skonto removed a comment on pull request #30675: URL: https://github.com/apache/spark/pull/30675#issuecomment-745183719
@attilapiros if I understand correctly we dont have a full snapshot upfront and thus the race. If I remember correctly every `KUBERNETES_EXECUTOR_API_POLLING_INTERVAL` we do a re-sync but its too late, is that the problem you see? So having a lister without the pod watch we have now (part of what I propose is to remove it), it should not trigger that pod check with an inconsistent snapshot state because informers do the sync of their caches anyway behind the scenes before they start. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
