caishunfeng commented on pull request #6697: URL: https://github.com/apache/dolphinscheduler/pull/6697#issuecomment-962417376
> This problem occurs when all nodes are deployed on k8s, and all registered addresses are svc hostname.This problem occurs when worker pods have some problem and frequently restart. At this time, the underlying ip of the same svc host has changed, but the channel in the cache is still connected to the previous pods.So here only need to clean up when the sending fails and exceed the limit, next time you will go to the channel creation logic again, at this time, the svc will be parsed into the pods ip of the latest worker. Hi @Bulls1986, when the host change, it will get the new channel from channel cache `NettyRemotingClient.getChannel`. Why it still connected to the previous pod host? -- 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]
