I ran into kind of a similar discrepancy, but about UnderReplicatedPartitions.
kafka-topics.sh and zookeeper were saying that we had underreplicated partitions. But JMX said that there were none. I took one of the partitions that ZK was saying was under-replicated and I ran DumpLogSegments on the leader broker and the broker that was out of the ISR, and it showed that the partition was fully replicated to the follower. So JMX+DumpLogSegments disagreed with kafka-topics.sh+zookeeper. So that is the opposite scenario of yours. Mine was on 0.9.0. -James > On Feb 9, 2016, at 12:02 PM, Eric Ogren <eog...@gmail.com> wrote: > > Hello - > > I am seeing that the node in our Kafka cluster currently elected as > controller is reporting 1 offline partition via JMX > (kafka.controller:type=KafkaController,name=OfflinePartitionsCount). > However, when I use kaka-topics to find the offline partition > (bin/kafka-topics.sh > --zookeeper localhost:2181 --unavailable-partitions --describe), nothing is > dumped out. We also haven't been seeing any errors around unavailable > partitions in any of our producers. > > Is there a good way to introspect the controller's state and figure out > what it thinks the offline partition is? > > This is for Kafka 0.8.2.1. > > thanks > Eric ________________________________ This email and any attachments may contain confidential and privileged material for the sole use of the intended recipient. Any review, copying, or distribution of this email (or any attachments) by others is prohibited. If you are not the intended recipient, please contact the sender immediately and permanently delete this email and any attachments. No employee or agent of TiVo Inc. is authorized to conclude any binding agreement on behalf of TiVo Inc. by email. Binding agreements with TiVo Inc. may only be made by a signed written agreement.
