Hi all! One of our broker went out of sync for one partition in one topic and we were wondering how this can happen. The broker is still in the list of replica and registered in zookeeper with the right ip address.
I noticed that the offset in replication-offset-checkpoint is stuck to a certain value. Moreover there is a lot of traffic between the leader and this replica, so I guess it is still trying to synchronize with the leader. Is there a way to fix this error? Or at least to get more information about it? There is no specific errors in the broker and controller logs. We are using kafka 0.8.1.1 Thanks, Dorian
