sandeepvinayak opened a new pull request #2960: URL: https://github.com/apache/hbase/pull/2960
On hbase remove peer, znodes are deleted for peer and then zk listeners do rest of the work: 1. Terminate the replication source 2. Close the queues and delete znodes for them. However, if the zk listerner somehow do not get invoked (may be a zk issue), all the region servers can crash while delete the znode for queue with `KeeperException.NoNodeException`. This could bring all the RSes down if they are participating in replication. Also, there can be race condition between queue znode delete from replication source vs replication source termination through zk listerner. In this patch, we are not crashing the region server on `NoNodeException` but instead, we can look at if the replication peer does not exist, we can terminate the source. ---------------------------------------------------------------- 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]
