Files
neon/.github/actions/run-python-test-set/action.yml
Alexander Bayandin e65f0fe874 CI(benchmarks): make job split consistent across reruns (#6614)
## Problem

We've got several issues with the current `benchmarks` job setup:
- `benchmark_durations.json` file (that we generate in runtime to
split tests into several jobs[0]) is not consistent between these
jobs (and very not consistent with the file if we rerun the job). I.e.
test selection for each job can be different, which could end up in
missed tests in a test run.
- `scripts/benchmark_durations` doesn't fetch all tests from the
database (it doesn't expect any extra directories inside
`test_runner/performance`)
- For some reason, currently split into 4 groups ends up with the 4th
group has no tests to run, which fails the job[1]

- [0] https://github.com/neondatabase/neon/pull/4683
- [1] https://github.com/neondatabase/neon/issues/6629

## Summary of changes
- Generate `benchmark_durations.json` file once before we start
`benchmarks` jobs (this makes it consistent across the jobs) and pass
the file content through the GitHub Actions input (this makes it
consistent for reruns)
- `scripts/benchmark_durations` fix SQL query for getting all required
tests
- Split benchmarks into 5 jobs instead of 4 jobs.
2024-02-06 17:00:55 +00:00

224 lines
8.6 KiB
YAML

name: 'Run python test'
description: 'Runs a Neon python test set, performing all the required preparations before'
inputs:
build_type:
description: 'Type of Rust (neon) and C (postgres) builds. Must be "release" or "debug", or "remote" for the remote cluster'
required: true
test_selection:
description: 'A python test suite to run'
required: true
extra_params:
description: 'Arbitrary parameters to pytest. For example "-s" to prevent capturing stdout/stderr'
required: false
default: ''
needs_postgres_source:
description: 'Set to true if the test suite requires postgres source checked out'
required: false
default: 'false'
run_in_parallel:
description: 'Whether to run tests in parallel'
required: false
default: 'true'
save_perf_report:
description: 'Whether to upload the performance report, if true PERF_TEST_RESULT_CONNSTR env variable should be set'
required: false
default: 'false'
run_with_real_s3:
description: 'Whether to pass real s3 credentials to the test suite'
required: false
default: 'false'
real_s3_bucket:
description: 'Bucket name for real s3 tests'
required: false
default: ''
real_s3_region:
description: 'Region name for real s3 tests'
required: false
default: ''
rerun_flaky:
description: 'Whether to rerun flaky tests'
required: false
default: 'false'
pg_version:
description: 'Postgres version to use for tests'
required: false
default: 'v14'
benchmark_durations:
description: 'benchmark durations JSON'
required: false
default: '{}'
runs:
using: "composite"
steps:
- name: Get Neon artifact
if: inputs.build_type != 'remote'
uses: ./.github/actions/download
with:
name: neon-${{ runner.os }}-${{ inputs.build_type }}-artifact
path: /tmp/neon
- name: Download Neon binaries for the previous release
if: inputs.build_type != 'remote'
uses: ./.github/actions/download
with:
name: neon-${{ runner.os }}-${{ inputs.build_type }}-artifact
path: /tmp/neon-previous
prefix: latest
- name: Download compatibility snapshot
if: inputs.build_type != 'remote'
uses: ./.github/actions/download
with:
name: compatibility-snapshot-${{ inputs.build_type }}-pg${{ inputs.pg_version }}
path: /tmp/compatibility_snapshot_pg${{ inputs.pg_version }}
prefix: latest
# The lack of compatibility snapshot (for example, for the new Postgres version)
# shouldn't fail the whole job. Only relevant test should fail.
skip-if-does-not-exist: true
- name: Checkout
if: inputs.needs_postgres_source == 'true'
uses: actions/checkout@v3
with:
submodules: true
fetch-depth: 1
- name: Cache poetry deps
id: cache_poetry
uses: actions/cache@v3
with:
path: ~/.cache/pypoetry/virtualenvs
key: v1-${{ runner.os }}-python-deps-${{ hashFiles('poetry.lock') }}
- name: Install Python deps
shell: bash -euxo pipefail {0}
run: ./scripts/pysync
- name: Run pytest
env:
NEON_BIN: /tmp/neon/bin
COMPATIBILITY_NEON_BIN: /tmp/neon-previous/bin
COMPATIBILITY_POSTGRES_DISTRIB_DIR: /tmp/neon-previous/pg_install
TEST_OUTPUT: /tmp/test_output
BUILD_TYPE: ${{ inputs.build_type }}
