Matthew, can you share some logs ? That issue is quite generic, and it describes a case that usually works as expected (= system works fine)
It is very common to bounce a server in a cluster of three, so there must be a cause for your problem. In a 3 nodes clusters usually you can continue working until up to 2 servers are up and running. Enrico Il giorno lun 28 ott 2019 alle ore 19:55 Matthew Knight < matthewryankni...@gmail.com> ha scritto: > Hi, > > I am setting up a 3-node Zookeeper cluster and have encountered the issue > described in https://issues.apache.org/jira/browse/ZOOKEEPER-2164 > It seems like Zookeeper is unable to recover if the leader node shuts down > unexpectedly. Is this truly a general problem with Zookeeper? Is there a > way to configure/work around this issue besides restarting all of the > Zookeeper nodes in the cluster? > > Thanks, > Matthew Knight >