Propagate gcs-connector options to GcsUtil by clairemcginty · Pull Request #32769 · apache/beam

clairemcginty · 2024-10-14T14:52:02Z

Context: I was reading GCS Parquet files via SplittableDoFn and noticed that ReadableFile#openSeekable does not propagate any of the gcs-connector options specified in my core-site.xml file. Particularly, I wanted to turn off fs.gs.inputstream.fast.fail.on.not.found.enable, which is redundant in a SDF with default empty-match treatment, and tweak fs.gs.inputstream.fadvise. It looks like these GoogleCloudStorageReadOptions options need to be set explicitly in GcsUtil, and passed to any GoogleCloudStorage#open calls (see reference).

The big downside of this PR is of course, pulling in Hadoop :( The alternative is to manually copy-paste all the Configuration keys manually into GcsUtil, which seems harder to maintain. Or, I could omit the GcsReadOptionsFactory factory logic entirely and leave it 100% up to the user to construct GoogleCloudStorageReadOptions instances.

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

Mention the appropriate issue in your description (for example: addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment fixes #<ISSUE NUMBER> instead.
Update CHANGES.md with noteworthy changes.
If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.

clairemcginty · 2024-10-14T14:53:37Z

...oogle-cloud-platform-core/src/main/java/org/apache/beam/sdk/extensions/gcp/util/GcsUtil.java

+      try {
+        // Check if gcs-connector-hadoop is loaded into classpath
+        Class.forName("com.google.cloud.hadoop.fs.gcs.GoogleHadoopFileSystemConfiguration");
+        Configuration config = new Configuration();
+        return GoogleCloudStorageReadOptions.builder()
+            .setFastFailOnNotFound(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_FAST_FAIL_ON_NOT_FOUND_ENABLE
+                    .get(config, config::getBoolean))
+            .setSupportGzipEncoding(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_SUPPORT_GZIP_ENCODING_ENABLE
+                    .get(config, config::getBoolean))
+            .setInplaceSeekLimit(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_INPLACE_SEEK_LIMIT.get(
+                    config, config::getLong))
+            .setFadvise(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_FADVISE.get(
+                    config, config::getEnum))
+            .setMinRangeRequestSize(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_MIN_RANGE_REQUEST_SIZE.get(
+                    config, config::getInt))
+            .setGrpcChecksumsEnabled(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_CHECKSUMS_ENABLE.get(
+                    config, config::getBoolean))
+            .setGrpcReadTimeoutMillis(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_TIMEOUT_MS.get(
+                    config, config::getLong))
+            .setGrpcReadMessageTimeoutMillis(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_MESSAGE_TIMEOUT_MS.get(
+                    config, config::getLong))
+            .setGrpcReadMetadataTimeoutMillis(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_METADATA_TIMEOUT_MS.get(
+                    config, config::getLong))
+            .setGrpcReadZeroCopyEnabled(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_ZEROCOPY_ENABLE.get(
+                    config, config::getBoolean))
+            .setTraceLogEnabled(
+                GoogleHadoopFileSystemConfiguration.GCS_TRACE_LOG_ENABLE.get(
+                    config, config::getBoolean))
+            .setTraceLogTimeThreshold(
+                GoogleHadoopFileSystemConfiguration.GCS_TRACE_LOG_TIME_THRESHOLD_MS.get(
+                    config, config::getLong))
+            .build();
+      } catch (ClassNotFoundException e) {


Copy-pasted from here: https://github.com/GoogleCloudDataproc/hadoop-connectors/blob/v2.2.25/gcs/src/main/java/com/google/cloud/hadoop/fs/gcs/GoogleHadoopFileSystemConfiguration.java#L656-L677

I think we could make a case to make that method public in a future release so we're not pulling in Hadoop explicitly here.

Or, I could omit this if/else branch entirely and always return GoogleCloudStorageReadOptions.DEFAULT, and leave it up to the user to supply a GoogleCloudStorageReadOptions instance (thus passing the Hadoop dependency down to the user-end).

