Apache9 commented on issue #437: HBASE-22539 WAL corruption due to early DBBs re-use when Durability.A… URL: https://github.com/apache/hbase/pull/437#issuecomment-518005747 > I still think we should release the BB in the exceptional case for AsyncFSWAL#appendAndSync, > > ``` > try { > appended = append(writer, entry); > } catch (IOException e) { > throw new AssertionError("should not happen", e); > } > ``` > > There're coprocessorHost.preWALWrite inside the append method, although the Assert message say it shouldn't happen, but user may write an incorrect CP, which would throw an IOException.. > Also, I think better to change the message "Shouldn't happen..." If this happens we just let the region server die, I think this is the correct way to deal with this. The design here can not deal with the IOException.
---------------------------------------------------------------- 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] With regards, Apache Git Services
