cluster --timeout 7200 -k test_perf_oltp_large_tenant_growth
pg_version: ${{ env.PG_VERSION }}
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
env:
BENCHMARK_CONNSTR: ${{ steps.set-up-connstr.outputs.connstr }}
VIP_VAP_ACCESS_TOKEN: "${{ secrets.VIP_VAP_ACCESS_TOKEN }}"
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
- name: Create Allure report
id: create-allure-report
if: ${{ !cancelled() }}
uses: ./.github/actions/allure-report-generate
with:
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
- name: Post to a Slack channel
if: ${{ github.event.schedule && failure() }}
uses: slackapi/slack-github-action@fcfb566f8b0aab22203f066d80ca1d7e4b5d05b3 # v1.27.1
with:
channel-id: "C06KHQVQ7U3" # on-call-qa-staging-stream
slack-message: |
Periodic large oltp tenant growth increase: ${{ job.status }}
<${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}|GitHub Run>
<${{ steps.create-allure-report.outputs.report-url }}|Allure report>
env:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}

View File

@@ -1,4 +1,4 @@
name: Periodic pagebench performance test on unit-perf hetzner runner
name: Periodic pagebench performance test on dedicated EC2 machine in eu-central-1 region
on:
schedule:
@@ -8,7 +8,7 @@ on:
# │ │ ┌───────────── day of the month (1 - 31)
# │ │ │ ┌───────────── month (1 - 12 or JAN-DEC)
# │ │ │ │ ┌───────────── day of the week (0 - 6 or SUN-SAT)
- cron: '0 */4 * * *' # Runs every 4 hours
- cron: '0 */3 * * *' # Runs every 3 hours
workflow_dispatch: # Allows manual triggering of the workflow
inputs:
commit_hash:
@@ -16,11 +16,6 @@ on:
description: 'The long neon repo commit hash for the system under test (pageserver) to be tested.'
required: false
default: ''
recreate_snapshots:
type: boolean
description: 'Recreate snapshots - !!!WARNING!!! We should only recreate snapshots if the previous ones are no longer compatible. Otherwise benchmarking results are not comparable across runs.'
required: false
default: false
defaults:
run:
@@ -34,13 +29,13 @@ permissions:
contents: read
jobs:
run_periodic_pagebench_test:
trigger_bench_on_ec2_machine_in_eu_central_1:
permissions:
id-token: write # aws-actions/configure-aws-credentials
statuses: write
contents: write
pull-requests: write
runs-on: [ self-hosted, unit-perf ]
runs-on: [ self-hosted, small ]
container:
image: ghcr.io/neondatabase/build-tools:pinned-bookworm
credentials:
@@ -49,13 +44,10 @@ jobs:
options: --init
timeout-minutes: 360 # Set the timeout to 6 hours
env:
API_KEY: ${{ secrets.PERIODIC_PAGEBENCH_EC2_RUNNER_API_KEY }}
RUN_ID: ${{ github.run_id }}
DEFAULT_PG_VERSION: 16
BUILD_TYPE: release
RUST_BACKTRACE: 1
# NEON_ENV_BUILDER_USE_OVERLAYFS_FOR_SNAPSHOTS: 1 - doesn't work without root in container
S3_BUCKET: neon-github-public-dev
PERF_TEST_RESULT_CONNSTR: "${{ secrets.PERF_TEST_RESULT_CONNSTR }}"
AWS_DEFAULT_REGION : "eu-central-1"
AWS_INSTANCE_ID : "i-02a59a3bf86bc7e74"
steps:
# we don't need the neon source code because we run everything remotely
# however we still need the local github actions to run the allure step below
@@ -64,194 +56,99 @@ jobs:
with:
egress-policy: audit
- name: Set up the environment which depends on $RUNNER_TEMP on nvme drive
id: set-env
shell: bash -euxo pipefail {0}
run: |
{
echo "NEON_DIR=${RUNNER_TEMP}/neon"
echo "NEON_BIN=${RUNNER_TEMP}/neon/bin"
echo "POSTGRES_DISTRIB_DIR=${RUNNER_TEMP}/neon/pg_install"
echo "LD_LIBRARY_PATH=${RUNNER_TEMP}/neon/pg_install/v${DEFAULT_PG_VERSION}/lib"
echo "BACKUP_DIR=${RUNNER_TEMP}/instance_store/saved_snapshots"
echo "TEST_OUTPUT=${RUNNER_TEMP}/neon/test_output"
echo "PERF_REPORT_DIR=${RUNNER_TEMP}/neon/test_output/perf-report-local"
echo "ALLURE_DIR=${RUNNER_TEMP}/neon/test_output/allure-results"
echo "ALLURE_RESULTS_DIR=${RUNNER_TEMP}/neon/test_output/allure-results/results"
} >> "$GITHUB_ENV"
- uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
echo "allure_results_dir=${RUNNER_TEMP}/neon/test_output/allure-results/results" >> "$GITHUB_OUTPUT"
- name: Show my own (github runner) external IP address - usefull for IP allowlisting
run: curl https://ifconfig.me
- uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2
- name: Assume AWS OIDC role that allows to manage (start/stop/describe... EC machine)
uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2
with:
aws-region: eu-central-1
role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
role-duration-seconds: 18000 # max 5 hours (needed in case commit hash is still being built)
role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_MANAGE_BENCHMARK_EC2_VMS_ARN }}
role-duration-seconds: 3600
- name: Start EC2 instance and wait for the instance to boot up
run: |
aws ec2 start-instances --instance-ids $AWS_INSTANCE_ID
aws ec2 wait instance-running --instance-ids $AWS_INSTANCE_ID
sleep 60 # sleep some time to allow cloudinit and our API server to start up
- name: Determine public IP of the EC2 instance and set env variable EC2_MACHINE_URL_US
run: |
public_ip=$(aws ec2 describe-instances --instance-ids $AWS_INSTANCE_ID --query 'Reservations[*].Instances[*].PublicIpAddress' --output text)
echo "Public IP of the EC2 instance: $public_ip"
echo "EC2_MACHINE_URL_US=https://${public_ip}:8443" >> $GITHUB_ENV
- name: Determine commit hash
id: commit_hash
shell: bash -euxo pipefail {0}
env:
INPUT_COMMIT_HASH: ${{ github.event.inputs.commit_hash }}
run: |
if [[ -z "${INPUT_COMMIT_HASH}" ]]; then
COMMIT_HASH=$(curl -s https://api.github.com/repos/neondatabase/neon/commits/main | jq -r '.sha')
echo "COMMIT_HASH=$COMMIT_HASH" >> $GITHUB_ENV
echo "commit_hash=$COMMIT_HASH" >> "$GITHUB_OUTPUT"
if [ -z "$INPUT_COMMIT_HASH" ]; then
echo "COMMIT_HASH=$(curl -s https://api.github.com/repos/neondatabase/neon/commits/main | jq -r '.sha')" >> $GITHUB_ENV
echo "COMMIT_HASH_TYPE=latest" >> $GITHUB_ENV
else
COMMIT_HASH="${INPUT_COMMIT_HASH}"
echo "COMMIT_HASH=$COMMIT_HASH" >> $GITHUB_ENV
echo "commit_hash=$COMMIT_HASH" >> "$GITHUB_OUTPUT"
echo "COMMIT_HASH=$INPUT_COMMIT_HASH" >> $GITHUB_ENV
echo "COMMIT_HASH_TYPE=manual" >> $GITHUB_ENV
fi
- name: Checkout the neon repository at given commit hash
uses: actions/checkout@11bd71901bbe5b1630ceea73d27597364c9af683 # v4.2.2
with:
ref: ${{ steps.commit_hash.outputs.commit_hash }}
# does not reuse ./.github/actions/download because we need to download the artifact for the given commit hash
# example artifact
# s3://neon-github-public-dev/artifacts/48b870bc078bd2c450eb7b468e743b9c118549bf/15036827400/1/neon-Linux-X64-release-artifact.tar.zst /instance_store/artifacts/neon-Linux-release-artifact.tar.zst
- name: Determine artifact S3_KEY for given commit hash and download and extract artifact
id: artifact_prefix
shell: bash -euxo pipefail {0}
env:
ARCHIVE: ${{ runner.temp }}/downloads/neon-${{ runner.os }}-${{ runner.arch }}-release-artifact.tar.zst
COMMIT_HASH: ${{ env.COMMIT_HASH }}
COMMIT_HASH_TYPE: ${{ env.COMMIT_HASH_TYPE }}
- name: Start Bench with run_id
run: |
attempt=0
max_attempts=24 # 5 minutes * 24 = 2 hours
curl -k -X 'POST' \
"${EC2_MACHINE_URL_US}/start_test/${GITHUB_RUN_ID}" \
-H 'accept: application/json' \
-H 'Content-Type: application/json' \
-H "Authorization: Bearer $API_KEY" \
-d "{\"neonRepoCommitHash\": \"${COMMIT_HASH}\", \"neonRepoCommitHashType\": \"${COMMIT_HASH_TYPE}\"}"
while [[ $attempt -lt $max_attempts ]]; do
# the following command will fail until the artifacts are available ...
S3_KEY=$(aws s3api list-objects-v2 --bucket "$S3_BUCKET" --prefix "artifacts/$COMMIT_HASH/" \
| jq -r '.Contents[]?.Key' \
| grep "neon-${{ runner.os }}-${{ runner.arch }}-release-artifact.tar.zst" \
| sort --version-sort \
| tail -1) || true # ... thus ignore errors from the command
if [[ -n "${S3_KEY}" ]]; then
echo "Artifact found: $S3_KEY"
echo "S3_KEY=$S3_KEY" >> $GITHUB_ENV
- name: Poll Test Status
id: poll_step
run: |
status=""
while [[ "$status" != "failure" && "$status" != "success" ]]; do
response=$(curl -k -X 'GET' \
"${EC2_MACHINE_URL_US}/test_status/${GITHUB_RUN_ID}" \
-H 'accept: application/json' \
-H "Authorization: Bearer $API_KEY")
echo "Response: $response"
set +x
status=$(echo $response | jq -r '.status')
echo "Test status: $status"
if [[ "$status" == "failure" ]]; then
echo "Test failed"
exit 1 # Fail the job step if status is failure
elif [[ "$status" == "success" || "$status" == "null" ]]; then
break
elif [[ "$status" == "too_many_runs" ]]; then
echo "Too many runs already running"
echo "too_many_runs=true" >> "$GITHUB_OUTPUT"
exit 1
fi
# Increment attempt counter and sleep for 5 minutes
attempt=$((attempt + 1))
echo "Attempt $attempt of $max_attempts to find artifacts in S3 bucket s3://$S3_BUCKET/artifacts/$COMMIT_HASH failed. Retrying in 5 minutes..."
sleep 300 # Sleep for 5 minutes
sleep 60 # Poll every 60 seconds
done
if [[ -z "${S3_KEY}" ]]; then
echo "Error: artifact not found in S3 bucket s3://$S3_BUCKET/artifacts/$COMMIT_HASH" after 2 hours
else
mkdir -p $(dirname $ARCHIVE)
time aws s3 cp --only-show-errors s3://$S3_BUCKET/${S3_KEY} ${ARCHIVE}
mkdir -p ${NEON_DIR}
time tar -xf ${ARCHIVE} -C ${NEON_DIR}
rm -f ${ARCHIVE}
fi
- name: Download snapshots from S3
if: ${{ github.event_name != 'workflow_dispatch' || github.event.inputs.recreate_snapshots == 'false' || github.event.inputs.recreate_snapshots == '' }}
id: download_snapshots
shell: bash -euxo pipefail {0}
- name: Retrieve Test Logs
if: always() && steps.poll_step.outputs.too_many_runs != 'true'
run: |
# Download the snapshots from S3
mkdir -p ${TEST_OUTPUT}
mkdir -p $BACKUP_DIR
cd $BACKUP_DIR
mkdir parts
cd parts
PART=$(aws s3api list-objects-v2 --bucket $S3_BUCKET --prefix performance/pagebench/ \
| jq -r '.Contents[]?.Key' \
| grep -E 'shared-snapshots-[0-9]{4}-[0-9]{2}-[0-9]{2}' \
| sort \
| tail -1)
echo "Latest PART: $PART"
if [[ -z "$PART" ]]; then
echo "ERROR: No matching S3 key found" >&2
exit 1
fi
S3_KEY=$(dirname $PART)
time aws s3 cp --only-show-errors --recursive s3://${S3_BUCKET}/$S3_KEY/ .
cd $TEST_OUTPUT
time cat $BACKUP_DIR/parts/* | zstdcat | tar --extract --preserve-permissions
rm -rf ${BACKUP_DIR}
curl -k -X 'GET' \
"${EC2_MACHINE_URL_US}/test_log/${GITHUB_RUN_ID}" \
-H 'accept: application/gzip' \
-H "Authorization: Bearer $API_KEY" \
--output "test_log_${GITHUB_RUN_ID}.gz"
- name: Cache poetry deps
uses: actions/cache@v4
with:
path: ~/.cache/pypoetry/virtualenvs
key: v2-${{ runner.os }}-${{ runner.arch }}-python-deps-bookworm-${{ hashFiles('poetry.lock') }}
- name: Install Python deps
shell: bash -euxo pipefail {0}
run: ./scripts/pysync
# we need high number of open files for pagebench
- name: show ulimits
shell: bash -euxo pipefail {0}
- name: Unzip Test Log and Print it into this job's log
if: always() && steps.poll_step.outputs.too_many_runs != 'true'
run: |
ulimit -a
- name: Run pagebench testcase
shell: bash -euxo pipefail {0}
env:
CI: false # need to override this env variable set by github to enforce using snapshots
run: |
export PLATFORM=hetzner-unit-perf-${COMMIT_HASH_TYPE}
# report the commit hash of the neon repository in the revision of the test results
export GITHUB_SHA=${COMMIT_HASH}
rm -rf ${PERF_REPORT_DIR}
rm -rf ${ALLURE_RESULTS_DIR}
mkdir -p ${PERF_REPORT_DIR}
mkdir -p ${ALLURE_RESULTS_DIR}
PARAMS="--alluredir=${ALLURE_RESULTS_DIR} --tb=short --verbose -rA"
EXTRA_PARAMS="--out-dir ${PERF_REPORT_DIR} --durations-path $TEST_OUTPUT/benchmark_durations.json"
# run only two selected tests
# environment set by parent:
# RUST_BACKTRACE=1 DEFAULT_PG_VERSION=16 BUILD_TYPE=release
./scripts/pytest ${PARAMS} test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py::test_pageserver_characterize_throughput_with_n_tenants ${EXTRA_PARAMS}
./scripts/pytest ${PARAMS} test_runner/performance/pageserver/pagebench/test_pageserver_max_throughput_getpage_at_latest_lsn.py::test_pageserver_characterize_latencies_with_1_client_and_throughput_with_many_clients_one_tenant ${EXTRA_PARAMS}
- name: upload the performance metrics to the Neon performance database which is used by grafana dashboards to display the results
shell: bash -euxo pipefail {0}
run: |
export REPORT_FROM="$PERF_REPORT_DIR"
export GITHUB_SHA=${COMMIT_HASH}
time ./scripts/generate_and_push_perf_report.sh
- name: Upload test results
if: ${{ !cancelled() }}
uses: ./.github/actions/allure-report-store
with:
report-dir: ${{ steps.set-env.outputs.allure_results_dir }}
unique-key: ${{ env.BUILD_TYPE }}-${{ env.DEFAULT_PG_VERSION }}-${{ runner.arch }}
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
gzip -d "test_log_${GITHUB_RUN_ID}.gz"
cat "test_log_${GITHUB_RUN_ID}"
- name: Create Allure report
id: create-allure-report
if: ${{ !cancelled() }}
uses: ./.github/actions/allure-report-generate
with:
aws-oidc-role-arn: ${{ vars.DEV_AWS_OIDC_ROLE_ARN }}
- name: Upload snapshots
if: ${{ github.event_name == 'workflow_dispatch' && github.event.inputs.recreate_snapshots != 'false' && github.event.inputs.recreate_snapshots != '' }}
id: upload_snapshots
shell: bash -euxo pipefail {0}
run: |
mkdir -p $BACKUP_DIR
cd $TEST_OUTPUT
tar --create --preserve-permissions --file - shared-snapshots | zstd -o $BACKUP_DIR/shared_snapshots.tar.zst
cd $BACKUP_DIR
mkdir parts
split -b 1G shared_snapshots.tar.zst ./parts/shared_snapshots.tar.zst.part.
SNAPSHOT_DATE=$(date +%F) # YYYY-MM-DD
cd parts
time aws s3 cp --recursive . s3://${S3_BUCKET}/performance/pagebench/shared-snapshots-${SNAPSHOT_DATE}/
- name: Post to a Slack channel
if: ${{ github.event.schedule && failure() }}
uses: slackapi/slack-github-action@fcfb566f8b0aab22203f066d80ca1d7e4b5d05b3 # v1.27.1
@@ -260,22 +157,26 @@ jobs:
slack-message: "Periodic pagebench testing on dedicated hardware: ${{ job.status }}\n${{ github.server_url }}/${{ github.repository }}/actions/runs/${{ github.run_id }}"
env:
SLACK_BOT_TOKEN: ${{ secrets.SLACK_BOT_TOKEN }}
- name: Cleanup Test Resources
if: always()
shell: bash -euxo pipefail {0}
env:
ARCHIVE: ${{ runner.temp }}/downloads/neon-${{ runner.os }}-${{ runner.arch }}-release-artifact.tar.zst
run: |
# Cleanup the test resources
if [[ -d "${BACKUP_DIR}" ]]; then
rm -rf ${BACKUP_DIR}
fi
if [[ -d "${TEST_OUTPUT}" ]]; then
rm -rf ${TEST_OUTPUT}
fi
if [[ -d "${NEON_DIR}" ]]; then
rm -rf ${NEON_DIR}
fi
rm -rf $(dirname $ARCHIVE)
curl -k -X 'POST' \
"${EC2_MACHINE_URL_US}/cleanup_test/${GITHUB_RUN_ID}" \
-H 'accept: application/json' \
-H "Authorization: Bearer $API_KEY" \
-d ''
- name: Assume AWS OIDC role that allows to manage (start/stop/describe... EC machine)
if: always() && steps.poll_step.outputs.too_many_runs != 'true'
uses: aws-actions/configure-aws-credentials@e3dd6a429d7300a6a4c196c26e071d42e0343502 # v4.0.2
with:
aws-region: eu-central-1
role-to-assume: ${{ vars.DEV_AWS_OIDC_ROLE_MANAGE_BENCHMARK_EC2_VMS_ARN }}
role-duration-seconds: 3600
- name: Stop EC2 instance and wait for the instance to be stopped
if: always() && steps.poll_step.outputs.too_many_runs != 'true'
run: |
aws ec2 stop-instances --instance-ids $AWS_INSTANCE_ID
aws ec2 wait instance-stopped --instance-ids $AWS_INSTANCE_ID

2
.gitignore vendored
View File

@@ -1,5 +1,4 @@
/artifact_cache
/build
/pg_install
/target
/tmp_check
@@ -14,7 +13,6 @@ neon.iml
/.neon
/integration_tests/.neon
compaction-suite-results.*
pgxn/neon/communicator/communicator_bindings.h
# Coverage
*.profraw

3513
Cargo.lock generated

File diff suppressed because it is too large Load Diff

View File

@@ -8,9 +8,7 @@ members = [
"pageserver/compaction",
"pageserver/ctl",
"pageserver/client",
"pageserver/client_grpc",
"pageserver/pagebench",
"pageserver/page_api",
"proxy",
"safekeeper",
"safekeeper/client",
@@ -23,7 +21,6 @@ members = [
"libs/http-utils",
"libs/pageserver_api",
"libs/postgres_ffi",
"libs/postgres_ffi_types",
"libs/safekeeper_api",
"libs/desim",
"libs/neon-shmem",
@@ -34,7 +31,6 @@ members = [
"libs/pq_proto",
"libs/tenant_size_model",
"libs/metrics",
"libs/neonart",
"libs/postgres_connection",
"libs/remote_storage",
"libs/tracing-utils",
@@ -47,7 +43,6 @@ members = [
"libs/proxy/postgres-types2",
"libs/proxy/tokio-postgres2",
"endpoint_storage",
"pgxn/neon/communicator",
]
[workspace.package]
@@ -75,8 +70,8 @@ aws-credential-types = "1.2.0"
aws-sigv4 = { version = "1.2", features = ["sign-http"] }
aws-types = "1.3"
axum = { version = "0.8.1", features = ["ws"] }
axum-extra = { version = "0.10.0", features = ["typed-header", "query"] }
base64 = "0.22"
axum-extra = { version = "0.10.0", features = ["typed-header"] }
base64 = "0.13.0"
bincode = "1.3"
bindgen = "0.71"
bit_field = "0.10.2"
@@ -91,7 +86,6 @@ clap = { version = "4.0", features = ["derive", "env"] }
clashmap = { version = "1.0", features = ["raw-api"] }
comfy-table = "7.1"
const_format = "0.2"
crossbeam-utils = "0.8.21"
crc32c = "0.6"
diatomic-waker = { version = "0.2.3" }
either = "1.8"
@@ -150,12 +144,11 @@ parquet = { version = "53", default-features = false, features = ["zstd"] }
parquet_derive = "53"
pbkdf2 = { version = "0.12.1", features = ["simple", "std"] }
pem = "3.0.3"
peekable = "0.3.0"
pin-project-lite = "0.2"
pprof = { version = "0.14", features = ["criterion", "flamegraph", "frame-pointer", "prost-codec"] }
procfs = "0.16"
prometheus = {version = "0.13", default-features=false, features = ["process"]} # removes protobuf dependency
prost = "0.13.5"
prost = "0.13"
rand = "0.8"
redis = { version = "0.29.2", features = ["tokio-rustls-comp", "keep-alive"] }
regex = "1.10.2"
@@ -177,7 +170,7 @@ sentry = { version = "0.37", default-features = false, features = ["backtrace",
serde = { version = "1.0", features = ["derive"] }
serde_json = "1"
serde_path_to_error = "0.1"
serde_with = { version = "3", features = [ "base64" ] }
serde_with = { version = "2.0", features = [ "base64" ] }
serde_assert = "0.5.0"
sha2 = "0.10.2"
signal-hook = "0.3"
@@ -185,7 +178,6 @@ smallvec = "1.11"
smol_str = { version = "0.2.0", features = ["serde"] }
socket2 = "0.5"
spki = "0.7.3"
spin = "0.9.8"
strum = "0.26"
strum_macros = "0.26"
"subtle" = "2.5.0"
@@ -197,16 +189,16 @@ thiserror = "1.0"
tikv-jemallocator = { version = "0.6", features = ["profiling", "stats", "unprefixed_malloc_on_supported_platforms"] }
tikv-jemalloc-ctl = { version = "0.6", features = ["stats"] }
tokio = { version = "1.43.1", features = ["macros"] }
tokio-epoll-uring = { git = "https://github.com/neondatabase/tokio-epoll-uring.git" , branch = "main" }
tokio-io-timeout = "1.2.0"
tokio-postgres-rustls = "0.12.0"
tokio-rustls = { version = "0.26.0", default-features = false, features = ["tls12", "ring"]}
tokio-stream = "0.1"
tokio-tar = "0.3"
tokio-util = { version = "0.7.10", features = ["io", "io-util", "rt"] }
tokio-util = { version = "0.7.10", features = ["io", "rt"] }
toml = "0.8"
toml_edit = "0.22"
tonic = { version = "0.13.1", default-features = false, features = ["channel", "codegen", "gzip", "prost", "router", "server", "tls-ring", "tls-native-roots", "zstd"] }
tonic-reflection = { version = "0.13.1", features = ["server"] }
tonic = {version = "0.12.3", default-features = false, features = ["channel", "tls", "tls-roots"]}
tower = { version = "0.5.2", default-features = false }
tower-http = { version = "0.6.2", features = ["auth", "request-id", "trace"] }
@@ -238,9 +230,6 @@ x509-cert = { version = "0.2.5" }
env_logger = "0.11"
log = "0.4"
tokio-epoll-uring = { git = "https://github.com/neondatabase/tokio-epoll-uring.git" , branch = "main" }
uring-common = { git = "https://github.com/neondatabase/tokio-epoll-uring.git" , branch = "main" }
## Libraries from neondatabase/ git forks, ideally with changes to be upstreamed
postgres = { git = "https://github.com/neondatabase/rust-postgres.git", branch = "neon" }
postgres-protocol = { git = "https://github.com/neondatabase/rust-postgres.git", branch = "neon" }
@@ -256,47 +245,40 @@ azure_storage_blobs = { git = "https://github.com/neondatabase/azure-sdk-for-rus
## Local libraries
compute_api = { version = "0.1", path = "./libs/compute_api/" }
consumption_metrics = { version = "0.1", path = "./libs/consumption_metrics/" }
desim = { version = "0.1", path = "./libs/desim" }
endpoint_storage = { version = "0.0.1", path = "./endpoint_storage/" }
http-utils = { version = "0.1", path = "./libs/http-utils/" }
metrics = { version = "0.1", path = "./libs/metrics/" }
neonart = { version = "0.1", path = "./libs/neonart/" }
neon-shmem = { version = "0.1", path = "./libs/neon-shmem/" }
pageserver = { path = "./pageserver" }
pageserver_api = { version = "0.1", path = "./libs/pageserver_api/" }
pageserver_client = { path = "./pageserver/client" }
pageserver_client_grpc = { path = "./pageserver/client_grpc" }
pageserver_compaction = { version = "0.1", path = "./pageserver/compaction/" }
pageserver_page_api = { path = "./pageserver/page_api" }
postgres_backend = { version = "0.1", path = "./libs/postgres_backend/" }
postgres_connection = { version = "0.1", path = "./libs/postgres_connection/" }
postgres_ffi = { version = "0.1", path = "./libs/postgres_ffi/" }
postgres_ffi_types = { version = "0.1", path = "./libs/postgres_ffi_types/" }
postgres_initdb = { path = "./libs/postgres_initdb" }
posthog_client_lite = { version = "0.1", path = "./libs/posthog_client_lite" }
pq_proto = { version = "0.1", path = "./libs/pq_proto/" }
remote_storage = { version = "0.1", path = "./libs/remote_storage/" }
safekeeper_api = { version = "0.1", path = "./libs/safekeeper_api" }
safekeeper_client = { path = "./safekeeper/client" }
desim = { version = "0.1", path = "./libs/desim" }
storage_broker = { version = "0.1", path = "./storage_broker/" } # Note: main broker code is inside the binary crate, so linking with the library shouldn't be heavy.
storage_controller_client = { path = "./storage_controller/client" }
tenant_size_model = { version = "0.1", path = "./libs/tenant_size_model/" }
tracing-utils = { version = "0.1", path = "./libs/tracing-utils/" }
utils = { version = "0.1", path = "./libs/utils/" }
vm_monitor = { version = "0.1", path = "./libs/vm_monitor/" }
wal_decoder = { version = "0.1", path = "./libs/wal_decoder" }
walproposer = { version = "0.1", path = "./libs/walproposer/" }
wal_decoder = { version = "0.1", path = "./libs/wal_decoder" }
## Common library dependency
workspace_hack = { version = "0.1", path = "./workspace_hack/" }
## Build dependencies
cbindgen = "0.28.0"
criterion = "0.5.1"
rcgen = "0.13"
rstest = "0.18"
camino-tempfile = "1.0.2"
tonic-build = "0.13.1"
tonic-build = "0.12"
[patch.crates-io]

View File

@@ -5,6 +5,8 @@
ARG REPOSITORY=ghcr.io/neondatabase
ARG IMAGE=build-tools
ARG TAG=pinned
ARG DEFAULT_PG_VERSION=17
ARG STABLE_PG_VERSION=16
ARG DEBIAN_VERSION=bookworm
ARG DEBIAN_FLAVOR=${DEBIAN_VERSION}-slim
@@ -45,6 +47,7 @@ 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 pg_install/build \
&& tar -C pg_install -czf /home/nonroot/postgres_install.tar.gz .
# Prepare cargo-chef recipe
@@ -60,11 +63,14 @@ FROM $REPOSITORY/$IMAGE:$TAG AS build
WORKDIR /home/nonroot
ARG GIT_VERSION=local
ARG BUILD_TAG
ARG STABLE_PG_VERSION
COPY --from=pg-build /home/nonroot/pg_install/v14/include/postgresql/server pg_install/v14/include/postgresql/server
COPY --from=pg-build /home/nonroot/pg_install/v15/include/postgresql/server pg_install/v15/include/postgresql/server
COPY --from=pg-build /home/nonroot/pg_install/v16/include/postgresql/server pg_install/v16/include/postgresql/server
COPY --from=pg-build /home/nonroot/pg_install/v17/include/postgresql/server pg_install/v17/include/postgresql/server
COPY --from=pg-build /home/nonroot/pg_install/v16/lib pg_install/v16/lib
COPY --from=pg-build /home/nonroot/pg_install/v17/lib pg_install/v17/lib
COPY --from=plan /home/nonroot/recipe.json recipe.json
ARG ADDITIONAL_RUSTFLAGS=""
@@ -91,6 +97,7 @@ RUN set -e \
# Build final image
#
FROM $BASE_IMAGE_SHA
ARG DEFAULT_PG_VERSION
WORKDIR /data
RUN set -e \
@@ -100,20 +107,9 @@ RUN set -e \
libreadline-dev \
libseccomp-dev \
ca-certificates \
# System postgres for use with client libraries (e.g. in storage controller)
postgresql-15 \
openssl \
unzip \
curl \
&& ARCH=$(uname -m) \
&& if [ "$ARCH" = "x86_64" ]; then \
curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"; \
elif [ "$ARCH" = "aarch64" ]; then \
curl "https://awscli.amazonaws.com/awscli-exe-linux-aarch64.zip" -o "awscliv2.zip"; \
else \
echo "Unsupported architecture: $ARCH" && exit 1; \
fi \
&& unzip awscliv2.zip \
&& ./aws/install \
&& rm -rf aws awscliv2.zip \
&& rm -f /etc/apt/apt.conf.d/80-retries \
&& rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* \
&& useradd -d /data neon \

172
Makefile
View File

@@ -1,18 +1,8 @@
ROOT_PROJECT_DIR := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
# Where to install Postgres, default is ./pg_install, maybe useful for package
# managers.
# Where to install Postgres, default is ./pg_install, maybe useful for package managers
POSTGRES_INSTALL_DIR ?= $(ROOT_PROJECT_DIR)/pg_install/
# CARGO_BUILD_FLAGS: Extra flags to pass to `cargo build`. `--locked`
# and `--features testing` are popular examples.
#
# CARGO_PROFILE: You can also set to override the cargo profile to
# use. By default, it is derived from BUILD_TYPE.
# All intermediate build artifacts are stored here.
BUILD_DIR := build
ICU_PREFIX_DIR := /usr/local/icu
#
@@ -26,19 +16,12 @@ ifeq ($(BUILD_TYPE),release)
PG_CONFIGURE_OPTS = --enable-debug --with-openssl
PG_CFLAGS += -O2 -g3 $(CFLAGS)
PG_LDFLAGS = $(LDFLAGS)
CARGO_PROFILE ?= --profile=release
# NEON_CARGO_ARTIFACT_TARGET_DIR is the directory where `cargo build` places
# the final build artifacts. There is unfortunately no easy way of changing
# it to a fully predictable path, nor to extract the path with a simple
# command. See https://github.com/rust-lang/cargo/issues/9661 and
# https://github.com/rust-lang/cargo/issues/6790.
NEON_CARGO_ARTIFACT_TARGET_DIR = $(ROOT_PROJECT_DIR)/target/release
# Unfortunately, `--profile=...` is a nightly feature
CARGO_BUILD_FLAGS += --release
else ifeq ($(BUILD_TYPE),debug)
PG_CONFIGURE_OPTS = --enable-debug --with-openssl --enable-cassert --enable-depend
PG_CFLAGS += -O0 -g3 $(CFLAGS)
PG_LDFLAGS = $(LDFLAGS)
CARGO_PROFILE ?= --profile=dev
NEON_CARGO_ARTIFACT_TARGET_DIR = $(ROOT_PROJECT_DIR)/target/debug
else
$(error Bad build type '$(BUILD_TYPE)', see Makefile for options)
endif
@@ -110,8 +93,7 @@ all: neon postgres neon-pg-ext
.PHONY: neon
neon: postgres-headers walproposer-lib cargo-target-dir
+@echo "Compiling Neon"
$(CARGO_CMD_PREFIX) cargo build $(CARGO_BUILD_FLAGS) $(CARGO_PROFILE)
$(CARGO_CMD_PREFIX) cargo build $(CARGO_BUILD_FLAGS)
.PHONY: cargo-target-dir
cargo-target-dir:
# https://github.com/rust-lang/cargo/issues/14281
@@ -122,20 +104,21 @@ cargo-target-dir:
# Some 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.
#
$(BUILD_DIR)/%/config.status:
mkdir -p $(BUILD_DIR)
test -e $(BUILD_DIR)/CACHEDIR.TAG || echo "$(CACHEDIR_TAG_CONTENTS)" > $(BUILD_DIR)/CACHEDIR.TAG
$(POSTGRES_INSTALL_DIR)/build/%/config.status:
mkdir -p $(POSTGRES_INSTALL_DIR)
test -e $(POSTGRES_INSTALL_DIR)/CACHEDIR.TAG || echo "$(CACHEDIR_TAG_CONTENTS)" > $(POSTGRES_INSTALL_DIR)/CACHEDIR.TAG
+@echo "Configuring Postgres $* build"
@test -s $(ROOT_PROJECT_DIR)/vendor/postgres-$*/configure || { \
echo "\nPostgres submodule not found in $(ROOT_PROJECT_DIR)/vendor/postgres-$*/, execute "; \
echo "'git submodule update --init --recursive --depth 2 --progress .' in project root.\n"; \
exit 1; }
mkdir -p $(BUILD_DIR)/$*
mkdir -p $(POSTGRES_INSTALL_DIR)/build/$*
VERSION=$*; \
EXTRA_VERSION=$$(cd $(ROOT_PROJECT_DIR)/vendor/postgres-$$VERSION && git rev-parse HEAD); \
(cd $(BUILD_DIR)/$$VERSION && \
(cd $(POSTGRES_INSTALL_DIR)/build/$$VERSION && \
env PATH="$(EXTRA_PATH_OVERRIDES):$$PATH" $(ROOT_PROJECT_DIR)/vendor/postgres-$$VERSION/configure \
CFLAGS='$(PG_CFLAGS)' LDFLAGS='$(PG_LDFLAGS)' \
$(PG_CONFIGURE_OPTS) --with-extra-version=" ($$EXTRA_VERSION)" \
@@ -147,57 +130,96 @@ $(BUILD_DIR)/%/config.status:
# the "build-all-versions" entry points) where direct mention of PostgreSQL
# versions is used.
.PHONY: postgres-configure-v17
postgres-configure-v17: $(BUILD_DIR)/v17/config.status
postgres-configure-v17: $(POSTGRES_INSTALL_DIR)/build/v17/config.status
.PHONY: postgres-configure-v16
postgres-configure-v16: $(BUILD_DIR)/v16/config.status
postgres-configure-v16: $(POSTGRES_INSTALL_DIR)/build/v16/config.status
.PHONY: postgres-configure-v15
postgres-configure-v15: $(BUILD_DIR)/v15/config.status
postgres-configure-v15: $(POSTGRES_INSTALL_DIR)/build/v15/config.status
.PHONY: postgres-configure-v14
postgres-configure-v14: $(BUILD_DIR)/v14/config.status
postgres-configure-v14: $(POSTGRES_INSTALL_DIR)/build/v14/config.status
# Install the PostgreSQL header files into $(POSTGRES_INSTALL_DIR)/<version>/include
.PHONY: postgres-headers-%
postgres-headers-%: postgres-configure-%
+@echo "Installing PostgreSQL $* headers"
$(MAKE) -C $(BUILD_DIR)/$*/src/include MAKELEVEL=0 install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/src/include MAKELEVEL=0 install
# Compile and install PostgreSQL
.PHONY: postgres-%
postgres-%: postgres-configure-% \
postgres-headers-% # to prevent `make install` conflicts with neon's `postgres-headers`
+@echo "Compiling PostgreSQL $*"
$(MAKE) -C $(BUILD_DIR)/$* MAKELEVEL=0 install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$* MAKELEVEL=0 install
+@echo "Compiling libpq $*"
$(MAKE) -C $(BUILD_DIR)/$*/src/interfaces/libpq install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/src/interfaces/libpq install
+@echo "Compiling pg_prewarm $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/pg_prewarm install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pg_prewarm install
+@echo "Compiling pg_buffercache $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/pg_buffercache install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pg_buffercache install
+@echo "Compiling pg_visibility $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/pg_visibility install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pg_visibility install
+@echo "Compiling pageinspect $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/pageinspect install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pageinspect install
+@echo "Compiling pg_trgm $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/pg_trgm install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/pg_trgm install
+@echo "Compiling amcheck $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/amcheck install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/amcheck install
+@echo "Compiling test_decoding $*"
$(MAKE) -C $(BUILD_DIR)/$*/contrib/test_decoding install
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/contrib/test_decoding install
.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
.PHONY: postgres-check-%
postgres-check-%: postgres-%
$(MAKE) -C $(BUILD_DIR)/$* MAKELEVEL=0 check
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$* MAKELEVEL=0 check
.PHONY: neon-pg-ext-%
neon-pg-ext-%: postgres-% cargo-target-dir
+@echo "Compiling neon-specific Postgres extensions for $*"
mkdir -p $(BUILD_DIR)/pgxn-$*
$(MAKE) PG_CONFIG="$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config" COPT='$(COPT)' \
NEON_CARGO_ARTIFACT_TARGET_DIR="$(NEON_CARGO_ARTIFACT_TARGET_DIR)" \
CARGO_BUILD_FLAGS="$(CARGO_BUILD_FLAGS)" \
CARGO_PROFILE="$(CARGO_PROFILE)" \
-C $(BUILD_DIR)/pgxn-$*\
-f $(ROOT_PROJECT_DIR)/pgxn/Makefile install
neon-pg-ext-%: postgres-%
+@echo "Compiling neon $*"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-$*
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config COPT='$(COPT)' \
-C $(POSTGRES_INSTALL_DIR)/build/neon-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile install
+@echo "Compiling neon_walredo $*"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-walredo-$*
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config COPT='$(COPT)' \
-C $(POSTGRES_INSTALL_DIR)/build/neon-walredo-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_walredo/Makefile install
+@echo "Compiling neon_rmgr $*"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-rmgr-$*
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config COPT='$(COPT)' \
-C $(POSTGRES_INSTALL_DIR)/build/neon-rmgr-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_rmgr/Makefile install
+@echo "Compiling neon_test_utils $*"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-$*
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config COPT='$(COPT)' \
-C $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_test_utils/Makefile install
+@echo "Compiling neon_utils $*"
mkdir -p $(POSTGRES_INSTALL_DIR)/build/neon-utils-$*
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config COPT='$(COPT)' \
-C $(POSTGRES_INSTALL_DIR)/build/neon-utils-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_utils/Makefile install
.PHONY: neon-pg-clean-ext-%
neon-pg-clean-ext-%:
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config \
-C $(POSTGRES_INSTALL_DIR)/build/neon-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile clean
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config \
-C $(POSTGRES_INSTALL_DIR)/build/neon-walredo-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_walredo/Makefile clean
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config \
-C $(POSTGRES_INSTALL_DIR)/build/neon-test-utils-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_test_utils/Makefile clean
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/$*/bin/pg_config \
-C $(POSTGRES_INSTALL_DIR)/build/neon-utils-$* \
-f $(ROOT_PROJECT_DIR)/pgxn/neon_utils/Makefile clean
# Build walproposer as a static library. walproposer source code is located
# in the pgxn/neon directory.
@@ -211,15 +233,15 @@ neon-pg-ext-%: postgres-% cargo-target-dir
.PHONY: walproposer-lib
walproposer-lib: neon-pg-ext-v17
+@echo "Compiling walproposer-lib"
mkdir -p $(BUILD_DIR)/walproposer-lib
mkdir -p $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v17/bin/pg_config COPT='$(COPT)' \
-C $(BUILD_DIR)/walproposer-lib \
-C $(POSTGRES_INSTALL_DIR)/build/walproposer-lib \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile walproposer-lib
cp $(POSTGRES_INSTALL_DIR)/v17/lib/libpgport.a $(BUILD_DIR)/walproposer-lib
cp $(POSTGRES_INSTALL_DIR)/v17/lib/libpgcommon.a $(BUILD_DIR)/walproposer-lib
$(AR) d $(BUILD_DIR)/walproposer-lib/libpgport.a \
cp $(POSTGRES_INSTALL_DIR)/v17/lib/libpgport.a $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
cp $(POSTGRES_INSTALL_DIR)/v17/lib/libpgcommon.a $(POSTGRES_INSTALL_DIR)/build/walproposer-lib
$(AR) d $(POSTGRES_INSTALL_DIR)/build/walproposer-lib/libpgport.a \
pg_strong_random.o
$(AR) d $(BUILD_DIR)/walproposer-lib/libpgcommon.a \
$(AR) d $(POSTGRES_INSTALL_DIR)/build/walproposer-lib/libpgcommon.a \
checksum_helper.o \
cryptohash_openssl.o \
hmac_openssl.o \
@@ -227,10 +249,16 @@ walproposer-lib: neon-pg-ext-v17
parse_manifest.o \
scram-common.o
ifeq ($(UNAME_S),Linux)
$(AR) d $(BUILD_DIR)/walproposer-lib/libpgcommon.a \
$(AR) d $(POSTGRES_INSTALL_DIR)/build/walproposer-lib/libpgcommon.a \
pg_crc32c.o
endif
.PHONY: walproposer-lib-clean
walproposer-lib-clean:
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v17/bin/pg_config \
-C $(POSTGRES_INSTALL_DIR)/build/walproposer-lib \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile clean
.PHONY: neon-pg-ext
neon-pg-ext: \
neon-pg-ext-v14 \
@@ -238,6 +266,13 @@ neon-pg-ext: \
neon-pg-ext-v16 \
neon-pg-ext-v17
.PHONY: neon-pg-clean-ext
neon-pg-clean-ext: \
neon-pg-clean-ext-v14 \
neon-pg-clean-ext-v15 \
neon-pg-clean-ext-v16 \
neon-pg-clean-ext-v17
# shorthand to build all Postgres versions
.PHONY: postgres
postgres: \
@@ -253,6 +288,13 @@ postgres-headers: \
postgres-headers-v16 \
postgres-headers-v17
.PHONY: postgres-clean
postgres-clean: \
postgres-clean-v14 \
postgres-clean-v15 \
postgres-clean-v16 \
postgres-clean-v17
.PHONY: postgres-check
postgres-check: \
postgres-check-v14 \
@@ -260,6 +302,12 @@ postgres-check: \
postgres-check-v16 \
postgres-check-v17
# This doesn't remove the effects of 'configure'.
.PHONY: clean
clean: postgres-clean neon-pg-clean-ext
$(MAKE) -C compute clean
$(CARGO_CMD_PREFIX) cargo clean
# This removes everything
.PHONY: distclean
distclean:
@@ -272,7 +320,7 @@ fmt:
postgres-%-pg-bsd-indent: postgres-%
+@echo "Compiling pg_bsd_indent"
$(MAKE) -C $(BUILD_DIR)/$*/src/tools/pg_bsd_indent/
$(MAKE) -C $(POSTGRES_INSTALL_DIR)/build/$*/src/tools/pg_bsd_indent/
# Create typedef list for the core. Note that generally it should be combined with
# buildfarm one to cover platform specific stuff.
@@ -291,7 +339,7 @@ postgres-%-pgindent: postgres-%-pg-bsd-indent postgres-%-typedefs.list
cat $(ROOT_PROJECT_DIR)/vendor/postgres-$*/src/tools/pgindent/typedefs.list |\
cat - postgres-$*-typedefs.list | sort | uniq > postgres-$*-typedefs-full.list
+@echo note: you might want to run it on selected files/dirs instead.
INDENT=$(BUILD_DIR)/$*/src/tools/pg_bsd_indent/pg_bsd_indent \
INDENT=$(POSTGRES_INSTALL_DIR)/build/$*/src/tools/pg_bsd_indent/pg_bsd_indent \
$(ROOT_PROJECT_DIR)/vendor/postgres-$*/src/tools/pgindent/pgindent --typedefs postgres-$*-typedefs-full.list \
$(ROOT_PROJECT_DIR)/vendor/postgres-$*/src/ \
--excludes $(ROOT_PROJECT_DIR)/vendor/postgres-$*/src/tools/pgindent/exclude_file_patterns
@@ -302,9 +350,9 @@ postgres-%-pgindent: postgres-%-pg-bsd-indent postgres-%-typedefs.list
neon-pgindent: postgres-v17-pg-bsd-indent neon-pg-ext-v17
$(MAKE) PG_CONFIG=$(POSTGRES_INSTALL_DIR)/v17/bin/pg_config COPT='$(COPT)' \
FIND_TYPEDEF=$(ROOT_PROJECT_DIR)/vendor/postgres-v17/src/tools/find_typedef \
INDENT=$(BUILD_DIR)/v17/src/tools/pg_bsd_indent/pg_bsd_indent \
INDENT=$(POSTGRES_INSTALL_DIR)/build/v17/src/tools/pg_bsd_indent/pg_bsd_indent \
PGINDENT_SCRIPT=$(ROOT_PROJECT_DIR)/vendor/postgres-v17/src/tools/pgindent/pgindent \
-C $(BUILD_DIR)/neon-v17 \
-C $(POSTGRES_INSTALL_DIR)/build/neon-v17 \
-f $(ROOT_PROJECT_DIR)/pgxn/neon/Makefile pgindent

