vincentpoon commented on PR #5874: URL: https://github.com/apache/hbase/pull/5874#issuecomment-2095068099
> It is just a bit strange that in ZKUtil we eat the NoNodeException while at the upper layer we create a new NoNodeException and throw it... I think the original author of `getMetaReplicaNodesAndWatchChildren` never envisioned base znode not existing. If there were a `ZKUtil` method that didn't swallow the exception, we would use that instead. IMO `getMetaReplicaNodesAndWatchChildren` is doing something very specific and should in fact throw in that situation. But happy to change `ZKWatcher` to return null and have `MetaRegionLocationCache` stay in the loop when null if folks are in agreement here that that's the best path forward. Note that this would mean that before `/hbase/meta-region-server` gets created, all Regionservers in cluster would be retrying listChildren calls against ZK, instead of waiting for watcher notifications. Whereas with the original patch, all the behavior stays the same except under the race condition where base `/hbase` znode hasn't been created yet. -- 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: issues-unsubscr...@hbase.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org