lhotari commented on pull request #11343: URL: https://github.com/apache/pulsar/pull/11343#issuecomment-882275064
> @lhotari Can you provide a case that there would be a concurrency issue? I looked through the code and we always use tmp files downloading the archive for each instance. This should prevent any concurrency issues. For example: > > 12:29:37.588 INFO [worker-scheduler-0] o.a.p.f.w.FunctionActioner@241 - Function package file is linked from /var/folders/ds/3dnmfjkd6gnfqgvbjgwwbm640000gn/T/PulsarSourceE2ETest3142222813321573090/downloadDirectory/external-repl-prop/io/PulsarSource-test/0/pulsar-io-data-generator-2.9.0-SNAPSHOT.nar.0.d65f1667-d711-4fbb-a6bd-d8b1b5c804ad to /var/folders/ds/3dnmfjkd6gnfqgvbjgwwbm640000gn/T/PulsarSourceE2ETest3142222813321573090/downloadDirectory/external-repl-prop/io/PulsarSource-test/0/pulsar-io-data-generator-2.9.0-SNAPSHOT.nar I have explained the issue in #11340 description. The problem isn't in downloading the function jar. It's the NarUnpacker that contains the issues explained in #11340 description. I didn't duplicate this information in this PR description. The problem can be reproduced when using the thread or process runtime and when running a function with parallelism > brokers/function workers. The problem cannot be reproduced on kubernetes runtime since the different function instance processes won't be sharing the same /tmp/pulsar-nar directory in that case. Please check the #11340 description and provide some follow-up questions so that I can fill the possible missing information to the issue description. -- 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: commits-unsubscr...@pulsar.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org