View File

@@ -155,7 +155,7 @@ RUN set -e \
# Keep the version the same as in compute/compute-node.Dockerfile and
# test_runner/regress/test_compute_metrics.py.
ENV SQL_EXPORTER_VERSION=0.17.3
ENV SQL_EXPORTER_VERSION=0.17.0
RUN curl -fsSL \
"https://github.com/burningalchemist/sql_exporter/releases/download/${SQL_EXPORTER_VERSION}/sql_exporter-${SQL_EXPORTER_VERSION}.linux-$(case "$(uname -m)" in x86_64) echo amd64;; aarch64) echo arm64;; esac).tar.gz" \
--output sql_exporter.tar.gz \
@@ -310,13 +310,13 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
. "$HOME/.cargo/env" && \
cargo --version && rustup --version && \
rustup component add llvm-tools rustfmt clippy && \
cargo install rustfilt --version ${RUSTFILT_VERSION} --locked && \
cargo install cargo-hakari --version ${CARGO_HAKARI_VERSION} --locked && \
cargo install cargo-deny --version ${CARGO_DENY_VERSION} --locked && \
cargo install cargo-hack --version ${CARGO_HACK_VERSION} --locked && \
cargo install cargo-nextest --version ${CARGO_NEXTEST_VERSION} --locked && \
cargo install cargo-chef --version ${CARGO_CHEF_VERSION} --locked && \
cargo install diesel_cli --version ${CARGO_DIESEL_CLI_VERSION} --locked \
cargo install rustfilt --version ${RUSTFILT_VERSION} && \
cargo install cargo-hakari --version ${CARGO_HAKARI_VERSION} && \
cargo install cargo-deny --locked --version ${CARGO_DENY_VERSION} && \
cargo install cargo-hack --version ${CARGO_HACK_VERSION} && \
cargo install cargo-nextest --version ${CARGO_NEXTEST_VERSION} && \
cargo install cargo-chef --locked --version ${CARGO_CHEF_VERSION} && \
cargo install diesel_cli --version ${CARGO_DIESEL_CLI_VERSION} \
--features postgres-bundled --no-default-features && \
rm -rf /home/nonroot/.cargo/registry && \
rm -rf /home/nonroot/.cargo/git

3
compute/.gitignore vendored
View File

@@ -3,6 +3,3 @@ etc/neon_collector.yml
etc/neon_collector_autoscaling.yml
etc/sql_exporter.yml
etc/sql_exporter_autoscaling.yml
# Node.js dependencies
node_modules/

View File

@@ -48,11 +48,3 @@ jsonnetfmt-test:
.PHONY: jsonnetfmt-format
jsonnetfmt-format:
jsonnetfmt --in-place $(jsonnet_files)
.PHONY: manifest-schema-validation
manifest-schema-validation: node_modules
node_modules/.bin/jsonschema validate -d https://json-schema.org/draft/2020-12/schema manifest.schema.json manifest.yaml
node_modules: package.json
npm install
touch node_modules

View File

@@ -77,6 +77,9 @@
# build_and_test.yml github workflow for how that's done.
ARG PG_VERSION
ARG REPOSITORY=ghcr.io/neondatabase
ARG IMAGE=build-tools
ARG TAG=pinned
ARG BUILD_TAG
ARG DEBIAN_VERSION=bookworm
ARG DEBIAN_FLAVOR=${DEBIAN_VERSION}-slim
@@ -146,11 +149,8 @@ RUN case $DEBIAN_VERSION in \
ninja-build git autoconf automake libtool build-essential bison flex libreadline-dev \
zlib1g-dev libxml2-dev libcurl4-openssl-dev libossp-uuid-dev wget ca-certificates pkg-config libssl-dev \
libicu-dev libxslt1-dev liblz4-dev libzstd-dev zstd curl unzip g++ \
libclang-dev \
jsonnet \
$VERSION_INSTALLS \
&& apt clean && rm -rf /var/lib/apt/lists/* && \
useradd -ms /bin/bash nonroot -b /home
&& apt clean && rm -rf /var/lib/apt/lists/*
#########################################################################################
#
@@ -297,7 +297,6 @@ RUN ./autogen.sh && \
./configure --with-sfcgal=/usr/local/bin/sfcgal-config && \
make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
make staged-install && \
cd extensions/postgis && \
make clean && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
@@ -583,38 +582,6 @@ RUN make -j $(getconf _NPROCESSORS_ONLN) && \
make -j $(getconf _NPROCESSORS_ONLN) install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/hypopg.control
#########################################################################################
#
# Layer "online_advisor-build"
# compile online_advisor extension
#
#########################################################################################
FROM build-deps AS online_advisor-src
ARG PG_VERSION
# online_advisor supports all Postgres version starting from PG14, but prior to PG17 has to be included in preload_shared_libraries
# last release 1.0 - May 15, 2025
WORKDIR /ext-src
RUN case "${PG_VERSION:?}" in \
"v17") \
;; \
*) \
echo "skipping the version of online_advistor for $PG_VERSION" && exit 0 \
;; \
esac && \
wget https://github.com/knizhnik/online_advisor/archive/refs/tags/1.0.tar.gz -O online_advisor.tar.gz && \
echo "37dcadf8f7cc8d6cc1f8831276ee245b44f1b0274f09e511e47a67738ba9ed0f online_advisor.tar.gz" | sha256sum --check && \
mkdir online_advisor-src && cd online_advisor-src && tar xzf ../online_advisor.tar.gz --strip-components=1 -C .
FROM pg-build AS online_advisor-build
COPY --from=online_advisor-src /ext-src/ /ext-src/
WORKDIR /ext-src/
RUN if [ -d online_advisor-src ]; then \
cd online_advisor-src && \
make -j install && \
echo 'trusted = true' >> /usr/local/pgsql/share/extension/online_advisor.control; \
fi
#########################################################################################
#
# Layer "pg_hashids-build"
@@ -1057,10 +1024,17 @@ RUN make -j $(getconf _NPROCESSORS_ONLN) && \
#########################################################################################
#
# Layer "build-deps with Rust toolchain installed"
# Layer "pg build with nonroot user and cargo installed"
# This layer is base and common for layers with `pgrx`
#
#########################################################################################
FROM build-deps AS build-deps-with-cargo
FROM pg-build AS pg-build-nonroot-with-cargo
ARG PG_VERSION
RUN apt update && \
apt install --no-install-recommends --no-install-suggests -y curl libclang-dev && \
apt clean && rm -rf /var/lib/apt/lists/* && \
useradd -ms /bin/bash nonroot -b /home
ENV HOME=/home/nonroot
ENV PATH="/home/nonroot/.cargo/bin:$PATH"
@@ -1075,29 +1049,13 @@ RUN curl -sSO https://static.rust-lang.org/rustup/dist/$(uname -m)-unknown-linux
./rustup-init -y --no-modify-path --profile minimal --default-toolchain stable && \
rm rustup-init
#########################################################################################
#
# Layer "pg-build with Rust toolchain installed"
# This layer is base and common for layers with `pgrx`
#
#########################################################################################
FROM pg-build AS pg-build-with-cargo
ARG PG_VERSION
ENV HOME=/home/nonroot
ENV PATH="/home/nonroot/.cargo/bin:$PATH"
USER nonroot
WORKDIR /home/nonroot
COPY --from=build-deps-with-cargo /home/nonroot /home/nonroot
#########################################################################################
#
# Layer "rust extensions"
# This layer is used to build `pgrx` deps
#
#########################################################################################
FROM pg-build-with-cargo AS rust-extensions-build
FROM pg-build-nonroot-with-cargo AS rust-extensions-build
ARG PG_VERSION
RUN case "${PG_VERSION:?}" in \
@@ -1119,7 +1077,7 @@ USER root
# and eventually get merged with `rust-extensions-build`
#
#########################################################################################
FROM pg-build-with-cargo AS rust-extensions-build-pgrx12
FROM pg-build-nonroot-with-cargo AS rust-extensions-build-pgrx12
ARG PG_VERSION
RUN cargo install --locked --version 0.12.9 cargo-pgrx && \
@@ -1136,7 +1094,7 @@ USER root
# and eventually get merged with `rust-extensions-build`
#
#########################################################################################
FROM pg-build-with-cargo AS rust-extensions-build-pgrx14
FROM pg-build-nonroot-with-cargo AS rust-extensions-build-pgrx14
ARG PG_VERSION
RUN cargo install --locked --version 0.14.1 cargo-pgrx && \
@@ -1153,12 +1111,10 @@ USER root
FROM build-deps AS pgrag-src
ARG PG_VERSION
WORKDIR /ext-src
COPY compute/patches/onnxruntime.patch .
WORKDIR /ext-src
RUN wget https://github.com/microsoft/onnxruntime/archive/refs/tags/v1.18.1.tar.gz -O onnxruntime.tar.gz && \
mkdir onnxruntime-src && cd onnxruntime-src && tar xzf ../onnxruntime.tar.gz --strip-components=1 -C . && \
patch -p1 < /ext-src/onnxruntime.patch && \
echo "#nothing to test here" > neon-test.sh
RUN wget https://github.com/neondatabase-labs/pgrag/archive/refs/tags/v0.1.2.tar.gz -O pgrag.tar.gz && \
@@ -1192,14 +1148,14 @@ RUN cd exts/rag && \
RUN cd exts/rag_bge_small_en_v15 && \
sed -i 's/pgrx = "0.14.1"/pgrx = { version = "0.14.1", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
ORT_LIB_LOCATION=/ext-src/onnxruntime-src/build/Linux \
REMOTE_ONNX_URL=http://pg-ext-s3-gateway.pg-ext-s3-gateway.svc.cluster.local/pgrag-data/bge_small_en_v15.onnx \
REMOTE_ONNX_URL=http://pg-ext-s3-gateway/pgrag-data/bge_small_en_v15.onnx \
cargo pgrx install --release --features remote_onnx && \
echo "trusted = true" >> /usr/local/pgsql/share/extension/rag_bge_small_en_v15.control
RUN cd exts/rag_jina_reranker_v1_tiny_en && \
sed -i 's/pgrx = "0.14.1"/pgrx = { version = "0.14.1", features = [ "unsafe-postgres" ] }/g' Cargo.toml && \
ORT_LIB_LOCATION=/ext-src/onnxruntime-src/build/Linux \
REMOTE_ONNX_URL=http://pg-ext-s3-gateway.pg-ext-s3-gateway.svc.cluster.local/pgrag-data/jina_reranker_v1_tiny_en.onnx \
REMOTE_ONNX_URL=http://pg-ext-s3-gateway/pgrag-data/jina_reranker_v1_tiny_en.onnx \
cargo pgrx install --release --features remote_onnx && \
echo "trusted = true" >> /usr/local/pgsql/share/extension/rag_jina_reranker_v1_tiny_en.control
@@ -1632,7 +1588,18 @@ FROM pg-build AS neon-ext-build
ARG PG_VERSION
COPY pgxn/ pgxn/
RUN make -j $(getconf _NPROCESSORS_ONLN) -C pgxn -s install-compute
RUN make -j $(getconf _NPROCESSORS_ONLN) \
-C pgxn/neon \
-s install && \
make -j $(getconf _NPROCESSORS_ONLN) \
-C pgxn/neon_utils \
-s install && \
make -j $(getconf _NPROCESSORS_ONLN) \
-C pgxn/neon_test_utils \
-s install && \
make -j $(getconf _NPROCESSORS_ONLN) \
-C pgxn/neon_rmgr \
-s install
#########################################################################################
#
@@ -1681,7 +1648,6 @@ COPY --from=pg_jsonschema-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=pg_graphql-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=pg_tiktoken-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=hypopg-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=online_advisor-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=pg_hashids-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=rum-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=pgtap-build /usr/local/pgsql/ /usr/local/pgsql/
@@ -1722,7 +1688,7 @@ FROM extensions-${EXTENSIONS} AS neon-pg-ext-build
# Compile the Neon-specific `compute_ctl`, `fast_import`, and `local_proxy` binaries
#
#########################################################################################
FROM build-deps-with-cargo AS compute-tools
FROM $REPOSITORY/$IMAGE:$TAG AS compute-tools
ARG BUILD_TAG
ENV BUILD_TAG=$BUILD_TAG
@@ -1732,7 +1698,7 @@ COPY --chown=nonroot . .
RUN --mount=type=cache,uid=1000,target=/home/nonroot/.cargo/registry \
--mount=type=cache,uid=1000,target=/home/nonroot/.cargo/git \
--mount=type=cache,uid=1000,target=/home/nonroot/target \
cargo build --locked --profile release-line-debug-size-lto --bin compute_ctl --bin fast_import --bin local_proxy && \
mold -run cargo build --locked --profile release-line-debug-size-lto --bin compute_ctl --bin fast_import --bin local_proxy && \
mkdir target-bin && \
cp target/release-line-debug-size-lto/compute_ctl \
target/release-line-debug-size-lto/fast_import \
@@ -1785,17 +1751,17 @@ ARG TARGETARCH
RUN if [ "$TARGETARCH" = "amd64" ]; then\
postgres_exporter_sha256='59aa4a7bb0f7d361f5e05732f5ed8c03cc08f78449cef5856eadec33a627694b';\
pgbouncer_exporter_sha256='c9f7cf8dcff44f0472057e9bf52613d93f3ffbc381ad7547a959daa63c5e84ac';\
sql_exporter_sha256='9a41127a493e8bfebfe692bf78c7ed2872a58a3f961ee534d1b0da9ae584aaab';\
sql_exporter_sha256='38e439732bbf6e28ca4a94d7bc3686d3fa1abdb0050773d5617a9efdb9e64d08';\
else\
postgres_exporter_sha256='d1dedea97f56c6d965837bfd1fbb3e35a3b4a4556f8cccee8bd513d8ee086124';\
pgbouncer_exporter_sha256='217c4afd7e6492ae904055bc14fe603552cf9bac458c063407e991d68c519da3';\
sql_exporter_sha256='530e6afc77c043497ed965532c4c9dfa873bc2a4f0b3047fad367715c0081d6a';\
sql_exporter_sha256='11918b00be6e2c3a67564adfdb2414fdcbb15a5db76ea17d1d1a944237a893c6';\
fi\
&& curl -sL https://github.com/prometheus-community/postgres_exporter/releases/download/v0.17.1/postgres_exporter-0.17.1.linux-${TARGETARCH}.tar.gz\
| tar xzf - --strip-components=1 -C.\
&& curl -sL https://github.com/prometheus-community/pgbouncer_exporter/releases/download/v0.10.2/pgbouncer_exporter-0.10.2.linux-${TARGETARCH}.tar.gz\
| tar xzf - --strip-components=1 -C.\
&& curl -sL https://github.com/burningalchemist/sql_exporter/releases/download/0.17.3/sql_exporter-0.17.3.linux-${TARGETARCH}.tar.gz\
&& curl -sL https://github.com/burningalchemist/sql_exporter/releases/download/0.17.0/sql_exporter-0.17.0.linux-${TARGETARCH}.tar.gz\
| tar xzf - --strip-components=1 -C.\
&& echo "${postgres_exporter_sha256} postgres_exporter" | sha256sum -c -\
&& echo "${pgbouncer_exporter_sha256} pgbouncer_exporter" | sha256sum -c -\
@@ -1826,11 +1792,10 @@ RUN rm /usr/local/pgsql/lib/lib*.a
# Preprocess the sql_exporter configuration files
#
#########################################################################################
FROM build-deps AS sql_exporter_preprocessor
FROM $REPOSITORY/$IMAGE:$TAG AS sql_exporter_preprocessor
ARG PG_VERSION
USER nonroot
WORKDIR /home/nonroot
COPY --chown=nonroot compute compute
@@ -1844,27 +1809,12 @@ RUN make PG_VERSION="${PG_VERSION:?}" -C compute
FROM pg-build AS extension-tests
ARG PG_VERSION
# This is required for the PostGIS test
RUN apt-get update && case $DEBIAN_VERSION in \
bullseye) \
apt-get install -y libproj19 libgdal28 time; \
;; \
bookworm) \
apt-get install -y libgdal32 libproj25 time; \
;; \
*) \
echo "Unknown Debian version ${DEBIAN_VERSION}" && exit 1 \
;; \
esac
COPY docker-compose/ext-src/ /ext-src/
COPY --from=pg-build /postgres /postgres
COPY --from=postgis-build /usr/local/pgsql/ /usr/local/pgsql/
COPY --from=postgis-build /ext-src/postgis-src /ext-src/postgis-src
COPY --from=postgis-build /sfcgal/* /usr
#COPY --from=postgis-src /ext-src/ /ext-src/
COPY --from=plv8-src /ext-src/ /ext-src/
COPY --from=h3-pg-src /ext-src/h3-pg-src /ext-src/h3-pg-src
#COPY --from=h3-pg-src /ext-src/ /ext-src/
COPY --from=postgresql-unit-src /ext-src/ /ext-src/
COPY --from=pgvector-src /ext-src/ /ext-src/
COPY --from=pgjwt-src /ext-src/ /ext-src/
@@ -1873,7 +1823,6 @@ COPY --from=pgjwt-src /ext-src/ /ext-src/
COPY --from=pg_graphql-src /ext-src/ /ext-src/
#COPY --from=pg_tiktoken-src /ext-src/ /ext-src/
COPY --from=hypopg-src /ext-src/ /ext-src/
COPY --from=online_advisor-src /ext-src/ /ext-src/
COPY --from=pg_hashids-src /ext-src/ /ext-src/
COPY --from=rum-src /ext-src/ /ext-src/
COPY --from=pgtap-src /ext-src/ /ext-src/
@@ -1903,7 +1852,6 @@ COPY compute/patches/pg_repack.patch /ext-src
RUN cd /ext-src/pg_repack-src && patch -p1 </ext-src/pg_repack.patch && rm -f /ext-src/pg_repack.patch
COPY --chmod=755 docker-compose/run-tests.sh /run-tests.sh
RUN echo /usr/local/pgsql/lib > /etc/ld.so.conf.d/00-neon.conf && /sbin/ldconfig
RUN apt-get update && apt-get install -y libtap-parser-sourcehandler-pgtap-perl jq \
&& apt clean && rm -rf /ext-src/*.tar.gz /ext-src/*.patch /var/lib/apt/lists/*
ENV PATH=/usr/local/pgsql/bin:$PATH

View File

@@ -21,8 +21,6 @@ unix_socket_dir=/tmp/
unix_socket_mode=0777
; required for pgbouncer_exporter
ignore_startup_parameters=extra_float_digits
; pidfile for graceful termination
pidfile=/tmp/pgbouncer.pid
;; Disable connection logging. It produces a lot of logs that no one looks at,
;; and we can get similar log entries from the proxy too. We had incidents in

View File

@@ -1,209 +0,0 @@
{
"$schema": "https://json-schema.org/draft/2020-12/schema",
"title": "Neon Compute Manifest Schema",
"description": "Schema for Neon compute node configuration manifest",
"type": "object",
"properties": {
"pg_settings": {
"type": "object",
"properties": {
"common": {
"type": "object",
"properties": {
"client_connection_check_interval": {
"type": "string",
"description": "Check for client disconnection interval in milliseconds"
},
"effective_io_concurrency": {
"type": "string",
"description": "Effective IO concurrency setting"
},
"fsync": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether to force fsync to disk"
},
"hot_standby": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether hot standby is enabled"
},
"idle_in_transaction_session_timeout": {
"type": "string",
"description": "Timeout for idle transactions in milliseconds"
},
"listen_addresses": {
"type": "string",
"description": "Addresses to listen on"
},
"log_connections": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether to log connections"
},
"log_disconnections": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether to log disconnections"
},
"log_temp_files": {
"type": "string",
"description": "Size threshold for logging temporary files in KB"
},
"log_error_verbosity": {
"type": "string",
"enum": ["terse", "verbose", "default"],
"description": "Error logging verbosity level"
},
"log_min_error_statement": {
"type": "string",
"description": "Minimum error level for statement logging"
},
"maintenance_io_concurrency": {
"type": "string",
"description": "Maintenance IO concurrency setting"
},
"max_connections": {
"type": "string",
"description": "Maximum number of connections"
},
"max_replication_flush_lag": {
"type": "string",
"description": "Maximum replication flush lag"
},
"max_replication_slots": {
"type": "string",
"description": "Maximum number of replication slots"
},
"max_replication_write_lag": {
"type": "string",
"description": "Maximum replication write lag"
},
"max_wal_senders": {
"type": "string",
"description": "Maximum number of WAL senders"
},
"max_wal_size": {
"type": "string",
"description": "Maximum WAL size"
},
"neon.unstable_extensions": {
"type": "string",
"description": "List of unstable extensions"
},
"neon.protocol_version": {
"type": "string",
"description": "Neon protocol version"
},
"password_encryption": {
"type": "string",
"description": "Password encryption method"
},
"restart_after_crash": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether to restart after crash"
},
"superuser_reserved_connections": {
"type": "string",
"description": "Number of reserved connections for superuser"
},
"synchronous_standby_names": {
"type": "string",
"description": "Names of synchronous standby servers"
},
"wal_keep_size": {
"type": "string",
"description": "WAL keep size"
},
"wal_level": {
"type": "string",
"description": "WAL level"
},
"wal_log_hints": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether to log hints in WAL"
},
"wal_sender_timeout": {
"type": "string",
"description": "WAL sender timeout in milliseconds"
}
},
"required": [
"client_connection_check_interval",
"effective_io_concurrency",
"fsync",
"hot_standby",
"idle_in_transaction_session_timeout",
"listen_addresses",
"log_connections",
"log_disconnections",
"log_temp_files",
"log_error_verbosity",
"log_min_error_statement",
"maintenance_io_concurrency",
"max_connections",
"max_replication_flush_lag",
"max_replication_slots",
"max_replication_write_lag",
"max_wal_senders",
"max_wal_size",
"neon.unstable_extensions",
"neon.protocol_version",
"password_encryption",
"restart_after_crash",
"superuser_reserved_connections",
"synchronous_standby_names",
"wal_keep_size",
"wal_level",
"wal_log_hints",
"wal_sender_timeout"
]
},
"replica": {
"type": "object",
"properties": {
"hot_standby": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether hot standby is enabled for replicas"
}
},
"required": ["hot_standby"]
},
"per_version": {
"type": "object",
"patternProperties": {
"^1[4-7]$": {
"type": "object",
"properties": {
"common": {
"type": "object",
"properties": {
"io_combine_limit": {
"type": "string",
"description": "IO combine limit"
}
}
},
"replica": {
"type": "object",
"properties": {
"recovery_prefetch": {
"type": "string",
"enum": ["on", "off"],
"description": "Whether to enable recovery prefetch for PostgreSQL replicas"
}
}
}
}
}
}
}
},
"required": ["common", "replica", "per_version"]
}
},
"required": ["pg_settings"]
}

View File

@@ -1,121 +0,0 @@
pg_settings:
# Common settings for primaries and replicas of all versions.
common:
# Check for client disconnection every 1 minute. By default, Postgres will detect the
# loss of the connection only at the next interaction with the socket, when it waits
# for, receives or sends data, so it will likely waste resources till the end of the
# query execution. There should be no drawbacks in setting this for everyone, so enable
# it by default. If anyone will complain, we can allow editing it.
# https://www.postgresql.org/docs/16/runtime-config-connection.html#GUC-CLIENT-CONNECTION-CHECK-INTERVAL
client_connection_check_interval: "60000" # 1 minute
# ---- IO ----
effective_io_concurrency: "20"
maintenance_io_concurrency: "100"
fsync: "off"
hot_standby: "off"
# We allow users to change this if needed, but by default we
# just don't want to see long-lasting idle transactions, as they
# prevent activity monitor from suspending projects.
idle_in_transaction_session_timeout: "300000" # 5 minutes
listen_addresses: "*"
# --- LOGGING ---- helps investigations
log_connections: "on"
log_disconnections: "on"
# 1GB, unit is KB
log_temp_files: "1048576"
# Disable dumping customer data to logs, both to increase data privacy
# and to reduce the amount the logs.
log_error_verbosity: "terse"
log_min_error_statement: "panic"
max_connections: "100"
# --- WAL ---
# - flush lag is the max amount of WAL that has been generated but not yet stored
# to disk in the page server. A smaller value means less delay after a pageserver
# restart, but if you set it too small you might again need to slow down writes if the
# pageserver cannot flush incoming WAL to disk fast enough. This must be larger
# than the pageserver's checkpoint interval, currently 1 GB! Otherwise you get a
# a deadlock where the compute node refuses to generate more WAL before the
# old WAL has been uploaded to S3, but the pageserver is waiting for more WAL
# to be generated before it is uploaded to S3.
max_replication_flush_lag: "10GB"
max_replication_slots: "10"
# Backpressure configuration:
# - write lag is the max amount of WAL that has been generated by Postgres but not yet
# processed by the page server. Making this smaller reduces the worst case latency
# of a GetPage request, if you request a page that was recently modified. On the other
# hand, if this is too small, the compute node might need to wait on a write if there is a
# hiccup in the network or page server so that the page server has temporarily fallen
# behind.
#
# Previously it was set to 500 MB, but it caused compute being unresponsive under load
# https://github.com/neondatabase/neon/issues/2028
max_replication_write_lag: "500MB"
max_wal_senders: "10"
# A Postgres checkpoint is cheap in storage, as doesn't involve any significant amount
# of real I/O. Only the SLRU buffers and some other small files are flushed to disk.
# However, as long as we have full_page_writes=on, page updates after a checkpoint
# include full-page images which bloats the WAL. So may want to bump max_wal_size to
# reduce the WAL bloating, but at the same it will increase pg_wal directory size on
# compute and can lead to out of disk error on k8s nodes.
max_wal_size: "1024"
wal_keep_size: "0"
wal_level: "replica"
# Reduce amount of WAL generated by default.
wal_log_hints: "off"
# - without wal_sender_timeout set we don't get feedback messages,
# required for backpressure.
wal_sender_timeout: "10000"
# We have some experimental extensions, which we don't want users to install unconsciously.
# To install them, users would need to set the `neon.allow_unstable_extensions` setting.
# There are two of them currently:
# - `pgrag` - https://github.com/neondatabase-labs/pgrag - extension is actually called just `rag`,
# and two dependencies:
# - `rag_bge_small_en_v15`
# - `rag_jina_reranker_v1_tiny_en`
# - `pg_mooncake` - https://github.com/Mooncake-Labs/pg_mooncake/
neon.unstable_extensions: "rag,rag_bge_small_en_v15,rag_jina_reranker_v1_tiny_en,pg_mooncake,anon"
neon.protocol_version: "3"
password_encryption: "scram-sha-256"
# This is important to prevent Postgres from trying to perform
# a local WAL redo after backend crash. It should exit and let
# the systemd or k8s to do a fresh startup with compute_ctl.
restart_after_crash: "off"
# By default 3. We have the following persistent connections in the VM:
# * compute_activity_monitor (from compute_ctl)
# * postgres-exporter (metrics collector; it has 2 connections)
# * sql_exporter (metrics collector; we have 2 instances [1 for us & users; 1 for autoscaling])
# * vm-monitor (to query & change file cache size)
# i.e. total of 6. Let's reserve 7, so there's still at least one left over.
superuser_reserved_connections: "7"
synchronous_standby_names: "walproposer"
replica:
hot_standby: "on"
per_version:
17:
common:
# PostgreSQL 17 has a new IO system called "read stream", which can combine IOs up to some
# size. It still has some issues with readahead, though, so we default to disabled/
# "no combining of IOs" to make sure we get the maximum prefetch depth.
# See also: https://github.com/neondatabase/neon/pull/9860
io_combine_limit: "1"
replica:
# prefetching of blocks referenced in WAL doesn't make sense for us
# Neon hot standby ignores pages that are not in the shared_buffers
recovery_prefetch: "off"
16:
common: {}
replica:
# prefetching of blocks referenced in WAL doesn't make sense for us
# Neon hot standby ignores pages that are not in the shared_buffers
recovery_prefetch: "off"
15:
common: {}
replica:
# prefetching of blocks referenced in WAL doesn't make sense for us
# Neon hot standby ignores pages that are not in the shared_buffers
recovery_prefetch: "off"
14:
common: {}
replica: {}

View File

@@ -1,37 +0,0 @@
{
"name": "neon-compute",
"lockfileVersion": 3,
"requires": true,
"packages": {
"": {
"name": "neon-compute",
"dependencies": {
"@sourcemeta/jsonschema": "9.3.4"
}
},
"node_modules/@sourcemeta/jsonschema": {
"version": "9.3.4",
"resolved": "https://registry.npmjs.org/@sourcemeta/jsonschema/-/jsonschema-9.3.4.tgz",
"integrity": "sha512-hkujfkZAIGXUs4U//We9faZW8LZ4/H9LqagRYsFSulH/VLcKPNhZyCTGg7AhORuzm27zqENvKpnX4g2FzudYFw==",
"cpu": [
"x64",
"arm64"
],
"license": "AGPL-3.0",
"os": [
"darwin",
"linux",
"win32"
],
"bin": {
"jsonschema": "cli.js"
},
"engines": {
"node": ">=16"
},
"funding": {
"url": "https://github.com/sponsors/sourcemeta"
}
}
}
}

View File

@@ -1,7 +0,0 @@
{
"name": "neon-compute",
"private": true,
"dependencies": {
"@sourcemeta/jsonschema": "9.3.4"
}
}

View File

@@ -1,15 +0,0 @@
diff --git a/cmake/deps.txt b/cmake/deps.txt
index d213b09034..229de2ebf0 100644
--- a/cmake/deps.txt
+++ b/cmake/deps.txt
@@ -22,7 +22,9 @@ dlpack;https://github.com/dmlc/dlpack/archive/refs/tags/v0.6.zip;4d565dd2e5b3132
# it contains changes on top of 3.4.0 which are required to fix build issues.
# Until the 3.4.1 release this is the best option we have.
# Issue link: https://gitlab.com/libeigen/eigen/-/issues/2744
-eigen;https://gitlab.com/libeigen/eigen/-/archive/e7248b26a1ed53fa030c5c459f7ea095dfd276ac/eigen-e7248b26a1ed53fa030c5c459f7ea095dfd276ac.zip;be8be39fdbc6e60e94fa7870b280707069b5b81a
+# Moved to github mirror to avoid gitlab issues.Add commentMore actions
+# Issue link: https://github.com/bazelbuild/bazel-central-registry/issues/4355
+eigen;https://github.com/eigen-mirror/eigen/archive/e7248b26a1ed53fa030c5c459f7ea095dfd276ac/eigen-e7248b26a1ed53fa030c5c459f7ea095dfd276ac.zip;61418a349000ba7744a3ad03cf5071f22ebf860a
flatbuffers;https://github.com/google/flatbuffers/archive/refs/tags/v23.5.26.zip;59422c3b5e573dd192fead2834d25951f1c1670c
fp16;https://github.com/Maratyszcza/FP16/archive/0a92994d729ff76a58f692d3028ca1b64b145d91.zip;b985f6985a05a1c03ff1bb71190f66d8f98a1494
fxdiv;https://github.com/Maratyszcza/FXdiv/archive/63058eff77e11aa15bf531df5dd34395ec3017c8.zip;a5658f4036402dbca7cebee32be57fb8149811e1

View File

@@ -7,7 +7,7 @@ index 255e616..1c6edb7 100644
RelationGetRelationName(index));
+#ifdef NEON_SMGR
+ smgr_start_unlogged_build(RelationGetSmgr(index));
+ smgr_start_unlogged_build(index->rd_smgr);
+#endif
+
initRumState(&buildstate.rumstate, index);
@@ -18,7 +18,7 @@ index 255e616..1c6edb7 100644
rumUpdateStats(index, &buildstate.buildStats, buildstate.rumstate.isBuild);
+#ifdef NEON_SMGR
+ smgr_finish_unlogged_build_phase_1(RelationGetSmgr(index));
+ smgr_finish_unlogged_build_phase_1(index->rd_smgr);
+#endif
+
/*
@@ -29,7 +29,7 @@ index 255e616..1c6edb7 100644
}
+#ifdef NEON_SMGR
+ smgr_end_unlogged_build(RelationGetSmgr(index));
+ smgr_end_unlogged_build(index->rd_smgr);
+#endif
+
/*

View File

@@ -38,7 +38,6 @@ once_cell.workspace = true
opentelemetry.workspace = true
opentelemetry_sdk.workspace = true
p256 = { version = "0.13", features = ["pem"] }
pageserver_page_api.workspace = true
postgres.workspace = true
regex.workspace = true
reqwest = { workspace = true, features = ["json"] }
@@ -54,7 +53,6 @@ tokio = { workspace = true, features = ["rt", "rt-multi-thread"] }
tokio-postgres.workspace = true
tokio-util.workspace = true
tokio-stream.workspace = true
tonic.workspace = true
tower-otel.workspace = true
tracing.workspace = true
tracing-opentelemetry.workspace = true

View File

@@ -40,7 +40,7 @@ use std::sync::mpsc;
use std::thread;
use std::time::Duration;
use anyhow::{Context, Result, bail};
use anyhow::{Context, Result};
use clap::Parser;
use compute_api::responses::ComputeConfig;
use compute_tools::compute::{
@@ -57,15 +57,31 @@ use tracing::{error, info};
use url::Url;
use utils::failpoint_support;
#[derive(Debug, Parser)]
// Compatibility hack: if the control plane specified any remote-ext-config
// use the default value for extension storage proxy gateway.
// Remove this once the control plane is updated to pass the gateway URL
fn parse_remote_ext_base_url(arg: &str) -> Result<String> {
const FALLBACK_PG_EXT_GATEWAY_BASE_URL: &str =
"http://pg-ext-s3-gateway.pg-ext-s3-gateway.svc.cluster.local";
Ok(if arg.starts_with("http") {
arg
} else {
FALLBACK_PG_EXT_GATEWAY_BASE_URL
}
.to_owned())
}
#[derive(Parser)]
#[command(rename_all = "kebab-case")]
struct Cli {
#[arg(short = 'b', long, default_value = "postgres", env = "POSTGRES_PATH")]
pub pgbin: String,
/// The base URL for the remote extension storage proxy gateway.
#[arg(short = 'r', long, value_parser = Self::parse_remote_ext_base_url)]
pub remote_ext_base_url: Option<Url>,
/// Should be in the form of `http(s)://<gateway-hostname>[:<port>]`.
#[arg(short = 'r', long, value_parser = parse_remote_ext_base_url, alias = "remote-ext-config")]
pub remote_ext_base_url: Option<String>,
/// The port to bind the external listening HTTP server to. Clients running
/// outside the compute will talk to the compute through this port. Keep
@@ -120,33 +136,6 @@ struct Cli {
requires = "compute-id"
)]
pub control_plane_uri: Option<String>,
/// Interval in seconds for collecting installed extensions statistics
#[arg(long, default_value = "3600")]
pub installed_extensions_collection_interval: u64,
/// Run in development mode, skipping VM-specific operations like process termination
#[arg(long, action = clap::ArgAction::SetTrue)]
pub dev: bool,
}
impl Cli {
/// Parse a URL from an argument. By default, this isn't necessary, but we
/// want to do some sanity checking.
fn parse_remote_ext_base_url(value: &str) -> Result<Url> {
// Remove extra trailing slashes, and add one. We use Url::join() later
// when downloading remote extensions. If the base URL is something like
// http://example.com/pg-ext-s3-gateway, and join() is called with
// something like "xyz", the resulting URL is http://example.com/xyz.
let value = value.trim_end_matches('/').to_owned() + "/";
let url = Url::parse(&value)?;
if url.query_pairs().count() != 0 {
bail!("parameters detected in remote extensions base URL")
}
Ok(url)
}
}
fn main() -> Result<()> {
@@ -163,7 +152,7 @@ fn main() -> Result<()> {
.build()?;
let _rt_guard = runtime.enter();
runtime.block_on(init(cli.dev))?;
runtime.block_on(init())?;
// enable core dumping for all child processes
setrlimit(Resource::CORE, rlimit::INFINITY, rlimit::INFINITY)?;
@@ -190,7 +179,6 @@ fn main() -> Result<()> {
cgroup: cli.cgroup,
#[cfg(target_os = "linux")]
vm_monitor_addr: cli.vm_monitor_addr,
installed_extensions_collection_interval: cli.installed_extensions_collection_interval,
},
config,
)?;
@@ -202,13 +190,13 @@ fn main() -> Result<()> {
deinit_and_exit(exit_code);
}
async fn init(dev_mode: bool) -> Result<()> {
async fn init() -> Result<()> {
init_tracing_and_logging(DEFAULT_LOG_LEVEL).await?;
let mut signals = Signals::new([SIGINT, SIGTERM, SIGQUIT])?;
thread::spawn(move || {
for sig in signals.forever() {
handle_exit_signal(sig, dev_mode);
handle_exit_signal(sig);
}
});
@@ -267,16 +255,15 @@ fn deinit_and_exit(exit_code: Option<i32>) -> ! {
/// When compute_ctl is killed, send also termination signal to sync-safekeepers
/// to prevent leakage. TODO: it is better to convert compute_ctl to async and
/// wait for termination which would be easy then.
fn handle_exit_signal(sig: i32, dev_mode: bool) {
fn handle_exit_signal(sig: i32) {
info!("received {sig} termination signal");
forward_termination_signal(dev_mode);
forward_termination_signal();
exit(1);
}
#[cfg(test)]
mod test {
use clap::{CommandFactory, Parser};
use url::Url;
use clap::CommandFactory;
use super::Cli;
@@ -286,41 +273,16 @@ mod test {
}
#[test]
fn verify_remote_ext_base_url() {
let cli = Cli::parse_from([
"compute_ctl",
"--pgdata=test",
"--connstr=test",
"--compute-id=test",
"--remote-ext-base-url",
"https://example.com/subpath",
]);
assert_eq!(
cli.remote_ext_base_url.unwrap(),
Url::parse("https://example.com/subpath/").unwrap()
);
fn parse_pg_ext_gateway_base_url() {
let arg = "http://pg-ext-s3-gateway2";
let result = super::parse_remote_ext_base_url(arg).unwrap();
assert_eq!(result, arg);
let cli = Cli::parse_from([
"compute_ctl",
"--pgdata=test",
"--connstr=test",
"--compute-id=test",
"--remote-ext-base-url",
"https://example.com//",
]);
let arg = "pg-ext-s3-gateway";
let result = super::parse_remote_ext_base_url(arg).unwrap();
assert_eq!(
cli.remote_ext_base_url.unwrap(),
Url::parse("https://example.com").unwrap()
result,
"http://pg-ext-s3-gateway.pg-ext-s3-gateway.svc.cluster.local"
);
Cli::try_parse_from([
"compute_ctl",
"--pgdata=test",
"--connstr=test",
"--compute-id=test",
"--remote-ext-base-url",
"https://example.com?hello=world",
])
.expect_err("URL parameters are not allowed");
}
}

View File

@@ -339,8 +339,6 @@ async fn run_dump_restore(
destination_connstring: String,
) -> Result<(), anyhow::Error> {
let dumpdir = workdir.join("dumpdir");
let num_jobs = num_cpus::get().to_string();
info!("using {num_jobs} jobs for dump/restore");
let common_args = [
// schema mapping (prob suffices to specify them on one side)
@@ -356,7 +354,7 @@ async fn run_dump_restore(
"directory".to_string(),
// concurrency
"--jobs".to_string(),
num_jobs,
num_cpus::get().to_string(),
// progress updates
"--verbose".to_string(),
];

