Hello, I've noticed that when a broker in the target cluster restarts (i.e. immediately comes back online), the MirrorMaker runs into a FATAL exception but the process hangs indefinitely instead of exiting. This seems to me like undesirable behavior but I'm not sure that it's a bug. I'd like for the MirrorMaker process to exit so that it can be restarted, since it is run under supervision.
We're on Kafka 0.7.2, and this MirrorMaker uses the sync producer and is configured with a broker list. What I typically find in the logs is something like this: [2013-07-02 20:56:35,431] 99283617 [mirrormaker-0] INFO kafka.producer.SyncProducer - Disconnecting from localhost:9094 [2013-07-02 20:56:35,434] 99283620 [mirrormaker-0] FATAL kafka.tools.MirrorMaker$MirrorMakerThread - %s stream unexpectedly exited. java.io.IOException: Broken pipe ... [2013-07-02 20:56:35,469] 99283655 [mirrormaker-0] INFO kafka.tools.MirrorMaker$MirrorMakerThread - Stopped thread mirrormaker-0. and then the process hangs! Only when I kill it does the MirrorMaker complete the shutdown process, e.g.: [2013-07-02 22:34:53,478] 508405 [Thread-2] INFO kafka.consumer.ZookeeperConsumerConnector - MirrorMaker_zkkafka-0.uswest2n.yb0t.com-1372803985231-10bd3698 ZKConsumerConnector shutting down [2013-07-02 22:34:53,479] 508406 [Thread-2] INFO kafka.consumer.ZookeeperTopicEventWatcher - Shutting down topic event watcher. [2013-07-02 22:34:53,479] 508406 [ZkClient-EventThread-19-localhost:2181] INFO org.I0Itec.zkclient.ZkEventThread - Terminate ZkClient event thread. [2013-07-02 22:34:53,482] 508409 [main-EventThread] INFO org.apache.zookeeper.ClientCnxn - EventThread shut down [2013-07-02 22:34:53,482] 508409 [Thread-2] INFO org.apache.zookeeper.ZooKeeper - Session: 0x13ea3d122570085 closed [2013-07-02 22:34:53,484] 508411 [Thread-2] INFO kafka.utils.KafkaScheduler - force shutdown scheduler Kafka-consumer-autocommit- [2013-07-02 22:34:53,485] 508412 [FetchRunnable-0] INFO kafka.consumer.FetcherRunnable - FecherRunnable Thread[FetchRunnable-0,5,main] interrupted [2013-07-02 22:34:53,486] 508413 [FetchRunnable-0] INFO kafka.consumer.FetcherRunnable - stopping fetcher FetchRunnable-0 to host zkkafka-0.uswest2n.yb0t.com [2013-07-02 22:34:53,490] 4254397417 [kafka-processor-0] INFO kafka.network.Processor - Closing socket connection to /10.249.66.10. [2013-07-02 22:34:53,499] 508426 [ZkClient-EventThread-12-localhost:2181] INFO org.I0Itec.zkclient.ZkEventThread - Terminate ZkClient event thread. [2013-07-02 22:34:53,503] 508430 [Thread-2] INFO org.apache.zookeeper.ZooKeeper - Session: 0x13ea3d122570084 closed [2013-07-02 22:34:53,503] 508430 [main-EventThread] INFO org.apache.zookeeper.ClientCnxn - EventThread shut down [2013-07-02 22:34:53,504] 508431 [Thread-2] INFO kafka.consumer.ZookeeperConsumerConnector - MirrorMaker_zkkafka-0.uswest2n.yb0t.com-1372803985231-10bd3698 ZKConsumerConnector shut down completed [2013-07-02 22:34:53,505] 508432 [Thread-2] INFO kafka.producer.ProducerPool - Closing all sync producers [2013-07-02 22:34:53,505] 508432 [Thread-2] INFO kafka.producer.SyncProducer - Disconnecting from localhost:9094 Suggestions appreciated. Thanks in advance! Homer
