1. It seems that you have 3 replicas and only 2 of them are caught up. Is broker 2 still alive?
2. We have jmx beans that measure the lag per follower replica. Updated http://kafka.apache.org/documentation.html#monitoring with this info. Thanks, Jun On Tue, Oct 22, 2013 at 3:24 PM, Vadim Keylis <[email protected]> wrote: > Good afternoon. I have a few questions about kafka monitoring. > > 1. In the kafka monitoring section ( > http://kafka.apache.org/documentation.html#monitoring) > <http://kafka.apache.org/documentation.html#monitoring > >"kafka.server":name="UnderReplicatedPartitions", > type="ReplicaManager" sets normal value to be as 0. I check jmx and my > value is set to 1. I am pretty sure I set replication factor to be as 1. > /usr/local/kafka/bin/kafka-list-topic.sh --zookeeper > zoo01.tagged.com:2181/kafka > topic: meetme partition: 0 leader: 1 replicas: 2,1,3 isr: 1,3 > > Would you please advise? > > 2. How programmatically know that partition has come up online and caught > up on replication after kafka service has been started. > > Thanks, > Vadim >