View File

@@ -1,9 +1,9 @@
use anyhow::{Context, Result, anyhow};
use anyhow::{Context, Result};
use chrono::{DateTime, Utc};
use compute_api::privilege::Privilege;
use compute_api::responses::{
ComputeConfig, ComputeCtlConfig, ComputeMetrics, ComputeStatus, LfcOffloadState,
LfcPrewarmState, TlsConfig,
LfcPrewarmState,
};
use compute_api::spec::{
ComputeAudit, ComputeFeature, ComputeMode, ComputeSpec, ExtVersion, PgIdent,
@@ -15,7 +15,6 @@ use itertools::Itertools;
use nix::sys::signal::{Signal, kill};
use nix::unistd::Pid;
use once_cell::sync::Lazy;
use pageserver_page_api as page_api;
use postgres;
use postgres::NoTls;
use postgres::error::SqlState;
@@ -31,13 +30,10 @@ use std::sync::{Arc, Condvar, Mutex, RwLock};
use std::time::{Duration, Instant};
use std::{env, fs};
use tokio::spawn;
use tokio_util::io::StreamReader;
use tracing::{Instrument, debug, error, info, instrument, warn};
use url::Url;
use utils::id::{TenantId, TimelineId};
use utils::lsn::Lsn;
use utils::measured_stream::MeasuredReader;
use utils::pid_file;
use crate::configurator::launch_configurator;
use crate::disk_quota::set_disk_quota;
@@ -47,7 +43,6 @@ use crate::lsn_lease::launch_lsn_lease_bg_task_for_static;
use crate::metrics::COMPUTE_CTL_UP;
use crate::monitor::launch_monitor;
use crate::pg_helpers::*;
use crate::pgbouncer::*;
use crate::rsyslog::{
PostgresLogsRsyslogConfig, configure_audit_rsyslog, configure_postgres_logs_export,
launch_pgaudit_gc,
@@ -101,10 +96,7 @@ pub struct ComputeNodeParams {
pub internal_http_port: u16,
/// the address of extension storage proxy gateway
pub remote_ext_base_url: Option<Url>,
/// Interval for installed extensions collection
pub installed_extensions_collection_interval: u64,
pub remote_ext_base_url: Option<String>,
}
/// Compute node info shared across several `compute_ctl` threads.
@@ -165,10 +157,6 @@ pub struct ComputeState {
pub lfc_prewarm_state: LfcPrewarmState,
pub lfc_offload_state: LfcOffloadState,
/// WAL flush LSN that is set after terminating Postgres and syncing safekeepers if
/// mode == ComputeMode::Primary. None otherwise
pub terminate_flush_lsn: Option<Lsn>,
pub metrics: ComputeMetrics,
}
@@ -184,7 +172,6 @@ impl ComputeState {
metrics: ComputeMetrics::default(),
lfc_prewarm_state: LfcPrewarmState::default(),
lfc_offload_state: LfcOffloadState::default(),
terminate_flush_lsn: None,
}
}
@@ -224,46 +211,6 @@ pub struct ParsedSpec {
pub endpoint_storage_token: Option<String>,
}
impl ParsedSpec {
pub fn validate(&self) -> Result<(), String> {
// Only Primary nodes are using safekeeper_connstrings, and at the moment
// this method only validates that part of the specs.
if self.spec.mode != ComputeMode::Primary {
return Ok(());
}
// While it seems like a good idea to check for an odd number of entries in
// the safekeepers connection string, changes to the list of safekeepers might
// incur appending a new server to a list of 3, in which case a list of 4
// entries is okay in production.
//
// Still we want unique entries, and at least one entry in the vector
if self.safekeeper_connstrings.is_empty() {
return Err(String::from("safekeeper_connstrings is empty"));
}
// check for uniqueness of the connection strings in the set
let mut connstrings = self.safekeeper_connstrings.clone();
connstrings.sort();
let mut previous = &connstrings[0];
for current in connstrings.iter().skip(1) {
// duplicate entry?
if current == previous {
return Err(format!(
"duplicate entry in safekeeper_connstrings: {}!",
current,
));
}
previous = current;
}
Ok(())
}
}
impl TryFrom<ComputeSpec> for ParsedSpec {
type Error = String;
fn try_from(spec: ComputeSpec) -> Result<Self, String> {
@@ -293,7 +240,6 @@ impl TryFrom<ComputeSpec> for ParsedSpec {
} else {
spec.safekeeper_connstrings.clone()
};
let storage_auth_token = spec.storage_auth_token.clone();
let tenant_id: TenantId = if let Some(tenant_id) = spec.tenant_id {
tenant_id
@@ -328,7 +274,7 @@ impl TryFrom<ComputeSpec> for ParsedSpec {
.clone()
.or_else(|| spec.cluster.settings.find("neon.endpoint_storage_token"));
let res = ParsedSpec {
Ok(ParsedSpec {
spec,
pageserver_connstr,
safekeeper_connstrings,
@@ -337,11 +283,7 @@ impl TryFrom<ComputeSpec> for ParsedSpec {
timeline_id,
endpoint_storage_addr,
endpoint_storage_token,
};
// Now check validity of the parsed specification
res.validate()?;
Ok(res)
})
}
}
@@ -408,11 +350,14 @@ impl ComputeNode {
// that can affect `compute_ctl` and prevent it from properly configuring the database schema.
// Unset them via connection string options before connecting to the database.
// N.B. keep it in sync with `ZENITH_OPTIONS` in `get_maintenance_client()`.
//
// TODO(ololobus): we currently pass `-c default_transaction_read_only=off` from control plane
// as well. After rolling out this code, we can remove this parameter from control plane.
// In the meantime, double-passing is fine, the last value is applied.
// See: <https://github.com/neondatabase/cloud/blob/133dd8c4dbbba40edfbad475bf6a45073ca63faf/goapp/controlplane/internal/pkg/compute/provisioner/provisioner_common.go#L70>
const EXTRA_OPTIONS: &str = "-c role=cloud_admin -c default_transaction_read_only=off -c search_path=public -c statement_timeout=0";
let options = match conn_conf.get_options() {
// Allow the control plane to override any options set by the
// compute
Some(options) => format!("{} {}", EXTRA_OPTIONS, options),
Some(options) => format!("{} {}", options, EXTRA_OPTIONS),
None => EXTRA_OPTIONS.to_string(),
};
conn_conf.options(&options);
@@ -420,7 +365,7 @@ impl ComputeNode {
let mut new_state = ComputeState::new();
if let Some(spec) = config.spec {
let pspec = ParsedSpec::try_from(spec).map_err(|msg| anyhow!(msg))?;
let pspec = ParsedSpec::try_from(spec).map_err(|msg| anyhow::anyhow!(msg))?;
new_state.pspec = Some(pspec);
}
@@ -447,7 +392,7 @@ impl ComputeNode {
// because QEMU will already have its memory allocated from the host, and
// the necessary binaries will already be cached.
if cli_spec.is_none() {
this.prewarm_postgres_vm_memory()?;
this.prewarm_postgres()?;
}
// Set the up metric with Empty status before starting the HTTP server.
@@ -540,21 +485,12 @@ impl ComputeNode {
// Reap the postgres process
delay_exit |= this.cleanup_after_postgres_exit()?;
// /terminate returns LSN. If we don't sleep at all, connection will break and we
// won't get result. If we sleep too much, tests will take significantly longer
// and Github Action run will error out
let sleep_duration = if delay_exit {
Duration::from_secs(30)
} else {
Duration::from_millis(300)
};
// If launch failed, keep serving HTTP requests for a while, so the cloud
// control plane can get the actual error.
if delay_exit {
info!("giving control plane 30s to collect the error before shutdown");
std::thread::sleep(Duration::from_secs(30));
}
std::thread::sleep(sleep_duration);
Ok(exit_code)
}
@@ -663,8 +599,6 @@ impl ComputeNode {
});
}
let tls_config = self.tls_config(&pspec.spec);
// If there are any remote extensions in shared_preload_libraries, start downloading them
if pspec.spec.remote_extensions.is_some() {
let (this, spec) = (self.clone(), pspec.spec.clone());
@@ -721,7 +655,7 @@ impl ComputeNode {
info!("tuning pgbouncer");
let pgbouncer_settings = pgbouncer_settings.clone();
let tls_config = tls_config.clone();
let tls_config = self.compute_ctl_config.tls.clone();
// Spawn a background task to do the tuning,
// so that we don't block the main thread that starts Postgres.
@@ -740,10 +674,7 @@ impl ComputeNode {
// Spawn a background task to do the configuration,
// so that we don't block the main thread that starts Postgres.
let mut local_proxy = local_proxy.clone();
local_proxy.tls = tls_config.clone();
let local_proxy = local_proxy.clone();
let _handle = tokio::spawn(async move {
if let Err(err) = local_proxy::configure(&local_proxy) {
error!("error while configuring local_proxy: {err:?}");
@@ -764,18 +695,25 @@ impl ComputeNode {
let log_directory_path = Path::new(&self.params.pgdata).join("log");
let log_directory_path = log_directory_path.to_string_lossy().to_string();
// Add project_id,endpoint_id to identify the logs.
// Add project_id,endpoint_id tag to identify the logs.
//
// These ids are passed from cplane,
let endpoint_id = pspec.spec.endpoint_id.as_deref().unwrap_or("");
let project_id = pspec.spec.project_id.as_deref().unwrap_or("");
// for backwards compatibility (old computes that don't have them),
// we set them to None.
// TODO: Clean up this code when all computes have them.
let tag: Option<String> = match (
pspec.spec.project_id.as_deref(),
pspec.spec.endpoint_id.as_deref(),
) {
(Some(project_id), Some(endpoint_id)) => {
Some(format!("{project_id}/{endpoint_id}"))
}
(Some(project_id), None) => Some(format!("{project_id}/None")),
(None, Some(endpoint_id)) => Some(format!("None,{endpoint_id}")),
(None, None) => None,
};
configure_audit_rsyslog(
log_directory_path.clone(),
endpoint_id,
project_id,
&remote_endpoint,
)?;
configure_audit_rsyslog(log_directory_path.clone(), tag, &remote_endpoint)?;
// Launch a background task to clean up the audit logs
launch_pgaudit_gc(log_directory_path);
@@ -811,7 +749,17 @@ impl ComputeNode {
let conf = self.get_tokio_conn_conf(None);
tokio::task::spawn(async {
let _ = installed_extensions(conf).await;
let res = get_installed_extensions(conf).await;
match res {
Ok(extensions) => {
info!(
"[NEON_EXT_STAT] {}",
serde_json::to_string(&extensions)
.expect("failed to serialize extensions list")
);
}
Err(err) => error!("could not get installed extensions: {err:?}"),
}
});
}
@@ -841,11 +789,8 @@ impl ComputeNode {
// Log metrics so that we can search for slow operations in logs
info!(?metrics, postmaster_pid = %postmaster_pid, "compute start finished");
// Spawn the extension stats background task
self.spawn_extension_stats_task();
if pspec.spec.autoprewarm {
self.prewarm_lfc(None);
if pspec.spec.prewarm_lfc_on_startup {
self.prewarm_lfc();
}
Ok(())
}
@@ -926,25 +871,20 @@ impl ComputeNode {
// Maybe sync safekeepers again, to speed up next startup
let compute_state = self.state.lock().unwrap().clone();
let pspec = compute_state.pspec.as_ref().expect("spec must be set");
let lsn = if matches!(pspec.spec.mode, compute_api::spec::ComputeMode::Primary) {
if matches!(pspec.spec.mode, compute_api::spec::ComputeMode::Primary) {
info!("syncing safekeepers on shutdown");
let storage_auth_token = pspec.storage_auth_token.clone();
let lsn = self.sync_safekeepers(storage_auth_token)?;
info!(%lsn, "synced safekeepers");
Some(lsn)
} else {
info!("not primary, not syncing safekeepers");
None
};
info!("synced safekeepers at lsn {lsn}");
}
let mut delay_exit = false;
let mut state = self.state.lock().unwrap();
state.terminate_flush_lsn = lsn;
if let ComputeStatus::TerminationPending { mode } = state.status {
if state.status == ComputeStatus::TerminationPending {
state.status = ComputeStatus::Terminated;
self.state_changed.notify_all();
// we were asked to terminate gracefully, don't exit to avoid restart
delay_exit = mode == compute_api::responses::TerminateMode::Fast
delay_exit = true
}
drop(state);
@@ -1006,75 +946,6 @@ impl ComputeNode {
#[instrument(skip_all, fields(%lsn))]
fn try_get_basebackup(&self, compute_state: &ComputeState, lsn: Lsn) -> Result<()> {
let spec = compute_state.pspec.as_ref().expect("spec must be set");
let shard0_connstr = spec.pageserver_connstr.split(',').next().unwrap();
match Url::parse(shard0_connstr)?.scheme() {
"postgres" | "postgresql" => self.try_get_basebackup_libpq(spec, lsn),
"grpc" => self.try_get_basebackup_grpc(spec, lsn),
scheme => return Err(anyhow!("unknown URL scheme {scheme}")),
}
}
fn try_get_basebackup_grpc(&self, spec: &ParsedSpec, lsn: Lsn) -> Result<()> {
let start_time = Instant::now();
let shard0_connstr = spec
.pageserver_connstr
.split(',')
.next()
.unwrap()
.to_string();
let chunks = tokio::runtime::Handle::current().block_on(async move {
let mut client = page_api::proto::PageServiceClient::connect(shard0_connstr).await?;
let req = page_api::proto::GetBaseBackupRequest {
lsn: lsn.0,
replica: false, // TODO: handle replicas, with LSN 0
full: false,
};
let mut req = tonic::Request::new(req);
let metadata = req.metadata_mut();
metadata.insert("neon-tenant-id", spec.tenant_id.to_string().parse()?);
metadata.insert("neon-timeline-id", spec.timeline_id.to_string().parse()?);
metadata.insert("neon-shard-id", "0000".to_string().parse()?); // TODO: shard count
if let Some(auth) = spec.storage_auth_token.as_ref() {
metadata.insert("authorization", format!("Bearer {auth}").parse()?);
}
let chunks = client.get_base_backup(req).await?.into_inner();
anyhow::Ok(chunks)
})?;
let pageserver_connect_micros = start_time.elapsed().as_micros() as u64;
// Convert the chunks stream into an AsyncRead
let stream_reader = StreamReader::new(
chunks.map(|chunk| chunk.map(|c| c.chunk).map_err(std::io::Error::other)),
);
// Wrap the AsyncRead into a blocking reader for compatibility with tar::Archive
let reader = tokio_util::io::SyncIoBridge::new(stream_reader);
let mut measured_reader = MeasuredReader::new(reader);
let mut bufreader = std::io::BufReader::new(&mut measured_reader);
// Read the archive directly from the `CopyOutReader`
//
// Set `ignore_zeros` so that unpack() reads all the Copy data and
// doesn't stop at the end-of-archive marker. Otherwise, if the server
// sends an Error after finishing the tarball, we will not notice it.
let mut ar = tar::Archive::new(&mut bufreader);
ar.set_ignore_zeros(true);
ar.unpack(&self.params.pgdata)?;
// Report metrics
let mut state = self.state.lock().unwrap();
state.metrics.pageserver_connect_micros = pageserver_connect_micros;
state.metrics.basebackup_bytes = measured_reader.get_byte_count() as u64;
state.metrics.basebackup_ms = start_time.elapsed().as_millis() as u64;
Ok(())
}
fn try_get_basebackup_libpq(&self, spec: &ParsedSpec, lsn: Lsn) -> Result<()> {
let start_time = Instant::now();
let shard0_connstr = spec.pageserver_connstr.split(',').next().unwrap();
@@ -1090,10 +961,12 @@ impl ComputeNode {
}
config.application_name("compute_ctl");
config.options(&format!(
"-c neon.compute_mode={}",
spec.spec.mode.to_type_str()
));
if let Some(spec) = &compute_state.pspec {
config.options(&format!(
"-c neon.compute_mode={}",
spec.spec.mode.to_type_str()
));
}
// Connect to pageserver
let mut client = config.connect(NoTls)?;
@@ -1167,7 +1040,10 @@ impl ComputeNode {
return result;
}
Err(ref e) if attempts < max_attempts => {
warn!("Failed to get basebackup: {e:?} (attempt {attempts}/{max_attempts})");
warn!(
"Failed to get basebackup: {} (attempt {}/{})",
e, attempts, max_attempts
);
std::thread::sleep(std::time::Duration::from_millis(retry_period_ms as u64));
retry_period_ms *= 1.5;
}
@@ -1339,15 +1215,13 @@ impl ComputeNode {
let spec = &pspec.spec;
let pgdata_path = Path::new(&self.params.pgdata);
let tls_config = self.tls_config(&pspec.spec);
// Remove/create an empty pgdata directory and put configuration there.
self.create_pgdata()?;
config::write_postgres_conf(
pgdata_path,
&pspec.spec,
self.params.internal_http_port,
tls_config,
&self.compute_ctl_config.tls,
)?;
// Syncing safekeepers is only safe with primary nodes: if a primary
@@ -1443,8 +1317,8 @@ impl ComputeNode {
}
/// Start and stop a postgres process to warm up the VM for startup.
pub fn prewarm_postgres_vm_memory(&self) -> Result<()> {
info!("prewarming VM memory");
pub fn prewarm_postgres(&self) -> Result<()> {
info!("prewarming");
// Create pgdata
let pgdata = &format!("{}.warmup", self.params.pgdata);
@@ -1486,7 +1360,7 @@ impl ComputeNode {
kill(pm_pid, Signal::SIGQUIT)?;
info!("sent SIGQUIT signal");
pg.wait()?;
info!("done prewarming vm memory");
info!("done prewarming");
// clean up
let _ok = fs::remove_dir_all(pgdata);
@@ -1672,22 +1546,14 @@ impl ComputeNode {
.clone(),
);
let mut tls_config = None::<TlsConfig>;
if spec.features.contains(&ComputeFeature::TlsExperimental) {
tls_config = self.compute_ctl_config.tls.clone();
}
let max_concurrent_connections = self.max_service_connections(compute_state, &spec);
// Merge-apply spec & changes to PostgreSQL state.
self.apply_spec_sql(spec.clone(), conf.clone(), max_concurrent_connections)?;
if let Some(local_proxy) = &spec.clone().local_proxy_config {
let mut local_proxy = local_proxy.clone();
local_proxy.tls = tls_config.clone();
info!("configuring local_proxy");
local_proxy::configure(&local_proxy).context("apply_config local_proxy")?;
local_proxy::configure(local_proxy).context("apply_config local_proxy")?;
}
// Run migrations separately to not hold up cold starts
@@ -1739,13 +1605,11 @@ impl ComputeNode {
pub fn reconfigure(&self) -> Result<()> {
let spec = self.state.lock().unwrap().pspec.clone().unwrap().spec;
let tls_config = self.tls_config(&spec);
if let Some(ref pgbouncer_settings) = spec.pgbouncer_settings {
info!("tuning pgbouncer");
let pgbouncer_settings = pgbouncer_settings.clone();
let tls_config = tls_config.clone();
let tls_config = self.compute_ctl_config.tls.clone();
// Spawn a background task to do the tuning,
// so that we don't block the main thread that starts Postgres.
@@ -1763,7 +1627,7 @@ impl ComputeNode {
// Spawn a background task to do the configuration,
// so that we don't block the main thread that starts Postgres.
let mut local_proxy = local_proxy.clone();
local_proxy.tls = tls_config.clone();
local_proxy.tls = self.compute_ctl_config.tls.clone();
tokio::spawn(async move {
if let Err(err) = local_proxy::configure(&local_proxy) {
error!("error while configuring local_proxy: {err:?}");
@@ -1781,7 +1645,7 @@ impl ComputeNode {
pgdata_path,
&spec,
self.params.internal_http_port,
tls_config,
&self.compute_ctl_config.tls,
)?;
if !spec.skip_pg_catalog_updates {
@@ -1879,7 +1743,7 @@ impl ComputeNode {
// exit loop
ComputeStatus::Failed
| ComputeStatus::TerminationPending { .. }
| ComputeStatus::TerminationPending
| ComputeStatus::Terminated => break 'cert_update,
// wait
@@ -1901,14 +1765,6 @@ impl ComputeNode {
}
}
pub fn tls_config(&self, spec: &ComputeSpec) -> &Option<TlsConfig> {
if spec.features.contains(&ComputeFeature::TlsExperimental) {
&self.compute_ctl_config.tls
} else {
&None::<TlsConfig>
}
}
/// Update the `last_active` in the shared state, but ensure that it's a more recent one.
pub fn update_last_active(&self, last_active: Option<DateTime<Utc>>) {
let mut state = self.state.lock().unwrap();
@@ -2045,7 +1901,7 @@ LIMIT 100",
self.params
.remote_ext_base_url
.as_ref()
.ok_or(DownloadError::BadInput(anyhow!(
.ok_or(DownloadError::BadInput(anyhow::anyhow!(
"Remote extensions storage is not configured",
)))?;
@@ -2241,7 +2097,7 @@ LIMIT 100",
let remote_extensions = spec
.remote_extensions
.as_ref()
.ok_or(anyhow!("Remote extensions are not configured"))?;
.ok_or(anyhow::anyhow!("Remote extensions are not configured"))?;
info!("parse shared_preload_libraries from spec.cluster.settings");
let mut libs_vec = Vec::new();
@@ -2343,105 +2199,14 @@ LIMIT 100",
info!("Pageserver config changed");
}
}
pub fn spawn_extension_stats_task(&self) {
let conf = self.tokio_conn_conf.clone();
let installed_extensions_collection_interval =
self.params.installed_extensions_collection_interval;
tokio::spawn(async move {
// An initial sleep is added to ensure that two collections don't happen at the same time.
// The first collection happens during compute startup.
tokio::time::sleep(tokio::time::Duration::from_secs(
installed_extensions_collection_interval,
))
.await;
let mut interval = tokio::time::interval(tokio::time::Duration::from_secs(
installed_extensions_collection_interval,
));
loop {
interval.tick().await;
let _ = installed_extensions(conf.clone()).await;
}
});
}
}
pub async fn installed_extensions(conf: tokio_postgres::Config) -> Result<()> {
let res = get_installed_extensions(conf).await;
match res {
Ok(extensions) => {
info!(
"[NEON_EXT_STAT] {}",
serde_json::to_string(&extensions).expect("failed to serialize extensions list")
);
}
Err(err) => error!("could not get installed extensions: {err:?}"),
}
Ok(())
}
pub fn forward_termination_signal(dev_mode: bool) {
pub fn forward_termination_signal() {
let ss_pid = SYNC_SAFEKEEPERS_PID.load(Ordering::SeqCst);
if ss_pid != 0 {
let ss_pid = nix::unistd::Pid::from_raw(ss_pid as i32);
kill(ss_pid, Signal::SIGTERM).ok();
}
if !dev_mode {
// Terminate pgbouncer with SIGKILL
match pid_file::read(PGBOUNCER_PIDFILE.into()) {
Ok(pid_file::PidFileRead::LockedByOtherProcess(pid)) => {
info!("sending SIGKILL to pgbouncer process pid: {}", pid);
if let Err(e) = kill(pid, Signal::SIGKILL) {
error!("failed to terminate pgbouncer: {}", e);
}
}
// pgbouncer does not lock the pid file, so we read and kill the process directly
Ok(pid_file::PidFileRead::NotHeldByAnyProcess(_)) => {
if let Ok(pid_str) = std::fs::read_to_string(PGBOUNCER_PIDFILE) {
if let Ok(pid) = pid_str.trim().parse::<i32>() {
info!(
"sending SIGKILL to pgbouncer process pid: {} (from unlocked pid file)",
pid
);
if let Err(e) = kill(Pid::from_raw(pid), Signal::SIGKILL) {
error!("failed to terminate pgbouncer: {}", e);
}
}
} else {
info!("pgbouncer pid file exists but process not running");
}
}
Ok(pid_file::PidFileRead::NotExist) => {
info!("pgbouncer pid file not found, process may not be running");
}
Err(e) => {
error!("error reading pgbouncer pid file: {}", e);
}
}
// Terminate local_proxy
match pid_file::read("/etc/local_proxy/pid".into()) {
Ok(pid_file::PidFileRead::LockedByOtherProcess(pid)) => {
info!("sending SIGTERM to local_proxy process pid: {}", pid);
if let Err(e) = kill(pid, Signal::SIGTERM) {
error!("failed to terminate local_proxy: {}", e);
}
}
Ok(pid_file::PidFileRead::NotHeldByAnyProcess(_)) => {
info!("local_proxy PID file exists but process not running");
}
Ok(pid_file::PidFileRead::NotExist) => {
info!("local_proxy PID file not found, process may not be running");
}
Err(e) => {
error!("error reading local_proxy PID file: {}", e);
}
}
} else {
info!("Skipping pgbouncer and local_proxy termination because in dev mode");
}
let pg_pid = PG_PID.load(Ordering::SeqCst);
if pg_pid != 0 {
let pg_pid = nix::unistd::Pid::from_raw(pg_pid as i32);
@@ -2474,21 +2239,3 @@ impl<T: 'static> JoinSetExt<T> for tokio::task::JoinSet<T> {
})
}
}
#[cfg(test)]
mod tests {
use std::fs::File;
use super::*;
#[test]
fn duplicate_safekeeper_connstring() {
let file = File::open("tests/cluster_spec.json").unwrap();
let spec: ComputeSpec = serde_json::from_reader(file).unwrap();
match ParsedSpec::try_from(spec.clone()) {
Ok(_p) => panic!("Failed to detect duplicate entry"),
Err(e) => assert!(e.starts_with("duplicate entry in safekeeper_connstrings:")),
};
}
}

View File

@@ -25,16 +25,11 @@ struct EndpointStoragePair {
}
const KEY: &str = "lfc_state";
impl EndpointStoragePair {
/// endpoint_id is set to None while prewarming from other endpoint, see replica promotion
/// If not None, takes precedence over pspec.spec.endpoint_id
fn from_spec_and_endpoint(
pspec: &crate::compute::ParsedSpec,
endpoint_id: Option<String>,
) -> Result<Self> {
let endpoint_id = endpoint_id.as_ref().or(pspec.spec.endpoint_id.as_ref());
let Some(ref endpoint_id) = endpoint_id else {
bail!("pspec.endpoint_id missing, other endpoint_id not provided")
impl TryFrom<&crate::compute::ParsedSpec> for EndpointStoragePair {
type Error = anyhow::Error;
fn try_from(pspec: &crate::compute::ParsedSpec) -> Result<Self, Self::Error> {
let Some(ref endpoint_id) = pspec.spec.endpoint_id else {
bail!("pspec.endpoint_id missing")
};
let Some(ref base_uri) = pspec.endpoint_storage_addr else {
bail!("pspec.endpoint_storage_addr missing")
@@ -89,7 +84,7 @@ impl ComputeNode {
}
/// Returns false if there is a prewarm request ongoing, true otherwise
pub fn prewarm_lfc(self: &Arc<Self>, from_endpoint: Option<String>) -> bool {
pub fn prewarm_lfc(self: &Arc<Self>) -> bool {
crate::metrics::LFC_PREWARM_REQUESTS.inc();
{
let state = &mut self.state.lock().unwrap().lfc_prewarm_state;
@@ -102,7 +97,7 @@ impl ComputeNode {
let cloned = self.clone();
spawn(async move {
let Err(err) = cloned.prewarm_impl(from_endpoint).await else {
let Err(err) = cloned.prewarm_impl().await else {
cloned.state.lock().unwrap().lfc_prewarm_state = LfcPrewarmState::Completed;
return;
};
@@ -114,14 +109,13 @@ impl ComputeNode {
true
}
/// from_endpoint: None for endpoint managed by this compute_ctl
fn endpoint_storage_pair(&self, from_endpoint: Option<String>) -> Result<EndpointStoragePair> {
fn endpoint_storage_pair(&self) -> Result<EndpointStoragePair> {
let state = self.state.lock().unwrap();
EndpointStoragePair::from_spec_and_endpoint(state.pspec.as_ref().unwrap(), from_endpoint)
state.pspec.as_ref().unwrap().try_into()
}
async fn prewarm_impl(&self, from_endpoint: Option<String>) -> Result<()> {
let EndpointStoragePair { url, token } = self.endpoint_storage_pair(from_endpoint)?;
async fn prewarm_impl(&self) -> Result<()> {
let EndpointStoragePair { url, token } = self.endpoint_storage_pair()?;
info!(%url, "requesting LFC state from endpoint storage");
let request = Client::new().get(&url).bearer_auth(token);
@@ -179,7 +173,7 @@ impl ComputeNode {
}
async fn offload_lfc_impl(&self) -> Result<()> {
let EndpointStoragePair { url, token } = self.endpoint_storage_pair(None)?;
let EndpointStoragePair { url, token } = self.endpoint_storage_pair()?;
info!(%url, "requesting LFC state from postgres");
let mut compressed = Vec::new();

View File

@@ -2,24 +2,10 @@
module(load="imfile")
# Input configuration for log files in the specified directory
# The messages can be multiline. The start of the message is a timestamp
# in "%Y-%m-%d %H:%M:%S.%3N GMT" (so timezone hardcoded).
# Replace log_directory with the directory containing the log files
input(type="imfile" File="{log_directory}/*.log"
Tag="pgaudit_log" Severity="info" Facility="local5"
startmsg.regex="^[[:digit:]]{{4}}-[[:digit:]]{{2}}-[[:digit:]]{{2}} [[:digit:]]{{2}}:[[:digit:]]{{2}}:[[:digit:]]{{2}}.[[:digit:]]{{3}} GMT,")
# Replace {log_directory} with the directory containing the log files
input(type="imfile" File="{log_directory}/*.log" Tag="{tag}" Severity="info" Facility="local0")
# the directory to store rsyslog state files
global(workDirectory="/var/log/rsyslog")
# Construct json, endpoint_id and project_id as additional metadata
set $.json_log!endpoint_id = "{endpoint_id}";
set $.json_log!project_id = "{project_id}";
set $.json_log!msg = $msg;
# Template suitable for rfc5424 syslog format
template(name="PgAuditLog" type="string"
string="<%PRI%>1 %TIMESTAMP:::date-rfc3339% %HOSTNAME% - - - - %$.json_log%")
# Forward to remote syslog receiver (@@<hostname>:<port>;format
local5.info @@{remote_endpoint};PgAuditLog
# Forward logs to remote syslog server
*.* @@{remote_endpoint}

View File

@@ -83,7 +83,6 @@ use reqwest::StatusCode;
use tar::Archive;
use tracing::info;
use tracing::log::warn;
use url::Url;
use zstd::stream::read::Decoder;
use crate::metrics::{REMOTE_EXT_REQUESTS_TOTAL, UNKNOWN_HTTP_STATUS};
@@ -159,7 +158,7 @@ fn parse_pg_version(human_version: &str) -> PostgresMajorVersion {
pub async fn download_extension(
ext_name: &str,
ext_path: &RemotePath,
remote_ext_base_url: &Url,
remote_ext_base_url: &str,
pgbin: &str,
) -> Result<u64> {
info!("Download extension {:?} from {:?}", ext_name, ext_path);
@@ -271,14 +270,10 @@ pub fn create_control_files(remote_extensions: &RemoteExtSpec, pgbin: &str) {
}
// Do request to extension storage proxy, e.g.,
// curl http://pg-ext-s3-gateway.pg-ext-s3-gateway.svc.cluster.local/latest/v15/extensions/anon.tar.zst
// curl http://pg-ext-s3-gateway/latest/v15/extensions/anon.tar.zst
// using HTTP GET and return the response body as bytes.
async fn download_extension_tar(remote_ext_base_url: &Url, ext_path: &str) -> Result<Bytes> {
let uri = remote_ext_base_url.join(ext_path).with_context(|| {
format!(
"failed to create the remote extension URI for {ext_path} using {remote_ext_base_url}"
)
})?;
async fn download_extension_tar(remote_ext_base_url: &str, ext_path: &str) -> Result<Bytes> {
let uri = format!("{}/{}", remote_ext_base_url, ext_path);
let filename = Path::new(ext_path)
.file_name()
.unwrap_or_else(|| std::ffi::OsStr::new("unknown"))
@@ -288,7 +283,7 @@ async fn download_extension_tar(remote_ext_base_url: &Url, ext_path: &str) -> Re
info!("Downloading extension file '{}' from uri {}", filename, uri);
match do_extension_server_request(uri).await {
match do_extension_server_request(&uri).await {
Ok(resp) => {
info!("Successfully downloaded remote extension data {}", ext_path);
REMOTE_EXT_REQUESTS_TOTAL
@@ -307,7 +302,7 @@ async fn download_extension_tar(remote_ext_base_url: &Url, ext_path: &str) -> Re
// Do a single remote extensions server request.
// Return result or (error message + stringified status code) in case of any failures.
async fn do_extension_server_request(uri: Url) -> Result<Bytes, (String, String)> {
async fn do_extension_server_request(uri: &str) -> Result<Bytes, (String, String)> {
let resp = reqwest::get(uri).await.map_err(|e| {
(
format!(

View File

@@ -48,9 +48,11 @@ impl JsonResponse {
/// Create an error response related to the compute being in an invalid state
pub(self) fn invalid_status(status: ComputeStatus) -> Response {
Self::error(
Self::create_response(
StatusCode::PRECONDITION_FAILED,
format!("invalid compute status: {status}"),
&GenericAPIError {
error: format!("invalid compute status: {status}"),
},
)
}
}

View File

@@ -22,7 +22,7 @@ pub(in crate::http) async fn configure(
State(compute): State<Arc<ComputeNode>>,
request: Json<ConfigurationRequest>,
) -> Response {
let pspec = match ParsedSpec::try_from(request.0.spec) {
let pspec = match ParsedSpec::try_from(request.spec.clone()) {
Ok(p) => p,
Err(e) => return JsonResponse::error(StatusCode::BAD_REQUEST, e),
};

View File

@@ -2,7 +2,6 @@ use crate::compute_prewarm::LfcPrewarmStateWithProgress;
use crate::http::JsonResponse;
use axum::response::{IntoResponse, Response};
use axum::{Json, http::StatusCode};
use axum_extra::extract::OptionalQuery;
use compute_api::responses::LfcOffloadState;
type Compute = axum::extract::State<std::sync::Arc<crate::compute::ComputeNode>>;
@@ -17,16 +16,8 @@ pub(in crate::http) async fn offload_state(compute: Compute) -> Json<LfcOffloadS
Json(compute.lfc_offload_state())
}
#[derive(serde::Deserialize)]
pub struct PrewarmQuery {
pub from_endpoint: String,
}
pub(in crate::http) async fn prewarm(
compute: Compute,
OptionalQuery(query): OptionalQuery<PrewarmQuery>,
) -> Response {
if compute.prewarm_lfc(query.map(|q| q.from_endpoint)) {
pub(in crate::http) async fn prewarm(compute: Compute) -> Response {
if compute.prewarm_lfc() {
StatusCode::ACCEPTED.into_response()
} else {
JsonResponse::error(

View File

@@ -1,42 +1,32 @@
use crate::compute::{ComputeNode, forward_termination_signal};
use crate::http::JsonResponse;
use axum::extract::State;
use axum::response::Response;
use axum_extra::extract::OptionalQuery;
use compute_api::responses::{ComputeStatus, TerminateResponse};
use http::StatusCode;
use serde::Deserialize;
use std::sync::Arc;
use axum::extract::State;
use axum::response::{IntoResponse, Response};
use compute_api::responses::ComputeStatus;
use http::StatusCode;
use tokio::task;
use tracing::info;
#[derive(Deserialize, Default)]
pub struct TerminateQuery {
mode: compute_api::responses::TerminateMode,
}
use crate::compute::{ComputeNode, forward_termination_signal};
use crate::http::JsonResponse;
/// Terminate the compute.
pub(in crate::http) async fn terminate(
State(compute): State<Arc<ComputeNode>>,
OptionalQuery(terminate): OptionalQuery<TerminateQuery>,
) -> Response {
let mode = terminate.unwrap_or_default().mode;
pub(in crate::http) async fn terminate(State(compute): State<Arc<ComputeNode>>) -> Response {
{
let mut state = compute.state.lock().unwrap();
if state.status == ComputeStatus::Terminated {
return JsonResponse::success(StatusCode::CREATED, state.terminate_flush_lsn);
return StatusCode::CREATED.into_response();
}
if !matches!(state.status, ComputeStatus::Empty | ComputeStatus::Running) {
return JsonResponse::invalid_status(state.status);
}
state.set_status(
ComputeStatus::TerminationPending { mode },
&compute.state_changed,
);
state.set_status(ComputeStatus::TerminationPending, &compute.state_changed);
drop(state);
}
forward_termination_signal(false);
forward_termination_signal();
info!("sent signal and notified waiters");
// Spawn a blocking thread to wait for compute to become Terminated.
@@ -44,7 +34,7 @@ pub(in crate::http) async fn terminate(
// be able to serve other requests while some particular request
// is waiting for compute to finish configuration.
let c = compute.clone();
let lsn = task::spawn_blocking(move || {
task::spawn_blocking(move || {
let mut state = c.state.lock().unwrap();
while state.status != ComputeStatus::Terminated {
state = c.state_changed.wait(state).unwrap();
@@ -54,10 +44,11 @@ pub(in crate::http) async fn terminate(
state.status
);
}
state.terminate_flush_lsn
})
.await
.unwrap();
info!("terminated Postgres");
JsonResponse::success(StatusCode::OK, TerminateResponse { lsn })
StatusCode::OK.into_response()
}

View File

@@ -22,7 +22,6 @@ mod migration;
pub mod monitor;
pub mod params;
pub mod pg_helpers;
pub mod pgbouncer;
pub mod rsyslog;
pub mod spec;
mod spec_apply;

View File

@@ -13,12 +13,6 @@ use crate::metrics::{PG_CURR_DOWNTIME_MS, PG_TOTAL_DOWNTIME_MS};
const MONITOR_CHECK_INTERVAL: Duration = Duration::from_millis(500);
/// Struct to store runtime state of the compute monitor thread.
/// In theory, this could be a part of `Compute`, but i)
/// this state is expected to be accessed only by single thread,
/// so we don't need to care about locking; ii) `Compute` is
/// already quite big. Thus, it seems to be a good idea to keep
/// all the activity/health monitoring parts here.
struct ComputeMonitor {
compute: Arc<ComputeNode>,
@@ -76,38 +70,12 @@ impl ComputeMonitor {
)
}
/// Check if compute is in some terminal or soon-to-be-terminal
/// state, then return `true`, signalling the caller that it
/// should exit gracefully. Otherwise, return `false`.
fn check_interrupts(&mut self) -> bool {
let compute_status = self.compute.get_status();
if matches!(
compute_status,
ComputeStatus::Terminated
| ComputeStatus::TerminationPending { .. }
| ComputeStatus::Failed
) {
info!(
"compute is in {} status, stopping compute monitor",
compute_status
);
return true;
}
false
}
/// Spin in a loop and figure out the last activity time in the Postgres.
/// Then update it in the shared state. This function currently never
/// errors out explicitly, but there is a graceful termination path.
/// Every time we receive an error trying to check Postgres, we use
/// [`ComputeMonitor::check_interrupts()`] because it could be that
/// compute is being terminated already, then we can exit gracefully
/// to not produce errors' noise in the log.
/// Then update it in the shared state. This function never errors out.
/// NB: the only expected panic is at `Mutex` unwrap(), all other errors
/// should be handled gracefully.
#[instrument(skip_all)]
pub fn run(&mut self) -> anyhow::Result<()> {
pub fn run(&mut self) {
// Suppose that `connstr` doesn't change
let connstr = self.compute.params.connstr.clone();
let conf = self
@@ -125,10 +93,6 @@ impl ComputeMonitor {
info!("starting compute monitor for {}", connstr);
loop {
if self.check_interrupts() {
break;
}
match &mut client {
Ok(cli) => {
if cli.is_closed() {
@@ -136,10 +100,6 @@ impl ComputeMonitor {
downtime_info = self.downtime_info(),
"connection to Postgres is closed, trying to reconnect"
);
if self.check_interrupts() {
break;
}
self.report_down();
// Connection is closed, reconnect and try again.
@@ -151,19 +111,15 @@ impl ComputeMonitor {
self.compute.update_last_active(self.last_active);
}
Err(e) => {
error!(
downtime_info = self.downtime_info(),
"could not check Postgres: {}", e
);
if self.check_interrupts() {
break;
}
// Although we have many places where we can return errors in `check()`,
// normally it shouldn't happen. I.e., we will likely return error if
// connection got broken, query timed out, Postgres returned invalid data, etc.
// In all such cases it's suspicious, so let's report this as downtime.
self.report_down();
error!(
downtime_info = self.downtime_info(),
"could not check Postgres: {}", e
);
// Reconnect to Postgres just in case. During tests, I noticed
// that queries in `check()` can fail with `connection closed`,
@@ -180,10 +136,6 @@ impl ComputeMonitor {
downtime_info = self.downtime_info(),
"could not connect to Postgres: {}, retrying", e
);
if self.check_interrupts() {
break;
}
self.report_down();
// Establish a new connection and try again.
@@ -195,9 +147,6 @@ impl ComputeMonitor {
self.last_checked = Utc::now();
thread::sleep(MONITOR_CHECK_INTERVAL);
}
// Graceful termination path
Ok(())
}
#[instrument(skip_all)]
@@ -480,10 +429,7 @@ pub fn launch_monitor(compute: &Arc<ComputeNode>) -> thread::JoinHandle<()> {
.spawn(move || {
let span = span!(Level::INFO, "compute_monitor");
let _enter = span.enter();
match monitor.run() {
Ok(_) => info!("compute monitor thread terminated gracefully"),
Err(err) => error!("compute monitor thread terminated abnormally {:?}", err),
}
monitor.run();
})
.expect("cannot launch compute monitor thread")
}

View File

@@ -213,10 +213,8 @@ impl Escaping for PgIdent {
// Find the first suitable tag that is not present in the string.
// Postgres' max role/DB name length is 63 bytes, so even in the
// worst case it won't take long. Outer tag is always `tag + "x"`,
// so if `tag` is not present in the string, `outer_tag` is not
// present in the string either.
while self.contains(&tag.to_string()) {
// worst case it won't take long.
while self.contains(&format!("${tag}$")) || self.contains(&format!("${outer_tag}$")) {
tag += "x";
outer_tag = tag.clone() + "x";
}

View File

@@ -1 +0,0 @@
pub const PGBOUNCER_PIDFILE: &str = "/tmp/pgbouncer.pid";

View File

@@ -27,40 +27,6 @@ fn get_rsyslog_pid() -> Option<String> {
}
}
fn wait_for_rsyslog_pid() -> Result<String, anyhow::Error> {
const MAX_WAIT: Duration = Duration::from_secs(5);
const INITIAL_SLEEP: Duration = Duration::from_millis(2);
let mut sleep_duration = INITIAL_SLEEP;
let start = std::time::Instant::now();
let mut attempts = 1;
for attempt in 1.. {
attempts = attempt;
match get_rsyslog_pid() {
Some(pid) => return Ok(pid),
None => {
if start.elapsed() >= MAX_WAIT {
break;
}
info!(
"rsyslogd is not running, attempt {}. Sleeping for {} ms",
attempt,
sleep_duration.as_millis()
);
std::thread::sleep(sleep_duration);
sleep_duration *= 2;
}
}
}
Err(anyhow::anyhow!(
"rsyslogd is not running after waiting for {} seconds and {} attempts",
attempts,
start.elapsed().as_secs()
))
}
// Restart rsyslogd to apply the new configuration.
// This is necessary, because there is no other way to reload the rsyslog configuration.
//
@@ -70,29 +36,27 @@ fn wait_for_rsyslog_pid() -> Result<String, anyhow::Error> {
// TODO: test it properly
//
fn restart_rsyslog() -> Result<()> {
let old_pid = get_rsyslog_pid().context("rsyslogd is not running")?;
info!("rsyslogd is running with pid: {}, restart it", old_pid);
// kill it to restart
let _ = Command::new("pkill")
.arg("rsyslogd")
.output()
.context("Failed to restart rsyslogd")?;
// ensure rsyslogd is running
wait_for_rsyslog_pid()?;
.context("Failed to stop rsyslogd")?;
Ok(())
}
pub fn configure_audit_rsyslog(
log_directory: String,
endpoint_id: &str,
project_id: &str,
tag: Option<String>,
remote_endpoint: &str,
) -> Result<()> {
let config_content: String = format!(
include_str!("config_template/compute_audit_rsyslog_template.conf"),
log_directory = log_directory,
endpoint_id = endpoint_id,
project_id = project_id,
tag = tag.unwrap_or("".to_string()),
remote_endpoint = remote_endpoint
);
@@ -167,11 +131,15 @@ pub fn configure_postgres_logs_export(conf: PostgresLogsRsyslogConfig) -> Result
return Ok(());
}
// Nothing to configure
// When new config is empty we can simply remove the configuration file.
if new_config.is_empty() {
// When the configuration is removed, PostgreSQL will stop sending data
// to the files watched by rsyslog, so restarting rsyslog is more effort
// than just ignoring this change.
info!("removing rsyslog config file: {}", POSTGRES_LOGS_CONF_PATH);
match std::fs::remove_file(POSTGRES_LOGS_CONF_PATH) {
Ok(_) => {}
Err(err) if err.kind() == ErrorKind::NotFound => {}
Err(err) => return Err(err.into()),
}
restart_rsyslog()?;
return Ok(());
}

View File

@@ -1,6 +0,0 @@
### Test files
The file `cluster_spec.json` has been copied over from libs/compute_api
tests, with some edits:
- the neon.safekeepers setting contains a duplicate value

View File

@@ -1,245 +0,0 @@
{
"format_version": 1.0,
"timestamp": "2021-05-23T18:25:43.511Z",
"operation_uuid": "0f657b36-4b0f-4a2d-9c2e-1dcd615e7d8b",
"cluster": {
"cluster_id": "test-cluster-42",
"name": "Zenith Test",
"state": "restarted",
"roles": [
{
"name": "postgres",
"encrypted_password": "6b1d16b78004bbd51fa06af9eda75972",
"options": null
},
{
"name": "alexk",
"encrypted_password": null,
"options": null
},
{
"name": "zenith \"new\"",
"encrypted_password": "5b1d16b78004bbd51fa06af9eda75972",
"options": null
},
{
"name": "zen",
"encrypted_password": "9b1d16b78004bbd51fa06af9eda75972"
},
{
"name": "\"name\";\\n select 1;",
"encrypted_password": "5b1d16b78004bbd51fa06af9eda75972"
},
{
"name": "MyRole",
"encrypted_password": "5b1d16b78004bbd51fa06af9eda75972"
}
],
"databases": [
{
"name": "DB2",
"owner": "alexk",
"options": [
{
"name": "LC_COLLATE",
"value": "C",
"vartype": "string"
},
{
"name": "LC_CTYPE",
"value": "C",
"vartype": "string"
},
{
"name": "TEMPLATE",
"value": "template0",
"vartype": "enum"
}
]
},
{
"name": "zenith",
"owner": "MyRole"
},
{
"name": "zen",
"owner": "zen"
}
],
"settings": [
{
"name": "fsync",
"value": "off",
"vartype": "bool"
},
{
"name": "wal_level",
"value": "logical",
"vartype": "enum"
},
{
"name": "hot_standby",
"value": "on",
"vartype": "bool"
},
{
"name": "prewarm_lfc_on_startup",
"value": "off",
"vartype": "bool"
},
{
"name": "neon.safekeepers",
"value": "127.0.0.1:6502,127.0.0.1:6503,127.0.0.1:6501,127.0.0.1:6502",
"vartype": "string"
},
{
"name": "wal_log_hints",
"value": "on",
"vartype": "bool"
},
{
"name": "log_connections",
"value": "on",
"vartype": "bool"
},
{
"name": "shared_buffers",
"value": "32768",
"vartype": "integer"
},
{
"name": "port",
"value": "55432",
"vartype": "integer"
},
{
"name": "max_connections",
"value": "100",
"vartype": "integer"
},
{
"name": "max_wal_senders",
"value": "10",
"vartype": "integer"
},
{
"name": "listen_addresses",
"value": "0.0.0.0",
"vartype": "string"
},
{
"name": "wal_sender_timeout",
"value": "0",
"vartype": "integer"
},
{
"name": "password_encryption",
"value": "md5",
"vartype": "enum"
},
{
"name": "maintenance_work_mem",
"value": "65536",
"vartype": "integer"
},
{
"name": "max_parallel_workers",
"value": "8",
"vartype": "integer"
},
{
"name": "max_worker_processes",
"value": "8",
"vartype": "integer"
},
{
"name": "neon.tenant_id",
"value": "b0554b632bd4d547a63b86c3630317e8",
"vartype": "string"
},
{
"name": "max_replication_slots",
"value": "10",
"vartype": "integer"
},
{
"name": "neon.timeline_id",
"value": "2414a61ffc94e428f14b5758fe308e13",
"vartype": "string"
},
{
"name": "shared_preload_libraries",
"value": "neon",
"vartype": "string"
},
{
"name": "synchronous_standby_names",
"value": "walproposer",
"vartype": "string"
},
{
"name": "neon.pageserver_connstring",
"value": "host=127.0.0.1 port=6400",
"vartype": "string"
},
{
"name": "test.escaping",
"value": "here's a backslash \\ and a quote ' and a double-quote \" hooray",
"vartype": "string"
}
]
},
"delta_operations": [
{
"action": "delete_db",
"name": "zenith_test"
},
{
"action": "rename_db",
"name": "DB",
"new_name": "DB2"
},
{
"action": "delete_role",
"name": "zenith2"
},
{
"action": "rename_role",
"name": "zenith new",
"new_name": "zenith \"new\""
}
],
"remote_extensions": {
"library_index": {
"postgis-3": "postgis",
"libpgrouting-3.4": "postgis",
"postgis_raster-3": "postgis",
"postgis_sfcgal-3": "postgis",
"postgis_topology-3": "postgis",
"address_standardizer-3": "postgis"
},
"extension_data": {
"postgis": {
"archive_path": "5834329303/v15/extensions/postgis.tar.zst",
"control_data": {
"postgis.control": "# postgis extension\ncomment = ''PostGIS geometry and geography spatial types and functions''\ndefault_version = ''3.3.2''\nmodule_pathname = ''$libdir/postgis-3''\nrelocatable = false\ntrusted = true\n",
"pgrouting.control": "# pgRouting Extension\ncomment = ''pgRouting Extension''\ndefault_version = ''3.4.2''\nmodule_pathname = ''$libdir/libpgrouting-3.4''\nrelocatable = true\nrequires = ''plpgsql''\nrequires = ''postgis''\ntrusted = true\n",
"postgis_raster.control": "# postgis_raster extension\ncomment = ''PostGIS raster types and functions''\ndefault_version = ''3.3.2''\nmodule_pathname = ''$libdir/postgis_raster-3''\nrelocatable = false\nrequires = postgis\ntrusted = true\n",
"postgis_sfcgal.control": "# postgis topology extension\ncomment = ''PostGIS SFCGAL functions''\ndefault_version = ''3.3.2''\nrelocatable = true\nrequires = postgis\ntrusted = true\n",
"postgis_topology.control": "# postgis topology extension\ncomment = ''PostGIS topology spatial types and functions''\ndefault_version = ''3.3.2''\nrelocatable = false\nschema = topology\nrequires = postgis\ntrusted = true\n",
"address_standardizer.control": "# address_standardizer extension\ncomment = ''Used to parse an address into constituent elements. Generally used to support geocoding address normalization step.''\ndefault_version = ''3.3.2''\nrelocatable = true\ntrusted = true\n",
"postgis_tiger_geocoder.control": "# postgis tiger geocoder extension\ncomment = ''PostGIS tiger geocoder and reverse geocoder''\ndefault_version = ''3.3.2''\nrelocatable = false\nschema = tiger\nrequires = ''postgis,fuzzystrmatch''\nsuperuser= false\ntrusted = true\n",
"address_standardizer_data_us.control": "# address standardizer us dataset\ncomment = ''Address Standardizer US dataset example''\ndefault_version = ''3.3.2''\nrelocatable = true\ntrusted = true\n"
}
}
},
"custom_extensions": [],
"public_extensions": ["postgis"]
},
"pgbouncer_settings": {
"default_pool_size": "42",
"pool_mode": "session"
}
}

View File

@@ -30,7 +30,7 @@ mod pg_helpers_tests {
r#"fsync = off
wal_level = logical
hot_standby = on
autoprewarm = off
prewarm_lfc_on_startup = off
neon.safekeepers = '127.0.0.1:6502,127.0.0.1:6503,127.0.0.1:6501'
wal_log_hints = on
log_connections = on
@@ -71,14 +71,6 @@ test.escaping = 'here''s a backslash \\ and a quote '' and a double-quote " hoor
("name$$$", ("$x$name$$$$x$", "xx")),
("name$$$$", ("$x$name$$$$$x$", "xx")),
("name$x$", ("$xx$name$x$$xx$", "xxx")),
("x", ("$xx$x$xx$", "xxx")),
("xx", ("$xxx$xx$xxx$", "xxxx")),
("$x", ("$xx$$x$xx$", "xxx")),
("x$", ("$xx$x$$xx$", "xxx")),
("$x$", ("$xx$$x$$xx$", "xxx")),
("xx$", ("$xxx$xx$$xxx$", "xxxx")),
("$xx", ("$xxx$$xx$xxx$", "xxxx")),
("$xx$", ("$xxx$$xx$$xxx$", "xxxx")),
];
for (input, expected) in test_cases {

