Apache9 commented on pull request #2824: URL: https://github.com/apache/hbase/pull/2824#issuecomment-752297484
> > In general, if there is an OOME, we can not assume that our handler can still be executed correctly, as we do not know where is the OOME thrown out, it may break the assumptions in our code and cause unexpected behavior. > > IIRC, we have '-XX:OnOutOfMemoryError="kill -9 %p"' in our command line when start daemons? It does not work in your deployment? > > Thanks. > > and in our env, when OOME occurs, -XX:OnOutOfMemoryError="kill -9 %p" exists, but active master is not stopped. > https://github.com/apache/hbase/blob/c96fbf04077fa37555033f88939fdd69ac810b35/bin/hbase#L811 So we need to find out why this opinion does not work first? OOME can happen at any place and break the assumptions in code, there is no way to fix it cleanly just by adding more checks in code... ---------------------------------------------------------------- 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: [email protected]
