wilfred-s commented on PR #1074: URL: https://github.com/apache/yunikorn-k8shim/pull/1074#issuecomment-5404332485
> Can you also check the `placeholder` & `dispatcher` stop operation and ensure we are doing right job? The placeholder code has issue, we need a new jira to cleanup. There are tiny race conditions as we do a check and set of the atomic in two separate calls, should use `mgr.running.CompareAndSwap(false, true)` etc. It also does not set running to false on shutdown. The dispatcher has a similar check and set issue and maybe more. @andyhuangdev could you file follow up jiras: one for the placeholder and one for the dispatcher to clean all that up? > Also, Why `apifactory` is not being stopped? I can see `start() `being triggerred from `scheduler` but not stop from nowhere. @tigerquoll Any idea why is this not caught through leaks detector? That is part of the informers and they are known to leak go routines. They are excluded from the checks as we can not fix that. That said: we should call `ss.apiFactory.Stop()` as soon as part of the `KubernetesShim.Stop()` same way as we do the dispatcher and the phManager to get that done. @andyhuangdev Please add that one line to the change -- 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]