View File

@@ -36,7 +36,6 @@ pageserver_api.workspace = true
pageserver_client.workspace = true
postgres_backend.workspace = true
safekeeper_api.workspace = true
safekeeper_client.workspace = true
postgres_connection.workspace = true
storage_broker.workspace = true
http-utils.workspace = true

View File

@@ -2,10 +2,8 @@
[pageserver]
listen_pg_addr = '127.0.0.1:64000'
listen_http_addr = '127.0.0.1:9898'
listen_grpc_addr = '127.0.0.1:51051'
pg_auth_type = 'Trust'
http_auth_type = 'Trust'
grpc_auth_type = 'Trust'
[[safekeepers]]
id = 1

View File

@@ -4,10 +4,8 @@
id=1
listen_pg_addr = '127.0.0.1:64000'
listen_http_addr = '127.0.0.1:9898'
listen_grpc_addr = '127.0.0.1:51051'
pg_auth_type = 'Trust'
http_auth_type = 'Trust'
grpc_auth_type = 'Trust'
[[safekeepers]]
id = 1

View File

@@ -18,7 +18,7 @@ use clap::Parser;
use compute_api::requests::ComputeClaimsScope;
use compute_api::spec::ComputeMode;
use control_plane::broker::StorageBroker;
use control_plane::endpoint::{ComputeControlPlane, EndpointTerminateMode, PageserverProtocol};
use control_plane::endpoint::ComputeControlPlane;
use control_plane::endpoint_storage::{ENDPOINT_STORAGE_DEFAULT_ADDR, EndpointStorage};
use control_plane::local_env;
use control_plane::local_env::{
@@ -32,7 +32,6 @@ use control_plane::storage_controller::{
};
use nix::fcntl::{Flock, FlockArg};
use pageserver_api::config::{
DEFAULT_GRPC_LISTEN_PORT as DEFAULT_PAGESERVER_GRPC_PORT,
DEFAULT_HTTP_LISTEN_PORT as DEFAULT_PAGESERVER_HTTP_PORT,
DEFAULT_PG_LISTEN_PORT as DEFAULT_PAGESERVER_PG_PORT,
};
@@ -45,7 +44,7 @@ use pageserver_api::models::{
use pageserver_api::shard::{DEFAULT_STRIPE_SIZE, ShardCount, ShardStripeSize, TenantShardId};
use postgres_backend::AuthType;
use postgres_connection::parse_host_port;
use safekeeper_api::membership::{SafekeeperGeneration, SafekeeperId};
use safekeeper_api::membership::SafekeeperGeneration;
use safekeeper_api::{
DEFAULT_HTTP_LISTEN_PORT as DEFAULT_SAFEKEEPER_HTTP_PORT,
DEFAULT_PG_LISTEN_PORT as DEFAULT_SAFEKEEPER_PG_PORT,
@@ -605,14 +604,6 @@ struct EndpointCreateCmdArgs {
#[clap(long, help = "Postgres version")]
pg_version: u32,
/// Use gRPC to communicate with Pageservers, by generating grpc:// connstrings.
///
/// Specified on creation such that it's retained across reconfiguration and restarts.
///
/// NB: not yet supported by computes.
#[clap(long)]
grpc: bool,
#[clap(
long,
help = "If set, the node will be a hot replica on the specified timeline",
@@ -672,13 +663,6 @@ struct EndpointStartCmdArgs {
#[clap(short = 't', long, value_parser= humantime::parse_duration, help = "timeout until we fail the command")]
#[arg(default_value = "90s")]
start_timeout: Duration,
#[clap(
long,
help = "Run in development mode, skipping VM-specific operations like process termination",
action = clap::ArgAction::SetTrue
)]
dev: bool,
}
#[derive(clap::Args)]
@@ -711,9 +695,10 @@ struct EndpointStopCmdArgs {
)]
destroy: bool,
#[clap(long, help = "Postgres shutdown mode")]
#[clap(default_value = "fast")]
mode: EndpointTerminateMode,
#[clap(long, help = "Postgres shutdown mode, passed to \"pg_ctl -m <mode>\"")]
#[arg(value_parser(["smart", "fast", "immediate"]))]
#[arg(default_value = "fast")]
mode: String,
}
#[derive(clap::Args)]
@@ -1022,16 +1007,13 @@ fn handle_init(args: &InitCmdArgs) -> anyhow::Result<LocalEnv> {
let pageserver_id = NodeId(DEFAULT_PAGESERVER_ID.0 + i as u64);
let pg_port = DEFAULT_PAGESERVER_PG_PORT + i;
let http_port = DEFAULT_PAGESERVER_HTTP_PORT + i;
let grpc_port = DEFAULT_PAGESERVER_GRPC_PORT + i;
NeonLocalInitPageserverConf {
id: pageserver_id,
listen_pg_addr: format!("127.0.0.1:{pg_port}"),
listen_http_addr: format!("127.0.0.1:{http_port}"),
listen_https_addr: None,
listen_grpc_addr: Some(format!("127.0.0.1:{grpc_port}")),
pg_auth_type: AuthType::Trust,
http_auth_type: AuthType::Trust,
grpc_auth_type: AuthType::Trust,
other: Default::default(),
// Typical developer machines use disks with slow fsync, and we don't care
// about data integrity: disable disk syncs.
@@ -1269,45 +1251,6 @@ async fn handle_timeline(cmd: &TimelineCmd, env: &mut local_env::LocalEnv) -> Re
pageserver
.timeline_import(tenant_id, timeline_id, base, pg_wal, args.pg_version)
.await?;
if env.storage_controller.timelines_onto_safekeepers {
println!("Creating timeline on safekeeper ...");
let timeline_info = pageserver
.timeline_info(
TenantShardId::unsharded(tenant_id),
timeline_id,
pageserver_client::mgmt_api::ForceAwaitLogicalSize::No,
)
.await?;
let default_sk = SafekeeperNode::from_env(env, env.safekeepers.first().unwrap());
let default_host = default_sk
.conf
.listen_addr
.clone()
.unwrap_or_else(|| "localhost".to_string());
let mconf = safekeeper_api::membership::Configuration {
generation: SafekeeperGeneration::new(1),
members: safekeeper_api::membership::MemberSet {
m: vec![SafekeeperId {
host: default_host,
id: default_sk.conf.id,
pg_port: default_sk.conf.pg_port,
}],
},
new_members: None,
};
let pg_version = args.pg_version * 10000;
let req = safekeeper_api::models::TimelineCreateRequest {
tenant_id,
timeline_id,
mconf,
pg_version,
system_id: None,
wal_seg_size: None,
start_lsn: timeline_info.last_record_lsn,
commit_lsn: None,
};
default_sk.create_timeline(&req).await?;
}
env.register_branch_mapping(branch_name.to_string(), tenant_id, timeline_id)?;
println!("Done");
}
@@ -1332,7 +1275,6 @@ async fn handle_timeline(cmd: &TimelineCmd, env: &mut local_env::LocalEnv) -> Re
mode: pageserver_api::models::TimelineCreateRequestMode::Branch {
ancestor_timeline_id,
ancestor_start_lsn: start_lsn,
read_only: false,
pg_version: None,
},
};
@@ -1465,7 +1407,6 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
args.internal_http_port,
args.pg_version,
mode,
args.grpc,
!args.update_catalog,
false,
)?;
@@ -1506,20 +1447,13 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
let (pageservers, stripe_size) = if let Some(pageserver_id) = pageserver_id {
let conf = env.get_pageserver_conf(pageserver_id).unwrap();
// Use gRPC if requested.
let pageserver = if endpoint.grpc {
let grpc_addr = conf.listen_grpc_addr.as_ref().expect("bad config");
let (host, port) = parse_host_port(grpc_addr)?;
let port = port.unwrap_or(DEFAULT_PAGESERVER_GRPC_PORT);
(PageserverProtocol::Grpc, host, port)
} else {
let (host, port) = parse_host_port(&conf.listen_pg_addr)?;
let port = port.unwrap_or(5432);
(PageserverProtocol::Libpq, host, port)
};
// If caller is telling us what pageserver to use, this is not a tenant which is
// fully managed by storage controller, therefore not sharded.
(vec![pageserver], DEFAULT_STRIPE_SIZE)
let parsed = parse_host_port(&conf.listen_pg_addr).expect("Bad config");
(
vec![(parsed.0, parsed.1.unwrap_or(5432))],
// If caller is telling us what pageserver to use, this is not a tenant which is
// full managed by storage controller, therefore not sharded.
DEFAULT_STRIPE_SIZE,
)
} else {
// Look up the currently attached location of the tenant, and its striping metadata,
// to pass these on to postgres.
@@ -1538,20 +1472,11 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
.await?;
}
let pageserver = if endpoint.grpc {
(
PageserverProtocol::Grpc,
Host::parse(&shard.listen_grpc_addr.expect("no gRPC address"))?,
shard.listen_grpc_port.expect("no gRPC port"),
)
} else {
(
PageserverProtocol::Libpq,
Host::parse(&shard.listen_pg_addr)?,
shard.listen_pg_port,
)
};
anyhow::Ok(pageserver)
anyhow::Ok((
Host::parse(&shard.listen_pg_addr)
.expect("Storage controller reported bad hostname"),
shard.listen_pg_port,
))
}),
)
.await?;
@@ -1596,7 +1521,6 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
stripe_size.0 as usize,
args.create_test_user,
args.start_timeout,
args.dev,
)
.await?;
}
@@ -1607,19 +1531,11 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
.get(endpoint_id.as_str())
.with_context(|| format!("postgres endpoint {endpoint_id} is not found"))?;
let pageservers = if let Some(ps_id) = args.endpoint_pageserver_id {
let conf = env.get_pageserver_conf(ps_id)?;
// Use gRPC if requested.
let pageserver = if endpoint.grpc {
let grpc_addr = conf.listen_grpc_addr.as_ref().expect("bad config");
let (host, port) = parse_host_port(grpc_addr)?;
let port = port.unwrap_or(DEFAULT_PAGESERVER_GRPC_PORT);
(PageserverProtocol::Grpc, host, port)
} else {
let (host, port) = parse_host_port(&conf.listen_pg_addr)?;
let port = port.unwrap_or(5432);
(PageserverProtocol::Libpq, host, port)
};
vec![pageserver]
let pageserver = PageServerNode::from_env(env, env.get_pageserver_conf(ps_id)?);
vec![(
pageserver.pg_connection_config.host().clone(),
pageserver.pg_connection_config.port(),
)]
} else {
let storage_controller = StorageController::from_env(env);
storage_controller
@@ -1628,21 +1544,11 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
.shards
.into_iter()
.map(|shard| {
// Use gRPC if requested.
if endpoint.grpc {
(
PageserverProtocol::Grpc,
Host::parse(&shard.listen_grpc_addr.expect("no gRPC address"))
.expect("bad hostname"),
shard.listen_grpc_port.expect("no gRPC port"),
)
} else {
(
PageserverProtocol::Libpq,
Host::parse(&shard.listen_pg_addr).expect("bad hostname"),
shard.listen_pg_port,
)
}
(
Host::parse(&shard.listen_pg_addr)
.expect("Storage controller reported malformed host"),
shard.listen_pg_port,
)
})
.collect::<Vec<_>>()
};
@@ -1657,10 +1563,7 @@ async fn handle_endpoint(subcmd: &EndpointCmd, env: &local_env::LocalEnv) -> Res
.endpoints
.get(endpoint_id)
.with_context(|| format!("postgres endpoint {endpoint_id} is not found"))?;
match endpoint.stop(args.mode, args.destroy).await?.lsn {
Some(lsn) => println!("{lsn}"),
None => println!("null"),
}
endpoint.stop(&args.mode, args.destroy)?;
}
EndpointCmd::GenerateJwt(args) => {
let endpoint = {
@@ -2092,16 +1995,11 @@ async fn handle_stop_all(args: &StopCmdArgs, env: &local_env::LocalEnv) -> Resul
}
async fn try_stop_all(env: &local_env::LocalEnv, immediate: bool) {
let mode = if immediate {
EndpointTerminateMode::Immediate
} else {
EndpointTerminateMode::Fast
};
// Stop all endpoints
match ComputeControlPlane::load(env.clone()) {
Ok(cplane) => {
for (_k, node) in cplane.endpoints {
if let Err(e) = node.stop(mode, false).await {
if let Err(e) = node.stop(if immediate { "immediate" } else { "fast" }, false) {
eprintln!("postgres stop failed: {e:#}");
}
}

View File

@@ -37,7 +37,6 @@
//! ```
//!
use std::collections::BTreeMap;
use std::fmt::Display;
use std::net::{IpAddr, Ipv4Addr, SocketAddr, TcpStream};
use std::path::PathBuf;
use std::process::Command;
@@ -46,14 +45,11 @@ use std::sync::Arc;
use std::time::{Duration, Instant};
use anyhow::{Context, Result, anyhow, bail};
use base64::Engine;
use base64::prelude::BASE64_URL_SAFE_NO_PAD;
use compute_api::requests::{
COMPUTE_AUDIENCE, ComputeClaims, ComputeClaimsScope, ConfigurationRequest,
};
use compute_api::responses::{
ComputeConfig, ComputeCtlConfig, ComputeStatus, ComputeStatusResponse, TerminateResponse,
TlsConfig,
ComputeConfig, ComputeCtlConfig, ComputeStatus, ComputeStatusResponse, TlsConfig,
};
use compute_api::spec::{
Cluster, ComputeAudit, ComputeFeature, ComputeMode, ComputeSpec, Database, PgIdent,
@@ -78,6 +74,7 @@ use utils::id::{NodeId, TenantId, TimelineId};
use crate::local_env::LocalEnv;
use crate::postgresql_conf::PostgresConf;
use crate::storage_controller::StorageController;
// contents of a endpoint.json file
#[derive(Serialize, Deserialize, PartialEq, Eq, Clone, Debug)]
@@ -90,7 +87,6 @@ pub struct EndpointConf {
external_http_port: u16,
internal_http_port: u16,
pg_version: u32,
grpc: bool,
skip_pg_catalog_updates: bool,
reconfigure_concurrency: usize,
drop_subscriptions_before_start: bool,
@@ -168,7 +164,7 @@ impl ComputeControlPlane {
public_key_use: Some(PublicKeyUse::Signature),
key_operations: Some(vec![KeyOperations::Verify]),
key_algorithm: Some(KeyAlgorithm::EdDSA),
key_id: Some(BASE64_URL_SAFE_NO_PAD.encode(key_hash)),
key_id: Some(base64::encode_config(key_hash, base64::URL_SAFE_NO_PAD)),
x509_url: None::<String>,
x509_chain: None::<Vec<String>>,
x509_sha1_fingerprint: None::<String>,
@@ -177,7 +173,7 @@ impl ComputeControlPlane {
algorithm: AlgorithmParameters::OctetKeyPair(OctetKeyPairParameters {
key_type: OctetKeyPairType::OctetKeyPair,
curve: EllipticCurve::Ed25519,
x: BASE64_URL_SAFE_NO_PAD.encode(public_key),
x: base64::encode_config(public_key, base64::URL_SAFE_NO_PAD),
}),
}],
})
@@ -194,7 +190,6 @@ impl ComputeControlPlane {
internal_http_port: Option<u16>,
pg_version: u32,
mode: ComputeMode,
grpc: bool,
skip_pg_catalog_updates: bool,
drop_subscriptions_before_start: bool,
) -> Result<Arc<Endpoint>> {
@@ -229,7 +224,6 @@ impl ComputeControlPlane {
// we also skip catalog updates in the cloud.
skip_pg_catalog_updates,
drop_subscriptions_before_start,
grpc,
reconfigure_concurrency: 1,
features: vec![],
cluster: None,
@@ -248,7 +242,6 @@ impl ComputeControlPlane {
internal_http_port,
pg_port,
pg_version,
grpc,
skip_pg_catalog_updates,
drop_subscriptions_before_start,
reconfigure_concurrency: 1,
@@ -303,8 +296,6 @@ pub struct Endpoint {
pub tenant_id: TenantId,
pub timeline_id: TimelineId,
pub mode: ComputeMode,
/// If true, the endpoint should use gRPC to communicate with Pageservers.
pub grpc: bool,
// port and address of the Postgres server and `compute_ctl`'s HTTP APIs
pub pg_address: SocketAddr,
@@ -340,58 +331,15 @@ pub enum EndpointStatus {
RunningNoPidfile,
}
impl Display for EndpointStatus {
impl std::fmt::Display for EndpointStatus {
fn fmt(&self, writer: &mut std::fmt::Formatter) -> std::fmt::Result {
writer.write_str(match self {
let s = match self {
Self::Running => "running",
Self::Stopped => "stopped",
Self::Crashed => "crashed",
Self::RunningNoPidfile => "running, no pidfile",
})
}
}
#[derive(Default, Clone, Copy, clap::ValueEnum)]
pub enum EndpointTerminateMode {
#[default]
/// Use pg_ctl stop -m fast
Fast,
/// Use pg_ctl stop -m immediate
Immediate,
/// Use /terminate?mode=immediate
ImmediateTerminate,
}
impl std::fmt::Display for EndpointTerminateMode {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.write_str(match &self {
EndpointTerminateMode::Fast => "fast",
EndpointTerminateMode::Immediate => "immediate",
EndpointTerminateMode::ImmediateTerminate => "immediate-terminate",
})
}
}
/// Protocol used to connect to a Pageserver.
#[derive(Clone, Copy, Debug)]
pub enum PageserverProtocol {
Libpq,
Grpc,
}
impl PageserverProtocol {
/// Returns the URL scheme for the protocol, used in connstrings.
pub fn scheme(&self) -> &'static str {
match self {
Self::Libpq => "postgresql",
Self::Grpc => "grpc",
}
}
}
impl Display for PageserverProtocol {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.write_str(self.scheme())
};
write!(writer, "{}", s)
}
}
@@ -430,7 +378,6 @@ impl Endpoint {
mode: conf.mode,
tenant_id: conf.tenant_id,
pg_version: conf.pg_version,
grpc: conf.grpc,
skip_pg_catalog_updates: conf.skip_pg_catalog_updates,
reconfigure_concurrency: conf.reconfigure_concurrency,
drop_subscriptions_before_start: conf.drop_subscriptions_before_start,
@@ -659,10 +606,10 @@ impl Endpoint {
}
}
fn build_pageserver_connstr(pageservers: &[(PageserverProtocol, Host, u16)]) -> String {
fn build_pageserver_connstr(pageservers: &[(Host, u16)]) -> String {
pageservers
.iter()
.map(|(scheme, host, port)| format!("{scheme}://no_user@{host}:{port}"))
.map(|(host, port)| format!("postgresql://no_user@{host}:{port}"))
.collect::<Vec<_>>()
.join(",")
}
@@ -707,12 +654,11 @@ impl Endpoint {
endpoint_storage_addr: String,
safekeepers_generation: Option<SafekeeperGeneration>,
safekeepers: Vec<NodeId>,
pageservers: Vec<(PageserverProtocol, Host, u16)>,
pageservers: Vec<(Host, u16)>,
remote_ext_base_url: Option<&String>,
shard_stripe_size: usize,
create_test_user: bool,
start_timeout: Duration,
dev: bool,
) -> Result<()> {
if self.status() == EndpointStatus::Running {
anyhow::bail!("The endpoint is already running");
@@ -801,7 +747,7 @@ impl Endpoint {
logs_export_host: None::<String>,
endpoint_storage_addr: Some(endpoint_storage_addr),
endpoint_storage_token: Some(endpoint_storage_token),
autoprewarm: false,
prewarm_lfc_on_startup: false,
};
// this strange code is needed to support respec() in tests
@@ -883,10 +829,6 @@ impl Endpoint {
cmd.args(["--remote-ext-base-url", remote_ext_base_url]);
}
if dev {
cmd.arg("--dev");
}
let child = cmd.spawn()?;
// set up a scopeguard to kill & wait for the child in case we panic or bail below
let child = scopeguard::guard(child, |mut child| {
@@ -939,7 +881,7 @@ impl Endpoint {
ComputeStatus::Empty
| ComputeStatus::ConfigurationPending
| ComputeStatus::Configuration
| ComputeStatus::TerminationPending { .. }
| ComputeStatus::TerminationPending
| ComputeStatus::Terminated => {
bail!("unexpected compute status: {:?}", state.status)
}
@@ -997,12 +939,10 @@ impl Endpoint {
pub async fn reconfigure(
&self,
pageservers: Vec<(PageserverProtocol, Host, u16)>,
mut pageservers: Vec<(Host, u16)>,
stripe_size: Option<ShardStripeSize>,
safekeepers: Option<Vec<NodeId>>,
) -> Result<()> {
anyhow::ensure!(!pageservers.is_empty(), "no pageservers provided");
let (mut spec, compute_ctl_config) = {
let config_path = self.endpoint_path().join("config.json");
let file = std::fs::File::open(config_path)?;
@@ -1014,7 +954,25 @@ impl Endpoint {
let postgresql_conf = self.read_postgresql_conf()?;
spec.cluster.postgresql_conf = Some(postgresql_conf);
// If we weren't given explicit pageservers, query the storage controller
if pageservers.is_empty() {
let storage_controller = StorageController::from_env(&self.env);
let locate_result = storage_controller.tenant_locate(self.tenant_id).await?;
pageservers = locate_result
.shards
.into_iter()
.map(|shard| {
(
Host::parse(&shard.listen_pg_addr)
.expect("Storage controller reported bad hostname"),
shard.listen_pg_port,
)
})
.collect::<Vec<_>>();
}
let pageserver_connstr = Self::build_pageserver_connstr(&pageservers);
assert!(!pageserver_connstr.is_empty());
spec.pageserver_connstring = Some(pageserver_connstr);
if stripe_size.is_some() {
spec.shard_stripe_size = stripe_size.map(|s| s.0 as usize);
@@ -1061,27 +1019,8 @@ impl Endpoint {
}
}
pub async fn stop(
&self,
mode: EndpointTerminateMode,
destroy: bool,
) -> Result<TerminateResponse> {
// pg_ctl stop is fast but doesn't allow us to collect LSN. /terminate is
// slow, and test runs time out. Solution: special mode "immediate-terminate"
// which uses /terminate
let response = if let EndpointTerminateMode::ImmediateTerminate = mode {
let ip = self.external_http_address.ip();
let port = self.external_http_address.port();
let url = format!("http://{ip}:{port}/terminate?mode=immediate");
let token = self.generate_jwt(Some(ComputeClaimsScope::Admin))?;
let request = reqwest::Client::new().post(url).bearer_auth(token);
let response = request.send().await.context("/terminate")?;
let text = response.text().await.context("/terminate result")?;
serde_json::from_str(&text).with_context(|| format!("deserializing {text}"))?
} else {
self.pg_ctl(&["-m", &mode.to_string(), "stop"], &None)?;
TerminateResponse { lsn: None }
};
pub fn stop(&self, mode: &str, destroy: bool) -> Result<()> {
self.pg_ctl(&["-m", mode, "stop"], &None)?;
// Also wait for the compute_ctl process to die. It might have some
// cleanup work to do after postgres stops, like syncing safekeepers,
@@ -1091,7 +1030,7 @@ impl Endpoint {
// waiting. Sometimes we do *not* want this cleanup: tests intentionally
// do stop when majority of safekeepers is down, so sync-safekeepers
// would hang otherwise. This could be a separate flag though.
let send_sigterm = destroy || !matches!(mode, EndpointTerminateMode::Fast);
let send_sigterm = destroy || mode == "immediate";
self.wait_for_compute_ctl_to_exit(send_sigterm)?;
if destroy {
println!(
@@ -1100,7 +1039,7 @@ impl Endpoint {
);
std::fs::remove_dir_all(self.endpoint_path())?;
}
Ok(response)
Ok(())
}
pub fn connstr(&self, user: &str, db_name: &str) -> String {

View File

@@ -209,8 +209,6 @@ pub struct NeonStorageControllerConf {
pub use_https_safekeeper_api: bool,
pub use_local_compute_notifications: bool,
pub timeline_safekeeper_count: Option<i64>,
}
impl NeonStorageControllerConf {
@@ -238,10 +236,9 @@ impl Default for NeonStorageControllerConf {
heartbeat_interval: Self::DEFAULT_HEARTBEAT_INTERVAL,
long_reconcile_threshold: None,
use_https_pageserver_api: false,
timelines_onto_safekeepers: true,
timelines_onto_safekeepers: false,
use_https_safekeeper_api: false,
use_local_compute_notifications: true,
timeline_safekeeper_count: None,
}
}
}
@@ -281,10 +278,8 @@ pub struct PageServerConf {
pub listen_pg_addr: String,
pub listen_http_addr: String,
pub listen_https_addr: Option<String>,
pub listen_grpc_addr: Option<String>,
pub pg_auth_type: AuthType,
pub http_auth_type: AuthType,
pub grpc_auth_type: AuthType,
pub no_sync: bool,
}
@@ -295,10 +290,8 @@ impl Default for PageServerConf {
listen_pg_addr: String::new(),
listen_http_addr: String::new(),
listen_https_addr: None,
listen_grpc_addr: None,
pg_auth_type: AuthType::Trust,
http_auth_type: AuthType::Trust,
grpc_auth_type: AuthType::Trust,
no_sync: false,
}
}
@@ -313,10 +306,8 @@ pub struct NeonLocalInitPageserverConf {
pub listen_pg_addr: String,
pub listen_http_addr: String,
pub listen_https_addr: Option<String>,
pub listen_grpc_addr: Option<String>,
pub pg_auth_type: AuthType,
pub http_auth_type: AuthType,
pub grpc_auth_type: AuthType,
#[serde(default, skip_serializing_if = "std::ops::Not::not")]
pub no_sync: bool,
#[serde(flatten)]
@@ -330,10 +321,8 @@ impl From<&NeonLocalInitPageserverConf> for PageServerConf {
listen_pg_addr,
listen_http_addr,
listen_https_addr,
listen_grpc_addr,
pg_auth_type,
http_auth_type,
grpc_auth_type,
no_sync,
other: _,
} = conf;
@@ -342,9 +331,7 @@ impl From<&NeonLocalInitPageserverConf> for PageServerConf {
listen_pg_addr: listen_pg_addr.clone(),
listen_http_addr: listen_http_addr.clone(),
listen_https_addr: listen_https_addr.clone(),
listen_grpc_addr: listen_grpc_addr.clone(),
pg_auth_type: *pg_auth_type,
grpc_auth_type: *grpc_auth_type,
http_auth_type: *http_auth_type,
no_sync: *no_sync,
}
@@ -720,10 +707,8 @@ impl LocalEnv {
listen_pg_addr: String,
listen_http_addr: String,
listen_https_addr: Option<String>,
listen_grpc_addr: Option<String>,
pg_auth_type: AuthType,
http_auth_type: AuthType,
grpc_auth_type: AuthType,
#[serde(default)]
no_sync: bool,
}
@@ -747,10 +732,8 @@ impl LocalEnv {
listen_pg_addr,
listen_http_addr,
listen_https_addr,
listen_grpc_addr,
pg_auth_type,
http_auth_type,
grpc_auth_type,
no_sync,
} = config_toml;
let IdentityTomlSubset {
@@ -767,10 +750,8 @@ impl LocalEnv {
listen_pg_addr,
listen_http_addr,
listen_https_addr,
listen_grpc_addr,
pg_auth_type,
http_auth_type,
grpc_auth_type,
no_sync,
};
pageservers.push(conf);

View File

@@ -16,7 +16,6 @@ use std::time::Duration;
use anyhow::{Context, bail};
use camino::Utf8PathBuf;
use pageserver_api::config::{DEFAULT_GRPC_LISTEN_PORT, DEFAULT_HTTP_LISTEN_PORT};
use pageserver_api::models::{self, TenantInfo, TimelineInfo};
use pageserver_api::shard::TenantShardId;
use pageserver_client::mgmt_api;
@@ -130,9 +129,7 @@ impl PageServerNode {
));
}
if [conf.http_auth_type, conf.pg_auth_type, conf.grpc_auth_type]
.contains(&AuthType::NeonJWT)
{
if conf.http_auth_type != AuthType::Trust || conf.pg_auth_type != AuthType::Trust {
// Keys are generated in the toplevel repo dir, pageservers' workdirs
// are one level below that, so refer to keys with ../
overrides.push("auth_validation_public_key_path='../auth_public_key.pem'".to_owned());
@@ -253,10 +250,9 @@ impl PageServerNode {
// the storage controller
let metadata_path = datadir.join("metadata.json");
let http_host = "localhost".to_string();
let (_, http_port) =
let (_http_host, http_port) =
parse_host_port(&self.conf.listen_http_addr).expect("Unable to parse listen_http_addr");
let http_port = http_port.unwrap_or(DEFAULT_HTTP_LISTEN_PORT);
let http_port = http_port.unwrap_or(9898);
let https_port = match self.conf.listen_https_addr.as_ref() {
Some(https_addr) => {
@@ -267,13 +263,6 @@ impl PageServerNode {
None => None,
};
let (mut grpc_host, mut grpc_port) = (None, None);
if let Some(grpc_addr) = &self.conf.listen_grpc_addr {
let (_, port) = parse_host_port(grpc_addr).expect("Unable to parse listen_grpc_addr");
grpc_host = Some("localhost".to_string());
grpc_port = Some(port.unwrap_or(DEFAULT_GRPC_LISTEN_PORT));
}
// Intentionally hand-craft JSON: this acts as an implicit format compat test
// in case the pageserver-side structure is edited, and reflects the real life
// situation: the metadata is written by some other script.
@@ -282,9 +271,7 @@ impl PageServerNode {
serde_json::to_vec(&pageserver_api::config::NodeMetadata {
postgres_host: "localhost".to_string(),
postgres_port: self.pg_connection_config.port(),
grpc_host,
grpc_port,
http_host,
http_host: "localhost".to_string(),
http_port,
https_port,
other: HashMap::from([(
@@ -524,6 +511,11 @@ impl PageServerNode {
.map(|x| x.parse::<bool>())
.transpose()
.context("Failed to parse 'timeline_offloading' as bool")?,
wal_receiver_protocol_override: settings
.remove("wal_receiver_protocol_override")
.map(serde_json::from_str)
.transpose()
.context("parse `wal_receiver_protocol_override` from json")?,
rel_size_v2_enabled: settings
.remove("rel_size_v2_enabled")
.map(|x| x.parse::<bool>())
@@ -554,16 +546,6 @@ impl PageServerNode {
.map(serde_json::from_str)
.transpose()
.context("Falied to parse 'sampling_ratio'")?,
relsize_snapshot_cache_capacity: settings
.remove("relsize snapshot cache capacity")
.map(|x| x.parse::<usize>())
.transpose()
.context("Falied to parse 'relsize_snapshot_cache_capacity' as integer")?,
basebackup_cache_enabled: settings
.remove("basebackup_cache_enabled")
.map(|x| x.parse::<bool>())
.transpose()
.context("Failed to parse 'basebackup_cache_enabled' as bool")?,
};
if !settings.is_empty() {
bail!("Unrecognized tenant settings: {settings:?}")
@@ -646,16 +628,4 @@ impl PageServerNode {
Ok(())
}
pub async fn timeline_info(
&self,
tenant_shard_id: TenantShardId,
timeline_id: TimelineId,
force_await_logical_size: mgmt_api::ForceAwaitLogicalSize,
) -> anyhow::Result<TimelineInfo> {
let timeline_info = self
.http_client
.timeline_info(tenant_shard_id, timeline_id, force_await_logical_size)
.await?;
Ok(timeline_info)
}
}

View File

@@ -6,6 +6,7 @@
//! .neon/safekeepers/<safekeeper id>
//! ```
use std::error::Error as _;
use std::future::Future;
use std::io::Write;
use std::path::PathBuf;
use std::time::Duration;
@@ -13,9 +14,9 @@ use std::{io, result};
use anyhow::Context;
use camino::Utf8PathBuf;
use http_utils::error::HttpErrorBody;
use postgres_connection::PgConnectionConfig;
use safekeeper_api::models::TimelineCreateRequest;
use safekeeper_client::mgmt_api;
use reqwest::{IntoUrl, Method};
use thiserror::Error;
use utils::auth::{Claims, Scope};
use utils::id::NodeId;
@@ -34,14 +35,25 @@ pub enum SafekeeperHttpError {
type Result<T> = result::Result<T, SafekeeperHttpError>;
fn err_from_client_err(err: mgmt_api::Error) -> SafekeeperHttpError {
use mgmt_api::Error::*;
match err {
ApiError(_, str) => SafekeeperHttpError::Response(str),
Cancelled => SafekeeperHttpError::Response("Cancelled".to_owned()),
ReceiveBody(err) => SafekeeperHttpError::Transport(err),
ReceiveErrorBody(err) => SafekeeperHttpError::Response(err),
Timeout(str) => SafekeeperHttpError::Response(format!("timeout: {str}")),
pub(crate) trait ResponseErrorMessageExt: Sized {
fn error_from_body(self) -> impl Future<Output = Result<Self>> + Send;
}
impl ResponseErrorMessageExt for reqwest::Response {
async fn error_from_body(self) -> Result<Self> {
let status = self.status();
if !(status.is_client_error() || status.is_server_error()) {
return Ok(self);
}
// 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>().await {
Ok(err_body) => format!("Error: {}", err_body.msg),
Err(_) => format!("Http error ({}) at {}.", status.as_u16(), url),
},
))
}
}
@@ -58,8 +70,9 @@ pub struct SafekeeperNode {
pub pg_connection_config: PgConnectionConfig,
pub env: LocalEnv,
pub http_client: mgmt_api::Client,
pub http_client: reqwest::Client,
pub listen_addr: String,
pub http_base_url: String,
}
impl SafekeeperNode {
@@ -69,14 +82,13 @@ impl SafekeeperNode {
} else {
"127.0.0.1".to_string()
};
let jwt = None;
let http_base_url = format!("http://{}:{}", listen_addr, conf.http_port);
SafekeeperNode {
id: conf.id,
conf: conf.clone(),
pg_connection_config: Self::safekeeper_connection_config(&listen_addr, conf.pg_port),
env: env.clone(),
http_client: mgmt_api::Client::new(env.create_http_client(), http_base_url, jwt),
http_client: env.create_http_client(),
http_base_url: format!("http://{}:{}/v1", listen_addr, conf.http_port),
listen_addr,
}
}
@@ -266,19 +278,20 @@ impl SafekeeperNode {
)
}
pub async fn check_status(&self) -> Result<()> {
self.http_client
.status()
.await
.map_err(err_from_client_err)?;
Ok(())
fn http_request<U: IntoUrl>(&self, method: Method, url: U) -> reqwest::RequestBuilder {
// TODO: authentication
//if self.env.auth_type == AuthType::NeonJWT {
// builder = builder.bearer_auth(&self.env.safekeeper_auth_token)
//}
self.http_client.request(method, url)
}
pub async fn create_timeline(&self, req: &TimelineCreateRequest) -> Result<()> {
self.http_client
.create_timeline(req)
.await
.map_err(err_from_client_err)?;
pub async fn check_status(&self) -> Result<()> {
self.http_request(Method::GET, format!("{}/{}", self.http_base_url, "status"))
.send()
.await?
.error_from_body()
.await?;
Ok(())
}
}

View File

@@ -628,10 +628,6 @@ impl StorageController {
args.push("--timelines-onto-safekeepers".to_string());
}
if let Some(sk_cnt) = self.config.timeline_safekeeper_count {
args.push(format!("--timeline-safekeeper-count={sk_cnt}"));
}
println!("Starting storage controller");
background_process::start_process(

View File

@@ -36,10 +36,6 @@ enum Command {
listen_pg_addr: String,
#[arg(long)]
listen_pg_port: u16,
#[arg(long)]
listen_grpc_addr: Option<String>,
#[arg(long)]
listen_grpc_port: Option<u16>,
#[arg(long)]
listen_http_addr: String,
@@ -65,16 +61,10 @@ enum Command {
#[arg(long)]
scheduling: Option<NodeSchedulingPolicy>,
},
// Set a node status as deleted.
NodeDelete {
#[arg(long)]
node_id: NodeId,
},
/// Delete a tombstone of node from the storage controller.
NodeDeleteTombstone {
#[arg(long)]
node_id: NodeId,
},
/// Modify a tenant's policies in the storage controller
TenantPolicy {
#[arg(long)]
@@ -92,8 +82,6 @@ enum Command {
},
/// List nodes known to the storage controller
Nodes {},
/// List soft deleted nodes known to the storage controller
NodeTombstones {},
/// List tenants known to the storage controller
Tenants {
/// If this field is set, it will list the tenants on a specific node
@@ -422,8 +410,6 @@ async fn main() -> anyhow::Result<()> {
node_id,
listen_pg_addr,
listen_pg_port,
listen_grpc_addr,
listen_grpc_port,
listen_http_addr,
listen_http_port,
listen_https_port,
@@ -437,8 +423,6 @@ async fn main() -> anyhow::Result<()> {
node_id,
listen_pg_addr,
listen_pg_port,
listen_grpc_addr,
listen_grpc_port,
listen_http_addr,
listen_http_port,
listen_https_port,
@@ -916,39 +900,6 @@ async fn main() -> anyhow::Result<()> {
.dispatch::<(), ()>(Method::DELETE, format!("control/v1/node/{node_id}"), None)
.await?;
}
Command::NodeDeleteTombstone { node_id } => {
storcon_client
.dispatch::<(), ()>(
Method::DELETE,
format!("debug/v1/tombstone/{node_id}"),
None,
)
.await?;
}
Command::NodeTombstones {} => {
let mut resp = storcon_client
.dispatch::<(), Vec<NodeDescribeResponse>>(
Method::GET,
"debug/v1/tombstone".to_string(),
None,
)
.await?;
resp.sort_by(|a, b| a.listen_http_addr.cmp(&b.listen_http_addr));
let mut table = comfy_table::Table::new();
table.set_header(["Id", "Hostname", "AZ", "Scheduling", "Availability"]);
for node in resp {
table.add_row([
format!("{}", node.id),
node.listen_http_addr,
node.availability_zone_id,
format!("{:?}", node.scheduling),
format!("{:?}", node.availability),
]);
}
println!("{table}");
}
Command::TenantSetTimeBasedEviction {
tenant_id,
period,

View File

@@ -13,6 +13,6 @@ RUN echo 'Acquire::Retries "5";' > /etc/apt/apt.conf.d/80-retries && \
jq \
netcat-openbsd
#This is required for the pg_hintplan test
RUN mkdir -p /ext-src/pg_hint_plan-src /postgres/contrib/file_fdw /ext-src/postgis-src/ && chown postgres /ext-src/pg_hint_plan-src /postgres/contrib/file_fdw /ext-src/postgis-src
RUN mkdir -p /ext-src/pg_hint_plan-src /postgres/contrib/file_fdw && chown postgres /ext-src/pg_hint_plan-src /postgres/contrib/file_fdw
USER postgres

View File

@@ -1,18 +1,18 @@
#!/usr/bin/env bash
#!/bin/bash
set -eux
# Generate a random tenant or timeline ID
#
# Takes a variable name as argument. The result is stored in that variable.
generate_id() {
local -n resvar=${1}
printf -v resvar '%08x%08x%08x%08x' ${SRANDOM} ${SRANDOM} ${SRANDOM} ${SRANDOM}
local -n resvar=$1
printf -v resvar '%08x%08x%08x%08x' $SRANDOM $SRANDOM $SRANDOM $SRANDOM
}
PG_VERSION=${PG_VERSION:-14}
readonly CONFIG_FILE_ORG=/var/db/postgres/configs/config.json
readonly CONFIG_FILE=/tmp/config.json
CONFIG_FILE_ORG=/var/db/postgres/configs/config.json
CONFIG_FILE=/tmp/config.json
# Test that the first library path that the dynamic loader looks in is the path
# that we use for custom compiled software
@@ -20,17 +20,17 @@ first_path="$(ldconfig --verbose 2>/dev/null \
| grep --invert-match ^$'\t' \
| cut --delimiter=: --fields=1 \
| head --lines=1)"
test "${first_path}" = '/usr/local/lib'
test "$first_path" == '/usr/local/lib' || true # Remove the || true in a follow-up PR. Needed for backwards compat.
echo "Waiting pageserver become ready."
while ! nc -z pageserver 6400; do
sleep 1
sleep 1;
done
echo "Page server is ready."
cp "${CONFIG_FILE_ORG}" "${CONFIG_FILE}"
cp ${CONFIG_FILE_ORG} ${CONFIG_FILE}
if [[ -n "${TENANT_ID:-}" && -n "${TIMELINE_ID:-}" ]]; then
if [ -n "${TENANT_ID:-}" ] && [ -n "${TIMELINE_ID:-}" ]; then
tenant_id=${TENANT_ID}
timeline_id=${TIMELINE_ID}
else
@@ -41,7 +41,7 @@ else
"http://pageserver:9898/v1/tenant"
)
tenant_id=$(curl "${PARAMS[@]}" | jq -r .[0].id)
if [[ -z "${tenant_id}" || "${tenant_id}" = null ]]; then
if [ -z "${tenant_id}" ] || [ "${tenant_id}" = null ]; then
echo "Create a tenant"
generate_id tenant_id
PARAMS=(
@@ -51,7 +51,7 @@ else
"http://pageserver:9898/v1/tenant/${tenant_id}/location_config"
)
result=$(curl "${PARAMS[@]}")
printf '%s\n' "${result}" | jq .
echo $result | jq .
fi
echo "Check if a timeline present"
@@ -61,7 +61,7 @@ else
"http://pageserver:9898/v1/tenant/${tenant_id}/timeline"
)
timeline_id=$(curl "${PARAMS[@]}" | jq -r .[0].timeline_id)
if [[ -z "${timeline_id}" || "${timeline_id}" = null ]]; then
if [ -z "${timeline_id}" ] || [ "${timeline_id}" = null ]; then
generate_id timeline_id
PARAMS=(
-sbf
@@ -71,7 +71,7 @@ else
"http://pageserver:9898/v1/tenant/${tenant_id}/timeline/"
)
result=$(curl "${PARAMS[@]}")
printf '%s\n' "${result}" | jq .
echo $result | jq .
fi
fi
@@ -82,10 +82,10 @@ else
fi
echo "Adding pgx_ulid"
shared_libraries=$(jq -r '.spec.cluster.settings[] | select(.name=="shared_preload_libraries").value' ${CONFIG_FILE})
sed -i "s|${shared_libraries}|${shared_libraries},${ulid_extension}|" ${CONFIG_FILE}
sed -i "s/${shared_libraries}/${shared_libraries},${ulid_extension}/" ${CONFIG_FILE}
echo "Overwrite tenant id and timeline id in spec file"
sed -i "s|TENANT_ID|${tenant_id}|" ${CONFIG_FILE}
sed -i "s|TIMELINE_ID|${timeline_id}|" ${CONFIG_FILE}
sed -i "s/TENANT_ID/${tenant_id}/" ${CONFIG_FILE}
sed -i "s/TIMELINE_ID/${timeline_id}/" ${CONFIG_FILE}
cat ${CONFIG_FILE}
@@ -93,6 +93,5 @@ echo "Start compute node"
/usr/local/bin/compute_ctl --pgdata /var/db/postgres/compute \
-C "postgresql://cloud_admin@localhost:55433/postgres" \
-b /usr/local/bin/postgres \
--compute-id "compute-${RANDOM}" \
--config "${CONFIG_FILE}"
--dev
--compute-id "compute-$RANDOM" \
--config "$CONFIG_FILE"

View File

@@ -186,14 +186,13 @@ services:
neon-test-extensions:
profiles: ["test-extensions"]
image: ${REPOSITORY:-ghcr.io/neondatabase}/neon-test-extensions-v${PG_TEST_VERSION:-${PG_VERSION:-16}}:${TEST_EXTENSIONS_TAG:-${TAG:-latest}}
image: ${REPOSITORY:-ghcr.io/neondatabase}/neon-test-extensions-v${PG_TEST_VERSION:-16}:${TEST_EXTENSIONS_TAG:-${TAG:-latest}}
environment:
- PGUSER=${PGUSER:-cloud_admin}
- PGPASSWORD=${PGPASSWORD:-cloud_admin}
- PGPASSWORD=cloud_admin
entrypoint:
- "/bin/bash"
- "-c"
command:
- sleep 3600
- sleep 1800
depends_on:
- compute

View File

@@ -54,15 +54,6 @@ for pg_version in ${TEST_VERSION_ONLY-14 15 16 17}; do
# It cannot be moved to Dockerfile now because the database directory is created after the start of the container
echo Adding dummy config
docker compose exec compute touch /var/db/postgres/compute/compute_ctl_temp_override.conf
# Prepare for the PostGIS test
docker compose exec compute mkdir -p /tmp/pgis_reg/pgis_reg_tmp
TMPDIR=$(mktemp -d)
docker compose cp neon-test-extensions:/ext-src/postgis-src/raster/test "${TMPDIR}"
docker compose cp neon-test-extensions:/ext-src/postgis-src/regress/00-regress-install "${TMPDIR}"
docker compose exec compute mkdir -p /ext-src/postgis-src/raster /ext-src/postgis-src/regress /ext-src/postgis-src/regress/00-regress-install
docker compose cp "${TMPDIR}/test" compute:/ext-src/postgis-src/raster/test
docker compose cp "${TMPDIR}/00-regress-install" compute:/ext-src/postgis-src/regress
rm -rf "${TMPDIR}"
# The following block copies the files for the pg_hintplan test to the compute node for the extension test in an isolated docker-compose environment
TMPDIR=$(mktemp -d)
docker compose cp neon-test-extensions:/ext-src/pg_hint_plan-src/data "${TMPDIR}/data"
@@ -77,7 +68,7 @@ for pg_version in ${TEST_VERSION_ONLY-14 15 16 17}; do
docker compose exec -T neon-test-extensions bash -c "(cd /postgres && patch -p1)" <"../compute/patches/contrib_pg${pg_version}.patch"
# We are running tests now
rm -f testout.txt testout_contrib.txt
docker compose exec -e USE_PGXS=1 -e SKIP=timescaledb-src,rdkit-src,pg_jsonschema-src,kq_imcx-src,wal2json_2_5-src,rag_jina_reranker_v1_tiny_en-src,rag_bge_small_en_v15-src \
docker compose exec -e USE_PGXS=1 -e SKIP=timescaledb-src,rdkit-src,postgis-src,pg_jsonschema-src,kq_imcx-src,wal2json_2_5-src,rag_jina_reranker_v1_tiny_en-src,rag_bge_small_en_v15-src \
neon-test-extensions /run-tests.sh /ext-src | tee testout.txt && EXT_SUCCESS=1 || EXT_SUCCESS=0
docker compose exec -e SKIP=start-scripts,postgres_fdw,ltree_plpython,jsonb_plpython,jsonb_plperl,hstore_plpython,hstore_plperl,dblink,bool_plperl \
neon-test-extensions /run-tests.sh /postgres/contrib | tee testout_contrib.txt && CONTRIB_SUCCESS=1 || CONTRIB_SUCCESS=0

