lhotari commented on pull request #11343: URL: https://github.com/apache/pulsar/pull/11343#issuecomment-883208038
> If the error is easily reproducible, can you provide steps to reproduce? What are the errors you see when this issue happens? @jerrypeng There's a repro now in #11379 , thanks to @freeznet . There's also short repro instructions and one possible error message in the original issue description, #11340 . > > > the tmp directory will facing the race condition, > > Each instance has its own temp directory for downloading artifacts, why would we have concurrency issues here? Downloading of the artifacts isn't the issue. The NAR classloader will extract the downloaded jar file to a directory called /tmp/pulsar-nar by default. This is a fixed location currently. For e2e Pulsar Function tests, there's a [unique narExtractionDirectory in use](https://github.com/apache/pulsar/pull/10032), so perhaps that is causing the confusion when running e2e tests. A unique `narExtractionDirectory` isn't a suitable solution for Pulsar Functions in production mode for the reason that it doesn't solve the problem for thread runtime. It would also consume more disk space since each instance would have it's own copy of the files. Please take a closer look to the solution in this PR since it solves the problem in a fairly simple way. -- 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]
