zhuzhurk commented on pull request #15229: URL: https://github.com/apache/flink/pull/15229#issuecomment-900955843
@Thesharing thanks for the comments and the one about the complexity of `DefaultScheduler#getAllocationsToReserve` is really good catch. Need to mention the problem also happens to streaming jobs with multiple regions. I did experiments with a job which only has 10000 source tasks. The scheduling time(first SCHEDULED to last SCHEDULED) increased to 20s. I think the time cost is not acceptable. I will think about how we can reduce the complexity by maintaining the allocations to reserve in ahead and update here when there is a clear idea. -- 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]
