[ https://issues.apache.org/jira/browse/KAFKA-17263?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17871560#comment-17871560 ]
Kuradeon commented on KAFKA-17263: ---------------------------------- It's fine after I made these changes in org.apache.kafka.clients.consumer.internals.AbstractCoordinator. Except FencedInstanceIdException, other exceptions in heartbeatThread always mean that the thread is end of loop, and will create a new heartbeatThread next time. But for static membership consumer, got FencedInstanceIdException wouldn't stop the current heartbeatThread, but just disable the heartbeat and continue then await(). I think there are two fix way, one is to initialize only one heartbeatThread, and make sure it's running. While the other way is like my fix (easier), it won't create a new heartbeatThread after got FencedInstanceIdException. !screenshot-3.png! !screenshot-4.png! !screenshot-5.png! > Sometimes KafkaConsumer.close will block the thread > --------------------------------------------------- > > Key: KAFKA-17263 > URL: https://issues.apache.org/jira/browse/KAFKA-17263 > Project: Kafka > Issue Type: Bug > Components: clients, consumer > Affects Versions: 2.8.2 > Environment: Java 8 > Reporter: Kuradeon > Priority: Major > Attachments: screenshot-1.png, screenshot-2.png, screenshot-3.png, > screenshot-4.png, screenshot-5.png > > > Sometimes, the thread will be blocked and never continue (state: waiting) > after invoke KafkaConsumer.close(). Here are the related thread dump: > {code:java} > Thread 73 (pool-5-thread-1): > State: WAITING > Blocked count: 25 > Waited count: 22 > Waiting on > org.apache.kafka.clients.consumer.internals.AbstractCoordinator$HeartbeatThread@565a2927 > Stack: > java.lang.Object.wait(Native Method) > java.lang.Thread.join(Thread.java:1257) > java.lang.Thread.join(Thread.java:1331) > > org.apache.kafka.clients.consumer.internals.AbstractCoordinator.closeHeartbeatThread(AbstractCoordinator.java:385) > > org.apache.kafka.clients.consumer.internals.AbstractCoordinator.close(AbstractCoordinator.java:1010) > > org.apache.kafka.clients.consumer.internals.ConsumerCoordinator.close(ConsumerCoordinator.java:926) > > org.apache.kafka.clients.consumer.KafkaConsumer.close(KafkaConsumer.java:2366) > > org.apache.kafka.clients.consumer.KafkaConsumer.close(KafkaConsumer.java:2333) > > org.apache.kafka.clients.consumer.KafkaConsumer.close(KafkaConsumer.java:2283) > Thread 687 (kafka-coordinator-heartbeat-thread | XXXX-st-c35): > State: WAITING > Blocked count: 0 > Waited count: 1 > Waiting on > org.apache.kafka.clients.consumer.internals.ConsumerCoordinator@51f1e632 > Stack: > java.lang.Object.wait(Native Method) > java.lang.Object.wait(Object.java:502) > > org.apache.kafka.clients.consumer.internals.AbstractCoordinator$HeartbeatThread.run(AbstractCoordinator.java:1360) > {code} -- This message was sent by Atlassian Jira (v8.20.10#820010)