View File

@@ -1,16 +0,0 @@
#!/usr/bin/env bash
set -ex
cd "$(dirname "${0}")"
PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
dropdb --if-exists contrib_regression
createdb contrib_regression
cd h3_postgis/test
psql -d contrib_regression -c "CREATE EXTENSION postgis" -c "CREATE EXTENSION postgis_raster" -c "CREATE EXTENSION h3" -c "CREATE EXTENSION h3_postgis"
TESTS=$(echo sql/* | sed 's|sql/||g; s|\.sql||g')
${PG_REGRESS} --use-existing --dbname contrib_regression ${TESTS}
cd ../../h3/test
TESTS=$(echo sql/* | sed 's|sql/||g; s|\.sql||g')
dropdb --if-exists contrib_regression
createdb contrib_regression
psql -d contrib_regression -c "CREATE EXTENSION h3"
${PG_REGRESS} --use-existing --dbname contrib_regression ${TESTS}

View File

@@ -1,7 +0,0 @@
#!/bin/sh
set -ex
cd "$(dirname ${0})"
PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
cd h3/test
TESTS=$(echo sql/* | sed 's|sql/||g; s|\.sql||g')
${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin' --dbname=contrib_regression ${TESTS}

View File

@@ -1,6 +0,0 @@
#!/bin/sh
set -ex
cd "$(dirname "${0}")"
if [ -f Makefile ]; then
make installcheck
fi

View File

@@ -1,9 +0,0 @@
#!/bin/sh
set -ex
cd "$(dirname ${0})"
[ -f Makefile ] || exit 0
dropdb --if-exist contrib_regression
createdb contrib_regression
PG_REGRESS=$(dirname "$(pg_config --pgxs)")/../test/regress/pg_regress
TESTS=$(echo sql/* | sed 's|sql/||g; s|\.sql||g')
${PG_REGRESS} --use-existing --inputdir=./ --bindir='/usr/local/pgsql/bin' --dbname=contrib_regression ${TESTS}

View File

@@ -1,70 +0,0 @@
# PostGIS Testing in Neon
This directory contains configuration files and patches for running PostGIS tests in the Neon database environment.
## Overview
PostGIS is a spatial database extension for PostgreSQL that adds support for geographic objects. Testing PostGIS compatibility ensures that Neon's modifications to PostgreSQL don't break compatibility with this critical extension.
## PostGIS Versions
- PostgreSQL v17: PostGIS 3.5.0
- PostgreSQL v14/v15/v16: PostGIS 3.3.3
## Test Configuration
The test setup includes:
- `postgis-no-upgrade-test.patch`: Disables upgrade tests by removing the upgrade test section from regress/runtest.mk
- `postgis-regular-v16.patch`: Version-specific patch for PostgreSQL v16
- `postgis-regular-v17.patch`: Version-specific patch for PostgreSQL v17
- `regular-test.sh`: Script to run PostGIS tests as a regular user
- `neon-test.sh`: Script to handle version-specific test configurations
- `raster_outdb_template.sql`: Template for raster tests with explicit file paths
## Excluded Tests
**Important Note:** The test exclusions listed below are specifically for regular-user tests against staging instances. These exclusions are necessary because staging instances run with limited privileges and cannot perform operations requiring superuser access. Docker-compose based tests are not affected by these exclusions.
### Tests Requiring Superuser Permissions
These tests cannot be run as a regular user:
- `estimatedextent`
- `regress/core/legacy`
- `regress/core/typmod`
- `regress/loader/TestSkipANALYZE`
- `regress/loader/TestANALYZE`
### Tests Requiring Filesystem Access
These tests need direct filesystem access that is only possible for superusers:
- `loader/load_outdb`
### Tests with Flaky Results
These tests have assumptions that don't always hold true:
- `regress/core/computed_columns` - Assumes computed columns always outperform alternatives, which is not consistently true
### Tests Requiring Tunable Parameter Modifications
These tests attempt to modify the `postgis.gdal_enabled_drivers` parameter, which is only accessible to superusers:
- `raster/test/regress/rt_wkb`
- `raster/test/regress/rt_addband`
- `raster/test/regress/rt_setbandpath`
- `raster/test/regress/rt_fromgdalraster`
- `raster/test/regress/rt_asgdalraster`
- `raster/test/regress/rt_astiff`
- `raster/test/regress/rt_asjpeg`
- `raster/test/regress/rt_aspng`
- `raster/test/regress/permitted_gdal_drivers`
- Loader tests: `BasicOutDB`, `Tiled10x10`, `Tiled10x10Copy`, `Tiled8x8`, `TiledAuto`, `TiledAutoSkipNoData`, `TiledAutoCopyn`
### Topology Tests (v17 only)
- `populate_topology_layer`
- `renametopogeometrycolumn`
## Other Modifications
- Binary.sql tests are modified to use explicit file paths
- Server-side SQL COPY commands (which require superuser privileges) are converted to client-side `\copy` commands
- Upgrade tests are disabled

View File

@@ -1,6 +0,0 @@
#!/bin/sh
set -ex
cd "$(dirname "$0")"
patch -p1 <"postgis-common-${PG_VERSION}.patch"
trap 'echo Cleaning up; patch -R -p1 <postgis-common-${PG_VERSION}.patch' EXIT
make installcheck-base

View File

@@ -1,37 +0,0 @@
diff --git a/regress/core/tests.mk b/regress/core/tests.mk
index 3abd7bc..64a9254 100644
--- a/regress/core/tests.mk
+++ b/regress/core/tests.mk
@@ -144,11 +144,6 @@ TESTS_SLOW = \
$(top_srcdir)/regress/core/concave_hull_hard \
$(top_srcdir)/regress/core/knn_recheck
-ifeq ($(shell expr "$(POSTGIS_PGSQL_VERSION)" ">=" 120),1)
- TESTS += \
- $(top_srcdir)/regress/core/computed_columns
-endif
-
ifeq ($(shell expr "$(POSTGIS_GEOS_VERSION)" ">=" 30700),1)
# GEOS-3.7 adds:
# ST_FrechetDistance
diff --git a/regress/runtest.mk b/regress/runtest.mk
index c051f03..010e493 100644
--- a/regress/runtest.mk
+++ b/regress/runtest.mk
@@ -24,16 +24,6 @@ check-regress:
POSTGIS_TOP_BUILD_DIR=$(abs_top_builddir) $(PERL) $(top_srcdir)/regress/run_test.pl $(RUNTESTFLAGS) $(RUNTESTFLAGS_INTERNAL) $(TESTS)
- @if echo "$(RUNTESTFLAGS)" | grep -vq -- --upgrade; then \
- echo "Running upgrade test as RUNTESTFLAGS did not contain that"; \
- POSTGIS_TOP_BUILD_DIR=$(abs_top_builddir) $(PERL) $(top_srcdir)/regress/run_test.pl \
- --upgrade \
- $(RUNTESTFLAGS) \
- $(RUNTESTFLAGS_INTERNAL) \
- $(TESTS); \
- else \
- echo "Skipping upgrade test as RUNTESTFLAGS already requested upgrades"; \
- fi
check-long:
$(PERL) $(top_srcdir)/regress/run_test.pl $(RUNTESTFLAGS) $(TESTS) $(TESTS_SLOW)

View File

@@ -1,35 +0,0 @@
diff --git a/regress/core/tests.mk b/regress/core/tests.mk
index 9e05244..90987df 100644
--- a/regress/core/tests.mk
+++ b/regress/core/tests.mk
@@ -143,8 +143,7 @@ TESTS += \
$(top_srcdir)/regress/core/oriented_envelope \
$(top_srcdir)/regress/core/point_coordinates \
$(top_srcdir)/regress/core/out_geojson \
- $(top_srcdir)/regress/core/wrapx \
- $(top_srcdir)/regress/core/computed_columns
+ $(top_srcdir)/regress/core/wrapx
# Slow slow tests
TESTS_SLOW = \
diff --git a/regress/runtest.mk b/regress/runtest.mk
index 4b95b7e..449d5a2 100644
--- a/regress/runtest.mk
+++ b/regress/runtest.mk
@@ -24,16 +24,6 @@ check-regress:
@POSTGIS_TOP_BUILD_DIR=$(abs_top_builddir) $(PERL) $(top_srcdir)/regress/run_test.pl $(RUNTESTFLAGS) $(RUNTESTFLAGS_INTERNAL) $(TESTS)
- @if echo "$(RUNTESTFLAGS)" | grep -vq -- --upgrade; then \
- echo "Running upgrade test as RUNTESTFLAGS did not contain that"; \
- POSTGIS_TOP_BUILD_DIR=$(abs_top_builddir) $(PERL) $(top_srcdir)/regress/run_test.pl \
- --upgrade \
- $(RUNTESTFLAGS) \
- $(RUNTESTFLAGS_INTERNAL) \
- $(TESTS); \
- else \
- echo "Skipping upgrade test as RUNTESTFLAGS already requested upgrades"; \
- fi
check-long:
$(PERL) $(top_srcdir)/regress/run_test.pl $(RUNTESTFLAGS) $(TESTS) $(TESTS_SLOW)

View File

@@ -1,186 +0,0 @@
diff --git a/raster/test/regress/tests.mk b/raster/test/regress/tests.mk
index 00918e1..7e2b6cd 100644
--- a/raster/test/regress/tests.mk
+++ b/raster/test/regress/tests.mk
@@ -17,9 +17,7 @@ override RUNTESTFLAGS_INTERNAL := \
$(RUNTESTFLAGS_INTERNAL) \
--after-upgrade-script $(top_srcdir)/raster/test/regress/hooks/hook-after-upgrade-raster.sql
-RASTER_TEST_FIRST = \
- $(top_srcdir)/raster/test/regress/check_gdal \
- $(top_srcdir)/raster/test/regress/loader/load_outdb
+RASTER_TEST_FIRST =
RASTER_TEST_LAST = \
$(top_srcdir)/raster/test/regress/clean
@@ -33,9 +31,7 @@ RASTER_TEST_IO = \
RASTER_TEST_BASIC_FUNC = \
$(top_srcdir)/raster/test/regress/rt_bytea \
- $(top_srcdir)/raster/test/regress/rt_wkb \
$(top_srcdir)/raster/test/regress/box3d \
- $(top_srcdir)/raster/test/regress/rt_addband \
$(top_srcdir)/raster/test/regress/rt_band \
$(top_srcdir)/raster/test/regress/rt_tile
@@ -73,16 +69,10 @@ RASTER_TEST_BANDPROPS = \
$(top_srcdir)/raster/test/regress/rt_neighborhood \
$(top_srcdir)/raster/test/regress/rt_nearestvalue \
$(top_srcdir)/raster/test/regress/rt_pixelofvalue \
- $(top_srcdir)/raster/test/regress/rt_polygon \
- $(top_srcdir)/raster/test/regress/rt_setbandpath
+ $(top_srcdir)/raster/test/regress/rt_polygon
RASTER_TEST_UTILITY = \
$(top_srcdir)/raster/test/regress/rt_utility \
- $(top_srcdir)/raster/test/regress/rt_fromgdalraster \
- $(top_srcdir)/raster/test/regress/rt_asgdalraster \
- $(top_srcdir)/raster/test/regress/rt_astiff \
- $(top_srcdir)/raster/test/regress/rt_asjpeg \
- $(top_srcdir)/raster/test/regress/rt_aspng \
$(top_srcdir)/raster/test/regress/rt_reclass \
$(top_srcdir)/raster/test/regress/rt_gdalwarp \
$(top_srcdir)/raster/test/regress/rt_gdalcontour \
@@ -120,21 +110,13 @@ RASTER_TEST_SREL = \
RASTER_TEST_BUGS = \
$(top_srcdir)/raster/test/regress/bug_test_car5 \
- $(top_srcdir)/raster/test/regress/permitted_gdal_drivers \
$(top_srcdir)/raster/test/regress/tickets
RASTER_TEST_LOADER = \
$(top_srcdir)/raster/test/regress/loader/Basic \
$(top_srcdir)/raster/test/regress/loader/Projected \
$(top_srcdir)/raster/test/regress/loader/BasicCopy \
- $(top_srcdir)/raster/test/regress/loader/BasicFilename \
- $(top_srcdir)/raster/test/regress/loader/BasicOutDB \
- $(top_srcdir)/raster/test/regress/loader/Tiled10x10 \
- $(top_srcdir)/raster/test/regress/loader/Tiled10x10Copy \
- $(top_srcdir)/raster/test/regress/loader/Tiled8x8 \
- $(top_srcdir)/raster/test/regress/loader/TiledAuto \
- $(top_srcdir)/raster/test/regress/loader/TiledAutoSkipNoData \
- $(top_srcdir)/raster/test/regress/loader/TiledAutoCopyn
+ $(top_srcdir)/raster/test/regress/loader/BasicFilename
RASTER_TESTS := $(RASTER_TEST_FIRST) \
$(RASTER_TEST_METADATA) $(RASTER_TEST_IO) $(RASTER_TEST_BASIC_FUNC) \
diff --git a/regress/core/binary.sql b/regress/core/binary.sql
index 7a36b65..ad78fc7 100644
--- a/regress/core/binary.sql
+++ b/regress/core/binary.sql
@@ -1,4 +1,5 @@
SET client_min_messages TO warning;
+
CREATE SCHEMA tm;
CREATE TABLE tm.geoms (id serial, g geometry);
@@ -31,24 +32,39 @@ SELECT st_force4d(g) FROM tm.geoms WHERE id < 15 ORDER BY id;
INSERT INTO tm.geoms(g)
SELECT st_setsrid(g,4326) FROM tm.geoms ORDER BY id;
-COPY tm.geoms TO :tmpfile WITH BINARY;
+-- define temp file path
+\set tmpfile '/tmp/postgis_binary_test.dat'
+
+-- export
+\set command '\\copy tm.geoms TO ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+-- import
CREATE TABLE tm.geoms_in AS SELECT * FROM tm.geoms LIMIT 0;
-COPY tm.geoms_in FROM :tmpfile WITH BINARY;
-SELECT 'geometry', count(*) FROM tm.geoms_in i, tm.geoms o WHERE i.id = o.id
- AND ST_OrderingEquals(i.g, o.g);
+\set command '\\copy tm.geoms_in FROM ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+SELECT 'geometry', count(*) FROM tm.geoms_in i, tm.geoms o
+WHERE i.id = o.id AND ST_OrderingEquals(i.g, o.g);
CREATE TABLE tm.geogs AS SELECT id,g::geography FROM tm.geoms
WHERE geometrytype(g) NOT LIKE '%CURVE%'
AND geometrytype(g) NOT LIKE '%CIRCULAR%'
AND geometrytype(g) NOT LIKE '%SURFACE%'
AND geometrytype(g) NOT LIKE 'TRIANGLE%'
- AND geometrytype(g) NOT LIKE 'TIN%'
-;
+ AND geometrytype(g) NOT LIKE 'TIN%';
-COPY tm.geogs TO :tmpfile WITH BINARY;
+-- export
+\set command '\\copy tm.geogs TO ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+-- import
CREATE TABLE tm.geogs_in AS SELECT * FROM tm.geogs LIMIT 0;
-COPY tm.geogs_in FROM :tmpfile WITH BINARY;
-SELECT 'geometry', count(*) FROM tm.geogs_in i, tm.geogs o WHERE i.id = o.id
- AND ST_OrderingEquals(i.g::geometry, o.g::geometry);
+\set command '\\copy tm.geogs_in FROM ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+SELECT 'geometry', count(*) FROM tm.geogs_in i, tm.geogs o
+WHERE i.id = o.id AND ST_OrderingEquals(i.g::geometry, o.g::geometry);
DROP SCHEMA tm CASCADE;
+
diff --git a/regress/core/tests.mk b/regress/core/tests.mk
index 64a9254..94903c3 100644
--- a/regress/core/tests.mk
+++ b/regress/core/tests.mk
@@ -23,7 +23,6 @@ current_dir := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
RUNTESTFLAGS_INTERNAL += \
--before-upgrade-script $(top_srcdir)/regress/hooks/hook-before-upgrade.sql \
--after-upgrade-script $(top_srcdir)/regress/hooks/hook-after-upgrade.sql \
- --after-create-script $(top_srcdir)/regress/hooks/hook-after-create.sql \
--before-uninstall-script $(top_srcdir)/regress/hooks/hook-before-uninstall.sql
TESTS += \
@@ -40,7 +39,6 @@ TESTS += \
$(top_srcdir)/regress/core/dumppoints \
$(top_srcdir)/regress/core/dumpsegments \
$(top_srcdir)/regress/core/empty \
- $(top_srcdir)/regress/core/estimatedextent \
$(top_srcdir)/regress/core/forcecurve \
$(top_srcdir)/regress/core/flatgeobuf \
$(top_srcdir)/regress/core/geography \
@@ -55,7 +53,6 @@ TESTS += \
$(top_srcdir)/regress/core/out_marc21 \
$(top_srcdir)/regress/core/in_encodedpolyline \
$(top_srcdir)/regress/core/iscollection \
- $(top_srcdir)/regress/core/legacy \
$(top_srcdir)/regress/core/letters \
$(top_srcdir)/regress/core/long_xact \
$(top_srcdir)/regress/core/lwgeom_regress \
@@ -112,7 +109,6 @@ TESTS += \
$(top_srcdir)/regress/core/temporal_knn \
$(top_srcdir)/regress/core/tickets \
$(top_srcdir)/regress/core/twkb \
- $(top_srcdir)/regress/core/typmod \
$(top_srcdir)/regress/core/wkb \
$(top_srcdir)/regress/core/wkt \
$(top_srcdir)/regress/core/wmsservers \
diff --git a/regress/loader/tests.mk b/regress/loader/tests.mk
index 1fc77ac..c3cb9de 100644
--- a/regress/loader/tests.mk
+++ b/regress/loader/tests.mk
@@ -38,7 +38,5 @@ TESTS += \
$(top_srcdir)/regress/loader/Latin1 \
$(top_srcdir)/regress/loader/Latin1-implicit \
$(top_srcdir)/regress/loader/mfile \
- $(top_srcdir)/regress/loader/TestSkipANALYZE \
- $(top_srcdir)/regress/loader/TestANALYZE \
$(top_srcdir)/regress/loader/CharNoWidth
diff --git a/regress/run_test.pl b/regress/run_test.pl
index 0ec5b2d..1c331f4 100755
--- a/regress/run_test.pl
+++ b/regress/run_test.pl
@@ -147,7 +147,6 @@ $ENV{"LANG"} = "C";
# Add locale info to the psql options
# Add pg12 precision suppression
my $PGOPTIONS = $ENV{"PGOPTIONS"};
-$PGOPTIONS .= " -c lc_messages=C";
$PGOPTIONS .= " -c client_min_messages=NOTICE";
$PGOPTIONS .= " -c extra_float_digits=0";
$ENV{"PGOPTIONS"} = $PGOPTIONS;

View File

@@ -1,208 +0,0 @@
diff --git a/raster/test/regress/tests.mk b/raster/test/regress/tests.mk
index 00918e1..7e2b6cd 100644
--- a/raster/test/regress/tests.mk
+++ b/raster/test/regress/tests.mk
@@ -17,9 +17,7 @@ override RUNTESTFLAGS_INTERNAL := \
$(RUNTESTFLAGS_INTERNAL) \
--after-upgrade-script $(top_srcdir)/raster/test/regress/hooks/hook-after-upgrade-raster.sql
-RASTER_TEST_FIRST = \
- $(top_srcdir)/raster/test/regress/check_gdal \
- $(top_srcdir)/raster/test/regress/loader/load_outdb
+RASTER_TEST_FIRST =
RASTER_TEST_LAST = \
$(top_srcdir)/raster/test/regress/clean
@@ -33,9 +31,7 @@ RASTER_TEST_IO = \
RASTER_TEST_BASIC_FUNC = \
$(top_srcdir)/raster/test/regress/rt_bytea \
- $(top_srcdir)/raster/test/regress/rt_wkb \
$(top_srcdir)/raster/test/regress/box3d \
- $(top_srcdir)/raster/test/regress/rt_addband \
$(top_srcdir)/raster/test/regress/rt_band \
$(top_srcdir)/raster/test/regress/rt_tile
@@ -73,16 +69,10 @@ RASTER_TEST_BANDPROPS = \
$(top_srcdir)/raster/test/regress/rt_neighborhood \
$(top_srcdir)/raster/test/regress/rt_nearestvalue \
$(top_srcdir)/raster/test/regress/rt_pixelofvalue \
- $(top_srcdir)/raster/test/regress/rt_polygon \
- $(top_srcdir)/raster/test/regress/rt_setbandpath
+ $(top_srcdir)/raster/test/regress/rt_polygon
RASTER_TEST_UTILITY = \
$(top_srcdir)/raster/test/regress/rt_utility \
- $(top_srcdir)/raster/test/regress/rt_fromgdalraster \
- $(top_srcdir)/raster/test/regress/rt_asgdalraster \
- $(top_srcdir)/raster/test/regress/rt_astiff \
- $(top_srcdir)/raster/test/regress/rt_asjpeg \
- $(top_srcdir)/raster/test/regress/rt_aspng \
$(top_srcdir)/raster/test/regress/rt_reclass \
$(top_srcdir)/raster/test/regress/rt_gdalwarp \
$(top_srcdir)/raster/test/regress/rt_gdalcontour \
@@ -120,21 +110,13 @@ RASTER_TEST_SREL = \
RASTER_TEST_BUGS = \
$(top_srcdir)/raster/test/regress/bug_test_car5 \
- $(top_srcdir)/raster/test/regress/permitted_gdal_drivers \
$(top_srcdir)/raster/test/regress/tickets
RASTER_TEST_LOADER = \
$(top_srcdir)/raster/test/regress/loader/Basic \
$(top_srcdir)/raster/test/regress/loader/Projected \
$(top_srcdir)/raster/test/regress/loader/BasicCopy \
- $(top_srcdir)/raster/test/regress/loader/BasicFilename \
- $(top_srcdir)/raster/test/regress/loader/BasicOutDB \
- $(top_srcdir)/raster/test/regress/loader/Tiled10x10 \
- $(top_srcdir)/raster/test/regress/loader/Tiled10x10Copy \
- $(top_srcdir)/raster/test/regress/loader/Tiled8x8 \
- $(top_srcdir)/raster/test/regress/loader/TiledAuto \
- $(top_srcdir)/raster/test/regress/loader/TiledAutoSkipNoData \
- $(top_srcdir)/raster/test/regress/loader/TiledAutoCopyn
+ $(top_srcdir)/raster/test/regress/loader/BasicFilename
RASTER_TESTS := $(RASTER_TEST_FIRST) \
$(RASTER_TEST_METADATA) $(RASTER_TEST_IO) $(RASTER_TEST_BASIC_FUNC) \
diff --git a/regress/core/binary.sql b/regress/core/binary.sql
index 7a36b65..ad78fc7 100644
--- a/regress/core/binary.sql
+++ b/regress/core/binary.sql
@@ -1,4 +1,5 @@
SET client_min_messages TO warning;
+
CREATE SCHEMA tm;
CREATE TABLE tm.geoms (id serial, g geometry);
@@ -31,24 +32,39 @@ SELECT st_force4d(g) FROM tm.geoms WHERE id < 15 ORDER BY id;
INSERT INTO tm.geoms(g)
SELECT st_setsrid(g,4326) FROM tm.geoms ORDER BY id;
-COPY tm.geoms TO :tmpfile WITH BINARY;
+-- define temp file path
+\set tmpfile '/tmp/postgis_binary_test.dat'
+
+-- export
+\set command '\\copy tm.geoms TO ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+-- import
CREATE TABLE tm.geoms_in AS SELECT * FROM tm.geoms LIMIT 0;
-COPY tm.geoms_in FROM :tmpfile WITH BINARY;
-SELECT 'geometry', count(*) FROM tm.geoms_in i, tm.geoms o WHERE i.id = o.id
- AND ST_OrderingEquals(i.g, o.g);
+\set command '\\copy tm.geoms_in FROM ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+SELECT 'geometry', count(*) FROM tm.geoms_in i, tm.geoms o
+WHERE i.id = o.id AND ST_OrderingEquals(i.g, o.g);
CREATE TABLE tm.geogs AS SELECT id,g::geography FROM tm.geoms
WHERE geometrytype(g) NOT LIKE '%CURVE%'
AND geometrytype(g) NOT LIKE '%CIRCULAR%'
AND geometrytype(g) NOT LIKE '%SURFACE%'
AND geometrytype(g) NOT LIKE 'TRIANGLE%'
- AND geometrytype(g) NOT LIKE 'TIN%'
-;
+ AND geometrytype(g) NOT LIKE 'TIN%';
-COPY tm.geogs TO :tmpfile WITH BINARY;
+-- export
+\set command '\\copy tm.geogs TO ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+-- import
CREATE TABLE tm.geogs_in AS SELECT * FROM tm.geogs LIMIT 0;
-COPY tm.geogs_in FROM :tmpfile WITH BINARY;
-SELECT 'geometry', count(*) FROM tm.geogs_in i, tm.geogs o WHERE i.id = o.id
- AND ST_OrderingEquals(i.g::geometry, o.g::geometry);
+\set command '\\copy tm.geogs_in FROM ':tmpfile' WITH (FORMAT BINARY)'
+:command
+
+SELECT 'geometry', count(*) FROM tm.geogs_in i, tm.geogs o
+WHERE i.id = o.id AND ST_OrderingEquals(i.g::geometry, o.g::geometry);
DROP SCHEMA tm CASCADE;
+
diff --git a/regress/core/tests.mk b/regress/core/tests.mk
index 90987df..74fe3f1 100644
--- a/regress/core/tests.mk
+++ b/regress/core/tests.mk
@@ -16,14 +16,13 @@ POSTGIS_PGSQL_VERSION=170
POSTGIS_GEOS_VERSION=31101
HAVE_JSON=yes
HAVE_SPGIST=yes
-INTERRUPTTESTS=yes
+INTERRUPTTESTS=no
current_dir := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
RUNTESTFLAGS_INTERNAL += \
--before-upgrade-script $(top_srcdir)/regress/hooks/hook-before-upgrade.sql \
--after-upgrade-script $(top_srcdir)/regress/hooks/hook-after-upgrade.sql \
- --after-create-script $(top_srcdir)/regress/hooks/hook-after-create.sql \
--before-uninstall-script $(top_srcdir)/regress/hooks/hook-before-uninstall.sql
TESTS += \
@@ -40,7 +39,6 @@ TESTS += \
$(top_srcdir)/regress/core/dumppoints \
$(top_srcdir)/regress/core/dumpsegments \
$(top_srcdir)/regress/core/empty \
- $(top_srcdir)/regress/core/estimatedextent \
$(top_srcdir)/regress/core/forcecurve \
$(top_srcdir)/regress/core/flatgeobuf \
$(top_srcdir)/regress/core/frechet \
@@ -60,7 +58,6 @@ TESTS += \
$(top_srcdir)/regress/core/out_marc21 \
$(top_srcdir)/regress/core/in_encodedpolyline \
$(top_srcdir)/regress/core/iscollection \
- $(top_srcdir)/regress/core/legacy \
$(top_srcdir)/regress/core/letters \
$(top_srcdir)/regress/core/lwgeom_regress \
$(top_srcdir)/regress/core/measures \
@@ -119,7 +116,6 @@ TESTS += \
$(top_srcdir)/regress/core/temporal_knn \
$(top_srcdir)/regress/core/tickets \
$(top_srcdir)/regress/core/twkb \
- $(top_srcdir)/regress/core/typmod \
$(top_srcdir)/regress/core/wkb \
$(top_srcdir)/regress/core/wkt \
$(top_srcdir)/regress/core/wmsservers \
diff --git a/regress/loader/tests.mk b/regress/loader/tests.mk
index ac4f8ad..4bad4fc 100644
--- a/regress/loader/tests.mk
+++ b/regress/loader/tests.mk
@@ -38,7 +38,5 @@ TESTS += \
$(top_srcdir)/regress/loader/Latin1 \
$(top_srcdir)/regress/loader/Latin1-implicit \
$(top_srcdir)/regress/loader/mfile \
- $(top_srcdir)/regress/loader/TestSkipANALYZE \
- $(top_srcdir)/regress/loader/TestANALYZE \
$(top_srcdir)/regress/loader/CharNoWidth \
diff --git a/regress/run_test.pl b/regress/run_test.pl
index cac4b2e..4c7c82b 100755
--- a/regress/run_test.pl
+++ b/regress/run_test.pl
@@ -238,7 +238,6 @@ $ENV{"LANG"} = "C";
# Add locale info to the psql options
# Add pg12 precision suppression
my $PGOPTIONS = $ENV{"PGOPTIONS"};
-$PGOPTIONS .= " -c lc_messages=C";
$PGOPTIONS .= " -c client_min_messages=NOTICE";
$PGOPTIONS .= " -c extra_float_digits=0";
$ENV{"PGOPTIONS"} = $PGOPTIONS;
diff --git a/topology/test/tests.mk b/topology/test/tests.mk
index cbe2633..2c7c18f 100644
--- a/topology/test/tests.mk
+++ b/topology/test/tests.mk
@@ -46,9 +46,7 @@ TESTS += \
$(top_srcdir)/topology/test/regress/legacy_query.sql \
$(top_srcdir)/topology/test/regress/legacy_validate.sql \
$(top_srcdir)/topology/test/regress/polygonize.sql \
- $(top_srcdir)/topology/test/regress/populate_topology_layer.sql \
$(top_srcdir)/topology/test/regress/removeunusedprimitives.sql \
- $(top_srcdir)/topology/test/regress/renametopogeometrycolumn.sql \
$(top_srcdir)/topology/test/regress/renametopology.sql \
$(top_srcdir)/topology/test/regress/share_sequences.sql \
$(top_srcdir)/topology/test/regress/sqlmm.sql \

File diff suppressed because one or more lines are too long

View File

@@ -1,17 +0,0 @@
#!/bin/bash
set -ex
cd "$(dirname "${0}")"
dropdb --if-exist contrib_regression
createdb contrib_regression
psql -d contrib_regression -c "ALTER DATABASE contrib_regression SET TimeZone='UTC'" \
-c "ALTER DATABASE contrib_regression SET DateStyle='ISO, MDY'" \
-c "CREATE EXTENSION postgis SCHEMA public" \
-c "CREATE EXTENSION postgis_topology" \
-c "CREATE EXTENSION postgis_tiger_geocoder CASCADE" \
-c "CREATE EXTENSION postgis_raster SCHEMA public" \
-c "CREATE EXTENSION postgis_sfcgal SCHEMA public"
patch -p1 <"postgis-common-${PG_VERSION}.patch"
patch -p1 <"postgis-regular-${PG_VERSION}.patch"
psql -d contrib_regression -f raster_outdb_template.sql
trap 'patch -R -p1 <postgis-regular-${PG_VERSION}.patch && patch -R -p1 <"postgis-common-${PG_VERSION}.patch"' EXIT
POSTGIS_REGRESS_DB=contrib_regression RUNTESTFLAGS=--nocreate make installcheck-base

View File

@@ -63,9 +63,5 @@ done
for d in ${FAILED}; do
cat "$(find $d -name regression.diffs)"
done
for postgis_diff in /tmp/pgis_reg/*_diff; do
echo "${postgis_diff}:"
cat "${postgis_diff}"
done
echo "${FAILED}"
exit 1

View File

@@ -82,8 +82,7 @@ EXTENSIONS='[
{"extname": "pg_ivm", "extdir": "pg_ivm-src"},
{"extname": "pgjwt", "extdir": "pgjwt-src"},
{"extname": "pgtap", "extdir": "pgtap-src"},
{"extname": "pg_repack", "extdir": "pg_repack-src"},
{"extname": "h3", "extdir": "h3-pg-src"}
{"extname": "pg_repack", "extdir": "pg_repack-src"}
]'
EXTNAMES=$(echo ${EXTENSIONS} | jq -r '.[].extname' | paste -sd ' ' -)
COMPUTE_TAG=${NEW_COMPUTE_TAG} docker compose --profile test-extensions up --quiet-pull --build -d

View File

@@ -8,7 +8,6 @@ anyhow.workspace = true
axum-extra.workspace = true
axum.workspace = true
camino.workspace = true
clap.workspace = true
futures.workspace = true
jsonwebtoken.workspace = true
prometheus.workspace = true

View File

@@ -462,8 +462,6 @@ MC4CAQAwBQYDK2VwBCIEID/Drmc1AA6U/znNRWpF3zEGegOATQxfkdWxitcOMsIH
if var(REAL_S3_ENV).is_ok() {
assert!(body.contains("remote_storage_s3_deleted_objects_total"));
}
#[cfg(target_os = "linux")]
assert!(body.contains("process_threads"));
}

View File

@@ -4,8 +4,6 @@
//! for large computes.
mod app;
use anyhow::Context;
use clap::Parser;
use std::net::{IpAddr, Ipv4Addr, SocketAddr};
use tracing::info;
use utils::logging;
@@ -14,29 +12,13 @@ const fn max_upload_file_limit() -> usize {
100 * 1024 * 1024
}
const fn listen() -> SocketAddr {
SocketAddr::new(IpAddr::V4(Ipv4Addr::new(0, 0, 0, 0)), 51243)
}
#[derive(Parser)]
struct Args {
#[arg(exclusive = true)]
config_file: Option<String>,
#[arg(long, default_value = "false", requires = "config")]
/// to allow testing k8s helm chart where we don't have s3 credentials
no_s3_check_on_startup: bool,
#[arg(long, value_name = "FILE")]
/// inline config mode for k8s helm chart
config: Option<String>,
}
#[derive(serde::Deserialize)]
#[serde(tag = "type")]
struct Config {
#[serde(default = "listen")]
listen: std::net::SocketAddr,
pemfile: camino::Utf8PathBuf,
#[serde(flatten)]
storage_kind: remote_storage::TypedRemoteStorageKind,
storage_config: remote_storage::RemoteStorageConfig,
#[serde(default = "max_upload_file_limit")]
max_upload_file_limit: usize,
}
@@ -49,18 +31,13 @@ async fn main() -> anyhow::Result<()> {
logging::Output::Stdout,
)?;
let args = Args::parse();
let config: Config = if let Some(config_path) = args.config_file {
info!("Reading config from {config_path}");
let config = std::fs::read_to_string(config_path)?;
serde_json::from_str(&config).context("parsing config")?
} else if let Some(config) = args.config {
info!("Reading inline config");
serde_json::from_str(&config).context("parsing config")?
} else {
anyhow::bail!("Supply either config file path or --config=inline-config");
};
let config: String = std::env::args().skip(1).take(1).collect();
if config.is_empty() {
anyhow::bail!("Usage: endpoint_storage config.json")
}
info!("Reading config from {config}");
let config = std::fs::read_to_string(config.clone())?;
let config: Config = serde_json::from_str(&config).context("parsing config")?;
info!("Reading pemfile from {}", config.pemfile.clone());
let pemfile = std::fs::read(config.pemfile.clone())?;
info!("Loading public key from {}", config.pemfile.clone());
@@ -69,12 +46,9 @@ async fn main() -> anyhow::Result<()> {
let listener = tokio::net::TcpListener::bind(config.listen).await.unwrap();
info!("listening on {}", listener.local_addr().unwrap());
let storage =
remote_storage::GenericRemoteStorage::from_storage_kind(config.storage_kind).await?;
let storage = remote_storage::GenericRemoteStorage::from_config(&config.storage_config).await?;
let cancel = tokio_util::sync::CancellationToken::new();
if !args.no_s3_check_on_startup {
app::check_storage_permissions(&storage, cancel.clone()).await?;
}
app::check_storage_permissions(&storage, cancel.clone()).await?;
let proxy = std::sync::Arc::new(endpoint_storage::Storage {
auth,

View File

@@ -16,7 +16,6 @@ pub static COMPUTE_AUDIENCE: &str = "compute";
pub enum ComputeClaimsScope {
/// An admin-scoped token allows access to all of `compute_ctl`'s authorized
/// facilities.
#[serde(rename = "compute_ctl:admin")]
Admin,
}
@@ -25,7 +24,7 @@ impl FromStr for ComputeClaimsScope {
fn from_str(s: &str) -> Result<Self, Self::Err> {
match s {
"compute_ctl:admin" => Ok(ComputeClaimsScope::Admin),
"admin" => Ok(ComputeClaimsScope::Admin),
_ => Err(anyhow::anyhow!("invalid compute claims scope \"{s}\"")),
}
}
@@ -81,23 +80,3 @@ pub struct SetRoleGrantsRequest {
pub privileges: Vec<Privilege>,
pub role: PgIdent,
}
#[cfg(test)]
mod test {
use std::str::FromStr;
use crate::requests::ComputeClaimsScope;
/// Confirm that whether we parse the scope by string or through serde, the
/// same values parse to the same enum variant.
#[test]
fn compute_request_scopes() {
const ADMIN_SCOPE: &str = "compute_ctl:admin";
let from_serde: ComputeClaimsScope =
serde_json::from_str(&format!("\"{ADMIN_SCOPE}\"")).unwrap();
let from_str = ComputeClaimsScope::from_str(ADMIN_SCOPE).unwrap();
assert_eq!(from_serde, from_str);
}
}

View File

@@ -83,16 +83,6 @@ pub struct ComputeStatusResponse {
pub error: Option<String>,
}
#[derive(Serialize, Clone, Copy, Debug, Deserialize, PartialEq, Eq, Default)]
#[serde(rename_all = "snake_case")]
pub enum TerminateMode {
#[default]
/// wait 30s till returning from /terminate to allow control plane to get the error
Fast,
/// return from /terminate immediately as soon as all components are terminated
Immediate,
}
#[derive(Serialize, Clone, Copy, Debug, Deserialize, PartialEq, Eq)]
#[serde(rename_all = "snake_case")]
pub enum ComputeStatus {
@@ -113,16 +103,11 @@ pub enum ComputeStatus {
// control-plane to terminate it.
Failed,
// Termination requested
TerminationPending { mode: TerminateMode },
TerminationPending,
// Terminated Postgres
Terminated,
}
#[derive(Deserialize, Serialize)]
pub struct TerminateResponse {
pub lsn: Option<utils::lsn::Lsn>,
}
impl Display for ComputeStatus {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
@@ -132,7 +117,7 @@ impl Display for ComputeStatus {
ComputeStatus::Running => f.write_str("running"),
ComputeStatus::Configuration => f.write_str("configuration"),
ComputeStatus::Failed => f.write_str("failed"),
ComputeStatus::TerminationPending { .. } => f.write_str("termination-pending"),
ComputeStatus::TerminationPending => f.write_str("termination-pending"),
ComputeStatus::Terminated => f.write_str("terminated"),
}
}

View File

@@ -4,7 +4,6 @@
//! provide it by calling the compute_ctl's `/compute_ctl` endpoint, or
//! compute_ctl can fetch it by calling the control plane's API.
use std::collections::HashMap;
use std::fmt::Display;
use indexmap::IndexMap;
use regex::Regex;
@@ -179,9 +178,9 @@ pub struct ComputeSpec {
/// JWT for authorizing requests to endpoint storage service
pub endpoint_storage_token: Option<String>,
/// Download LFC state from endpoint_storage and pass it to Postgres on startup
/// If true, download LFC state from endpoint_storage and pass it to Postgres on startup
#[serde(default)]
pub autoprewarm: bool,
pub prewarm_lfc_on_startup: bool,
}
/// Feature flag to signal `compute_ctl` to enable certain experimental functionality.
@@ -193,9 +192,6 @@ pub enum ComputeFeature {
/// track short-lived connections as user activity.
ActivityMonitorExperimental,
/// Enable TLS functionality.
TlsExperimental,
/// This is a special feature flag that is used to represent unknown feature flags.
/// Basically all unknown to enum flags are represented as this one. See unit test
/// `parse_unknown_features()` for more details.
@@ -254,44 +250,34 @@ impl RemoteExtSpec {
}
match self.extension_data.get(real_ext_name) {
Some(_ext_data) => Ok((
real_ext_name.to_string(),
Self::build_remote_path(build_tag, pg_major_version, real_ext_name)?,
)),
Some(_ext_data) => {
// We have decided to use the Go naming convention due to Kubernetes.
let arch = match std::env::consts::ARCH {
"x86_64" => "amd64",
"aarch64" => "arm64",
arch => arch,
};
// Construct the path to the extension archive
// BUILD_TAG/PG_MAJOR_VERSION/extensions/EXTENSION_NAME.tar.zst
//
// Keep it in sync with path generation in
// https://github.com/neondatabase/build-custom-extensions/tree/main
let archive_path_str = format!(
"{build_tag}/{arch}/{pg_major_version}/extensions/{real_ext_name}.tar.zst"
);
Ok((
real_ext_name.to_string(),
RemotePath::from_string(&archive_path_str)?,
))
}
None => Err(anyhow::anyhow!(
"real_ext_name {} is not found",
real_ext_name
)),
}
}
/// Get the architecture-specific portion of the remote extension path. We
/// use the Go naming convention due to Kubernetes.
fn get_arch() -> &'static str {
match std::env::consts::ARCH {
"x86_64" => "amd64",
"aarch64" => "arm64",
arch => arch,
}
}
/// Build a [`RemotePath`] for an extension.
fn build_remote_path(
build_tag: &str,
pg_major_version: &str,
ext_name: &str,
) -> anyhow::Result<RemotePath> {
let arch = Self::get_arch();
// Construct the path to the extension archive
// BUILD_TAG/PG_MAJOR_VERSION/extensions/EXTENSION_NAME.tar.zst
//
// Keep it in sync with path generation in
// https://github.com/neondatabase/build-custom-extensions/tree/main
RemotePath::from_string(&format!(
"{build_tag}/{arch}/{pg_major_version}/extensions/{ext_name}.tar.zst"
))
}
}
#[derive(Clone, Copy, Debug, Default, Eq, PartialEq, Deserialize, Serialize)]
@@ -320,12 +306,6 @@ impl ComputeMode {
}
}
impl Display for ComputeMode {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
f.write_str(self.to_type_str())
}
}
/// Log level for audit logging
#[derive(Clone, Debug, Default, Eq, PartialEq, Deserialize, Serialize)]
pub enum ComputeAudit {
@@ -538,37 +518,6 @@ mod tests {
.expect("Library should be found");
}
#[test]
fn remote_extension_path() {
let rspec: RemoteExtSpec = serde_json::from_value(serde_json::json!({
"public_extensions": ["ext"],
"custom_extensions": [],
"library_index": {
"extlib": "ext",
},
"extension_data": {
"ext": {
"control_data": {
"ext.control": ""
},
"archive_path": ""
}
},
}))
.unwrap();
let (_ext_name, ext_path) = rspec
.get_ext("ext", false, "latest", "v17")
.expect("Extension should be found");
// Starting with a forward slash would have consequences for the
// Url::join() that occurs when downloading a remote extension.
assert!(!ext_path.to_string().starts_with("/"));
assert_eq!(
ext_path,
RemoteExtSpec::build_remote_path("latest", "v17", "ext").unwrap()
);
}
#[test]
fn parse_spec_file() {
let file = File::open("tests/cluster_spec.json").unwrap();

View File

@@ -85,7 +85,7 @@
"vartype": "bool"
},
{
"name": "autoprewarm",
"name": "prewarm_lfc_on_startup",
"value": "off",
"vartype": "bool"
},

View File

@@ -419,13 +419,13 @@ pub fn now() -> u64 {
with_thread_context(|ctx| ctx.clock.get().unwrap().now())
}
pub fn exit(code: i32, msg: String) -> ! {
pub fn exit(code: i32, msg: String) {
with_thread_context(|ctx| {
ctx.allow_panic.store(true, Ordering::SeqCst);
let mut result = ctx.result.lock();
*result = (code, msg);
panic!("exit");
})
});
}
pub(crate) fn get_thread_ctx() -> Arc<ThreadContext> {

View File

@@ -107,7 +107,7 @@ impl<const N: usize> MetricType for HyperLogLogState<N> {
}
impl<const N: usize> HyperLogLogState<N> {
pub fn measure(&self, item: &(impl Hash + ?Sized)) {
pub fn measure(&self, item: &impl Hash) {
// changing the hasher will break compatibility with previous measurements.
self.record(BuildHasherDefault::<xxh3::Hash64>::default().hash_one(item));
}

View File

@@ -27,7 +27,6 @@ pub use prometheus::{
pub mod launch_timestamp;
mod wrappers;
pub use prometheus;
pub use wrappers::{CountedReader, CountedWriter};
mod hll;
pub use hll::{HyperLogLog, HyperLogLogState, HyperLogLogVec};

View File

@@ -6,27 +6,8 @@ license.workspace = true
[dependencies]
thiserror.workspace = true
nix.workspace = true
nix.workspace=true
workspace_hack = { version = "0.1", path = "../../workspace_hack" }
rustc-hash = { version = "2.1.1" }
rand = "0.9.1"
libc.workspace = true
lock_api = "0.4.13"
[dev-dependencies]
criterion = { workspace = true, features = ["html_reports"] }
rand_distr = "0.5.1"
xxhash-rust = { version = "0.8.15", features = ["xxh3"] }
ahash.workspace = true
twox-hash = { version = "2.1.1" }
seahash = "4.1.0"
hashbrown = { git = "https://github.com/quantumish/hashbrown.git", rev = "6610e6d" }
foldhash = "0.1.5"
[target.'cfg(target_os = "macos")'.dependencies]
tempfile = "3.14.0"
[[bench]]
name = "hmap_resize"
harness = false

