Partitions on 3rd nide was not replicated to other nodes. Those partitions are having only one copy. If the node goes down then it will be ended of having data loss.
This is not an expected behavior. I think you may need to check replica fetcher threads to see why partitions are not getting replicated. Can you confirm that ISRon node never increased? On Jan 24, 2016 12:33 PM, "Jason J. W. Williams" <jasonjwwilli...@gmail.com> wrote: > Hi Guys, > > I've set up a test cluster using 3 Vagrant nodes each running a Kafka > broker and zookeeper instance. Zookeepers are all linked properly and the > brokers appear to be registering properly. > > > After creating a new topic with 128 partitions and a replication factor of > 3 (https://gist.github.com/williamsjj/e52186da1ca708cf4f05), initially all > of the replicas show up under ISR...but within 30 seconds the 3rd node has > dropped from about 20% of the partitions: > > https://gist.github.com/williamsjj/7fdb931339263af5f401 > > Is that expected? > > -J >
