Skip to content

Conversation

@lostluck
Copy link
Contributor

@lostluck lostluck commented Aug 11, 2021

Executing all test binaries at the same time seems to be overloading the mini-local Flink runner,
setting a limit to avoid overwhelming it.

Only applies to the portable/validates runner tests.


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

  • Choose reviewer(s) and mention them in a comment (R: @username).
  • Format the pull request title like [BEAM-XXX] Fixes bug in ApproximateQuantiles, where you replace BEAM-XXX with the appropriate JIRA issue, if applicable. This will automatically link the pull request to the issue.
  • 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.

ValidatesRunner compliance status (on master branch)

Lang ULR Dataflow Flink Samza Spark Twister2
Go --- Build Status Build Status Build Status Build Status ---
Java Build Status Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Build Status
Python --- Build Status
Build Status
Build Status
Build Status
Build Status
Build Status Build Status ---
XLang Build Status Build Status Build Status Build Status Build Status ---

Examples testing status on various runners

Lang ULR Dataflow Flink Samza Spark Twister2
Go --- --- --- --- --- --- ---
Java --- Build Status
Build Status
Build Status
--- --- --- --- ---
Python --- --- --- --- --- --- ---
XLang --- --- --- --- --- --- ---

Post-Commit SDK/Transform Integration Tests Status (on master branch)

Go Java Python
Build Status Build Status Build Status
Build Status
Build Status

Pre-Commit Tests Status (on master branch)

--- Java Python Go Website Whitespace Typescript
Non-portable Build Status
Build Status
Build Status
Build Status
Build Status
Build Status Build Status Build Status Build Status
Portable --- Build Status Build Status --- --- ---

See .test-infra/jenkins/README for trigger phrase, status and link of all Jenkins jobs.

GitHub Actions Tests Status (on master branch)

Build python source distribution and wheels
Python tests
Java tests

See CI.md for more information about GitHub Actions CI.

Executing all test binaries at the same time seems to be overloading the mini-local Flink runner,
setting a limit to avoid overwhelming it.
@lostluck
Copy link
Contributor Author

Run Go Flink ValidatesRunner

@lostluck
Copy link
Contributor Author

Run Go PostCommit

@lostluck
Copy link
Contributor Author

R: @ibzib

TIMEOUT=1h

# Default limit on simultaneous test binaries/packages being executed.
PARALLEL=3
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I saw in the docs that the default is the number of cores, so it should be fine to leave it unset for the Dataflow runner, which doesn't need to do a lot of heavy-duty computation locally. But for Flink I think it should be 1. If nothing else, it makes reading the logs easier -- otherwise the logs for different jobs get interleaved and it becomes very difficult to tell which logs came from which test.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'd say that's overthinking the current issue.
Dataflow does have a per-project simultaneous job limit though, and they all run on apache-beam-testing.

TBH most of the output available from the Jenkins tests, due to gradle aren't actually useful in debugging anything in specific because gradle+ jenkins swallow the logs. It's more important at the moment that the tests can run to completion. We can always tool it downwards later.

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

TBH most of the output available from the Jenkins tests, due to gradle aren't actually useful in debugging anything in specific because gradle+ jenkins swallow the logs.

Flink logs don't get swallowed, so they'd be useful if only we could tell which job they came from.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the main source of conflict here is that I didn't see that Spark and Samza were already being hard limited to 1 test binary at a time anyway. I'm keeping the 3 limit for Dataflow and Portable, and have made it flaggable. Also changed things so they're only specifying the flag once, what I had would override it.

A later PR might change this kind of hacky overrides to happen at the script invocation level, but I'm currently making these changes via the web UI, and would rather keep it simple.

@lostluck
Copy link
Contributor Author

Run Go Flink ValidatesRunner

@lostluck
Copy link
Contributor Author

Run Go Flink ValidatesRunner

@lostluck
Copy link
Contributor Author

Run Go PostCommit

@lostluck
Copy link
Contributor Author

Run Go Spark ValidatesRunner

@lostluck
Copy link
Contributor Author

Run Go Samza ValidatesRunner

Copy link

@ibzib ibzib left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@lostluck lostluck merged commit 55f42a5 into master Aug 12, 2021
@lostluck lostluck deleted the lostluck-testp branch August 12, 2021 02:02
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants