chore: add optional CI flow for parquet writes#4696
Open
comphead wants to merge 2 commits into
Open
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
Comment on lines
+58
to
+154
| name: spark-sql-writer/spark-${{ inputs.spark-version }} | ||
| runs-on: ubuntu-24.04 | ||
| container: | ||
| image: amd64/rust | ||
| steps: | ||
| - uses: actions/checkout@v7 | ||
|
|
||
| - name: Resolve Spark full version and JDK | ||
| id: resolve | ||
| shell: bash | ||
| run: | | ||
| # Map each supported Spark minor version to its full version + JDK. | ||
| # Mirrors ci.yml's per-version reusable invocations (default-on PR | ||
| # versions only; 3.4 and 4.0 are label-gated and not offered here). | ||
| case "${{ inputs.spark-version }}" in | ||
| 3.5) spark_full=3.5.8; java=17 ;; | ||
| 4.1) spark_full=4.1.2; java=17 ;; | ||
| *) echo "Unsupported spark-version: ${{ inputs.spark-version }}" >&2; exit 1 ;; | ||
| esac | ||
| echo "spark-full=$spark_full" >> "$GITHUB_OUTPUT" | ||
| echo "java=$java" >> "$GITHUB_OUTPUT" | ||
|
|
||
| - name: Setup Rust & Java toolchain | ||
| uses: ./.github/actions/setup-builder | ||
| with: | ||
| rust-version: ${{ env.RUST_VERSION }} | ||
| jdk-version: ${{ steps.resolve.outputs.java }} | ||
|
|
||
| - name: Restore Cargo cache | ||
| uses: actions/cache/restore@v5 | ||
| with: | ||
| path: | | ||
| ~/.cargo/registry | ||
| ~/.cargo/git | ||
| native/target | ||
| key: ${{ runner.os }}-cargo-ci-${{ hashFiles('native/**/Cargo.lock', 'native/**/Cargo.toml') }}-${{ hashFiles('native/**/*.rs') }} | ||
| restore-keys: | | ||
| ${{ runner.os }}-cargo-ci-${{ hashFiles('native/**/Cargo.lock', 'native/**/Cargo.toml') }}- | ||
|
|
||
| - name: Build native library (CI profile) | ||
| run: | | ||
| cd native | ||
| cargo build --profile ci | ||
| env: | ||
| RUSTFLAGS: "-Ctarget-cpu=x86-64-v3 -Clink-arg=-fuse-ld=bfd" | ||
|
|
||
| - name: Stage native library at release path | ||
| run: | | ||
| # setup-spark-builder's `mvnw install -DskipTests` (skip-native-build | ||
| # path) bundles native/target/release/libcomet.so into the Comet JAR. | ||
| # We built with --profile ci to avoid LTO, so the file lives at | ||
| # native/target/ci/. Copy it to where the Maven build expects it. | ||
| mkdir -p native/target/release | ||
| cp native/target/ci/libcomet.so native/target/release/libcomet.so | ||
|
|
||
| - name: Setup Spark | ||
| uses: ./.github/actions/setup-spark-builder | ||
| with: | ||
| spark-version: ${{ steps.resolve.outputs.spark-full }} | ||
| spark-short-version: ${{ inputs.spark-version }} | ||
| skip-native-build: true | ||
|
|
||
| - name: Run Parquet writer tests | ||
| run: | | ||
| cd apache-spark | ||
| rm -rf /root/.m2/repository/org/apache/parquet # somehow parquet cache requires cleanups | ||
| # SERIAL_SBT_TESTS gates SparkParallelTestGrouping in | ||
| # project/SparkBuild.scala. We always set it to reduce peak memory | ||
| # on standard 7 GB runners (3.5 and 4.1 are unaffected by the | ||
| # 4.0+JDK 21 file-stream-leak case the reusable workflow handles). | ||
| export SERIAL_SBT_TESTS=1 | ||
| # Same forked-test-JVM caps as sql_core-* in spark_sql_test_reusable.yml. | ||
| export HEAP_SIZE=3g | ||
| export METASPACE_SIZE=1g | ||
| NOLINT_ON_COMPILE=true ENABLE_COMET=true ENABLE_COMET_ONHEAP=true ENABLE_COMET_WRITE=true \ | ||
| build/sbt -Dsbt.log.noformat=true -mem $SBT_MEM \ | ||
| 'set Global / concurrentRestrictions := Seq(Tags.limit(Tags.ForkedTestGroup, 1))' \ | ||
| "sql/testOnly org.apache.spark.sql.execution.datasources.parquet.ParquetIOSuite org.apache.spark.sql.execution.datasources.parquet.ParquetCommitterSuite org.apache.spark.sql.execution.datasources.parquet.ParquetEncodingSuite org.apache.spark.sql.execution.datasources.parquet.ParquetCompressionCodecPrecedenceSuite org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormatV1Suite org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormatV2Suite org.apache.spark.sql.execution.datasources.parquet.ParquetV1QuerySuite org.apache.spark.sql.execution.datasources.parquet.ParquetV2QuerySuite org.apache.spark.sql.execution.datasources.parquet.ParquetV1PartitionDiscoverySuite org.apache.spark.sql.execution.datasources.parquet.ParquetV2PartitionDiscoverySuite org.apache.spark.sql.execution.datasources.parquet.ParquetFieldIdIOSuite org.apache.spark.sql.execution.datasources.parquet.ParquetSchemaSuite org.apache.spark.sql.execution.datasources.FileFormatWriterSuite org.apache.spark.sql.sources.PartitionedWriteSuite" | ||
| env: | ||
| LC_ALL: "C.UTF-8" | ||
| # Cap SBT orchestrator heap so the freed RAM goes to the forked test | ||
| # JVM and OS/container overhead, fixing cgroup-OOM SIGKILLs under | ||
| # 7 GB runners. | ||
| SBT_MEM: "1024" | ||
| # G1GC + tuning for the SBT orchestrator JVM. -Xss4m replaces the | ||
| # launcher's -Xss64m default (no compile here, deep recursion not | ||
| # needed). UseStringDeduplication and MaxMetaspaceSize cap real and | ||
| # ceiling footprint. ExitOnOutOfMemoryError fails fast. | ||
| SBT_OPTS: >- | ||
| -Xss4m | ||
| -XX:+UseG1GC | ||
| -XX:+UseStringDeduplication | ||
| -XX:MaxMetaspaceSize=384m | ||
| -XX:G1HeapRegionSize=2m | ||
| -XX:InitiatingHeapOccupancyPercent=35 | ||
| -XX:+ParallelRefProcEnabled | ||
| -XX:+ExitOnOutOfMemoryError |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Which issue does this PR close?
Closes #3209 .
Rationale for this change
New
.github/workflows/spark_sql_writer_tests.yml,workflow_dispatchonly. Runs a focused set of upstream Parquet writer suites withENABLE_COMET_WRITE=true(spark.comet.parquet.write.enabled).What changes are included in this PR?
How are these changes tested?