Or, I could omit this if/else branch entirely and always return GoogleCloudStorageReadOptions.DEFAULT, and leave it up to the user to supply a GoogleCloudStorageReadOptions instance (thus passing the Hadoop dependency down to the user-end).

I think this would be preferable to avoid having to pull in the other packages but asking for some other Beam maintainers more familiar with dep management etc to take a look as well.

Agree with @scwhittle. It is also easier for maintainers and if new options are added, we don't need to change the code to support that.

Sounds good! I removed the Hadoop dep and this parsing block; this leaves it up to the user to pass in a GoogleCloudStorageReadOptions constructed however they prefer 👍

clairemcginty · 2024-10-14T14:54:25Z

...ud-platform-core/src/test/java/org/apache/beam/sdk/extensions/gcp/GcpCoreApiSurfaceTest.java

            classesInPackage("com.google.auth"),
            classesInPackage("com.fasterxml.jackson.annotation"),
+            classesInPackage("com.google.cloud.hadoop.gcsio"),
+            classesInPackage("com.google.common.collect"), // via GoogleCloudStorageReadOptions


this is unfortunate 😓 GoogleCloudStorageReadOptions has one property that's strictly an ImmutableSet: https://github.com/GoogleCloudDataproc/hadoop-connectors/blob/v2.2.25/gcs/src/main/java/com/google/cloud/hadoop/fs/gcs/GoogleHadoopFileSystemConfiguration.java#L519

github-actions · 2024-10-14T16:06:54Z

Checks are failing. Will not request review until checks are succeeding. If you'd like to override that behavior, comment assign set of reviewers

clairemcginty · 2024-10-14T16:24:33Z

assign set of reviewers

github-actions · 2024-10-14T16:25:45Z

Assigning reviewers. If you would like to opt out of this review, comment assign to next reviewer:

R: @damondouglas for label java.

Available commands:

stop reviewer notifications - opt out of the automated review tooling
remind me after tests pass - tag the comment author after tests pass
waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).

clairemcginty · 2024-10-14T17:05:10Z

assigned reviewers since at first glance, the GHA failures did not look related/might be transient? could be wrong though.

clairemcginty · 2024-10-22T12:18:28Z

bumping this PR -- cc @scwhittle since I saw you recently made changes to GcsUtil?

scwhittle · 2024-10-23T08:46:15Z

...oogle-cloud-platform-core/src/main/java/org/apache/beam/sdk/extensions/gcp/util/GcsUtil.java

+      try {
+        // Check if gcs-connector-hadoop is loaded into classpath
+        Class.forName("com.google.cloud.hadoop.fs.gcs.GoogleHadoopFileSystemConfiguration");
+        Configuration config = new Configuration();
+        return GoogleCloudStorageReadOptions.builder()
+            .setFastFailOnNotFound(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_FAST_FAIL_ON_NOT_FOUND_ENABLE
+                    .get(config, config::getBoolean))
+            .setSupportGzipEncoding(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_SUPPORT_GZIP_ENCODING_ENABLE
+                    .get(config, config::getBoolean))
+            .setInplaceSeekLimit(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_INPLACE_SEEK_LIMIT.get(
+                    config, config::getLong))
+            .setFadvise(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_FADVISE.get(
+                    config, config::getEnum))
+            .setMinRangeRequestSize(
+                GoogleHadoopFileSystemConfiguration.GCS_INPUT_STREAM_MIN_RANGE_REQUEST_SIZE.get(
+                    config, config::getInt))
+            .setGrpcChecksumsEnabled(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_CHECKSUMS_ENABLE.get(
+                    config, config::getBoolean))
+            .setGrpcReadTimeoutMillis(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_TIMEOUT_MS.get(
+                    config, config::getLong))
+            .setGrpcReadMessageTimeoutMillis(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_MESSAGE_TIMEOUT_MS.get(
+                    config, config::getLong))
+            .setGrpcReadMetadataTimeoutMillis(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_METADATA_TIMEOUT_MS.get(
+                    config, config::getLong))
+            .setGrpcReadZeroCopyEnabled(
+                GoogleHadoopFileSystemConfiguration.GCS_GRPC_READ_ZEROCOPY_ENABLE.get(
+                    config, config::getBoolean))
+            .setTraceLogEnabled(
+                GoogleHadoopFileSystemConfiguration.GCS_TRACE_LOG_ENABLE.get(
+                    config, config::getBoolean))
+            .setTraceLogTimeThreshold(
+                GoogleHadoopFileSystemConfiguration.GCS_TRACE_LOG_TIME_THRESHOLD_MS.get(
+                    config, config::getLong))
+            .build();
+      } catch (ClassNotFoundException e) {


