Please suggest on how to avoid restart of Kafka Consumer Client on one of the broker fail down in the cluster. Or do we need to update the offset manually if we put back the failed broker in the cluster.
On 5 January 2018 at 19:18, rAhul <c.rahulku...@gmail.com> wrote: > Hi, > > I have a Apache kafka cluster with 3 nodes(say 1,2,3) with replication > factor of 3 and partitions as 3. > > When my producer client, consumer client and the cluster are running, able > to transfer messages from producer to consumer without any issues. > > Now I stopped leader node say node 1 from the cluster and now say node 2 > is promoted as leader. > > Message flow from producer to consumer works fine without any issues. > > Now I started node 1 and stopped node 2, either node 1 or node 3 is > promoted as leader. > > Now producer able to send messages but consumer not able to receive > messages. > > I see consumer lag using kafka manager web console. > > Again if I start node 2, consumer able to receive messages. > > Please suggest how to overcome this issue and fix it. > > Thanks. >
