SEZ9 commented on issue #13246: URL: https://github.com/apache/seatunnel/issues/13246#issuecomment-6871188828
Thanks @goutamadwant for following up with #13247. The direction sounds right: treating `G1 Refine#N` as a JVM-native thread in `isSystemThread` is exactly the fix needed so `ContainerUtil.getJVMThreads` / `SeaTunnelContainer.doExecuteJob` stop reporting GC refinement workers as leaked job threads, while keeping the 120s cleanup wait and the application-thread check intact. A couple of things that would help when reviewing the PR: 1. Please match the name defensively (e.g. a prefix/regex on `G1 Refine` rather than a single fixed string), since the worker index varies with the GC configuration. 2. Add a note in the PR description (or a comment in the code) explaining why these threads are expected to survive job cleanup on Java 17, so the exclusion isn't mistaken for hiding a real leak later. 3. If you can, include a run of one of the previously failing Python source/transform E2E cases on Java 17 showing the detector now passes, and confirm that a genuinely leaked application thread is still flagged. I'll take a look at #13247 and leave any further comments there. <!-- streview-comment:2346 --> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