Or, I could omit this if/else branch entirely and always return GoogleCloudStorageReadOptions.DEFAULT, and leave it up to the user to supply a GoogleCloudStorageReadOptions instance (thus passing the Hadoop dependency down to the user-end).

I think this would be preferable to avoid having to pull in the other packages but asking for some other Beam maintainers more familiar with dep management etc to take a look as well.

scwhittle · 2024-10-23T08:52:35Z

...oogle-cloud-platform-core/src/main/java/org/apache/beam/sdk/extensions/gcp/util/GcsUtil.java

    googleCloudStorageOptions =
        GoogleCloudStorageOptions.builder()
            .setAppName("Beam")
+            .setReadChannelOptions(this.googleCloudStorageReadOptions)


This seems like a bug in GoogleCloudStorageImpl.open method that it doesn't use these options but you can remove the separate member variable and use googleCloudStorageOptions.getReadChannelOptions() in open below below.

yeah GoogleCloudStorageImpl is set up in a slightly strange way -- it's constructed with a GoogleCloudStorageOptions arg but also accepts a separate GoogleCloudStorageOptions as an argument to open, while ignoring the former instance variable. Dropping the separate member variable here makes sense, will do 👍

github-actions · 2024-10-31T12:14:28Z

Reminder, please take a look at this pr: @damondouglas

scwhittle · 2024-11-05T09:32:08Z

R: @shunping (XQ suggested you to help review this)

github-actions · 2024-11-05T09:33:28Z

Stopping reviewer notifications for this pull request: review requested by someone other than the bot, ceding control. If you'd like to restart, comment assign set of reviewers

shunping · 2024-11-05T16:59:36Z

R: @shunping (XQ suggested you to help review this)

ack. will take a look today

clairemcginty · 2024-11-14T15:13:56Z

@shunping , whenever you have a chance I'd appreciate any feedback on this!

sdks/java/extensions/google-cloud-platform-core/src/test/resources/test-hadoop-conf.xml

shunping

Thanks for adding this option. I left some minor comments there, and overall it looks good.

After you finish revising, could you please sync to the current HEAD and so the tests are re-triggered? Thanks!

shunping · 2024-11-26T16:16:41Z

Running the failed precommit test again, though the failure seems unrelated to the code change here.

shunping · 2024-11-27T02:24:47Z

...ud-platform-core/src/test/java/org/apache/beam/sdk/extensions/gcp/GcpCoreApiSurfaceTest.java

            classesInPackage("com.google.api.services.storage"),
            classesInPackage("com.google.auth"),
            classesInPackage("com.fasterxml.jackson.annotation"),
+            classesInPackage("com.google.cloud.hadoop.gcsio"),


Looks like this breaks a test:
https://ge.apache.org/s/5ubkbvlvrjzgi/console-log/task/:sdks:java:extensions:google-cloud-platform-core:test?anchor=5&page=1

thanks for tracking that down! Pushed a fix. one of the precommit tests is still failing though:

org.apache.beam.runners.dataflow.worker.streaming.harness.FanOutStreamingEngineWorkerHarnessTest > testOnNewWorkerMetadata_correctlyRemovesStaleWindmillServers FAILED java.lang.AssertionError at GrpcCleanupRule.java:201

Not sure if/how this could be related to my PR

shunping · 2024-12-02T15:32:03Z

Run Java PreCommit

shunping

LGTM. Thanks!

