[ https://issues.apache.org/jira/browse/KAFKA-6348?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
sikaijian updated KAFKA-6348: ----------------------------- Comment: was deleted (was: I got the same problem. In my case, I use kafka for logs. So I killed one broker. And it worked. But when I start the broker, I got failure agin. {code:java} 2018-02-27 11:09:07 DEBUG kafka-consumer-ElasticsearchStorage - Issuing group metadata request to broker 11 2018-02-27 11:09:07 DEBUG kafka-consumer-ElasticsearchStorage - Group metadata response ClientResponse(receivedTimeMs=1519700947797, disconnected=false, request=ClientRequest(expectResponse=true, callback=org.apache.kafka.clients.consumer.internals.ConsumerNetworkClient$RequestFutureCompletionHandler@4afb1ca2, request=RequestSend(header={api_key=10,api_version=0,correlation_id=196,client_id=consumer-1}, body={group_id=ElasticsearchStorage-group}), createdTimeMs=1519700947796, sendTimeMs=1519700947796), responseBody={error_code=0,coordinator={node_id=4,host=kafkabroker5.pajkdc.com,port=9092}}) 2018-02-27 11:09:07 DEBUG kafka-consumer-ElasticsearchStorage - (Re-)joining group ElasticsearchStorage-group 2018-02-27 11:09:07 DEBUG kafka-consumer-ElasticsearchStorage - Issuing request (JOIN_GROUP: {group_id=ElasticsearchStorage-group,session_timeout=30000,member_id=,protocol_type=consumer,group_protocols=[{protocol_name=range,protocol_metadata=java.nio.HeapByteBuffer[pos=0 lim=32 cap=32]}]}) to coordinator 2147483643 2018-02-27 11:09:07 INFO kafka-consumer-ElasticsearchStorage - Marking the coordinator 2147483643 dead. 2018-02-27 11:09:07 INFO kafka-consumer-ElasticsearchStorage - Attempt to join group ElasticsearchStorage-group failed due to obsolete coordinator information, retrying. {code} The coordinator for consumer group "ElasticsearchStorage-group" is broker 4. I just want to change the coordinator. So I killed it. As I know, only these brokers who own the partition of __consumer_offset can work as a coordinator for consumer group. And a consumer group's offset locate on only one partition of __consumer_offset. And in my case Math.abs("ElasticsearchStorage-group".hashCode()) % 100(partitionCount of __consumer_offset, in my case is 100) = 4. But I consume 3 topics using the same consumer group "ElasticsearchStorage-group". Only on topic met the problem. I still can't figure out the root cause. I am using 0.9.0.1, and the new consumer. ) > Kafka consumer can't restore from coordinator failure > ----------------------------------------------------- > > Key: KAFKA-6348 > URL: https://issues.apache.org/jira/browse/KAFKA-6348 > Project: Kafka > Issue Type: Bug > Components: consumer, core > Affects Versions: 0.10.1.1 > Reporter: Renjie Liu > Priority: Major > > Kafka consumer blocks and keep reporting coordinator is dead. I tried to > restart the process and it still can't work. Then we shutdown the broker and > restart consumer, but it still keep reporting coordinator is dead. This > situation continues until we change our group id and it works. -- This message was sent by Atlassian JIRA (v7.6.3#76005)