Hi Team,

The official document indicates that leader node election will block write
operations, as the leader is responsible for managing them. However, in my
ZooKeeper cluster deployed on Kubernetes using a StatefulSet, I found that
after restarting the leader pod, the leader election process began.
During this time, I encountered the following error logs:

* java.io.IOException: ZooKeeperServer not running.*

It appears that the entire ZooKeeper cluster, including follower pods,
becomes unavailable during the leader election and only resumes normal
operations once a new leader is elected. I would like to check with the
team to see if this behavior is expected.

Thanks,
Yifei

Reply via email to