CHANGES.md

codecov · 2024-12-03T16:07:30Z

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 57.39%. Comparing base (160ffd5) to head (27672ab).
Report is 4 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##             master   #32769   +/-   ##
=========================================
  Coverage     57.39%   57.39%           
  Complexity     1474     1474           
=========================================
  Files           970      970           
  Lines        154426   154426           
  Branches       1076     1076           
=========================================
  Hits          88637    88637           
  Misses        63585    63585           
  Partials       2204     2204

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

github-actions bot added java extensions gcp labels Oct 14, 2024

clairemcginty commented Oct 14, 2024

View reviewed changes

clairemcginty marked this pull request as ready for review October 14, 2024 15:48

github-actions bot added the Next Action: Reviewers label Oct 14, 2024

scwhittle reviewed Oct 23, 2024

View reviewed changes

github-actions bot added the slow-review label Oct 31, 2024

shunping reviewed Nov 20, 2024

View reviewed changes

sdks/java/extensions/google-cloud-platform-core/src/test/resources/test-hadoop-conf.xml Outdated Show resolved Hide resolved

shunping requested changes Nov 20, 2024

View reviewed changes

clairemcginty added 4 commits November 25, 2024 11:32

Propagate gcs-connector options to GcsUtil

00264cd

newline

fe24a1e

Update CHANGES.md

ed811a3

Remove Hadoop dependency

661a826

clairemcginty force-pushed the propagate-gcs-connector-opts branch from 57c0b50 to 661a826 Compare November 25, 2024 16:44

clairemcginty added 2 commits November 25, 2024 12:56

Remove unused deps

3a1b9a8

Drop googleCloudStorageReadOptions member variable

2432dba

clairemcginty changed the title ~~Proposal: Propagate gcs-connector options to GcsUtil~~ Propagate gcs-connector options to GcsUtil Nov 25, 2024

shunping reviewed Nov 27, 2024

View reviewed changes

clairemcginty added 2 commits November 26, 2024 22:39

add missing package

efb4035

Merge branch 'master' into propagate-gcs-connector-opts

c1b6fcd

clairemcginty added 2 commits December 3, 2024 09:52

Merge branch 'master' into propagate-gcs-connector-opts

eb9004a

fixup CHANGES.md

27672ab

shunping approved these changes Dec 3, 2024

View reviewed changes

damccorm approved these changes Dec 3, 2024

View reviewed changes

CHANGES.md Outdated Show resolved Hide resolved

damccorm merged commit 63d89cd into apache:master Dec 3, 2024

clairemcginty deleted the propagate-gcs-connector-opts branch December 3, 2024 16:17

clairemcginty mentioned this pull request Jan 10, 2025

Make {Read,Write}Options-parsing methods public GoogleCloudDataproc/hadoop-connectors#1294

Merged

clairemcginty mentioned this pull request Jan 22, 2025

Supply gcs-connector options from default Configuration spotify/scio#5549

Merged

Conversation

clairemcginty commented Oct 14, 2024

GitHub Actions Tests Status (on master branch)

Uh oh!

clairemcginty Oct 14, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Oct 14, 2024

Uh oh!

clairemcginty commented Oct 14, 2024

Uh oh!

github-actions bot commented Oct 14, 2024

Uh oh!

clairemcginty commented Oct 14, 2024

Uh oh!

clairemcginty commented Oct 22, 2024

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

clairemcginty Oct 23, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Oct 31, 2024

Uh oh!

scwhittle commented Nov 5, 2024

Uh oh!

github-actions bot commented Nov 5, 2024

Uh oh!

shunping commented Nov 5, 2024

Uh oh!

clairemcginty commented Nov 14, 2024

Uh oh!

Uh oh!

shunping left a comment

Choose a reason for hiding this comment

Uh oh!

shunping commented Nov 26, 2024

Uh oh!

shunping Nov 27, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

shunping commented Dec 2, 2024

Uh oh!

shunping left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

codecov bot commented Dec 3, 2024

Codecov Report

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

clairemcginty Oct 14, 2024 •

edited

Loading

clairemcginty Oct 23, 2024 •

edited

Loading

shunping Nov 27, 2024 •

edited

Loading