View File

@@ -1,282 +0,0 @@
use criterion::{criterion_group, criterion_main, BatchSize, Criterion, BenchmarkId};
use neon_shmem::hash::HashMapAccess;
use neon_shmem::hash::HashMapInit;
use neon_shmem::hash::entry::Entry;
use rand::prelude::*;
use rand::distr::{Distribution, StandardUniform};
use std::hash::BuildHasher;
use std::default::Default;
// Taken from bindings to C code
#[derive(Clone, Debug, Hash, Eq, PartialEq)]
#[repr(C)]
pub struct FileCacheKey {
pub _spc_id: u32,
pub _db_id: u32,
pub _rel_number: u32,
pub _fork_num: u32,
pub _block_num: u32,
}
impl Distribution<FileCacheKey> for StandardUniform {
// questionable, but doesn't need to be good randomness
fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> FileCacheKey {
FileCacheKey {
_spc_id: rng.random(),
_db_id: rng.random(),
_rel_number: rng.random(),
_fork_num: rng.random(),
_block_num: rng.random()
}
}
}
#[derive(Clone, Debug)]
#[repr(C)]
pub struct FileCacheEntry {
pub _offset: u32,
pub _access_count: u32,
pub _prev: *mut FileCacheEntry,
pub _next: *mut FileCacheEntry,
pub _state: [u32; 8],
}
impl FileCacheEntry {
fn dummy() -> Self {
Self {
_offset: 0,
_access_count: 0,
_prev: std::ptr::null_mut(),
_next: std::ptr::null_mut(),
_state: [0; 8]
}
}
}
// Utilities for applying operations.
#[derive(Clone, Debug)]
struct TestOp<K,V>(K, Option<V>);
fn apply_op<K: Clone + std::hash::Hash + Eq, V, S: std::hash::BuildHasher>(
op: TestOp<K,V>,
map: &mut HashMapAccess<K,V,S>,
) {
let entry = map.entry(op.0);
match op.1 {
Some(new) => {
match entry {
Entry::Occupied(mut e) => Some(e.insert(new)),
Entry::Vacant(e) => { _ = e.insert(new).unwrap(); None },
}
},
None => {
match entry {
Entry::Occupied(e) => Some(e.remove()),
Entry::Vacant(_) => None,
}
},
};
}
// Hash utilities
struct SeaRandomState {
k1: u64,
k2: u64,
k3: u64,
k4: u64
}
impl std::hash::BuildHasher for SeaRandomState {
type Hasher = seahash::SeaHasher;
fn build_hasher(&self) -> Self::Hasher {
seahash::SeaHasher::with_seeds(self.k1, self.k2, self.k3, self.k4)
}
}
impl SeaRandomState {
fn new() -> Self {
let mut rng = rand::rng();
Self { k1: rng.random(), k2: rng.random(), k3: rng.random(), k4: rng.random() }
}
}
fn small_benchs(c: &mut Criterion) {
let mut group = c.benchmark_group("Small maps");
group.sample_size(10);
group.bench_function("small_rehash", |b| {
let ideal_filled = 4_000_000;
let size = 5_000_000;
let mut writer = HashMapInit::new_resizeable(size, size * 2).attach_writer();
let mut rng = rand::rng();
while writer.get_num_buckets_in_use() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
apply_op(TestOp(key, Some(val)), &mut writer);
}
b.iter(|| writer.shuffle());
});
group.bench_function("small_rehash_xxhash", |b| {
let ideal_filled = 4_000_000;
let size = 5_000_000;
let mut writer = HashMapInit::new_resizeable(size, size * 2)
.with_hasher(twox_hash::xxhash64::RandomState::default())
.attach_writer();
let mut rng = rand::rng();
while writer.get_num_buckets_in_use() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
apply_op(TestOp(key, Some(val)), &mut writer);
}
b.iter(|| writer.shuffle());
});
group.bench_function("small_rehash_ahash", |b| {
let ideal_filled = 4_000_000;
let size = 5_000_000;
let mut writer = HashMapInit::new_resizeable(size, size * 2)
.with_hasher(ahash::RandomState::default())
.attach_writer();
let mut rng = rand::rng();
while writer.get_num_buckets_in_use() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
apply_op(TestOp(key, Some(val)), &mut writer);
}
b.iter(|| writer.shuffle());
});
group.bench_function("small_rehash_seahash", |b| {
let ideal_filled = 4_000_000;
let size = 5_000_000;
let mut writer = HashMapInit::new_resizeable(size, size * 2)
.with_hasher(SeaRandomState::new())
.attach_writer();
let mut rng = rand::rng();
while writer.get_num_buckets_in_use() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
apply_op(TestOp(key, Some(val)), &mut writer);
}
b.iter(|| writer.shuffle());
});
group.finish();
}
fn real_benchs(c: &mut Criterion) {
let mut group = c.benchmark_group("Realistic workloads");
group.sample_size(10);
group.bench_function("real_bulk_insert", |b| {
let size = 125_000_000;
let ideal_filled = 100_000_000;
let mut rng = rand::rng();
b.iter_batched(
|| HashMapInit::new_resizeable(size, size * 2).attach_writer(),
|writer| {
for _ in 0..ideal_filled {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
let entry = writer.entry(key);
std::hint::black_box(match entry {
Entry::Occupied(mut e) => { e.insert(val); },
Entry::Vacant(e) => { _ = e.insert(val).unwrap(); },
})
}
},
BatchSize::SmallInput,
)
});
group.bench_function("real_rehash", |b| {
let size = 125_000_000;
let ideal_filled = 100_000_000;
let mut writer = HashMapInit::new_resizeable(size, size).attach_writer();
let mut rng = rand::rng();
while writer.get_num_buckets_in_use() < ideal_filled {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
apply_op(TestOp(key, Some(val)), &mut writer);
}
b.iter(|| writer.shuffle());
});
group.bench_function("real_rehash_hashbrown", |b| {
let size = 125_000_000;
let ideal_filled = 100_000_000;
let mut writer = hashbrown::raw::RawTable::new();
let mut rng = rand::rng();
let hasher = rustc_hash::FxBuildHasher::default();
unsafe {
writer.resize(size, |(k,_)| hasher.hash_one(&k),
hashbrown::raw::Fallibility::Infallible).unwrap();
}
while writer.len() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
writer.insert(hasher.hash_one(&key), (key, val), |(k,_)| hasher.hash_one(&k));
}
b.iter(|| unsafe { writer.table.rehash_in_place(
&|table, index| hasher.hash_one(&table.bucket::<(FileCacheKey, FileCacheEntry)>(index).as_ref().0),
std::mem::size_of::<(FileCacheKey, FileCacheEntry)>(),
if std::mem::needs_drop::<(FileCacheKey, FileCacheEntry)>() {
Some(|ptr| std::ptr::drop_in_place(ptr as *mut (FileCacheKey, FileCacheEntry)))
} else {
None
},
) });
});
for elems in [2, 4, 8, 16, 32, 64, 96, 112] {
group.bench_with_input(BenchmarkId::new("real_rehash_varied", elems), &elems, |b, &size| {
let ideal_filled = size * 1_000_000;
let size = 125_000_000;
let mut writer = HashMapInit::new_resizeable(size, size).attach_writer();
let mut rng = rand::rng();
while writer.get_num_buckets_in_use() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
apply_op(TestOp(key, Some(val)), &mut writer);
}
b.iter(|| writer.shuffle());
});
group.bench_with_input(BenchmarkId::new("real_rehash_varied_hashbrown", elems), &elems, |b, &size| {
let ideal_filled = size * 1_000_000;
let size = 125_000_000;
let mut writer = hashbrown::raw::RawTable::new();
let mut rng = rand::rng();
let hasher = rustc_hash::FxBuildHasher::default();
unsafe {
writer.resize(size, |(k,_)| hasher.hash_one(&k),
hashbrown::raw::Fallibility::Infallible).unwrap();
}
while writer.len() < ideal_filled as usize {
let key: FileCacheKey = rng.random();
let val = FileCacheEntry::dummy();
writer.insert(hasher.hash_one(&key), (key, val), |(k,_)| hasher.hash_one(&k));
}
b.iter(|| unsafe { writer.table.rehash_in_place(
&|table, index| hasher.hash_one(&table.bucket::<(FileCacheKey, FileCacheEntry)>(index).as_ref().0),
std::mem::size_of::<(FileCacheKey, FileCacheEntry)>(),
if std::mem::needs_drop::<(FileCacheKey, FileCacheEntry)>() {
Some(|ptr| std::ptr::drop_in_place(ptr as *mut (FileCacheKey, FileCacheEntry)))
} else {
None
},
) });
});
}
group.finish();
}
criterion_group!(benches, small_benchs, real_benchs);
criterion_main!(benches);

View File

@@ -1,533 +0,0 @@
//! Resizable hash table implementation on top of byte-level storage (either a [`ShmemHandle`] or a fixed byte array).
//!
//! This hash table has two major components: the bucket array and the dictionary. Each bucket within the
//! bucket array contains a `Option<(K, V)>` and an index of another bucket. In this way there is both an
//! implicit freelist within the bucket array (`None` buckets point to other `None` entries) and various hash
//! chains within the bucket array (a Some bucket will point to other Some buckets that had the same hash).
//!
//! Buckets are never moved unless they are within a region that is being shrunk, and so the actual hash-
//! dependent component is done with the dictionary. When a new key is inserted into the map, a position
//! within the dictionary is decided based on its hash, the data is inserted into an empty bucket based
//! off of the freelist, and then the index of said bucket is placed in the dictionary.
//!
//! This map is resizable (if initialized on top of a [`ShmemHandle`]). Both growing and shrinking happen
//! in-place and are at a high level achieved by expanding/reducing the bucket array and rebuilding the
//! dictionary by rehashing all keys.
use std::hash::{Hash, BuildHasher};
use std::mem::MaybeUninit;
use crate::{shmem, sync::*};
use crate::shmem::ShmemHandle;
mod core;
pub mod entry;
#[cfg(test)]
mod tests;
use core::{Bucket, CoreHashMap, INVALID_POS};
use entry::{Entry, OccupiedEntry, VacantEntry, PrevPos};
/// Builder for a [`HashMapAccess`].
#[must_use]
pub struct HashMapInit<'a, K, V, S = rustc_hash::FxBuildHasher> {
shmem_handle: Option<ShmemHandle>,
shared_ptr: *mut RwLock<HashMapShared<'a, K, V>>,
shared_size: usize,
hasher: S,
num_buckets: u32,
}
/// Accessor for a hash table.
pub struct HashMapAccess<'a, K, V, S = rustc_hash::FxBuildHasher> {
shmem_handle: Option<ShmemHandle>,
shared_ptr: *mut HashMapShared<'a, K, V>,
hasher: S,
}
unsafe impl<K: Sync, V: Sync, S> Sync for HashMapAccess<'_, K, V, S> {}
unsafe impl<K: Send, V: Send, S> Send for HashMapAccess<'_, K, V, S> {}
impl<'a, K: Clone + Hash + Eq, V, S> HashMapInit<'a, K, V, S> {
pub fn with_hasher<T: BuildHasher>(self, hasher: T) -> HashMapInit<'a, K, V, T> {
HashMapInit {
hasher,
shmem_handle: self.shmem_handle,
shared_ptr: self.shared_ptr,
shared_size: self.shared_size,
num_buckets: self.num_buckets,
}
}
/// Loosely (over)estimate the size needed to store a hash table with `num_buckets` buckets.
pub fn estimate_size(num_buckets: u32) -> usize {
// add some margin to cover alignment etc.
CoreHashMap::<K, V>::estimate_size(num_buckets) + size_of::<HashMapShared<K, V>>() + 1000
}
/// Initialize a table for writing.
pub fn attach_writer(self) -> HashMapAccess<'a, K, V, S> {
let mut ptr: *mut u8 = self.shared_ptr.cast();
let end_ptr: *mut u8 = unsafe { ptr.add(self.shared_size) };
// carve out area for the One Big Lock (TM) and the HashMapShared.
ptr = unsafe { ptr.add(ptr.align_offset(align_of::<libc::pthread_rwlock_t>())) };
let raw_lock_ptr = ptr;
ptr = unsafe { ptr.add(size_of::<libc::pthread_rwlock_t>()) };
ptr = unsafe { ptr.add(ptr.align_offset(align_of::<HashMapShared<K, V>>())) };
let shared_ptr: *mut HashMapShared<K, V> = ptr.cast();
ptr = unsafe { ptr.add(size_of::<HashMapShared<K, V>>()) };
// carve out the buckets
ptr = unsafe { ptr.byte_add(ptr.align_offset(align_of::<core::Bucket<K, V>>())) };
let buckets_ptr = ptr;
ptr = unsafe { ptr.add(size_of::<core::Bucket<K, V>>() * self.num_buckets as usize) };
// use remaining space for the dictionary
ptr = unsafe { ptr.byte_add(ptr.align_offset(align_of::<u32>())) };
assert!(ptr.addr() < end_ptr.addr());
let dictionary_ptr = ptr;
let dictionary_size = unsafe { end_ptr.byte_offset_from(ptr) / size_of::<u32>() as isize };
assert!(dictionary_size > 0);
let buckets =
unsafe { std::slice::from_raw_parts_mut(buckets_ptr.cast(), self.num_buckets as usize) };
let dictionary = unsafe {
std::slice::from_raw_parts_mut(dictionary_ptr.cast(), dictionary_size as usize)
};
let hashmap = CoreHashMap::new(buckets, dictionary);
let lock = RwLock::from_raw(PthreadRwLock::new(raw_lock_ptr.cast()), hashmap);
unsafe {
std::ptr::write(shared_ptr, lock);
}
HashMapAccess {
shmem_handle: self.shmem_handle,
shared_ptr,
hasher: self.hasher,
}
}
/// Initialize a table for reading. Currently identical to [`HashMapInit::attach_writer`].
pub fn attach_reader(self) -> HashMapAccess<'a, K, V, S> {
self.attach_writer()
}
}
/// Hash table data that is actually stored in the shared memory area.
///
/// NOTE: We carve out the parts from a contiguous chunk. Growing and shrinking the hash table
/// relies on the memory layout! The data structures are laid out in the contiguous shared memory
/// area as follows:
///
/// [`libc::pthread_rwlock_t`]
/// [`HashMapShared`]
/// [buckets]
/// [dictionary]
///
/// In between the above parts, there can be padding bytes to align the parts correctly.
type HashMapShared<'a, K, V> = RwLock<CoreHashMap<'a, K, V>>;
impl<'a, K, V> HashMapInit<'a, K, V, rustc_hash::FxBuildHasher>
where
K: Clone + Hash + Eq
{
/// Place the hash table within a user-supplied fixed memory area.
pub fn with_fixed(
num_buckets: u32,
area: &'a mut [MaybeUninit<u8>],
) -> Self {
Self {
num_buckets,
shmem_handle: None,
shared_ptr: area.as_mut_ptr().cast(),
shared_size: area.len(),
hasher: rustc_hash::FxBuildHasher,
}
}
/// Place a new hash map in the given shared memory area
///
/// # Panics
/// Will panic on failure to resize area to expected map size.
pub fn with_shmem(num_buckets: u32, shmem: ShmemHandle) -> Self {
let size = Self::estimate_size(num_buckets);
shmem
.set_size(size)
.expect("could not resize shared memory area");
Self {
num_buckets,
shared_ptr: shmem.data_ptr.as_ptr().cast(),
shmem_handle: Some(shmem),
shared_size: size,
hasher: rustc_hash::FxBuildHasher
}
}
/// Make a resizable hash map within a new shared memory area with the given name.
pub fn new_resizeable_named(num_buckets: u32, max_buckets: u32, name: &str) -> Self {
let size = Self::estimate_size(num_buckets);
let max_size = Self::estimate_size(max_buckets);
let shmem = ShmemHandle::new(name, size, max_size)
.expect("failed to make shared memory area");
Self {
num_buckets,
shared_ptr: shmem.data_ptr.as_ptr().cast(),
shmem_handle: Some(shmem),
shared_size: size,
hasher: rustc_hash::FxBuildHasher
}
}
/// Make a resizable hash map within a new anonymous shared memory area.
pub fn new_resizeable(num_buckets: u32, max_buckets: u32) -> Self {
use std::sync::atomic::{AtomicUsize, Ordering};
static COUNTER: AtomicUsize = AtomicUsize::new(0);
let val = COUNTER.fetch_add(1, Ordering::Relaxed);
let name = format!("neon_shmem_hmap{val}");
Self::new_resizeable_named(num_buckets, max_buckets, &name)
}
}
impl<'a, K, V, S: BuildHasher> HashMapAccess<'a, K, V, S>
where
K: Clone + Hash + Eq,
{
/// Hash a key using the map's hasher.
#[inline]
fn get_hash_value(&self, key: &K) -> u64 {
self.hasher.hash_one(key)
}
fn entry_with_hash(&self, key: K, hash: u64) -> Entry<'a, '_, K, V> {
let mut map = unsafe { self.shared_ptr.as_ref() }.unwrap().write();
let dict_pos = hash as usize % map.dictionary.len();
let first = map.dictionary[dict_pos];
if first == INVALID_POS {
// no existing entry
return Entry::Vacant(VacantEntry {
map,
key,
dict_pos: dict_pos as u32,
});
}
let mut prev_pos = PrevPos::First(dict_pos as u32);
let mut next = first;
loop {
let bucket = &mut map.buckets[next as usize];
let (bucket_key, _bucket_value) = bucket.inner.as_mut().expect("entry is in use");
if *bucket_key == key {
// found existing entry
return Entry::Occupied(OccupiedEntry {
map,
_key: key,
prev_pos,
bucket_pos: next,
});
}
if bucket.next == INVALID_POS {
// No existing entry
return Entry::Vacant(VacantEntry {
map,
key,
dict_pos: dict_pos as u32,
});
}
prev_pos = PrevPos::Chained(next);
next = bucket.next;
}
}
/// Get a reference to the corresponding value for a key.
pub fn get<'e>(&'e self, key: &K) -> Option<ValueReadGuard<'e, V>> {
let hash = self.get_hash_value(key);
let map = unsafe { self.shared_ptr.as_ref() }.unwrap().read();
RwLockReadGuard::try_map(map, |m| m.get_with_hash(key, hash)).ok()
}
/// Get a reference to the entry containing a key.
pub fn entry(&self, key: K) -> Entry<'a, '_, K, V> {
let hash = self.get_hash_value(&key);
self.entry_with_hash(key, hash)
}
/// Remove a key given its hash. Returns the associated value if it existed.
pub fn remove(&self, key: &K) -> Option<V> {
let hash = self.get_hash_value(&key);
match self.entry_with_hash(key.clone(), hash) {
Entry::Occupied(e) => Some(e.remove()),
Entry::Vacant(_) => None
}
}
/// Insert/update a key. Returns the previous associated value if it existed.
///
/// # Errors
/// Will return [`core::FullError`] if there is no more space left in the map.
pub fn insert(&self, key: K, value: V) -> Result<Option<V>, core::FullError> {
let hash = self.get_hash_value(&key);
match self.entry_with_hash(key.clone(), hash) {
Entry::Occupied(mut e) => Ok(Some(e.insert(value))),
Entry::Vacant(e) => {
_ = e.insert(value)?;
Ok(None)
}
}
}
/// Optionally return the entry for a bucket at a given index if it exists.
///
/// Has more overhead than one would intuitively expect: performs both a clone of the key
/// due to the [`OccupiedEntry`] type owning the key and also a hash of the key in order
/// to enable repairing the hash chain if the entry is removed.
pub fn entry_at_bucket(&self, pos: usize) -> Option<OccupiedEntry<'a, '_, K, V>> {
let map = unsafe { self.shared_ptr.as_mut() }.unwrap().write();
if pos >= map.buckets.len() {
return None;
}
let entry = map.buckets[pos].inner.as_ref();
match entry {
Some((key, _)) => Some(OccupiedEntry {
_key: key.clone(),
bucket_pos: pos as u32,
prev_pos: entry::PrevPos::Unknown(
self.get_hash_value(&key)
),
map,
}),
_ => None,
}
}
/// Returns the number of buckets in the table.
pub fn get_num_buckets(&self) -> usize {
let map = unsafe { self.shared_ptr.as_ref() }.unwrap().read();
map.get_num_buckets()
}
/// Return the key and value stored in bucket with given index. This can be used to
/// iterate through the hash map.
// TODO: An Iterator might be nicer. The communicator's clock algorithm needs to
// _slowly_ iterate through all buckets with its clock hand, without holding a lock.
// If we switch to an Iterator, it must not hold the lock.
pub fn get_at_bucket(&self, pos: usize) -> Option<ValueReadGuard<(K, V)>> {
let map = unsafe { self.shared_ptr.as_ref() }.unwrap().read();
if pos >= map.buckets.len() {
return None;
}
RwLockReadGuard::try_map(map, |m| m.buckets[pos].inner.as_ref()).ok()
}
/// Returns the index of the bucket a given value corresponds to.
pub fn get_bucket_for_value(&self, val_ptr: *const V) -> usize {
let map = unsafe { self.shared_ptr.as_ref() }.unwrap().read();
let origin = map.buckets.as_ptr();
let idx = (val_ptr as usize - origin as usize) / size_of::<Bucket<K, V>>();
assert!(idx < map.buckets.len());
idx
}
/// Returns the number of occupied buckets in the table.
pub fn get_num_buckets_in_use(&self) -> usize {
let map = unsafe { self.shared_ptr.as_ref() }.unwrap().read();
map.buckets_in_use as usize
}
/// Clears all entries in a table. Does not reset any shrinking operations.
pub fn clear(&self) {
let mut map = unsafe { self.shared_ptr.as_mut() }.unwrap().write();
map.clear();
}
/// Perform an in-place rehash of some region (0..`rehash_buckets`) of the table and reset
/// the `buckets` and `dictionary` slices to be as long as `num_buckets`. Resets the freelist
/// in the process.
fn rehash_dict(
&self,
inner: &mut CoreHashMap<'a, K, V>,
buckets_ptr: *mut core::Bucket<K, V>,
end_ptr: *mut u8,
num_buckets: u32,
rehash_buckets: u32,
) {
inner.free_head = INVALID_POS;
let buckets;
let dictionary;
unsafe {
let buckets_end_ptr = buckets_ptr.add(num_buckets as usize);
let dictionary_ptr: *mut u32 = buckets_end_ptr
.byte_add(buckets_end_ptr.align_offset(align_of::<u32>()))
.cast();
let dictionary_size: usize =
end_ptr.byte_offset_from(buckets_end_ptr) as usize / size_of::<u32>();
buckets = std::slice::from_raw_parts_mut(buckets_ptr, num_buckets as usize);
dictionary = std::slice::from_raw_parts_mut(dictionary_ptr, dictionary_size);
}
for e in dictionary.iter_mut() {
*e = INVALID_POS;
}
for (i, bucket) in buckets.iter_mut().enumerate().take(rehash_buckets as usize) {
if bucket.inner.is_none() {
bucket.next = inner.free_head;
inner.free_head = i as u32;
continue;
}
let hash = self.hasher.hash_one(&bucket.inner.as_ref().unwrap().0);
let pos: usize = (hash % dictionary.len() as u64) as usize;
bucket.next = dictionary[pos];
dictionary[pos] = i as u32;
}
inner.dictionary = dictionary;
inner.buckets = buckets;
}
/// Rehash the map without growing or shrinking.
pub fn shuffle(&self) {
let mut map = unsafe { self.shared_ptr.as_mut() }.unwrap().write();
let num_buckets = map.get_num_buckets() as u32;
let size_bytes = HashMapInit::<K, V, S>::estimate_size(num_buckets);
let end_ptr: *mut u8 = unsafe { self.shared_ptr.byte_add(size_bytes).cast() };
let buckets_ptr = map.buckets.as_mut_ptr();
self.rehash_dict(&mut map, buckets_ptr, end_ptr, num_buckets, num_buckets);
}
/// Grow the number of buckets within the table.
///
/// 1. Grows the underlying shared memory area
/// 2. Initializes new buckets and overwrites the current dictionary
/// 3. Rehashes the dictionary
///
/// # Panics
/// Panics if called on a map initialized with [`HashMapInit::with_fixed`].
///
/// # Errors
/// Returns an [`shmem::Error`] if any errors occur resizing the memory region.
pub fn grow(&self, num_buckets: u32) -> Result<(), shmem::Error> {
let mut map = unsafe { self.shared_ptr.as_mut() }.unwrap().write();
let old_num_buckets = map.buckets.len() as u32;
assert!(num_buckets >= old_num_buckets, "grow called with a smaller number of buckets");
if num_buckets == old_num_buckets {
return Ok(());
}
let shmem_handle = self
.shmem_handle
.as_ref()
.expect("grow called on a fixed-size hash table");
let size_bytes = HashMapInit::<K, V, S>::estimate_size(num_buckets);
shmem_handle.set_size(size_bytes)?;
let end_ptr: *mut u8 = unsafe { shmem_handle.data_ptr.as_ptr().add(size_bytes) };
// Initialize new buckets. The new buckets are linked to the free list.
// NB: This overwrites the dictionary!
let buckets_ptr = map.buckets.as_mut_ptr();
unsafe {
for i in old_num_buckets..num_buckets {
let bucket = buckets_ptr.add(i as usize);
bucket.write(core::Bucket {
next: if i < num_buckets-1 {
i + 1
} else {
map.free_head
},
inner: None,
});
}
}
self.rehash_dict(&mut map, buckets_ptr, end_ptr, num_buckets, old_num_buckets);
map.free_head = old_num_buckets;
Ok(())
}
/// Begin a shrink, limiting all new allocations to be in buckets with index below `num_buckets`.
///
/// # Panics
/// Panics if called on a map initialized with [`HashMapInit::with_fixed`] or if `num_buckets` is
/// greater than the number of buckets in the map.
pub fn begin_shrink(&mut self, num_buckets: u32) {
let mut map = unsafe { self.shared_ptr.as_mut() }.unwrap().write();
assert!(
num_buckets <= map.get_num_buckets() as u32,
"shrink called with a larger number of buckets"
);
_ = self
.shmem_handle
.as_ref()
.expect("shrink called on a fixed-size hash table");
map.alloc_limit = num_buckets;
}
/// If a shrink operation is underway, returns the target size of the map. Otherwise, returns None.
pub fn shrink_goal(&self) -> Option<usize> {
let map = unsafe { self.shared_ptr.as_mut() }.unwrap().read();
let goal = map.alloc_limit;
if goal == INVALID_POS { None } else { Some(goal as usize) }
}
/// Complete a shrink after caller has evicted entries, removing the unused buckets and rehashing.
///
/// # Panics
/// The following cases result in a panic:
/// - Calling this function on a map initialized with [`HashMapInit::with_fixed`].
/// - Calling this function on a map when no shrink operation is in progress.
/// - Calling this function on a map with `shrink_mode` set to [`HashMapShrinkMode::Remap`] and
/// there are more buckets in use than the value returned by [`HashMapAccess::shrink_goal`].
///
/// # Errors
/// Returns an [`shmem::Error`] if any errors occur resizing the memory region.
pub fn finish_shrink(&self) -> Result<(), shmem::Error> {
let mut map = unsafe { self.shared_ptr.as_mut() }.unwrap().write();
assert!(
map.alloc_limit != INVALID_POS,
"called finish_shrink when no shrink is in progress"
);
let num_buckets = map.alloc_limit;
if map.get_num_buckets() == num_buckets as usize {
return Ok(());
}
assert!(
map.buckets_in_use <= num_buckets,
"called finish_shrink before enough entries were removed"
);
for i in (num_buckets as usize)..map.buckets.len() {
if let Some((k, v)) = map.buckets[i].inner.take() {
// alloc_bucket increases count, so need to decrease since we're just moving
map.buckets_in_use -= 1;
map.alloc_bucket(k, v).unwrap();
}
}
let shmem_handle = self
.shmem_handle
.as_ref()
.expect("shrink called on a fixed-size hash table");
let size_bytes = HashMapInit::<K, V, S>::estimate_size(num_buckets);
shmem_handle.set_size(size_bytes)?;
let end_ptr: *mut u8 = unsafe { shmem_handle.data_ptr.as_ptr().add(size_bytes) };
let buckets_ptr = map.buckets.as_mut_ptr();
self.rehash_dict(&mut map, buckets_ptr, end_ptr, num_buckets, num_buckets);
map.alloc_limit = INVALID_POS;
Ok(())
}
}

View File

@@ -1,177 +0,0 @@
//! Simple hash table with chaining.
use std::hash::Hash;
use std::mem::MaybeUninit;
use crate::hash::entry::*;
/// Invalid position within the map (either within the dictionary or bucket array).
pub(crate) const INVALID_POS: u32 = u32::MAX;
/// Fundamental storage unit within the hash table. Either empty or contains a key-value pair.
/// Always part of a chain of some kind (either a freelist if empty or a hash chain if full).
pub(crate) struct Bucket<K, V> {
/// Index of next bucket in the chain.
pub(crate) next: u32,
/// Key-value pair contained within bucket.
pub(crate) inner: Option<(K, V)>,
}
/// Core hash table implementation.
pub(crate) struct CoreHashMap<'a, K, V> {
/// Dictionary used to map hashes to bucket indices.
pub(crate) dictionary: &'a mut [u32],
/// Buckets containing key-value pairs.
pub(crate) buckets: &'a mut [Bucket<K, V>],
/// Head of the freelist.
pub(crate) free_head: u32,
/// Maximum index of a bucket allowed to be allocated. [`INVALID_POS`] if no limit.
pub(crate) alloc_limit: u32,
/// The number of currently occupied buckets.
pub(crate) buckets_in_use: u32,
// pub(crate) lock: libc::pthread_mutex_t,
// Unclear what the purpose of this is.
pub(crate) _user_list_head: u32,
}
/// Error for when there are no empty buckets left but one is needed.
#[derive(Debug, PartialEq)]
pub struct FullError();
impl<'a, K: Clone + Hash + Eq, V> CoreHashMap<'a, K, V> {
const FILL_FACTOR: f32 = 0.60;
/// Estimate the size of data contained within the the hash map.
pub fn estimate_size(num_buckets: u32) -> usize {
let mut size = 0;
// buckets
size += size_of::<Bucket<K, V>>() * num_buckets as usize;
// dictionary
size += (f32::ceil((size_of::<u32>() * num_buckets as usize) as f32 / Self::FILL_FACTOR))
as usize;
size
}
pub fn new(
buckets: &'a mut [MaybeUninit<Bucket<K, V>>],
dictionary: &'a mut [MaybeUninit<u32>],
) -> Self {
// Initialize the buckets
for i in 0..buckets.len() {
buckets[i].write(Bucket {
next: if i < buckets.len() - 1 {
i as u32 + 1
} else {
INVALID_POS
},
inner: None,
});
}
// Initialize the dictionary
for e in dictionary.iter_mut() {
e.write(INVALID_POS);
}
// TODO: use std::slice::assume_init_mut() once it stabilizes
let buckets =
unsafe { std::slice::from_raw_parts_mut(buckets.as_mut_ptr().cast(), buckets.len()) };
let dictionary = unsafe {
std::slice::from_raw_parts_mut(dictionary.as_mut_ptr().cast(), dictionary.len())
};
Self {
dictionary,
buckets,
free_head: 0,
buckets_in_use: 0,
_user_list_head: INVALID_POS,
alloc_limit: INVALID_POS,
}
}
/// Get the value associated with a key (if it exists) given its hash.
pub fn get_with_hash(&self, key: &K, hash: u64) -> Option<&V> {
let mut next = self.dictionary[hash as usize % self.dictionary.len()];
loop {
if next == INVALID_POS {
return None;
}
let bucket = &self.buckets[next as usize];
let (bucket_key, bucket_value) = bucket.inner.as_ref().expect("entry is in use");
if bucket_key == key {
return Some(bucket_value);
}
next = bucket.next;
}
}
/// Get number of buckets in map.
pub fn get_num_buckets(&self) -> usize {
self.buckets.len()
}
/// Clears all entries from the hashmap.
///
/// Does not reset any allocation limits, but does clear any entries beyond them.
pub fn clear(&mut self) {
for i in 0..self.buckets.len() {
self.buckets[i] = Bucket {
next: if i < self.buckets.len() - 1 {
i as u32 + 1
} else {
INVALID_POS
},
inner: None,
}
}
for i in 0..self.dictionary.len() {
self.dictionary[i] = INVALID_POS;
}
self.free_head = 0;
self.buckets_in_use = 0;
}
/// Find the position of an unused bucket via the freelist and initialize it.
pub(crate) fn alloc_bucket(&mut self, key: K, value: V) -> Result<u32, FullError> {
let mut pos = self.free_head;
// Find the first bucket we're *allowed* to use.
let mut prev = PrevPos::First(self.free_head);
while pos != INVALID_POS && pos >= self.alloc_limit {
let bucket = &mut self.buckets[pos as usize];
prev = PrevPos::Chained(pos);
pos = bucket.next;
}
if pos == INVALID_POS {
return Err(FullError());
}
// Repair the freelist.
match prev {
PrevPos::First(_) => {
let next_pos = self.buckets[pos as usize].next;
self.free_head = next_pos;
}
PrevPos::Chained(p) => if p != INVALID_POS {
let next_pos = self.buckets[pos as usize].next;
self.buckets[p as usize].next = next_pos;
},
_ => unreachable!()
}
// Initialize the bucket.
let bucket = &mut self.buckets[pos as usize];
self.buckets_in_use += 1;
bucket.next = INVALID_POS;
bucket.inner = Some((key, value));
Ok(pos)
}
}

View File

@@ -1,139 +0,0 @@
//! Equivalent of [`std::collections::hash_map::Entry`] for this hashmap.
use crate::hash::core::{CoreHashMap, FullError, INVALID_POS};
use crate::sync::{RwLockWriteGuard, ValueWriteGuard};
use std::hash::Hash;
use std::mem;
pub enum Entry<'a, 'b, K, V> {
Occupied(OccupiedEntry<'a, 'b, K, V>),
Vacant(VacantEntry<'a, 'b, K, V>),
}
/// Enum representing the previous position within a chain.
#[derive(Clone, Copy)]
pub(crate) enum PrevPos {
/// Starting index within the dictionary.
First(u32),
/// Regular index within the buckets.
Chained(u32),
/// Unknown - e.g. the associated entry was retrieved by index instead of chain.
Unknown(u64),
}
pub struct OccupiedEntry<'a, 'b, K, V> {
/// Mutable reference to the map containing this entry.
pub(crate) map: RwLockWriteGuard<'b, CoreHashMap<'a, K, V>>,
/// The key of the occupied entry
pub(crate) _key: K,
/// The index of the previous entry in the chain.
pub(crate) prev_pos: PrevPos,
/// The position of the bucket in the [`CoreHashMap`] bucket array.
pub(crate) bucket_pos: u32,
}
impl<K, V> OccupiedEntry<'_, '_, K, V> {
pub fn get(&self) -> &V {
&self.map.buckets[self.bucket_pos as usize]
.inner
.as_ref()
.unwrap()
.1
}
pub fn get_mut(&mut self) -> &mut V {
&mut self.map.buckets[self.bucket_pos as usize]
.inner
.as_mut()
.unwrap()
.1
}
/// Inserts a value into the entry, replacing (and returning) the existing value.
pub fn insert(&mut self, value: V) -> V {
let bucket = &mut self.map.buckets[self.bucket_pos as usize];
// This assumes inner is Some, which it must be for an OccupiedEntry
mem::replace(&mut bucket.inner.as_mut().unwrap().1, value)
}
/// Removes the entry from the hash map, returning the value originally stored within it.
///
/// This may result in multiple bucket accesses if the entry was obtained by index as the
/// previous chain entry needs to be discovered in this case.
///
/// # Panics
/// Panics if the `prev_pos` field is equal to [`PrevPos::Unknown`]. In practice, this means
/// the entry was obtained via calling something like [`CoreHashMap::entry_at_bucket`].
pub fn remove(mut self) -> V {
// If this bucket was queried by index, go ahead and follow its chain from the start.
let prev = if let PrevPos::Unknown(hash) = self.prev_pos {
let dict_idx = hash as usize % self.map.dictionary.len();
let mut prev = PrevPos::First(dict_idx as u32);
let mut curr = self.map.dictionary[dict_idx];
while curr != self.bucket_pos {
curr = self.map.buckets[curr as usize].next;
prev = PrevPos::Chained(curr);
}
prev
} else {
self.prev_pos
};
// CoreHashMap::remove returns Option<(K, V)>. We know it's Some for an OccupiedEntry.
let bucket = &mut self.map.buckets[self.bucket_pos as usize];
// unlink it from the chain
match prev {
PrevPos::First(dict_pos) => {
self.map.dictionary[dict_pos as usize] = bucket.next;
},
PrevPos::Chained(bucket_pos) => {
// println!("we think prev of {} is {bucket_pos}", self.bucket_pos);
self.map.buckets[bucket_pos as usize].next = bucket.next;
},
_ => unreachable!(),
}
// and add it to the freelist
let free = self.map.free_head;
let bucket = &mut self.map.buckets[self.bucket_pos as usize];
let old_value = bucket.inner.take();
bucket.next = free;
self.map.free_head = self.bucket_pos;
self.map.buckets_in_use -= 1;
old_value.unwrap().1
}
}
/// An abstract view into a vacant entry within the map.
pub struct VacantEntry<'a, 'b, K, V> {
/// Mutable reference to the map containing this entry.
pub(crate) map: RwLockWriteGuard<'b, CoreHashMap<'a, K, V>>,
/// The key to be inserted into this entry.
pub(crate) key: K,
/// The position within the dictionary corresponding to the key's hash.
pub(crate) dict_pos: u32,
}
impl<'b, K: Clone + Hash + Eq, V> VacantEntry<'_, 'b, K, V> {
/// Insert a value into the vacant entry, finding and populating an empty bucket in the process.
///
/// # Errors
/// Will return [`FullError`] if there are no unoccupied buckets in the map.
pub fn insert(mut self, value: V) -> Result<ValueWriteGuard<'b, V>, FullError> {
let pos = self.map.alloc_bucket(self.key, value)?;
if pos == INVALID_POS {
return Err(FullError());
}
self.map.buckets[pos as usize].next = self.map.dictionary[self.dict_pos as usize];
self.map.dictionary[self.dict_pos as usize] = pos;
Ok(RwLockWriteGuard::map(
self.map,
|m| &mut m.buckets[pos as usize].inner.as_mut().unwrap().1
))
}
}

View File

@@ -1,426 +0,0 @@
use std::collections::BTreeMap;
use std::collections::HashSet;
use std::fmt::Debug;
use std::mem::MaybeUninit;
use crate::hash::HashMapAccess;
use crate::hash::HashMapInit;
use crate::hash::Entry;
use crate::hash::core::FullError;
use rand::seq::SliceRandom;
use rand::{Rng, RngCore};
use rand_distr::Zipf;
const TEST_KEY_LEN: usize = 16;
#[derive(Clone, Copy, Debug, Hash, PartialEq, Eq, PartialOrd, Ord)]
struct TestKey([u8; TEST_KEY_LEN]);
impl From<&TestKey> for u128 {
fn from(val: &TestKey) -> u128 {
u128::from_be_bytes(val.0)
}
}
impl From<u128> for TestKey {
fn from(val: u128) -> TestKey {
TestKey(val.to_be_bytes())
}
}
impl<'a> From<&'a [u8]> for TestKey {
fn from(bytes: &'a [u8]) -> TestKey {
TestKey(bytes.try_into().unwrap())
}
}
fn test_inserts<K: Into<TestKey> + Copy>(keys: &[K]) {
let w = HashMapInit::<TestKey, usize>::new_resizeable_named(
100000, 120000, "test_inserts"
).attach_writer();
for (idx, k) in keys.iter().enumerate() {
let res = w.entry((*k).into());
match res {
Entry::Occupied(mut e) => { e.insert(idx); }
Entry::Vacant(e) => {
let res = e.insert(idx);
assert!(res.is_ok());
},
};
}
for (idx, k) in keys.iter().enumerate() {
let x = w.get(&(*k).into());
let value = x.as_deref().copied();
assert_eq!(value, Some(idx));
}
}
#[test]
fn dense() {
// This exercises splitting a node with prefix
let keys: &[u128] = &[0, 1, 2, 3, 256];
test_inserts(keys);
// Dense keys
let mut keys: Vec<u128> = (0..10000).collect();
test_inserts(&keys);
// Do the same in random orders
for _ in 1..10 {
keys.shuffle(&mut rand::rng());
test_inserts(&keys);
}
}
#[test]
fn sparse() {
// sparse keys
let mut keys: Vec<TestKey> = Vec::new();
let mut used_keys = HashSet::new();
for _ in 0..10000 {
loop {
let key = rand::random::<u128>();
if used_keys.get(&key).is_some() {
continue;
}
used_keys.insert(key);
keys.push(key.into());
break;
}
}
test_inserts(&keys);
}
#[derive(Clone, Debug)]
struct TestOp(TestKey, Option<usize>);
fn apply_op(
op: &TestOp,
map: &mut HashMapAccess<TestKey, usize>,
shadow: &mut BTreeMap<TestKey, usize>,
) {
// apply the change to the shadow tree first
let shadow_existing = if let Some(v) = op.1 {
shadow.insert(op.0, v)
} else {
shadow.remove(&op.0)
};
let entry = map.entry(op.0);
let hash_existing = match op.1 {
Some(new) => {
match entry {
Entry::Occupied(mut e) => Some(e.insert(new)),
Entry::Vacant(e) => { _ = e.insert(new).unwrap(); None },
}
},
None => {
match entry {
Entry::Occupied(e) => Some(e.remove()),
Entry::Vacant(_) => None,
}
},
};
assert_eq!(shadow_existing, hash_existing);
}
fn do_random_ops(
num_ops: usize,
size: u32,
del_prob: f64,
writer: &mut HashMapAccess<TestKey, usize>,
shadow: &mut BTreeMap<TestKey, usize>,
rng: &mut rand::rngs::ThreadRng,
) {
for i in 0..num_ops {
let key: TestKey = ((rng.next_u32() % size) as u128).into();
let op = TestOp(key, if rng.random_bool(del_prob) { Some(i) } else { None });
apply_op(&op, writer, shadow);
}
}
fn do_deletes(
num_ops: usize,
writer: &mut HashMapAccess<TestKey, usize>,
shadow: &mut BTreeMap<TestKey, usize>,
) {
for _ in 0..num_ops {
let (k, _) = shadow.pop_first().unwrap();
writer.remove(&k);
}
}
fn do_shrink(
writer: &mut HashMapAccess<TestKey, usize>,
shadow: &mut BTreeMap<TestKey, usize>,
to: u32
) {
assert!(writer.shrink_goal().is_none());
writer.begin_shrink(to);
assert_eq!(writer.shrink_goal(), Some(to as usize));
while writer.get_num_buckets_in_use() > to as usize {
let (k, _) = shadow.pop_first().unwrap();
let entry = writer.entry(k);
if let Entry::Occupied(e) = entry {
e.remove();
}
}
let old_usage = writer.get_num_buckets_in_use();
writer.finish_shrink().unwrap();
assert!(writer.shrink_goal().is_none());
assert_eq!(writer.get_num_buckets_in_use(), old_usage);
}
#[test]
fn random_ops() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
100000, 120000, "test_random"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let distribution = Zipf::new(u128::MAX as f64, 1.1).unwrap();
let mut rng = rand::rng();
for i in 0..100000 {
let key: TestKey = (rng.sample(distribution) as u128).into();
let op = TestOp(key, if rng.random_bool(0.75) { Some(i) } else { None });
apply_op(&op, &mut writer, &mut shadow);
}
}
#[test]
fn test_shuffle() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1000, 1200, "test_shuf"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(10000, 1000, 0.75, &mut writer, &mut shadow, &mut rng);
writer.shuffle();
do_random_ops(10000, 1000, 0.75, &mut writer, &mut shadow, &mut rng);
}
#[test]
fn test_grow() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1000, 2000, "test_grow"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(10000, 1000, 0.75, &mut writer, &mut shadow, &mut rng);
let old_usage = writer.get_num_buckets_in_use();
writer.grow(1500).unwrap();
assert_eq!(writer.get_num_buckets_in_use(), old_usage);
assert_eq!(writer.get_num_buckets(), 1500);
do_random_ops(10000, 1500, 0.75, &mut writer, &mut shadow, &mut rng);
}
#[test]
fn test_clear() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2000, "test_clear"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(2000, 1500, 0.75, &mut writer, &mut shadow, &mut rng);
writer.clear();
assert_eq!(writer.get_num_buckets_in_use(), 0);
assert_eq!(writer.get_num_buckets(), 1500);
while let Some((key, _)) = shadow.pop_first() {
assert!(writer.get(&key).is_none());
}
do_random_ops(2000, 1500, 0.75, &mut writer, &mut shadow, &mut rng);
for i in 0..(1500 - writer.get_num_buckets_in_use()) {
writer.insert((1500 + i as u128).into(), 0).unwrap();
}
assert_eq!(writer.insert(5000.into(), 0), Err(FullError {}));
writer.clear();
assert!(writer.insert(5000.into(), 0).is_ok());
}
#[test]
fn test_idx_remove() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2000, "test_clear"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(2000, 1500, 0.25, &mut writer, &mut shadow, &mut rng);
for _ in 0..100 {
let idx = (rng.next_u32() % 1500) as usize;
if let Some(e) = writer.entry_at_bucket(idx) {
shadow.remove(&e._key);
e.remove();
}
}
while let Some((key, val)) = shadow.pop_first() {
assert_eq!(*writer.get(&key).unwrap(), val);
}
}
#[test]
fn test_idx_get() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2000, "test_clear"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(2000, 1500, 0.25, &mut writer, &mut shadow, &mut rng);
for _ in 0..100 {
let idx = (rng.next_u32() % 1500) as usize;
if let Some(pair) = writer.get_at_bucket(idx) {
{
let v: *const usize = &pair.1;
assert_eq!(writer.get_bucket_for_value(v), idx);
}
{
let v: *const usize = &pair.1;
assert_eq!(writer.get_bucket_for_value(v), idx);
}
}
}
}
#[test]
fn test_shrink() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2000, "test_shrink"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(10000, 1500, 0.75, &mut writer, &mut shadow, &mut rng);
do_shrink(&mut writer, &mut shadow, 1000);
assert_eq!(writer.get_num_buckets(), 1000);
do_deletes(500, &mut writer, &mut shadow);
do_random_ops(10000, 500, 0.75, &mut writer, &mut shadow, &mut rng);
assert!(writer.get_num_buckets_in_use() <= 1000);
}
#[test]
fn test_shrink_grow_seq() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1000, 20000, "test_grow_seq"
).attach_writer();
let mut shadow: std::collections::BTreeMap<TestKey, usize> = BTreeMap::new();
let mut rng = rand::rng();
do_random_ops(500, 1000, 0.1, &mut writer, &mut shadow, &mut rng);
eprintln!("Shrinking to 750");
do_shrink(&mut writer, &mut shadow, 750);
do_random_ops(200, 1000, 0.5, &mut writer, &mut shadow, &mut rng);
eprintln!("Growing to 1500");
writer.grow(1500).unwrap();
do_random_ops(600, 1500, 0.1, &mut writer, &mut shadow, &mut rng);
eprintln!("Shrinking to 200");
while shadow.len() > 100 {
do_deletes(1, &mut writer, &mut shadow);
}
do_shrink(&mut writer, &mut shadow, 200);
do_random_ops(50, 1500, 0.25, &mut writer, &mut shadow, &mut rng);
eprintln!("Growing to 10k");
writer.grow(10000).unwrap();
do_random_ops(10000, 5000, 0.25, &mut writer, &mut shadow, &mut rng);
}
#[test]
fn test_bucket_ops() {
let writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1000, 1200, "test_bucket_ops"
).attach_writer();
match writer.entry(1.into()) {
Entry::Occupied(mut e) => { e.insert(2); },
Entry::Vacant(e) => { _ = e.insert(2).unwrap(); },
}
assert_eq!(writer.get_num_buckets_in_use(), 1);
assert_eq!(writer.get_num_buckets(), 1000);
assert_eq!(*writer.get(&1.into()).unwrap(), 2);
let pos = match writer.entry(1.into()) {
Entry::Occupied(e) => {
assert_eq!(e._key, 1.into());
let pos = e.bucket_pos as usize;
pos
},
Entry::Vacant(_) => { panic!("Insert didn't affect entry"); },
};
assert_eq!(writer.entry_at_bucket(pos).unwrap()._key, 1.into());
assert_eq!(*writer.get_at_bucket(pos).unwrap(), (1.into(), 2));
{
let ptr: *const usize = &*writer.get(&1.into()).unwrap();
assert_eq!(writer.get_bucket_for_value(ptr), pos);
}
writer.remove(&1.into());
assert!(writer.get(&1.into()).is_none());
}
#[test]
fn test_shrink_zero() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2000, "test_shrink_zero"
).attach_writer();
writer.begin_shrink(0);
for i in 0..1500 {
writer.entry_at_bucket(i).map(|x| x.remove());
}
writer.finish_shrink().unwrap();
assert_eq!(writer.get_num_buckets_in_use(), 0);
let entry = writer.entry(1.into());
if let Entry::Vacant(v) = entry {
assert!(v.insert(2).is_err());
} else {
panic!("Somehow got non-vacant entry in empty map.")
}
writer.grow(50).unwrap();
let entry = writer.entry(1.into());
if let Entry::Vacant(v) = entry {
assert!(v.insert(2).is_ok());
} else {
panic!("Somehow got non-vacant entry in empty map.")
}
assert_eq!(writer.get_num_buckets_in_use(), 1);
}
#[test]
#[should_panic]
fn test_grow_oom() {
let writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2000, "test_grow_oom"
).attach_writer();
writer.grow(20000).unwrap();
}
#[test]
#[should_panic]
fn test_shrink_bigger() {
let mut writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2500, "test_shrink_bigger"
).attach_writer();
writer.begin_shrink(2000);
}
#[test]
#[should_panic]
fn test_shrink_early_finish() {
let writer = HashMapInit::<TestKey, usize>::new_resizeable_named(
1500, 2500, "test_shrink_early_finish"
).attach_writer();
writer.finish_shrink().unwrap();
}
#[test]
#[should_panic]
fn test_shrink_fixed_size() {
let mut area = [MaybeUninit::uninit(); 10000];
let init_struct = HashMapInit::<TestKey, usize>::with_fixed(3, &mut area);
let mut writer = init_struct.attach_writer();
writer.begin_shrink(1);
}

View File

@@ -1,5 +1,418 @@
//! Shared memory utilities for neon communicator
pub mod hash;
pub mod shmem;
pub mod sync;
use std::num::NonZeroUsize;
use std::os::fd::{AsFd, BorrowedFd, OwnedFd};
use std::ptr::NonNull;
use std::sync::atomic::{AtomicUsize, Ordering};
use nix::errno::Errno;
use nix::sys::mman::MapFlags;
use nix::sys::mman::ProtFlags;
use nix::sys::mman::mmap as nix_mmap;
use nix::sys::mman::munmap as nix_munmap;
use nix::unistd::ftruncate as nix_ftruncate;
/// ShmemHandle represents a shared memory area that can be shared by processes over fork().
/// Unlike shared memory allocated by Postgres, this area is resizable, up to 'max_size' that's
/// specified at creation.
///
/// The area is backed by an anonymous file created with memfd_create(). The full address space for
/// 'max_size' is reserved up-front with mmap(), but whenever you call [`ShmemHandle::set_size`],
/// the underlying file is resized. Do not access the area beyond the current size. Currently, that
/// will cause the file to be expanded, but we might use mprotect() etc. to enforce that in the
/// future.
pub struct ShmemHandle {
/// memfd file descriptor
fd: OwnedFd,
max_size: usize,
// Pointer to the beginning of the shared memory area. The header is stored there.
shared_ptr: NonNull<SharedStruct>,
// Pointer to the beginning of the user data
pub data_ptr: NonNull<u8>,
}
/// This is stored at the beginning in the shared memory area.
struct SharedStruct {
max_size: usize,
/// Current size of the backing file. The high-order bit is used for the RESIZE_IN_PROGRESS flag
current_size: AtomicUsize,
}
const RESIZE_IN_PROGRESS: usize = 1 << 63;
const HEADER_SIZE: usize = std::mem::size_of::<SharedStruct>();
/// Error type returned by the ShmemHandle functions.
#[derive(thiserror::Error, Debug)]
#[error("{msg}: {errno}")]
pub struct Error {
pub msg: String,
pub errno: Errno,
}
impl Error {
fn new(msg: &str, errno: Errno) -> Error {
Error {
msg: msg.to_string(),
errno,
}
}
}
impl ShmemHandle {
/// Create a new shared memory area. To communicate between processes, the processes need to be
/// fork()'d after calling this, so that the ShmemHandle is inherited by all processes.
///
/// If the ShmemHandle is dropped, the memory is unmapped from the current process. Other
/// processes can continue using it, however.
pub fn new(name: &str, initial_size: usize, max_size: usize) -> Result<ShmemHandle, Error> {
// create the backing anonymous file.
let fd = create_backing_file(name)?;
Self::new_with_fd(fd, initial_size, max_size)
}
fn new_with_fd(
fd: OwnedFd,
initial_size: usize,
max_size: usize,
) -> Result<ShmemHandle, Error> {
// We reserve the high-order bit for the RESIZE_IN_PROGRESS flag, and the actual size
// is a little larger than this because of the SharedStruct header. Make the upper limit
// somewhat smaller than that, because with anything close to that, you'll run out of
// memory anyway.
if max_size >= 1 << 48 {
panic!("max size {} too large", max_size);
}
if initial_size > max_size {
panic!("initial size {initial_size} larger than max size {max_size}");
}
// The actual initial / max size is the one given by the caller, plus the size of
// 'SharedStruct'.
let initial_size = HEADER_SIZE + initial_size;
let max_size = NonZeroUsize::new(HEADER_SIZE + max_size).unwrap();
// Reserve address space for it with mmap
//
// TODO: Use MAP_HUGETLB if possible
let start_ptr = unsafe {
nix_mmap(
None,
max_size,
ProtFlags::PROT_READ | ProtFlags::PROT_WRITE,
MapFlags::MAP_SHARED,
&fd,
0,
)
}
.map_err(|e| Error::new("mmap failed: {e}", e))?;
// Reserve space for the initial size
enlarge_file(fd.as_fd(), initial_size as u64)?;
// Initialize the header
let shared: NonNull<SharedStruct> = start_ptr.cast();
unsafe {
shared.write(SharedStruct {
max_size: max_size.into(),
current_size: AtomicUsize::new(initial_size),
})
};
// The user data begins after the header
let data_ptr = unsafe { start_ptr.cast().add(HEADER_SIZE) };
Ok(ShmemHandle {
fd,
max_size: max_size.into(),
shared_ptr: shared,
data_ptr,
})
}
// return reference to the header
fn shared(&self) -> &SharedStruct {
unsafe { self.shared_ptr.as_ref() }
}
/// Resize the shared memory area. 'new_size' must not be larger than the 'max_size' specified
/// when creating the area.
///
/// This may only be called from one process/thread concurrently. We detect that case
/// and return an Error.
pub fn set_size(&self, new_size: usize) -> Result<(), Error> {
let new_size = new_size + HEADER_SIZE;
let shared = self.shared();
if new_size > self.max_size {
panic!(
"new size ({} is greater than max size ({})",
new_size, self.max_size
);
}
assert_eq!(self.max_size, shared.max_size);
// Lock the area by setting the bit in 'current_size'
//
// Ordering::Relaxed would probably be sufficient here, as we don't access any other memory
// and the posix_fallocate/ftruncate call is surely a synchronization point anyway. But
// since this is not performance-critical, better safe than sorry .
let mut old_size = shared.current_size.load(Ordering::Acquire);
loop {
if (old_size & RESIZE_IN_PROGRESS) != 0 {
return Err(Error::new(
"concurrent resize detected",
Errno::UnknownErrno,
));
}
match shared.current_size.compare_exchange(
old_size,
new_size,
Ordering::Acquire,
Ordering::Relaxed,
) {
Ok(_) => break,
Err(x) => old_size = x,
}
}
// Ok, we got the lock.
//
// NB: If anything goes wrong, we *must* clear the bit!
let result = {
use std::cmp::Ordering::{Equal, Greater, Less};
match new_size.cmp(&old_size) {
Less => nix_ftruncate(&self.fd, new_size as i64).map_err(|e| {
Error::new("could not shrink shmem segment, ftruncate failed: {e}", e)
}),
Equal => Ok(()),
Greater => enlarge_file(self.fd.as_fd(), new_size as u64),
}
};
// Unlock
shared.current_size.store(
if result.is_ok() { new_size } else { old_size },
Ordering::Release,
);
result
}
/// Returns the current user-visible size of the shared memory segment.
///
/// NOTE: a concurrent set_size() call can change the size at any time. It is the caller's
/// responsibility not to access the area beyond the current size.
pub fn current_size(&self) -> usize {
let total_current_size =
self.shared().current_size.load(Ordering::Relaxed) & !RESIZE_IN_PROGRESS;
total_current_size - HEADER_SIZE
}
}
impl Drop for ShmemHandle {
fn drop(&mut self) {
// SAFETY: The pointer was obtained from mmap() with the given size.
// We unmap the entire region.
let _ = unsafe { nix_munmap(self.shared_ptr.cast(), self.max_size) };
// The fd is dropped automatically by OwnedFd.
}
}
/// Create a "backing file" for the shared memory area. On Linux, use memfd_create(), to create an
/// anonymous in-memory file. One macos, fall back to a regular file. That's good enough for
/// development and testing, but in production we want the file to stay in memory.
///
/// disable 'unused_variables' warnings, because in the macos path, 'name' is unused.
#[allow(unused_variables)]
fn create_backing_file(name: &str) -> Result<OwnedFd, Error> {
#[cfg(not(target_os = "macos"))]
{
nix::sys::memfd::memfd_create(name, nix::sys::memfd::MFdFlags::empty())
.map_err(|e| Error::new("memfd_create failed: {e}", e))
}
#[cfg(target_os = "macos")]
{
let file = tempfile::tempfile().map_err(|e| {
Error::new(
"could not create temporary file to back shmem area: {e}",
nix::errno::Errno::from_raw(e.raw_os_error().unwrap_or(0)),
)
})?;
Ok(OwnedFd::from(file))
}
}
fn enlarge_file(fd: BorrowedFd, size: u64) -> Result<(), Error> {
// Use posix_fallocate() to enlarge the file. It reserves the space correctly, so that
// we don't get a segfault later when trying to actually use it.
#[cfg(not(target_os = "macos"))]
{
nix::fcntl::posix_fallocate(fd, 0, size as i64).map_err(|e| {
Error::new(
"could not grow shmem segment, posix_fallocate failed: {e}",
e,
)
})
}
// As a fallback on macos, which doesn't have posix_fallocate, use plain 'fallocate'
#[cfg(target_os = "macos")]
{
nix::unistd::ftruncate(fd, size as i64)
.map_err(|e| Error::new("could not grow shmem segment, ftruncate failed: {e}", e))
}
}
#[cfg(test)]
mod tests {
use super::*;
use nix::unistd::ForkResult;
use std::ops::Range;
/// check that all bytes in given range have the expected value.
fn assert_range(ptr: *const u8, expected: u8, range: Range<usize>) {
for i in range {
let b = unsafe { *(ptr.add(i)) };
assert_eq!(expected, b, "unexpected byte at offset {}", i);
}
}
/// Write 'b' to all bytes in the given range
fn write_range(ptr: *mut u8, b: u8, range: Range<usize>) {
unsafe { std::ptr::write_bytes(ptr.add(range.start), b, range.end - range.start) };
}
// simple single-process test of growing and shrinking
#[test]
fn test_shmem_resize() -> Result<(), Error> {
let max_size = 1024 * 1024;
let init_struct = ShmemHandle::new("test_shmem_resize", 0, max_size)?;
assert_eq!(init_struct.current_size(), 0);
// Initial grow
let size1 = 10000;
init_struct.set_size(size1).unwrap();
assert_eq!(init_struct.current_size(), size1);
// Write some data
let data_ptr = init_struct.data_ptr.as_ptr();
write_range(data_ptr, 0xAA, 0..size1);
assert_range(data_ptr, 0xAA, 0..size1);
// Shrink
let size2 = 5000;
init_struct.set_size(size2).unwrap();
assert_eq!(init_struct.current_size(), size2);
// Grow again
let size3 = 20000;
init_struct.set_size(size3).unwrap();
assert_eq!(init_struct.current_size(), size3);
// Try to read it. The area that was shrunk and grown again should read as all zeros now
assert_range(data_ptr, 0xAA, 0..5000);
assert_range(data_ptr, 0, 5000..size1);
// Try to grow beyond max_size
//let size4 = max_size + 1;
//assert!(init_struct.set_size(size4).is_err());
// Dropping init_struct should unmap the memory
drop(init_struct);
Ok(())
}
/// This is used in tests to coordinate between test processes. It's like std::sync::Barrier,
/// but is stored in the shared memory area and works across processes. It's implemented by
/// polling, because e.g. standard rust mutexes are not guaranteed to work across processes.
struct SimpleBarrier {
num_procs: usize,
count: AtomicUsize,
}
impl SimpleBarrier {
unsafe fn init(ptr: *mut SimpleBarrier, num_procs: usize) {
unsafe {
*ptr = SimpleBarrier {
num_procs,
count: AtomicUsize::new(0),
}
}
}
pub fn wait(&self) {
let old = self.count.fetch_add(1, Ordering::Relaxed);
let generation = old / self.num_procs;
let mut current = old + 1;
while current < (generation + 1) * self.num_procs {
std::thread::sleep(std::time::Duration::from_millis(10));
current = self.count.load(Ordering::Relaxed);
}
}
}
#[test]
fn test_multi_process() {
// Initialize
let max_size = 1_000_000_000_000;
let init_struct = ShmemHandle::new("test_multi_process", 0, max_size).unwrap();
let ptr = init_struct.data_ptr.as_ptr();
// Store the SimpleBarrier in the first 1k of the area.
init_struct.set_size(10000).unwrap();
let barrier_ptr: *mut SimpleBarrier = unsafe {
ptr.add(ptr.align_offset(std::mem::align_of::<SimpleBarrier>()))
.cast()
};
unsafe { SimpleBarrier::init(barrier_ptr, 2) };
let barrier = unsafe { barrier_ptr.as_ref().unwrap() };
// Fork another test process. The code after this runs in both processes concurrently.
let fork_result = unsafe { nix::unistd::fork().unwrap() };
// In the parent, fill bytes between 1000..2000. In the child, between 2000..3000
if fork_result.is_parent() {
write_range(ptr, 0xAA, 1000..2000);
} else {
write_range(ptr, 0xBB, 2000..3000);
}
barrier.wait();
// Verify the contents. (in both processes)
assert_range(ptr, 0xAA, 1000..2000);
assert_range(ptr, 0xBB, 2000..3000);
// Grow, from the child this time
let size = 10_000_000;
if !fork_result.is_parent() {
init_struct.set_size(size).unwrap();
}
barrier.wait();
// make some writes at the end
if fork_result.is_parent() {
write_range(ptr, 0xAA, (size - 10)..size);
} else {
write_range(ptr, 0xBB, (size - 20)..(size - 10));
}
barrier.wait();
// Verify the contents. (This runs in both processes)
assert_range(ptr, 0, (size - 1000)..(size - 20));
assert_range(ptr, 0xBB, (size - 20)..(size - 10));
assert_range(ptr, 0xAA, (size - 10)..size);
if let ForkResult::Parent { child } = fork_result {
nix::sys::wait::waitpid(child, None).unwrap();
}
}
}

View File

@@ -1,418 +0,0 @@
//! Dynamically resizable contiguous chunk of shared memory
use std::num::NonZeroUsize;
use std::os::fd::{AsFd, BorrowedFd, OwnedFd};
use std::ptr::NonNull;
use std::sync::atomic::{AtomicUsize, Ordering};
use nix::errno::Errno;
use nix::sys::mman::MapFlags;
use nix::sys::mman::ProtFlags;
use nix::sys::mman::mmap as nix_mmap;
use nix::sys::mman::munmap as nix_munmap;
use nix::unistd::ftruncate as nix_ftruncate;
/// `ShmemHandle` represents a shared memory area that can be shared by processes over `fork()`.
/// Unlike shared memory allocated by Postgres, this area is resizable, up to `max_size` that's
/// specified at creation.
///
/// The area is backed by an anonymous file created with `memfd_create()`. The full address space for
/// `max_size` is reserved up-front with `mmap()`, but whenever you call [`ShmemHandle::set_size`],
/// the underlying file is resized. Do not access the area beyond the current size. Currently, that
/// will cause the file to be expanded, but we might use `mprotect()` etc. to enforce that in the
/// future.
pub struct ShmemHandle {
/// memfd file descriptor
fd: OwnedFd,
max_size: usize,
// Pointer to the beginning of the shared memory area. The header is stored there.
shared_ptr: NonNull<SharedStruct>,
// Pointer to the beginning of the user data
pub data_ptr: NonNull<u8>,
}
/// This is stored at the beginning in the shared memory area.
struct SharedStruct {
max_size: usize,
/// Current size of the backing file. The high-order bit is used for the [`RESIZE_IN_PROGRESS`] flag.
current_size: AtomicUsize,
}
const RESIZE_IN_PROGRESS: usize = 1 << 63;
const HEADER_SIZE: usize = std::mem::size_of::<SharedStruct>();
/// Error type returned by the [`ShmemHandle`] functions.
#[derive(thiserror::Error, Debug)]
#[error("{msg}: {errno}")]
pub struct Error {
pub msg: String,
pub errno: Errno,
}
impl Error {
fn new(msg: &str, errno: Errno) -> Self {
Self {
msg: msg.to_string(),
errno,
}
}
}
impl ShmemHandle {
/// Create a new shared memory area. To communicate between processes, the processes need to be
/// `fork()`'d after calling this, so that the `ShmemHandle` is inherited by all processes.
///
/// If the `ShmemHandle` is dropped, the memory is unmapped from the current process. Other
/// processes can continue using it, however.
pub fn new(name: &str, initial_size: usize, max_size: usize) -> Result<Self, Error> {
// create the backing anonymous file.
let fd = create_backing_file(name)?;
Self::new_with_fd(fd, initial_size, max_size)
}
fn new_with_fd(
fd: OwnedFd,
initial_size: usize,
max_size: usize,
) -> Result<Self, Error> {
// We reserve the high-order bit for the `RESIZE_IN_PROGRESS` flag, and the actual size
// is a little larger than this because of the SharedStruct header. Make the upper limit
// somewhat smaller than that, because with anything close to that, you'll run out of
// memory anyway.
assert!(max_size < 1 << 48, "max size {max_size} too large");
assert!(
initial_size <= max_size,
"initial size {initial_size} larger than max size {max_size}"
);
// The actual initial / max size is the one given by the caller, plus the size of
// 'SharedStruct'.
let initial_size = HEADER_SIZE + initial_size;
let max_size = NonZeroUsize::new(HEADER_SIZE + max_size).unwrap();
// Reserve address space for it with mmap
//
// TODO: Use MAP_HUGETLB if possible
let start_ptr = unsafe {
nix_mmap(
None,
max_size,
ProtFlags::PROT_READ | ProtFlags::PROT_WRITE,
MapFlags::MAP_SHARED,
&fd,
0,
)
}
.map_err(|e| Error::new("mmap failed", e))?;
// Reserve space for the initial size
enlarge_file(fd.as_fd(), initial_size as u64)?;
// Initialize the header
let shared: NonNull<SharedStruct> = start_ptr.cast();
unsafe {
shared.write(SharedStruct {
max_size: max_size.into(),
current_size: AtomicUsize::new(initial_size),
});
}
// The user data begins after the header
let data_ptr = unsafe { start_ptr.cast().add(HEADER_SIZE) };
Ok(Self {
fd,
max_size: max_size.into(),
shared_ptr: shared,
data_ptr,
})
}
// return reference to the header
fn shared(&self) -> &SharedStruct {
unsafe { self.shared_ptr.as_ref() }
}
/// Resize the shared memory area. `new_size` must not be larger than the `max_size` specified
/// when creating the area.
///
/// This may only be called from one process/thread concurrently. We detect that case
/// and return an [`shmem::Error`](Error).
pub fn set_size(&self, new_size: usize) -> Result<(), Error> {
let new_size = new_size + HEADER_SIZE;
let shared = self.shared();
assert!(
new_size <= self.max_size,
"new size ({new_size}) is greater than max size ({})",
self.max_size
);
assert_eq!(self.max_size, shared.max_size);
// Lock the area by setting the bit in `current_size`
//
// Ordering::Relaxed would probably be sufficient here, as we don't access any other memory
// and the `posix_fallocate`/`ftruncate` call is surely a synchronization point anyway. But
// since this is not performance-critical, better safe than sorry.
let mut old_size = shared.current_size.load(Ordering::Acquire);
loop {
if (old_size & RESIZE_IN_PROGRESS) != 0 {
return Err(Error::new(
"concurrent resize detected",
Errno::UnknownErrno,
));
}
match shared.current_size.compare_exchange(
old_size,
new_size,
Ordering::Acquire,
Ordering::Relaxed,
) {
Ok(_) => break,
Err(x) => old_size = x,
}
}
// Ok, we got the lock.
//
// NB: If anything goes wrong, we *must* clear the bit!
let result = {
use std::cmp::Ordering::{Equal, Greater, Less};
match new_size.cmp(&old_size) {
Less => nix_ftruncate(&self.fd, new_size as i64).map_err(|e| {
Error::new("could not shrink shmem segment, ftruncate failed", e)
}),
Equal => Ok(()),
Greater => enlarge_file(self.fd.as_fd(), new_size as u64),
}
};
// Unlock
shared.current_size.store(
if result.is_ok() { new_size } else { old_size },
Ordering::Release,
);
result
}
/// Returns the current user-visible size of the shared memory segment.
///
/// NOTE: a concurrent [`ShmemHandle::set_size()`] call can change the size at any time.
/// It is the caller's responsibility not to access the area beyond the current size.
pub fn current_size(&self) -> usize {
let total_current_size =
self.shared().current_size.load(Ordering::Relaxed) & !RESIZE_IN_PROGRESS;
total_current_size - HEADER_SIZE
}
}
impl Drop for ShmemHandle {
fn drop(&mut self) {
// SAFETY: The pointer was obtained from mmap() with the given size.
// We unmap the entire region.
let _ = unsafe { nix_munmap(self.shared_ptr.cast(), self.max_size) };
// The fd is dropped automatically by OwnedFd.
}
}
/// Create a "backing file" for the shared memory area. On Linux, use `memfd_create()`, to create an
/// anonymous in-memory file. One macos, fall back to a regular file. That's good enough for
/// development and testing, but in production we want the file to stay in memory.
///
/// Disable unused variables warnings because `name` is unused in the macos path.
#[allow(unused_variables)]
fn create_backing_file(name: &str) -> Result<OwnedFd, Error> {
#[cfg(not(target_os = "macos"))]
{
nix::sys::memfd::memfd_create(name, nix::sys::memfd::MFdFlags::empty())
.map_err(|e| Error::new("memfd_create failed", e))
}
#[cfg(target_os = "macos")]
{
let file = tempfile::tempfile().map_err(|e| {
Error::new(
"could not create temporary file to back shmem area",
nix::errno::Errno::from_raw(e.raw_os_error().unwrap_or(0)),
)
})?;
Ok(OwnedFd::from(file))
}
}
fn enlarge_file(fd: BorrowedFd, size: u64) -> Result<(), Error> {
// Use posix_fallocate() to enlarge the file. It reserves the space correctly, so that
// we don't get a segfault later when trying to actually use it.
#[cfg(not(target_os = "macos"))]
{
nix::fcntl::posix_fallocate(fd, 0, size as i64).map_err(|e| {
Error::new(
"could not grow shmem segment, posix_fallocate failed",
e,
)
})
}
// As a fallback on macos, which doesn't have posix_fallocate, use plain 'fallocate'
#[cfg(target_os = "macos")]
{
nix::unistd::ftruncate(fd, size as i64)
.map_err(|e| Error::new("could not grow shmem segment, ftruncate failed", e))
}
}
#[cfg(test)]
mod tests {
use super::*;
use nix::unistd::ForkResult;
use std::ops::Range;
/// check that all bytes in given range have the expected value.
fn assert_range(ptr: *const u8, expected: u8, range: Range<usize>) {
for i in range {
let b = unsafe { *(ptr.add(i)) };
assert_eq!(expected, b, "unexpected byte at offset {}", i);
}
}
/// Write 'b' to all bytes in the given range
fn write_range(ptr: *mut u8, b: u8, range: Range<usize>) {
unsafe { std::ptr::write_bytes(ptr.add(range.start), b, range.end - range.start) };
}
// simple single-process test of growing and shrinking
#[test]
fn test_shmem_resize() -> Result<(), Error> {
let max_size = 1024 * 1024;
let init_struct = ShmemHandle::new("test_shmem_resize", 0, max_size)?;
assert_eq!(init_struct.current_size(), 0);
// Initial grow
let size1 = 10000;
init_struct.set_size(size1).unwrap();
assert_eq!(init_struct.current_size(), size1);
// Write some data
let data_ptr = init_struct.data_ptr.as_ptr();
write_range(data_ptr, 0xAA, 0..size1);
assert_range(data_ptr, 0xAA, 0..size1);
// Shrink
let size2 = 5000;
init_struct.set_size(size2).unwrap();
assert_eq!(init_struct.current_size(), size2);
// Grow again
let size3 = 20000;
init_struct.set_size(size3).unwrap();
assert_eq!(init_struct.current_size(), size3);
// Try to read it. The area that was shrunk and grown again should read as all zeros now
assert_range(data_ptr, 0xAA, 0..5000);
assert_range(data_ptr, 0, 5000..size1);
// Try to grow beyond max_size
//let size4 = max_size + 1;
//assert!(init_struct.set_size(size4).is_err());
// Dropping init_struct should unmap the memory
drop(init_struct);
Ok(())
}
/// This is used in tests to coordinate between test processes. It's like `std::sync::Barrier`,
/// but is stored in the shared memory area and works across processes. It's implemented by
/// polling, because e.g. standard rust mutexes are not guaranteed to work across processes.
struct SimpleBarrier {
num_procs: usize,
count: AtomicUsize,
}
impl SimpleBarrier {
unsafe fn init(ptr: *mut SimpleBarrier, num_procs: usize) {
unsafe {
*ptr = SimpleBarrier {
num_procs,
count: AtomicUsize::new(0),
}
}
}
pub fn wait(&self) {
let old = self.count.fetch_add(1, Ordering::Relaxed);
let generation = old / self.num_procs;
let mut current = old + 1;
while current < (generation + 1) * self.num_procs {
std::thread::sleep(std::time::Duration::from_millis(10));
current = self.count.load(Ordering::Relaxed);
}
}
}
#[test]
fn test_multi_process() {
// Initialize
let max_size = 1_000_000_000_000;
let init_struct = ShmemHandle::new("test_multi_process", 0, max_size).unwrap();
let ptr = init_struct.data_ptr.as_ptr();
// Store the SimpleBarrier in the first 1k of the area.
init_struct.set_size(10000).unwrap();
let barrier_ptr: *mut SimpleBarrier = unsafe {
ptr.add(ptr.align_offset(std::mem::align_of::<SimpleBarrier>()))
.cast()
};
unsafe { SimpleBarrier::init(barrier_ptr, 2) };
let barrier = unsafe { barrier_ptr.as_ref().unwrap() };
// Fork another test process. The code after this runs in both processes concurrently.
let fork_result = unsafe { nix::unistd::fork().unwrap() };
// In the parent, fill bytes between 1000..2000. In the child, between 2000..3000
if fork_result.is_parent() {
write_range(ptr, 0xAA, 1000..2000);
} else {
write_range(ptr, 0xBB, 2000..3000);
}
barrier.wait();
// Verify the contents. (in both processes)
assert_range(ptr, 0xAA, 1000..2000);
assert_range(ptr, 0xBB, 2000..3000);
// Grow, from the child this time
let size = 10_000_000;
if !fork_result.is_parent() {
init_struct.set_size(size).unwrap();
}
barrier.wait();
// make some writes at the end
if fork_result.is_parent() {
write_range(ptr, 0xAA, (size - 10)..size);
} else {
write_range(ptr, 0xBB, (size - 20)..(size - 10));
}
barrier.wait();
// Verify the contents. (This runs in both processes)
assert_range(ptr, 0, (size - 1000)..(size - 20));
assert_range(ptr, 0xBB, (size - 20)..(size - 10));
assert_range(ptr, 0xAA, (size - 10)..size);
if let ForkResult::Parent { child } = fork_result {
nix::sys::wait::waitpid(child, None).unwrap();
}
}
}

View File

@@ -1,105 +0,0 @@
//! Simple utilities akin to what's in [`std::sync`] but designed to work with shared memory.
use std::mem::MaybeUninit;
use std::ptr::NonNull;
use nix::errno::Errno;
pub type RwLock<T> = lock_api::RwLock<PthreadRwLock, T>;
pub(crate) type RwLockReadGuard<'a, T> = lock_api::RwLockReadGuard<'a, PthreadRwLock, T>;
pub type RwLockWriteGuard<'a, T> = lock_api::RwLockWriteGuard<'a, PthreadRwLock, T>;
pub type ValueReadGuard<'a, T> = lock_api::MappedRwLockReadGuard<'a, PthreadRwLock, T>;
pub type ValueWriteGuard<'a, T> = lock_api::MappedRwLockWriteGuard<'a, PthreadRwLock, T>;
/// Shared memory read-write lock.
pub struct PthreadRwLock(Option<NonNull<libc::pthread_rwlock_t>>);
impl PthreadRwLock {
pub fn new(lock: *mut libc::pthread_rwlock_t) -> Self {
unsafe {
let mut attrs = MaybeUninit::uninit();
// Ignoring return value here - only possible error is OOM.
libc::pthread_rwlockattr_init(attrs.as_mut_ptr());
libc::pthread_rwlockattr_setpshared(
attrs.as_mut_ptr(),
libc::PTHREAD_PROCESS_SHARED
);
// TODO(quantumish): worth making this function return Result?
libc::pthread_rwlock_init(lock, attrs.as_mut_ptr());
// Safety: POSIX specifies that "any function affecting the attributes
// object (including destruction) shall not affect any previously
// initialized read-write locks".
libc::pthread_rwlockattr_destroy(attrs.as_mut_ptr());
Self(Some(NonNull::new_unchecked(lock)))
}
}
fn inner(&self) -> NonNull<libc::pthread_rwlock_t> {
match self.0 {
None => panic!("PthreadRwLock constructed badly - something likely used RawMutex::INIT"),
Some(x) => x,
}
}
}
unsafe impl lock_api::RawRwLock for PthreadRwLock {
type GuardMarker = lock_api::GuardSend;
const INIT: Self = Self(None);
fn lock_shared(&self) {
unsafe {
let res = libc::pthread_rwlock_rdlock(self.inner().as_ptr());
if res != 0 {
panic!("rdlock failed with {}", Errno::from_raw(res));
}
}
}
fn try_lock_shared(&self) -> bool {
unsafe {
let res = libc::pthread_rwlock_tryrdlock(self.inner().as_ptr());
match res {
0 => true,
libc::EAGAIN => false,
o => panic!("try_rdlock failed with {}", Errno::from_raw(res)),
}
}
}
fn lock_exclusive(&self) {
unsafe {
let res = libc::pthread_rwlock_wrlock(self.inner().as_ptr());
if res != 0 {
panic!("wrlock failed with {}", Errno::from_raw(res));
}
}
}
fn try_lock_exclusive(&self) -> bool {
unsafe {
let res = libc::pthread_rwlock_trywrlock(self.inner().as_ptr());
match res {
0 => true,
libc::EAGAIN => false,
o => panic!("try_wrlock failed with {}", Errno::from_raw(res)),
}
}
}
unsafe fn unlock_exclusive(&self) {
unsafe {
let res = libc::pthread_rwlock_unlock(self.inner().as_ptr());
if res != 0 {
panic!("unlock failed with {}", Errno::from_raw(res));
}
}
}
unsafe fn unlock_shared(&self) {
unsafe {
let res = libc::pthread_rwlock_unlock(self.inner().as_ptr());
if res != 0 {
panic!("unlock failed with {}", Errno::from_raw(res));
}
}
}
}

View File

@@ -1,14 +0,0 @@
[package]
name = "neonart"
version = "0.1.0"
edition.workspace = true
license.workspace = true
[dependencies]
crossbeam-utils.workspace = true
spin.workspace = true
tracing.workspace = true
[dev-dependencies]
rand = "0.9.1"
rand_distr = "0.5.1"

View File

@@ -1,594 +0,0 @@
mod lock_and_version;
pub(crate) mod node_ptr;
mod node_ref;
use std::vec::Vec;
use crate::algorithm::lock_and_version::ConcurrentUpdateError;
use crate::algorithm::node_ptr::MAX_PREFIX_LEN;
use crate::algorithm::node_ref::{NewNodeRef, NodeRef, ReadLockedNodeRef, WriteLockedNodeRef};
use crate::allocator::OutOfMemoryError;
use crate::TreeWriteGuard;
use crate::UpdateAction;
use crate::allocator::ArtAllocator;
use crate::epoch::EpochPin;
use crate::{Key, Value};
pub(crate) type RootPtr<V> = node_ptr::NodePtr<V>;
#[derive(Debug)]
pub enum ArtError {
ConcurrentUpdate, // need to retry
OutOfMemory,
}
impl From<ConcurrentUpdateError> for ArtError {
fn from(_: ConcurrentUpdateError) -> ArtError {
ArtError::ConcurrentUpdate
}
}
impl From<OutOfMemoryError> for ArtError {
fn from(_: OutOfMemoryError) -> ArtError {
ArtError::OutOfMemory
}
}
pub fn new_root<V: Value>(
allocator: &impl ArtAllocator<V>,
) -> Result<RootPtr<V>, OutOfMemoryError> {
node_ptr::new_root(allocator)
}
pub(crate) fn search<'e, K: Key, V: Value>(
key: &K,
root: RootPtr<V>,
epoch_pin: &'e EpochPin,
) -> Option<&'e V> {
loop {
let root_ref = NodeRef::from_root_ptr(root);
if let Ok(result) = lookup_recurse(key.as_bytes(), root_ref, None, epoch_pin) {
break result;
}
// retry
}
}
pub(crate) fn iter_next<'e, V: Value>(
key: &[u8],
root: RootPtr<V>,
epoch_pin: &'e EpochPin,
) -> Option<(Vec<u8>, &'e V)> {
loop {
let mut path = Vec::new();
let root_ref = NodeRef::from_root_ptr(root);
match next_recurse(key, &mut path, root_ref, epoch_pin) {
Ok(Some(v)) => {
assert_eq!(path.len(), key.len());
break Some((path, v));
}
Ok(None) => break None,
Err(ConcurrentUpdateError()) => {
// retry
continue;
}
}
}
}
pub(crate) fn update_fn<'e, 'g, K: Key, V: Value, A: ArtAllocator<V>, F>(
key: &K,
value_fn: F,
root: RootPtr<V>,
guard: &'g mut TreeWriteGuard<'e, K, V, A>,
) -> Result<(), OutOfMemoryError>
where
F: FnOnce(Option<&V>) -> UpdateAction<V>,
{
let value_fn_cell = std::cell::Cell::new(Some(value_fn));
loop {
let root_ref = NodeRef::from_root_ptr(root);
let this_value_fn = |arg: Option<&V>| value_fn_cell.take().unwrap()(arg);
let key_bytes = key.as_bytes();
match update_recurse(
key_bytes,
this_value_fn,
root_ref,
None,
None,
guard,
0,
key_bytes,
) {
Ok(()) => break Ok(()),
Err(ArtError::ConcurrentUpdate) => {
continue; // retry
}
Err(ArtError::OutOfMemory) => break Err(OutOfMemoryError()),
}
}
}
// Error means you must retry.
//
// This corresponds to the 'lookupOpt' function in the paper
fn lookup_recurse<'e, V: Value>(
key: &[u8],
node: NodeRef<'e, V>,
parent: Option<ReadLockedNodeRef<V>>,
epoch_pin: &'e EpochPin,
) -> Result<Option<&'e V>, ConcurrentUpdateError> {
let rnode = node.read_lock_or_restart()?;
if let Some(parent) = parent {
parent.read_unlock_or_restart()?;
}
// check if the prefix matches, may increment level
let prefix_len = if let Some(prefix_len) = rnode.prefix_matches(key) {
prefix_len
} else {
rnode.read_unlock_or_restart()?;
return Ok(None);
};
if rnode.is_leaf() {
assert_eq!(key.len(), prefix_len);
let vptr = rnode.get_leaf_value_ptr()?;
// safety: It's OK to return a ref of the pointer because we checked the version
// and the lifetime of 'epoch_pin' enforces that the reference is only accessible
// as long as the epoch is pinned.
let v = unsafe { vptr.as_ref().unwrap() };
return Ok(Some(v));
}
let key = &key[prefix_len..];
// find child (or leaf value)
let next_node = rnode.find_child_or_restart(key[0])?;
match next_node {
None => Ok(None), // key not found
Some(child) => lookup_recurse(&key[1..], child, Some(rnode), epoch_pin),
}
}
fn next_recurse<'e, V: Value>(
min_key: &[u8],
path: &mut Vec<u8>,
node: NodeRef<'e, V>,
epoch_pin: &'e EpochPin,
) -> Result<Option<&'e V>, ConcurrentUpdateError> {
let rnode = node.read_lock_or_restart()?;
let prefix = rnode.get_prefix();
if prefix.len() != 0 {
path.extend_from_slice(prefix);
}
use std::cmp::Ordering;
let comparison = path.as_slice().cmp(&min_key[0..path.len()]);
if comparison == Ordering::Less {
rnode.read_unlock_or_restart()?;
return Ok(None);
}
if rnode.is_leaf() {
assert_eq!(path.len(), min_key.len());
let vptr = rnode.get_leaf_value_ptr()?;
// safety: It's OK to return a ref of the pointer because we checked the version
// and the lifetime of 'epoch_pin' enforces that the reference is only accessible
// as long as the epoch is pinned.
let v = unsafe { vptr.as_ref().unwrap() };
return Ok(Some(v));
}
let mut min_key_byte = match comparison {
Ordering::Less => unreachable!(), // checked this above already
Ordering::Equal => min_key[path.len()],
Ordering::Greater => 0,
};
loop {
match rnode.find_next_child_or_restart(min_key_byte)? {
None => {
return Ok(None);
}
Some((key_byte, child_ref)) => {
let path_len = path.len();
path.push(key_byte);
let result = next_recurse(min_key, path, child_ref, epoch_pin)?;
if result.is_some() {
return Ok(result);
}
if key_byte == u8::MAX {
return Ok(None);
}
path.truncate(path_len);
min_key_byte = key_byte + 1;
}
}
}
}
// This corresponds to the 'insertOpt' function in the paper
pub(crate) fn update_recurse<'e, 'g, K: Key, V: Value, A: ArtAllocator<V>, F>(
key: &[u8],
value_fn: F,
node: NodeRef<'e, V>,
rparent: Option<(ReadLockedNodeRef<V>, u8)>,
rgrandparent: Option<(ReadLockedNodeRef<V>, u8)>,
guard: &'g mut TreeWriteGuard<'e, K, V, A>,
level: usize,
orig_key: &[u8],
) -> Result<(), ArtError>
where
F: FnOnce(Option<&V>) -> UpdateAction<V>,
{
let rnode = node.read_lock_or_restart()?;
let prefix_match_len = rnode.prefix_matches(key);
if prefix_match_len.is_none() {
let (rparent, parent_key) = rparent.expect("direct children of the root have no prefix");
let mut wparent = rparent.upgrade_to_write_lock_or_restart()?;
let mut wnode = rnode.upgrade_to_write_lock_or_restart()?;
match value_fn(None) {
UpdateAction::Nothing => {}
UpdateAction::Insert(new_value) => {
insert_split_prefix(key, new_value, &mut wnode, &mut wparent, parent_key, guard)?;
}
UpdateAction::Remove => {
panic!("unexpected Remove action on insertion");
}
}
wnode.write_unlock();
wparent.write_unlock();
return Ok(());
}
let prefix_match_len = prefix_match_len.unwrap();
let key = &key[prefix_match_len as usize..];
let level = level + prefix_match_len as usize;
if rnode.is_leaf() {
assert_eq!(key.len(), 0);
let (rparent, parent_key) = rparent.expect("root cannot be leaf");
let mut wparent = rparent.upgrade_to_write_lock_or_restart()?;
let mut wnode = rnode.upgrade_to_write_lock_or_restart()?;
// safety: Now that we have acquired the write lock, we have exclusive access to the
// value. XXX: There might be concurrent reads though?
let value_mut = wnode.get_leaf_value_mut();
match value_fn(Some(value_mut)) {
UpdateAction::Nothing => {
wparent.write_unlock();
wnode.write_unlock();
}
UpdateAction::Insert(_) => panic!("cannot insert over existing value"),
UpdateAction::Remove => {
guard.remember_obsolete_node(wnode.as_ptr());
wparent.delete_child(parent_key);
wnode.write_unlock_obsolete();
if let Some(rgrandparent) = rgrandparent {
// FIXME: Ignore concurrency error. It doesn't lead to
// corruption, but it means we might leak something. Until
// another update cleans it up.
let _ = cleanup_parent(wparent, rgrandparent, guard);
}
}
}
return Ok(());
}
let next_node = rnode.find_child_or_restart(key[0])?;
if next_node.is_none() {
if rnode.is_full() {
let (rparent, parent_key) = rparent.expect("root node cannot become full");
let mut wparent = rparent.upgrade_to_write_lock_or_restart()?;
let wnode = rnode.upgrade_to_write_lock_or_restart()?;
match value_fn(None) {
UpdateAction::Nothing => {
wnode.write_unlock();
wparent.write_unlock();
}
UpdateAction::Insert(new_value) => {
insert_and_grow(key, new_value, wnode, &mut wparent, parent_key, guard)?;
wparent.write_unlock();
}
UpdateAction::Remove => {
panic!("unexpected Remove action on insertion");
}
};
} else {
let mut wnode = rnode.upgrade_to_write_lock_or_restart()?;
if let Some((rparent, _)) = rparent {
rparent.read_unlock_or_restart()?;
}
match value_fn(None) {
UpdateAction::Nothing => {}
UpdateAction::Insert(new_value) => {
insert_to_node(&mut wnode, key, new_value, guard)?;
}
UpdateAction::Remove => {
panic!("unexpected Remove action on insertion");
}
};
wnode.write_unlock();
}
return Ok(());
} else {
let next_child = next_node.unwrap(); // checked above it's not None
if let Some((ref rparent, _)) = rparent {
rparent.check_or_restart()?;
}
// recurse to next level
update_recurse(
&key[1..],
value_fn,
next_child,
Some((rnode, key[0])),
rparent,
guard,
level + 1,
orig_key,
)
}
}
#[derive(Clone)]
enum PathElement {
Prefix(Vec<u8>),
KeyByte(u8),
}
impl std::fmt::Debug for PathElement {
fn fmt(&self, fmt: &mut std::fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
match self {
PathElement::Prefix(prefix) => write!(fmt, "{:?}", prefix),
PathElement::KeyByte(key_byte) => write!(fmt, "{}", key_byte),
}
}
}
pub(crate) fn dump_tree<'e, V: Value + std::fmt::Debug>(
root: RootPtr<V>,
epoch_pin: &'e EpochPin,
dst: &mut dyn std::io::Write,
) {
let root_ref = NodeRef::from_root_ptr(root);
let _ = dump_recurse(&[], root_ref, &epoch_pin, 0, dst);
}
// TODO: return an Err if writeln!() returns error, instead of unwrapping
fn dump_recurse<'e, V: Value + std::fmt::Debug>(
path: &[PathElement],
node: NodeRef<'e, V>,
epoch_pin: &'e EpochPin,
level: usize,
dst: &mut dyn std::io::Write,
) -> Result<(), ConcurrentUpdateError> {
let indent = str::repeat(" ", level);
let rnode = node.read_lock_or_restart()?;
let mut path = Vec::from(path);
let prefix = rnode.get_prefix();
if prefix.len() != 0 {
path.push(PathElement::Prefix(Vec::from(prefix)));
}
if rnode.is_leaf() {
let vptr = rnode.get_leaf_value_ptr()?;
// safety: It's OK to return a ref of the pointer because we checked the version
// and the lifetime of 'epoch_pin' enforces that the reference is only accessible
// as long as the epoch is pinned.
let val = unsafe { vptr.as_ref().unwrap() };
writeln!(dst, "{} {:?}: {:?}", indent, path, val).unwrap();
return Ok(());
}
for key_byte in 0..=u8::MAX {
match rnode.find_child_or_restart(key_byte)? {
None => continue,
Some(child_ref) => {
let rchild = child_ref.read_lock_or_restart()?;
writeln!(
dst,
"{} {:?}, {}: prefix {:?}",
indent,
&path,
key_byte,
rchild.get_prefix()
)
.unwrap();
let mut child_path = path.clone();
child_path.push(PathElement::KeyByte(key_byte));
dump_recurse(&child_path, child_ref, epoch_pin, level + 1, dst)?;
}
}
}
Ok(())
}
///```text
/// [fooba]r -> value
///
/// [foo]b -> [a]r -> value
/// e -> [ls]e -> value
///```
fn insert_split_prefix<'e, K: Key, V: Value, A: ArtAllocator<V>>(
key: &[u8],
value: V,
node: &mut WriteLockedNodeRef<V>,
parent: &mut WriteLockedNodeRef<V>,
parent_key: u8,
guard: &'e TreeWriteGuard<K, V, A>,
) -> Result<(), OutOfMemoryError> {
let old_node = node;
let old_prefix = old_node.get_prefix();
let common_prefix_len = common_prefix(key, old_prefix);
// Allocate a node for the new value.
let new_value_node = allocate_node_for_value(
&key[common_prefix_len + 1..],
value,
guard.tree_writer.allocator,
)?;
// Allocate a new internal node with the common prefix
// FIXME: deallocate 'new_value_node' on OOM
let mut prefix_node =
node_ref::new_internal(&key[..common_prefix_len], guard.tree_writer.allocator)?;
// Add the old node and the new nodes to the new internal node
prefix_node.insert_old_child(old_prefix[common_prefix_len], old_node);
prefix_node.insert_new_child(key[common_prefix_len], new_value_node);
// Modify the prefix of the old child in place
old_node.truncate_prefix(old_prefix.len() - common_prefix_len - 1);
// replace the pointer in the parent
parent.replace_child(parent_key, prefix_node.into_ptr());
Ok(())
}
fn insert_to_node<'e, K: Key, V: Value, A: ArtAllocator<V>>(
wnode: &mut WriteLockedNodeRef<V>,
key: &[u8],
value: V,
guard: &'e TreeWriteGuard<K, V, A>,
) -> Result<(), OutOfMemoryError> {
let value_child = allocate_node_for_value(&key[1..], value, guard.tree_writer.allocator)?;
wnode.insert_child(key[0], value_child.into_ptr());
Ok(())
}
// On entry: 'parent' and 'node' are locked
fn insert_and_grow<'e, 'g, K: Key, V: Value, A: ArtAllocator<V>>(
key: &[u8],
value: V,
wnode: WriteLockedNodeRef<V>,
parent: &mut WriteLockedNodeRef<V>,
parent_key_byte: u8,
guard: &'g mut TreeWriteGuard<'e, K, V, A>,
) -> Result<(), ArtError> {
let mut bigger_node = wnode.grow(guard.tree_writer.allocator)?;
// FIXME: deallocate 'bigger_node' on OOM
let value_child = allocate_node_for_value(&key[1..], value, guard.tree_writer.allocator)?;
bigger_node.insert_new_child(key[0], value_child);
// Replace the pointer in the parent
parent.replace_child(parent_key_byte, bigger_node.into_ptr());
guard.remember_obsolete_node(wnode.as_ptr());
wnode.write_unlock_obsolete();
Ok(())
}
fn cleanup_parent<'e, 'g, K: Key, V: Value, A: ArtAllocator<V>>(
wparent: WriteLockedNodeRef<V>,
rgrandparent: (ReadLockedNodeRef<V>, u8),
guard: &'g mut TreeWriteGuard<'e, K, V, A>,
) -> Result<(), ArtError> {
let (rgrandparent, grandparent_key_byte) = rgrandparent;
// If the parent becomes completely empty after the deletion, remove the parent from the
// grandparent. (This case is possible because we reserve only 8 bytes for the prefix.)
// TODO: not implemented.
// If the parent has only one child, replace the parent with the remaining child. (This is not
// possible if the child's prefix field cannot absorb the parent's)
if wparent.num_children() == 1 {
// Try to lock the remaining child. This can fail if the child is updated
// concurrently.
let (key_byte, remaining_child) = wparent.find_remaining_child();
let mut wremaining_child = remaining_child.write_lock_or_restart()?;
if 1 + wremaining_child.get_prefix().len() + wparent.get_prefix().len() <= MAX_PREFIX_LEN {
let mut wgrandparent = rgrandparent.upgrade_to_write_lock_or_restart()?;
// Ok, we have locked the leaf, the parent, the grandparent, and the parent's only
// remaining leaf. Proceed with the updates.
// Update the prefix on the remaining leaf
wremaining_child.prepend_prefix(wparent.get_prefix(), key_byte);
// Replace the pointer in the grandparent to point directly to the remaining leaf
wgrandparent.replace_child(grandparent_key_byte, wremaining_child.as_ptr());
// Mark the parent as deleted.
guard.remember_obsolete_node(wparent.as_ptr());
wparent.write_unlock_obsolete();
return Ok(());
}
}
// If the parent's children would fit on a smaller node type after the deletion, replace it with
// a smaller node.
if wparent.can_shrink() {
let mut wgrandparent = rgrandparent.upgrade_to_write_lock_or_restart()?;
let smaller_node = wparent.shrink(guard.tree_writer.allocator)?;
// Replace the pointer in the grandparent
wgrandparent.replace_child(grandparent_key_byte, smaller_node.into_ptr());
guard.remember_obsolete_node(wparent.as_ptr());
wparent.write_unlock_obsolete();
return Ok(());
}
// nothing to do
wparent.write_unlock();
Ok(())
}
// Allocate a new leaf node to hold 'value'. If the key is long, we
// may need to allocate new internal nodes to hold it too
fn allocate_node_for_value<'a, V: Value, A: ArtAllocator<V>>(
key: &[u8],
value: V,
allocator: &'a A,
) -> Result<NewNodeRef<'a, V, A>, OutOfMemoryError> {
let mut prefix_off = key.len().saturating_sub(MAX_PREFIX_LEN);
let leaf_node = node_ref::new_leaf(&key[prefix_off..key.len()], value, allocator)?;
let mut node = leaf_node;
while prefix_off > 0 {
// Need another internal node
let remain_prefix = &key[0..prefix_off];
prefix_off = remain_prefix.len().saturating_sub(MAX_PREFIX_LEN + 1);
let mut internal_node = node_ref::new_internal(
&remain_prefix[prefix_off..remain_prefix.len() - 1],
allocator,
)?;
internal_node.insert_new_child(*remain_prefix.last().unwrap(), node);
node = internal_node;
}
Ok(node)
}
fn common_prefix(a: &[u8], b: &[u8]) -> usize {
for i in 0..MAX_PREFIX_LEN {
if a[i] != b[i] {
return i;
}
}
panic!("prefixes are equal");
}

