Apache9 commented on pull request #1593: URL: https://github.com/apache/hbase/pull/1593#issuecomment-624396591
> Ya, my question was about the word "retries". That will show up in the exception message (IIUC). These give a false impression that things are being retried, right (when in reality they are hedged, in some sense). I think query multiple masters could also be some types of retry? If this master is not available, we just try another, and if all are failed we will return the exception to you. > I was thinking about Call#setException(), which cleans up the caller state and propagates the exception to the the future callback? For example, if a master is hung and the RPC is hung, that state would be cleaned up quicker. Not a big deal? Only a small amount of memory. Implementing this will lead to a more complicated logic, we have to reference the CompletableFutures for other requests and call cancel on them, and we also need to deal with the cancel event and do the clean up work. Comparing to the benefits, I do not think it worths... ---------------------------------------------------------------- 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: us...@infra.apache.org
