Hi,
I'm running a Kafka cluster with 3 nodes.
I have a topic with a replication factor of 3.
When I stop node 1 running kafka-topics.sh shows me that node 2 and 3 have
successfully failed over the partitions for the topic.
The message producers are still sending messages and I can still consume
them using kafka-console-consumer.sh
However my 0.9.0.1 Java consumer stops consuming messages. ( I thought
maybe it was rebalancing but after an hour nothing)
The consumer logs simply contains lots of "Marking the coordinator
8147683647 dead."
If I start node 1 up the consumers detect this and start to consume
messages again.
On startup of the consumers I can see that all three nodes are listed in
the "bootstrap.servers".
Any idea whats wrong with the Java consumer?
Thanks for any help or pointers in the right direction.
Michael