Chris created ZOOKEEPER-4629: -------------------------------- Summary: Observers fail to restart when using quorum sasl Key: ZOOKEEPER-4629 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4629 Project: ZooKeeper Issue Type: Bug Affects Versions: 3.7.0 Reporter: Chris
Our observers have started failing to restart after introducing tls & sasl for quorum and client connections. We have narrowed the cause down to quorum sasl and disabling quorum sasl fixed the issue. I can recreate the incident as follows: # 6 server cluster of participants (this is our standard setup) # Reconfig server id 1 to observer ## id1 switches to observer mode and stays functional # restart id1 ## id1 starts and never manages to join the cluster. Note that it does not reach OBSERVING state. ## Subsequent restarts also fail # stop id=1 # (repair) dynamic reconfig id 1 to participant ## (You can try to start id 1 immediately, but it will not retrieve the new config and just get in the startup loop as before.) ## Give id 1 a zoo.cfg.dynamic.initial ## Start id 1 and it will connect, retrieve current config and function normally -- This message was sent by Atlassian Jira (v8.20.10#820010)