The GitHub Actions job "CI" on pekko-connectors-kafka.git/ci-increase-heap has 
succeeded.
Run started by GitHub user pjfanning (triggered by pjfanning).

Head commit for run:
b364f566425c82abc98d29c0e2f48b96eda12c9a / PJ Fanning 
<[email protected]>
increase CI heap so the binary-compatibility check does not run out of memory

Motivation:
The `Check Code Style` job, which runs `checkCodeStyle; mimaReportBinaryIssues`,
fails with `java.lang.OutOfMemoryError: Java heap space`. sbt runs with its
default 1g heap and spends most of the job in GC before dying:

    [warn] In the last 10 seconds, 24.282 (253.1%) were spent in GC.
           [Heap: 0.00GB free of 1.00GB, max 1.00GB]
    java.lang.OutOfMemoryError: Java heap space
    [error] [launcher] error during sbt launcher: java.lang.OutOfMemoryError

The unscoped `mimaReportBinaryIssues` loads current and previous classfiles for
every module in one JVM, so it is the peak consumer. This is not specific to any
one PR - `main` has failed the same way on every run since #652.

Modification:
Set `JAVA_OPTS: "-Xmx4g"` in the workflow-level `env` block of the CI workflow,
so every job in it gets a 4g heap. 4g matches what this repository already uses
for its other heavy sbt jobs (`link-validator.yml`, `publish-*-docs.yml`), and
is well inside the memory available on the GitHub-hosted runners.

This is deliberately confined to the CI workflow rather than `.sbtopts`, so
local sbt runs keep whatever heap the developer has configured.

Result:
The binary-compatibility check completes instead of exhausting the heap, and the
`Check / Binary Compatibility` job can pass again.

Tests:
- Not run - CI configuration only; the change is verified by the CI run on this
  PR, which is the job that was failing.

References:
Refs #653 - that PR is blocked by this same pre-existing failure

Report URL: 
https://github.com/apache/pekko-connectors-kafka/actions/runs/34546783195

With regards,
GitHub Actions via GitBox


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to