View File

@@ -1,117 +0,0 @@
//! Each node in the tree has contains one atomic word that stores three things:
//!
//! Bit 0: set if the node is "obsolete". An obsolete node has been removed from the tree,
//! but might still be accessed by concurrent readers until the epoch expires.
//! Bit 1: set if the node is currently write-locked. Used as a spinlock.
//! Bits 2-63: Version number, incremented every time the node is modified.
//!
//! AtomicLockAndVersion represents that.
use std::sync::atomic::{AtomicU64, Ordering};
pub(crate) struct ConcurrentUpdateError();
pub(crate) struct AtomicLockAndVersion {
inner: AtomicU64,
}
impl AtomicLockAndVersion {
pub(crate) fn new() -> AtomicLockAndVersion {
AtomicLockAndVersion {
inner: AtomicU64::new(0),
}
}
}
impl AtomicLockAndVersion {
pub(crate) fn read_lock_or_restart(&self) -> Result<u64, ConcurrentUpdateError> {
let version = self.await_node_unlocked();
if is_obsolete(version) {
return Err(ConcurrentUpdateError());
}
Ok(version)
}
pub(crate) fn check_or_restart(&self, version: u64) -> Result<(), ConcurrentUpdateError> {
self.read_unlock_or_restart(version)
}
pub(crate) fn read_unlock_or_restart(&self, version: u64) -> Result<(), ConcurrentUpdateError> {
if self.inner.load(Ordering::Acquire) != version {
return Err(ConcurrentUpdateError());
}
Ok(())
}
pub(crate) fn upgrade_to_write_lock_or_restart(
&self,
version: u64,
) -> Result<(), ConcurrentUpdateError> {
if self
.inner
.compare_exchange(
version,
set_locked_bit(version),
Ordering::Acquire,
Ordering::Relaxed,
)
.is_err()
{
return Err(ConcurrentUpdateError());
}
Ok(())
}
pub(crate) fn write_lock_or_restart(&self) -> Result<(), ConcurrentUpdateError> {
let old = self.inner.load(Ordering::Relaxed);
if is_obsolete(old) || is_locked(old) {
return Err(ConcurrentUpdateError());
}
if self
.inner
.compare_exchange(
old,
set_locked_bit(old),
Ordering::Acquire,
Ordering::Relaxed,
)
.is_err()
{
return Err(ConcurrentUpdateError());
}
Ok(())
}
pub(crate) fn write_unlock(&self) {
// reset locked bit and overflow into version
self.inner.fetch_add(2, Ordering::Release);
}
pub(crate) fn write_unlock_obsolete(&self) {
// set obsolete, reset locked, overflow into version
self.inner.fetch_add(3, Ordering::Release);
}
// Helper functions
fn await_node_unlocked(&self) -> u64 {
let mut version = self.inner.load(Ordering::Acquire);
while is_locked(version) {
// spinlock
std::thread::yield_now();
version = self.inner.load(Ordering::Acquire)
}
version
}
}
fn set_locked_bit(version: u64) -> u64 {
return version + 2;
}
fn is_obsolete(version: u64) -> bool {
return (version & 1) == 1;
}
fn is_locked(version: u64) -> bool {
return (version & 2) == 2;
}

File diff suppressed because it is too large Load Diff

View File

@@ -1,349 +0,0 @@
use std::fmt::Debug;
use std::marker::PhantomData;
use super::node_ptr;
use super::node_ptr::NodePtr;
use crate::EpochPin;
use crate::Value;
use crate::algorithm::lock_and_version::AtomicLockAndVersion;
use crate::algorithm::lock_and_version::ConcurrentUpdateError;
use crate::allocator::ArtAllocator;
use crate::allocator::OutOfMemoryError;
pub struct NodeRef<'e, V> {
ptr: NodePtr<V>,
phantom: PhantomData<&'e EpochPin<'e>>,
}
impl<'e, V> Debug for NodeRef<'e, V> {
fn fmt(&self, fmt: &mut std::fmt::Formatter<'_>) -> Result<(), std::fmt::Error> {
write!(fmt, "{:?}", self.ptr)
}
}
impl<'e, V: Value> NodeRef<'e, V> {
pub(crate) fn from_root_ptr(root_ptr: NodePtr<V>) -> NodeRef<'e, V> {
NodeRef {
ptr: root_ptr,
phantom: PhantomData,
}
}
pub(crate) fn read_lock_or_restart(
&self,
) -> Result<ReadLockedNodeRef<'e, V>, ConcurrentUpdateError> {
let version = self.lockword().read_lock_or_restart()?;
Ok(ReadLockedNodeRef {
ptr: self.ptr,
version,
phantom: self.phantom,
})
}
pub(crate) fn write_lock_or_restart(
&self,
) -> Result<WriteLockedNodeRef<'e, V>, ConcurrentUpdateError> {
self.lockword().write_lock_or_restart()?;
Ok(WriteLockedNodeRef {
ptr: self.ptr,
phantom: self.phantom,
})
}
fn lockword(&self) -> &AtomicLockAndVersion {
self.ptr.lockword()
}
}
/// A reference to a node that has been optimistically read-locked. The functions re-check
/// the version after each read.
pub struct ReadLockedNodeRef<'e, V> {
ptr: NodePtr<V>,
version: u64,
phantom: PhantomData<&'e EpochPin<'e>>,
}
impl<'e, V: Value> ReadLockedNodeRef<'e, V> {
pub(crate) fn is_leaf(&self) -> bool {
self.ptr.is_leaf()
}
pub(crate) fn is_full(&self) -> bool {
self.ptr.is_full()
}
pub(crate) fn get_prefix(&self) -> &[u8] {
self.ptr.get_prefix()
}
/// Note: because we're only holding a read lock, the prefix can change concurrently.
/// You must be prepared to restart, if read_unlock() returns error later.
///
/// Returns the length of the prefix, or None if it's not a match
pub(crate) fn prefix_matches(&self, key: &[u8]) -> Option<usize> {
self.ptr.prefix_matches(key)
}
pub(crate) fn find_child_or_restart(
&self,
key_byte: u8,
) -> Result<Option<NodeRef<'e, V>>, ConcurrentUpdateError> {
let child_or_value = self.ptr.find_child(key_byte);
self.ptr.lockword().check_or_restart(self.version)?;
match child_or_value {
None => Ok(None),
Some(child_ptr) => Ok(Some(NodeRef {
ptr: child_ptr,
phantom: self.phantom,
})),
}
}
pub(crate) fn find_next_child_or_restart(
&self,
min_key_byte: u8,
) -> Result<Option<(u8, NodeRef<'e, V>)>, ConcurrentUpdateError> {
let child_or_value = self.ptr.find_next_child(min_key_byte);
self.ptr.lockword().check_or_restart(self.version)?;
match child_or_value {
None => Ok(None),
Some((k, child_ptr)) => Ok(Some((
k,
NodeRef {
ptr: child_ptr,
phantom: self.phantom,
},
))),
}
}
pub(crate) fn get_leaf_value_ptr(&self) -> Result<*const V, ConcurrentUpdateError> {
let result = self.ptr.get_leaf_value();
self.ptr.lockword().check_or_restart(self.version)?;
// Extend the lifetime.
let result = std::ptr::from_ref(result);
Ok(result)
}
pub(crate) fn upgrade_to_write_lock_or_restart(
self,
) -> Result<WriteLockedNodeRef<'e, V>, ConcurrentUpdateError> {
self.ptr
.lockword()
.upgrade_to_write_lock_or_restart(self.version)?;
Ok(WriteLockedNodeRef {
ptr: self.ptr,
phantom: self.phantom,
})
}
pub(crate) fn read_unlock_or_restart(self) -> Result<(), ConcurrentUpdateError> {
self.ptr.lockword().check_or_restart(self.version)?;
Ok(())
}
pub(crate) fn check_or_restart(&self) -> Result<(), ConcurrentUpdateError> {
self.ptr.lockword().check_or_restart(self.version)?;
Ok(())
}
}
/// A reference to a node that has been optimistically read-locked. The functions re-check
/// the version after each read.
pub struct WriteLockedNodeRef<'e, V> {
ptr: NodePtr<V>,
phantom: PhantomData<&'e EpochPin<'e>>,
}
impl<'e, V: Value> WriteLockedNodeRef<'e, V> {
pub(crate) fn can_shrink(&self) -> bool {
self.ptr.can_shrink()
}
pub(crate) fn num_children(&self) -> usize {
self.ptr.num_children()
}
pub(crate) fn write_unlock(mut self) {
self.ptr.lockword().write_unlock();
self.ptr = NodePtr::null();
}
pub(crate) fn write_unlock_obsolete(mut self) {
self.ptr.lockword().write_unlock_obsolete();
self.ptr = NodePtr::null();
}
pub(crate) fn get_prefix(&self) -> &[u8] {
self.ptr.get_prefix()
}
pub(crate) fn truncate_prefix(&mut self, new_prefix_len: usize) {
self.ptr.truncate_prefix(new_prefix_len)
}
pub(crate) fn prepend_prefix(&mut self, prefix: &[u8], prefix_byte: u8) {
self.ptr.prepend_prefix(prefix, prefix_byte)
}
pub(crate) fn insert_child(&mut self, key_byte: u8, child: NodePtr<V>) {
self.ptr.insert_child(key_byte, child)
}
pub(crate) fn get_leaf_value_mut(&mut self) -> &mut V {
self.ptr.get_leaf_value_mut()
}
pub(crate) fn grow<'a, A>(
&self,
allocator: &'a A,
) -> Result<NewNodeRef<'a, V, A>, OutOfMemoryError>
where
A: ArtAllocator<V>,
{
let new_node = self.ptr.grow(allocator)?;
Ok(NewNodeRef {
ptr: new_node,
allocator,
extra_nodes: Vec::new(),
})
}
pub(crate) fn shrink<'a, A>(
&self,
allocator: &'a A,
) -> Result<NewNodeRef<'a, V, A>, OutOfMemoryError>
where
A: ArtAllocator<V>,
{
let new_node = self.ptr.shrink(allocator)?;
Ok(NewNodeRef {
ptr: new_node,
allocator,
extra_nodes: Vec::new(),
})
}
pub(crate) fn as_ptr(&self) -> NodePtr<V> {
self.ptr
}
pub(crate) fn replace_child(&mut self, key_byte: u8, replacement: NodePtr<V>) {
self.ptr.replace_child(key_byte, replacement);
}
pub(crate) fn delete_child(&mut self, key_byte: u8) {
self.ptr.delete_child(key_byte);
}
pub(crate) fn find_remaining_child(&self) -> (u8, NodeRef<'e, V>) {
assert_eq!(self.num_children(), 1);
let child_or_value = self.ptr.find_next_child(0);
match child_or_value {
None => panic!("could not find only child in node"),
Some((k, child_ptr)) => (
k,
NodeRef {
ptr: child_ptr,
phantom: self.phantom,
},
),
}
}
}
impl<'e, V> Drop for WriteLockedNodeRef<'e, V> {
fn drop(&mut self) {
if !self.ptr.is_null() {
self.ptr.lockword().write_unlock();
}
}
}
pub(crate) struct NewNodeRef<'a, V, A>
where
V: Value,
A: ArtAllocator<V>,
{
ptr: NodePtr<V>,
allocator: &'a A,
extra_nodes: Vec<NodePtr<V>>,
}
impl<'a, V, A> NewNodeRef<'a, V, A>
where
V: Value,
A: ArtAllocator<V>,
{
pub(crate) fn insert_old_child(&mut self, key_byte: u8, child: &WriteLockedNodeRef<V>) {
self.ptr.insert_child(key_byte, child.as_ptr())
}
pub(crate) fn into_ptr(mut self) -> NodePtr<V> {
let ptr = self.ptr;
self.ptr = NodePtr::null();
ptr
}
pub(crate) fn insert_new_child(&mut self, key_byte: u8, child: NewNodeRef<'a, V, A>) {
let child_ptr = child.into_ptr();
self.ptr.insert_child(key_byte, child_ptr);
self.extra_nodes.push(child_ptr);
}
}
impl<'a, V, A> Drop for NewNodeRef<'a, V, A>
where
V: Value,
A: ArtAllocator<V>,
{
/// This drop implementation deallocates the newly allocated node, if into_ptr() was not called.
fn drop(&mut self) {
if !self.ptr.is_null() {
self.ptr.deallocate(self.allocator);
for p in self.extra_nodes.iter() {
p.deallocate(self.allocator);
}
}
}
}
pub(crate) fn new_internal<'a, V, A>(
prefix: &[u8],
allocator: &'a A,
) -> Result<NewNodeRef<'a, V, A>, OutOfMemoryError>
where
V: Value,
A: ArtAllocator<V>,
{
Ok(NewNodeRef {
ptr: node_ptr::new_internal(prefix, allocator)?,
allocator,
extra_nodes: Vec::new(),
})
}
pub(crate) fn new_leaf<'a, V, A>(
prefix: &[u8],
value: V,
allocator: &'a A,
) -> Result<NewNodeRef<'a, V, A>, OutOfMemoryError>
where
V: Value,
A: ArtAllocator<V>,
{
Ok(NewNodeRef {
ptr: node_ptr::new_leaf(prefix, value, allocator)?,
allocator,
extra_nodes: Vec::new(),
})
}

View File

@@ -1,158 +0,0 @@
pub mod block;
mod multislab;
mod slab;
pub mod r#static;
use std::alloc::Layout;
use std::marker::PhantomData;
use std::mem::MaybeUninit;
use std::sync::atomic::Ordering;
use crate::allocator::multislab::MultiSlabAllocator;
use crate::allocator::r#static::alloc_from_slice;
use spin;
use crate::Tree;
pub use crate::algorithm::node_ptr::{
NodeInternal4, NodeInternal16, NodeInternal48, NodeInternal256, NodeLeaf,
};
#[derive(Debug)]
pub struct OutOfMemoryError();
pub trait ArtAllocator<V: crate::Value> {
fn alloc_tree(&self) -> *mut Tree<V>;
fn alloc_node_internal4(&self) -> *mut NodeInternal4<V>;
fn alloc_node_internal16(&self) -> *mut NodeInternal16<V>;
fn alloc_node_internal48(&self) -> *mut NodeInternal48<V>;
fn alloc_node_internal256(&self) -> *mut NodeInternal256<V>;
fn alloc_node_leaf(&self) -> *mut NodeLeaf<V>;
fn dealloc_node_internal4(&self, ptr: *mut NodeInternal4<V>);
fn dealloc_node_internal16(&self, ptr: *mut NodeInternal16<V>);
fn dealloc_node_internal48(&self, ptr: *mut NodeInternal48<V>);
fn dealloc_node_internal256(&self, ptr: *mut NodeInternal256<V>);
fn dealloc_node_leaf(&self, ptr: *mut NodeLeaf<V>);
}
pub struct ArtMultiSlabAllocator<'t, V>
where
V: crate::Value,
{
tree_area: spin::Mutex<Option<&'t mut MaybeUninit<Tree<V>>>>,
pub(crate) inner: MultiSlabAllocator<'t, 5>,
phantom_val: PhantomData<V>,
}
impl<'t, V: crate::Value> ArtMultiSlabAllocator<'t, V> {
const LAYOUTS: [Layout; 5] = [
Layout::new::<NodeInternal4<V>>(),
Layout::new::<NodeInternal16<V>>(),
Layout::new::<NodeInternal48<V>>(),
Layout::new::<NodeInternal256<V>>(),
Layout::new::<NodeLeaf<V>>(),
];
pub fn new(area: &'t mut [MaybeUninit<u8>]) -> &'t mut ArtMultiSlabAllocator<'t, V> {
let (allocator_area, remain) = alloc_from_slice::<ArtMultiSlabAllocator<V>>(area);
let (tree_area, remain) = alloc_from_slice::<Tree<V>>(remain);
let allocator = allocator_area.write(ArtMultiSlabAllocator {
tree_area: spin::Mutex::new(Some(tree_area)),
inner: MultiSlabAllocator::new(remain, &Self::LAYOUTS),
phantom_val: PhantomData,
});
allocator
}
}
impl<'t, V: crate::Value> ArtAllocator<V> for ArtMultiSlabAllocator<'t, V> {
fn alloc_tree(&self) -> *mut Tree<V> {
let mut t = self.tree_area.lock();
if let Some(tree_area) = t.take() {
return tree_area.as_mut_ptr().cast();
}
panic!("cannot allocate more than one tree");
}
fn alloc_node_internal4(&self) -> *mut NodeInternal4<V> {
self.inner.alloc_slab(0).cast()
}
fn alloc_node_internal16(&self) -> *mut NodeInternal16<V> {
self.inner.alloc_slab(1).cast()
}
fn alloc_node_internal48(&self) -> *mut NodeInternal48<V> {
self.inner.alloc_slab(2).cast()
}
fn alloc_node_internal256(&self) -> *mut NodeInternal256<V> {
self.inner.alloc_slab(3).cast()
}
fn alloc_node_leaf(&self) -> *mut NodeLeaf<V> {
self.inner.alloc_slab(4).cast()
}
fn dealloc_node_internal4(&self, ptr: *mut NodeInternal4<V>) {
self.inner.dealloc_slab(0, ptr.cast())
}
fn dealloc_node_internal16(&self, ptr: *mut NodeInternal16<V>) {
self.inner.dealloc_slab(1, ptr.cast())
}
fn dealloc_node_internal48(&self, ptr: *mut NodeInternal48<V>) {
self.inner.dealloc_slab(2, ptr.cast())
}
fn dealloc_node_internal256(&self, ptr: *mut NodeInternal256<V>) {
self.inner.dealloc_slab(3, ptr.cast())
}
fn dealloc_node_leaf(&self, ptr: *mut NodeLeaf<V>) {
self.inner.dealloc_slab(4, ptr.cast())
}
}
impl<'t, V: crate::Value> ArtMultiSlabAllocator<'t, V> {
pub(crate) fn get_statistics(&self) -> ArtMultiSlabStats {
ArtMultiSlabStats {
num_internal4: self.inner.slab_descs[0]
.num_allocated
.load(Ordering::Relaxed),
num_internal16: self.inner.slab_descs[1]
.num_allocated
.load(Ordering::Relaxed),
num_internal48: self.inner.slab_descs[2]
.num_allocated
.load(Ordering::Relaxed),
num_internal256: self.inner.slab_descs[3]
.num_allocated
.load(Ordering::Relaxed),
num_leaf: self.inner.slab_descs[4]
.num_allocated
.load(Ordering::Relaxed),
num_blocks_internal4: self.inner.slab_descs[0].num_blocks.load(Ordering::Relaxed),
num_blocks_internal16: self.inner.slab_descs[1].num_blocks.load(Ordering::Relaxed),
num_blocks_internal48: self.inner.slab_descs[2].num_blocks.load(Ordering::Relaxed),
num_blocks_internal256: self.inner.slab_descs[3].num_blocks.load(Ordering::Relaxed),
num_blocks_leaf: self.inner.slab_descs[4].num_blocks.load(Ordering::Relaxed),
}
}
}
#[derive(Clone, Debug)]
pub struct ArtMultiSlabStats {
pub num_internal4: u64,
pub num_internal16: u64,
pub num_internal48: u64,
pub num_internal256: u64,
pub num_leaf: u64,
pub num_blocks_internal4: u64,
pub num_blocks_internal16: u64,
pub num_blocks_internal48: u64,
pub num_blocks_internal256: u64,
pub num_blocks_leaf: u64,
}

View File

@@ -1,191 +0,0 @@
//! Simple allocator of fixed-size blocks
use std::mem::MaybeUninit;
use std::sync::atomic::{AtomicU64, Ordering};
use spin;
pub const BLOCK_SIZE: usize = 16 * 1024;
const INVALID_BLOCK: u64 = u64::MAX;
pub(crate) struct BlockAllocator<'t> {
blocks_ptr: &'t [MaybeUninit<u8>],
num_blocks: u64,
num_initialized: AtomicU64,
freelist_head: spin::Mutex<u64>,
}
struct FreeListBlock {
inner: spin::Mutex<FreeListBlockInner>,
}
struct FreeListBlockInner {
next: u64,
num_free_blocks: u64,
free_blocks: [u64; 100], // FIXME: fill the rest of the block
}
impl<'t> BlockAllocator<'t> {
pub(crate) fn new(area: &'t mut [MaybeUninit<u8>]) -> Self {
// Use all the space for the blocks
let padding = area.as_ptr().align_offset(BLOCK_SIZE);
let remain = &mut area[padding..];
let num_blocks = (remain.len() / BLOCK_SIZE) as u64;
BlockAllocator {
blocks_ptr: remain,
num_blocks,
num_initialized: AtomicU64::new(0),
freelist_head: spin::Mutex::new(INVALID_BLOCK),
}
}
/// safety: you must hold a lock on the pointer to this block, otherwise it might get
/// reused for another kind of block
fn read_freelist_block(&self, blkno: u64) -> &FreeListBlock {
let ptr: *const FreeListBlock = self.get_block_ptr(blkno).cast();
unsafe { ptr.as_ref().unwrap() }
}
fn get_block_ptr(&self, blkno: u64) -> *mut u8 {
assert!(blkno < self.num_blocks);
unsafe {
self.blocks_ptr
.as_ptr()
.byte_offset(blkno as isize * BLOCK_SIZE as isize)
}
.cast_mut()
.cast()
}
#[allow(clippy::mut_from_ref)]
pub(crate) fn alloc_block(&self) -> &mut [MaybeUninit<u8>] {
// FIXME: handle OOM
let blkno = self.alloc_block_internal();
if blkno == INVALID_BLOCK {
panic!("out of memory");
}
let ptr: *mut MaybeUninit<u8> = self.get_block_ptr(blkno).cast();
unsafe { std::slice::from_raw_parts_mut(ptr, BLOCK_SIZE) }
}
fn alloc_block_internal(&self) -> u64 {
// check the free list.
{
let mut freelist_head = self.freelist_head.lock();
if *freelist_head != INVALID_BLOCK {
let freelist_block = self.read_freelist_block(*freelist_head);
// acquire lock on the freelist block before releasing the lock on the parent (i.e. lock coupling)
let mut g = freelist_block.inner.lock();
if g.num_free_blocks > 0 {
g.num_free_blocks -= 1;
let result = g.free_blocks[g.num_free_blocks as usize];
return result;
} else {
// consume the freelist block itself
let result = *freelist_head;
*freelist_head = g.next;
// This freelist block is now unlinked and can be repurposed
drop(g);
return result;
}
}
}
// If there are some blocks left that we've never used, pick next such block
let mut next_uninitialized = self.num_initialized.load(Ordering::Relaxed);
while next_uninitialized < self.num_blocks {
match self.num_initialized.compare_exchange(
next_uninitialized,
next_uninitialized + 1,
Ordering::Relaxed,
Ordering::Relaxed,
) {
Ok(_) => {
return next_uninitialized;
}
Err(old) => {
next_uninitialized = old;
continue;
}
}
}
// out of blocks
return INVALID_BLOCK;
}
// TODO: this is currently unused. The slab allocator never releases blocks
#[allow(dead_code)]
pub(crate) fn release_block(&self, block_ptr: *mut u8) {
let blockno = unsafe { block_ptr.byte_offset_from(self.blocks_ptr) / BLOCK_SIZE as isize };
self.release_block_internal(blockno as u64);
}
fn release_block_internal(&self, blockno: u64) {
let mut freelist_head = self.freelist_head.lock();
if *freelist_head != INVALID_BLOCK {
let freelist_block = self.read_freelist_block(*freelist_head);
// acquire lock on the freelist block before releasing the lock on the parent (i.e. lock coupling)
let mut g = freelist_block.inner.lock();
let num_free_blocks = g.num_free_blocks;
if num_free_blocks < g.free_blocks.len() as u64 {
g.free_blocks[num_free_blocks as usize] = blockno;
g.num_free_blocks += 1;
return;
}
}
// Convert the block into a new freelist block
let block_ptr: *mut FreeListBlock = self.get_block_ptr(blockno).cast();
let init = FreeListBlock {
inner: spin::Mutex::new(FreeListBlockInner {
next: *freelist_head,
num_free_blocks: 0,
free_blocks: [INVALID_BLOCK; 100],
}),
};
unsafe { (*block_ptr) = init };
*freelist_head = blockno;
}
// for debugging
pub(crate) fn get_statistics(&self) -> BlockAllocatorStats {
let mut num_free_blocks = 0;
let mut _prev_lock = None;
let head_lock = self.freelist_head.lock();
let mut next_blk = *head_lock;
let mut _head_lock = Some(head_lock);
while next_blk != INVALID_BLOCK {
let freelist_block = self.read_freelist_block(next_blk);
let lock = freelist_block.inner.lock();
num_free_blocks += lock.num_free_blocks;
next_blk = lock.next;
_prev_lock = Some(lock); // hold the lock until we've read the next block
_head_lock = None;
}
BlockAllocatorStats {
num_blocks: self.num_blocks,
num_initialized: self.num_initialized.load(Ordering::Relaxed),
num_free_blocks,
}
}
}
#[derive(Clone, Debug)]
pub struct BlockAllocatorStats {
pub num_blocks: u64,
pub num_initialized: u64,
pub num_free_blocks: u64,
}

View File

@@ -1,33 +0,0 @@
use std::alloc::Layout;
use std::mem::MaybeUninit;
use crate::allocator::block::BlockAllocator;
use crate::allocator::slab::SlabDesc;
pub struct MultiSlabAllocator<'t, const N: usize> {
pub(crate) block_allocator: BlockAllocator<'t>,
pub(crate) slab_descs: [SlabDesc; N],
}
impl<'t, const N: usize> MultiSlabAllocator<'t, N> {
pub(crate) fn new(
area: &'t mut [MaybeUninit<u8>],
layouts: &[Layout; N],
) -> MultiSlabAllocator<'t, N> {
let block_allocator = BlockAllocator::new(area);
MultiSlabAllocator {
block_allocator,
slab_descs: std::array::from_fn(|i| SlabDesc::new(&layouts[i])),
}
}
pub(crate) fn alloc_slab(&self, slab_idx: usize) -> *mut u8 {
self.slab_descs[slab_idx].alloc_chunk(&self.block_allocator)
}
pub(crate) fn dealloc_slab(&self, slab_idx: usize, ptr: *mut u8) {
self.slab_descs[slab_idx].dealloc_chunk(ptr, &self.block_allocator)
}
}

View File

@@ -1,432 +0,0 @@
//! A slab allocator that carves out fixed-size chunks from larger blocks.
//!
//!
use std::alloc::Layout;
use std::mem::MaybeUninit;
use std::ops::Deref;
use std::sync::atomic::{AtomicU32, AtomicU64, Ordering};
use spin;
use super::alloc_from_slice;
use super::block::BlockAllocator;
use crate::allocator::block::BLOCK_SIZE;
pub(crate) struct SlabDesc {
pub(crate) layout: Layout,
block_lists: spin::RwLock<BlockLists>,
pub(crate) num_blocks: AtomicU64,
pub(crate) num_allocated: AtomicU64,
}
// FIXME: Not sure if SlabDesc is really Sync or Send. It probably is when it's empty, but
// 'block_lists' contains pointers when it's not empty. In the current use as part of the
// the art tree, SlabDescs are only moved during initialization.
unsafe impl Sync for SlabDesc {}
unsafe impl Send for SlabDesc {}
#[derive(Default, Debug)]
struct BlockLists {
full_blocks: BlockList,
nonfull_blocks: BlockList,
}
impl BlockLists {
// Unlink a node. It must be in either one of the two lists.
unsafe fn unlink(&mut self, elem: *mut SlabBlockHeader) {
let list = unsafe {
if (*elem).next.is_null() {
if self.full_blocks.tail == elem {
Some(&mut self.full_blocks)
} else {
Some(&mut self.nonfull_blocks)
}
} else if (*elem).prev.is_null() {
if self.full_blocks.head == elem {
Some(&mut self.full_blocks)
} else {
Some(&mut self.nonfull_blocks)
}
} else {
None
}
};
unsafe { unlink_slab_block(list, elem) };
}
}
unsafe fn unlink_slab_block(mut list: Option<&mut BlockList>, elem: *mut SlabBlockHeader) {
unsafe {
if (*elem).next.is_null() {
assert_eq!(list.as_ref().unwrap().tail, elem);
list.as_mut().unwrap().tail = (*elem).prev;
} else {
assert_eq!((*(*elem).next).prev, elem);
(*(*elem).next).prev = (*elem).prev;
}
if (*elem).prev.is_null() {
assert_eq!(list.as_ref().unwrap().head, elem);
list.as_mut().unwrap().head = (*elem).next;
} else {
assert_eq!((*(*elem).prev).next, elem);
(*(*elem).prev).next = (*elem).next;
}
}
}
#[derive(Debug)]
struct BlockList {
head: *mut SlabBlockHeader,
tail: *mut SlabBlockHeader,
}
impl Default for BlockList {
fn default() -> Self {
BlockList {
head: std::ptr::null_mut(),
tail: std::ptr::null_mut(),
}
}
}
impl BlockList {
unsafe fn push_head(&mut self, elem: *mut SlabBlockHeader) {
unsafe {
if self.is_empty() {
self.tail = elem;
(*elem).next = std::ptr::null_mut();
} else {
(*elem).next = self.head;
(*self.head).prev = elem;
}
(*elem).prev = std::ptr::null_mut();
self.head = elem;
}
}
fn is_empty(&self) -> bool {
self.head.is_null()
}
unsafe fn unlink(&mut self, elem: *mut SlabBlockHeader) {
unsafe { unlink_slab_block(Some(self), elem) }
}
#[cfg(test)]
fn dump(&self) {
let mut next = self.head;
while !next.is_null() {
let n = unsafe { next.as_ref() }.unwrap();
eprintln!(
" blk {:?} (free {}/{})",
next,
n.num_free_chunks.load(Ordering::Relaxed),
n.num_chunks
);
next = n.next;
}
}
}
impl SlabDesc {
pub(crate) fn new(layout: &Layout) -> SlabDesc {
SlabDesc {
layout: *layout,
block_lists: spin::RwLock::new(BlockLists::default()),
num_allocated: AtomicU64::new(0),
num_blocks: AtomicU64::new(0),
}
}
}
#[derive(Debug)]
struct SlabBlockHeader {
free_chunks_head: spin::Mutex<*mut FreeChunk>,
num_free_chunks: AtomicU32,
num_chunks: u32, // this is really a constant for a given Layout
// these fields are protected by the lock on the BlockLists
prev: *mut SlabBlockHeader,
next: *mut SlabBlockHeader,
}
struct FreeChunk {
next: *mut FreeChunk,
}
enum ReadOrWriteGuard<'a, T> {
Read(spin::RwLockReadGuard<'a, T>),
Write(spin::RwLockWriteGuard<'a, T>),
}
impl<'a, T> Deref for ReadOrWriteGuard<'a, T> {
type Target = T;
fn deref(&self) -> &<Self as Deref>::Target {
match self {
ReadOrWriteGuard::Read(g) => g.deref(),
ReadOrWriteGuard::Write(g) => g.deref(),
}
}
}
impl SlabDesc {
pub fn alloc_chunk(&self, block_allocator: &BlockAllocator) -> *mut u8 {
// Are there any free chunks?
let mut acquire_write = false;
'outer: loop {
let mut block_lists_guard = if acquire_write {
ReadOrWriteGuard::Write(self.block_lists.write())
} else {
ReadOrWriteGuard::Read(self.block_lists.read())
};
'inner: loop {
let block_ptr = block_lists_guard.nonfull_blocks.head;
if block_ptr.is_null() {
break 'outer;
}
unsafe {
let mut free_chunks_head = (*block_ptr).free_chunks_head.lock();
if !(*free_chunks_head).is_null() {
let result = *free_chunks_head;
(*free_chunks_head) = (*result).next;
let _old = (*block_ptr).num_free_chunks.fetch_sub(1, Ordering::Relaxed);
self.num_allocated.fetch_add(1, Ordering::Relaxed);
return result.cast();
}
}
// The block at the head of the list was full. Grab write lock and retry
match block_lists_guard {
ReadOrWriteGuard::Read(_) => {
acquire_write = true;
continue 'outer;
}
ReadOrWriteGuard::Write(ref mut g) => {
// move the node to the list of full blocks
unsafe {
g.nonfull_blocks.unlink(block_ptr);
g.full_blocks.push_head(block_ptr);
};
continue 'inner;
}
}
}
}
// no free chunks. Allocate a new block (and the chunk from that)
let (new_block, new_chunk) = self.alloc_block_and_chunk(block_allocator);
self.num_blocks.fetch_add(1, Ordering::Relaxed);
// Add the block to the list in the SlabDesc
unsafe {
let mut block_lists_guard = self.block_lists.write();
block_lists_guard.nonfull_blocks.push_head(new_block);
}
self.num_allocated.fetch_add(1, Ordering::Relaxed);
new_chunk
}
pub fn dealloc_chunk(&self, chunk_ptr: *mut u8, _block_allocator: &BlockAllocator) {
// Find the block it belongs to. You can find the block from the address. (And knowing the
// layout, you could calculate the chunk number too.)
let block_ptr: *mut SlabBlockHeader = {
let block_addr = (chunk_ptr.addr() / BLOCK_SIZE) * BLOCK_SIZE;
chunk_ptr.with_addr(block_addr).cast()
};
let chunk_ptr: *mut FreeChunk = chunk_ptr.cast();
// Mark the chunk as free in 'freechunks' list
let num_chunks;
let num_free_chunks;
unsafe {
let mut free_chunks_head = (*block_ptr).free_chunks_head.lock();
(*chunk_ptr).next = *free_chunks_head;
*free_chunks_head = chunk_ptr;
num_free_chunks = (*block_ptr).num_free_chunks.fetch_add(1, Ordering::Relaxed) + 1;
num_chunks = (*block_ptr).num_chunks;
}
if num_free_chunks == 1 {
// If the block was full previously, add it to the nonfull blocks list. Note that
// we're not holding the lock anymore, so it can immediately become full again.
// That's harmless, it will be moved back to the full list again when a call
// to alloc_chunk() sees it.
let mut block_lists = self.block_lists.write();
unsafe {
block_lists.unlink(block_ptr);
block_lists.nonfull_blocks.push_head(block_ptr);
};
} else if num_free_chunks == num_chunks {
// If the block became completely empty, move it to the free list
// TODO
// FIXME: we're still holding the spinlock. It's not exactly safe to return it to
// the free blocks list, is it? Defer it as garbage to wait out concurrent updates?
//block_allocator.release_block()
}
// update stats
self.num_allocated.fetch_sub(1, Ordering::Relaxed);
}
fn alloc_block_and_chunk(
&self,
block_allocator: &BlockAllocator,
) -> (*mut SlabBlockHeader, *mut u8) {
// fixme: handle OOM
let block_slice: &mut [MaybeUninit<u8>] = block_allocator.alloc_block();
let (block_header, remain) = alloc_from_slice::<SlabBlockHeader>(block_slice);
let padding = remain.as_ptr().align_offset(self.layout.align());
let num_chunks = (remain.len() - padding) / self.layout.size();
let first_chunk_ptr: *mut FreeChunk = remain[padding..].as_mut_ptr().cast();
unsafe {
let mut chunk_ptr = first_chunk_ptr;
for _ in 0..num_chunks - 1 {
let next_chunk_ptr = chunk_ptr.byte_add(self.layout.size());
(*chunk_ptr).next = next_chunk_ptr;
chunk_ptr = next_chunk_ptr;
}
(*chunk_ptr).next = std::ptr::null_mut();
let result_chunk = first_chunk_ptr;
let block_header = block_header.write(SlabBlockHeader {
free_chunks_head: spin::Mutex::new((*first_chunk_ptr).next),
prev: std::ptr::null_mut(),
next: std::ptr::null_mut(),
num_chunks: num_chunks as u32,
num_free_chunks: AtomicU32::new(num_chunks as u32 - 1),
});
(block_header, result_chunk.cast())
}
}
#[cfg(test)]
fn dump(&self) {
eprintln!(
"slab dump ({} blocks, {} allocated chunks)",
self.num_blocks.load(Ordering::Relaxed),
self.num_allocated.load(Ordering::Relaxed)
);
let lists = self.block_lists.read();
eprintln!("nonfull blocks:");
lists.nonfull_blocks.dump();
eprintln!("full blocks:");
lists.full_blocks.dump();
}
}
#[cfg(test)]
mod tests {
use super::*;
use rand::Rng;
use rand_distr::Zipf;
struct TestObject {
val: usize,
_dummy: [u8; BLOCK_SIZE / 4],
}
struct TestObjectSlab<'a>(SlabDesc, BlockAllocator<'a>);
impl<'a> TestObjectSlab<'a> {
fn new(block_allocator: BlockAllocator) -> TestObjectSlab {
TestObjectSlab(SlabDesc::new(&Layout::new::<TestObject>()), block_allocator)
}
fn alloc(&self, val: usize) -> *mut TestObject {
let obj: *mut TestObject = self.0.alloc_chunk(&self.1).cast();
unsafe { (*obj).val = val };
obj
}
fn dealloc(&self, obj: *mut TestObject) {
self.0.dealloc_chunk(obj.cast(), &self.1)
}
}
#[test]
fn test_slab_alloc() {
const MEM_SIZE: usize = 100000000;
let mut area = Box::new_uninit_slice(MEM_SIZE);
let block_allocator = BlockAllocator::new(&mut area);
let slab = TestObjectSlab::new(block_allocator);
let mut all: Vec<*mut TestObject> = Vec::new();
for i in 0..11 {
all.push(slab.alloc(i));
}
for i in 0..11 {
assert!(unsafe { (*all[i]).val == i });
}
let distribution = Zipf::new(10 as f64, 1.1).unwrap();
let mut rng = rand::rng();
for _ in 0..100000 {
slab.0.dump();
let idx = (rng.sample(distribution) as usize).into();
let ptr: *mut TestObject = all[idx];
if !ptr.is_null() {
assert_eq!(unsafe { (*ptr).val }, idx);
slab.dealloc(ptr);
all[idx] = std::ptr::null_mut();
} else {
all[idx] = slab.alloc(idx);
}
}
}
fn new_test_blk(i: u32) -> *mut SlabBlockHeader {
Box::into_raw(Box::new(SlabBlockHeader {
free_chunks_head: spin::Mutex::new(std::ptr::null_mut()),
num_free_chunks: AtomicU32::new(0),
num_chunks: i,
prev: std::ptr::null_mut(),
next: std::ptr::null_mut(),
}))
}
#[test]
fn test_block_linked_list() {
// note: these are leaked, but that's OK for tests
let a = new_test_blk(0);
let b = new_test_blk(1);
let mut list = BlockList::default();
assert!(list.is_empty());
unsafe {
list.push_head(a);
assert!(!list.is_empty());
list.unlink(a);
}
assert!(list.is_empty());
unsafe {
list.push_head(b);
list.push_head(a);
assert_eq!(list.head, a);
assert_eq!((*a).next, b);
assert_eq!((*b).prev, a);
assert_eq!(list.tail, b);
list.unlink(a);
list.unlink(b);
assert!(list.is_empty());
}
}
}

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