zuston commented on issue #263: URL: https://github.com/apache/incubator-uniffle/issues/263#issuecomment-1306994949
I think the shuffle-server will follow the next steps when upgrading ### Dump and stop 1. Stop the grpc server to avoid extra data sent or got (client will have to hold on waiting and retry due to the connection refused failure) 2. Make its state unhealthy, which should be recognized by coordinator to avoid extra assignments 3. Dumping all state to file ### Restore 1. Restore state from the file 2. Start the grpc server to be in service I want to use kryo to serialize the partial state into file, and then recover from it. Do you have some good ideas on it? @jerqi -- 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]