COMPATIBILITY_SNAPSHOT_DIR: /tmp/compatibility_snapshot_pg${{ inputs.pg_version }}
ALLOW_BACKWARD_COMPATIBILITY_BREAKAGE: contains(github.event.pull_request.labels.*.name, 'backward compatibility breakage')
ALLOW_FORWARD_COMPATIBILITY_BREAKAGE: contains(github.event.pull_request.labels.*.name, 'forward compatibility breakage')
RERUN_FLAKY: ${{ inputs.rerun_flaky }}
PG_VERSION: ${{ inputs.pg_version }}
shell: bash -euxo pipefail {0}
run: |
# PLATFORM will be embedded in the perf test report
# and it is needed to distinguish different environments
export PLATFORM=${PLATFORM:-github-actions-selfhosted}
export POSTGRES_DISTRIB_DIR=${POSTGRES_DISTRIB_DIR:-/tmp/neon/pg_install}
export DEFAULT_PG_VERSION=${PG_VERSION#v}
if [ "${BUILD_TYPE}" = "remote" ]; then
export REMOTE_ENV=1
fi
PERF_REPORT_DIR="$(realpath test_runner/perf-report-local)"
rm -rf $PERF_REPORT_DIR
TEST_SELECTION="test_runner/${{ inputs.test_selection }}"
EXTRA_PARAMS="${{ inputs.extra_params }}"
if [ -z "$TEST_SELECTION" ]; then
echo "test_selection must be set"
exit 1
fi
if [[ "${{ inputs.run_in_parallel }}" == "true" ]]; then
# -n16 uses sixteen processes to run tests via pytest-xdist
EXTRA_PARAMS="-n16 $EXTRA_PARAMS"
# --dist=loadgroup points tests marked with @pytest.mark.xdist_group
# to the same worker to make @pytest.mark.order work with xdist
EXTRA_PARAMS="--dist=loadgroup $EXTRA_PARAMS"
fi
if [[ "${{ inputs.run_with_real_s3 }}" == "true" ]]; then
echo "REAL S3 ENABLED"
export ENABLE_REAL_S3_REMOTE_STORAGE=nonempty
export REMOTE_STORAGE_S3_BUCKET=${{ inputs.real_s3_bucket }}
export REMOTE_STORAGE_S3_REGION=${{ inputs.real_s3_region }}
fi
if [[ "${{ inputs.save_perf_report }}" == "true" ]]; then
mkdir -p "$PERF_REPORT_DIR"
EXTRA_PARAMS="--out-dir $PERF_REPORT_DIR $EXTRA_PARAMS"
fi
if [ "${RERUN_FLAKY}" == "true" ]; then
mkdir -p $TEST_OUTPUT
poetry run ./scripts/flaky_tests.py "${TEST_RESULT_CONNSTR}" \
--days 7 \
--output "$TEST_OUTPUT/flaky.json" \
--pg-version "${DEFAULT_PG_VERSION}" \
--build-type "${BUILD_TYPE}"
EXTRA_PARAMS="--flaky-tests-json $TEST_OUTPUT/flaky.json $EXTRA_PARAMS"
fi
# We use pytest-split plugin to run benchmarks in parallel on different CI runners
if [ "${TEST_SELECTION}" = "test_runner/performance" ] && [ "${{ inputs.build_type }}" != "remote" ]; then
mkdir -p $TEST_OUTPUT
echo '${{ inputs.benchmark_durations || '{}' }}' > $TEST_OUTPUT/benchmark_durations.json
EXTRA_PARAMS="--durations-path $TEST_OUTPUT/benchmark_durations.json $EXTRA_PARAMS"
fi
if [[ "${{ inputs.build_type }}" == "debug" ]]; then
cov_prefix=(scripts/coverage "--profraw-prefix=$GITHUB_JOB" --dir=/tmp/coverage run)
elif [[ "${{ inputs.build_type }}" == "release" ]]; then
cov_prefix=()
else
cov_prefix=()
fi
# Wake up the cluster if we use remote neon instance
if [ "${{ inputs.build_type }}" = "remote" ] && [ -n "${BENCHMARK_CONNSTR}" ]; then
${POSTGRES_DISTRIB_DIR}/v${DEFAULT_PG_VERSION}/bin/psql ${BENCHMARK_CONNSTR} -c "SELECT version();"
fi
# Run the tests.
#
# The junit.xml file allows CI tools to display more fine-grained test information
# in its "Tests" tab in the results page.
# --verbose prints name of each test (helpful when there are
# multiple tests in one file)
# -rA prints summary in the end
# -s is not used to prevent pytest from capturing output, because tests are running
# in parallel and logs are mixed between different tests
#
mkdir -p $TEST_OUTPUT/allure/results
"${cov_prefix[@]}" ./scripts/pytest \
--junitxml=$TEST_OUTPUT/junit.xml \
--alluredir=$TEST_OUTPUT/allure/results \
--tb=short \
--verbose \
-rA $TEST_SELECTION $EXTRA_PARAMS
if [[ "${{ inputs.save_perf_report }}" == "true" ]]; then
export REPORT_FROM="$PERF_REPORT_DIR"
export REPORT_TO="$PLATFORM"
scripts/generate_and_push_perf_report.sh
fi
- name: Upload compatibility snapshot
if: github.ref_name == 'release'
uses: ./.github/actions/upload
with:
name: compatibility-snapshot-${{ inputs.build_type }}-pg${{ inputs.pg_version }}-${{ github.run_id }}
# Directory is created by test_compatibility.py::test_create_snapshot, keep the path in sync with the test
path: /tmp/test_output/compatibility_snapshot_pg${{ inputs.pg_version }}/
prefix: latest
- name: Upload test results
if: ${{ !cancelled() }}
uses: ./.github/actions/allure-report-store
with:
report-dir: /tmp/test_output/allure/results
unique-key: ${{ inputs.build_type }}-${{ inputs.pg_version }}