fox created YARN-8700:
-------------------------
Summary: Application cannot un-registered
Key: YARN-8700
URL: https://issues.apache.org/jira/browse/YARN-8700
Project: Hadoop YARN
Issue Type: Bug
Components: resourcemanager
Affects Versions: 2.7.3
Reporter: fox
Dear all,
I found a problem with application unregistration in AWS EMR environment
(emr-5.8.0, hadoop 2.7.3, spark 2.2.0).
Application Type: Both Yarn and Spark
State: RUNNING
Inside the job logs, I got
07:00:07.190 [main] INFO c.w.c.e.a.n.b.AbstractNormalBatchMain - [EDP2] Ready
to run Tear Down
07:00:07.192 [main] INFO c.w.c.e.a.n.b.AbstractNormalBatchMain - [EDP2] Ready
to run Tear Down
07:00:07.192 [main] INFO c.w.c.e.a.n.b.AbstractNormalBatchMain - [EDP2] Job
Finish
07:00:07.195 [main] INFO o.s.c.a.AnnotationConfigApplicationContext - Closing
org.springframework.context.annotation.AnnotationConfigApplicationContext@144ab54:
startup date [Tue Aug 21 06:59:23 UTC 2018]; root of context hierarchy
07:00:07.306 [main] INFO o.s.s.c.ThreadPoolTaskExecutor - Shutting down
ExecutorService 'redisClusterExecutor'
07:00:07.551 [main] INFO o.a.k.clients.producer.KafkaProducer - Closing the
Kafka producer with timeoutMillis = 9223372036854775807 ms.
07:00:07.565 [main] INFO c.w.c.f.m.MessageQueueKafkaProducerImpl - Closed all
the producer's connections for tenant: 7fd0356c-1258-11e8-abfd-0242ac110002.
07:00:09.869 [main] INFO c.w.c.edp2.normal.batch.AppMaster - finish run main
method
07:00:09.870 [main] INFO c.w.c.edp2.normal.batch.AppMaster - delete temp file
/tmp/aa33f388-f591-44a8-9aa3-13e2f8427c5d2802069659156113885.jar
07:00:10.112 [main] INFO o.a.h.y.c.api.impl.AMRMClientImpl - Waiting for
application to be successfully unregistered.
07:00:10.215 [main] INFO o.a.h.y.c.api.impl.AMRMClientImpl - Waiting for
application to be successfully unregistered.
07:00:10.319 [main] INFO o.a.h.y.c.api.impl.AMRMClientImpl - Waiting for
application to be successfully unregistered.
07:00:10.422 [main] INFO o.a.h.y.c.api.impl.AMRMClientImpl - Waiting for
application to be successfully unregistered.
07:00:10.528 [main] INFO o.a.h.y.c.api.impl.AMRMClientImpl - Waiting for
application to be successfully unregistered.
and it keeps more than one day until I stopped the whole cluster.
I also try to kill the application by yarn command, which also keeps forever
waiting for application to be killed.
hadoop@ip-10-100-2-124 ~]$ yarn application -kill application_1534810852740_0721
18/08/22 12:24:28 INFO impl.TimelineClientImpl: Timeline service address:
http://ip-10-100-2-124.ap-northeast-1.compute.internal:8188/ws/v1/timeline/
18/08/22 12:24:29 INFO client.RMProxy: Connecting to ResourceManager at
ip-10-100-2-124.ap-northeast-1.compute.internal/10.100.2.124:8032
Killing application application_1534810852740_0721
18/08/22 12:24:32 INFO impl.YarnClientImpl: Waiting for application
application_1534810852740_0721 to be killed.
18/08/22 12:24:34 INFO impl.YarnClientImpl: Waiting for application
application_1534810852740_0721 to be killed.
18/08/22 12:24:36 INFO impl.YarnClientImpl: Waiting for application
application_1534810852740_0721 to be killed.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
