hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Varun Saxena (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-5874) RM -format-state-store and -remove-application-from-state-store commands fail with NPE
Date Sat, 12 Nov 2016 18:31:58 GMT

    [ https://issues.apache.org/jira/browse/YARN-5874?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15660056#comment-15660056
] 

Varun Saxena commented on YARN-5874:
------------------------------------

Thanks [~rohithsharma] for the review.

bq. Did you test with this patch? 
Yes I tested it in an unsecure cluster with default ZK ACLs'.

bq. Any test can be added for this?
The code added is in a static method. I will have to use Powermockito to simulate it and RM
pom doesnt even have dependency for it. Do you want me to add ?

bq.  I think RM object should be initialized. 
Only bit of ResourceManager#init required in ZKRMStateStore is related to creation of curator.
So initialization of all other services by calling ResourceManager#serviceInit is unnecessary
here. And in ZKRMStateStore#initiInternal, we have code as under. So if curatorFramework is
not already initialized by RM init, it will be initialized here. So IMO just creating the
object should be enough. Thoughts ?
{code}
    curatorFramework = resourceManager.getCurator();
    if (curatorFramework == null) {
      curatorFramework = resourceManager.createAndStartCurator(conf);
    }
{code}


> RM -format-state-store and -remove-application-from-state-store commands fail with NPE
> --------------------------------------------------------------------------------------
>
>                 Key: YARN-5874
>                 URL: https://issues.apache.org/jira/browse/YARN-5874
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: resourcemanager
>            Reporter: Varun Saxena
>            Assignee: Varun Saxena
>            Priority: Critical
>         Attachments: YARN-5874.01.patch
>
>
> {noformat}
> 16/11/12 13:29:29 INFO resourcemanager.ResourceManager: registered UNIX signal handlers
for [TERM, HUP, INT]
> 16/11/12 13:29:29 INFO recovery.RMStateStoreFactory: Using RMStateStore implementation
- class org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
> 16/11/12 13:29:30 INFO event.AsyncDispatcher: Registering class org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStoreEventType
for class org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore$ForwardingEventHandler
> 16/11/12 13:29:30 INFO service.AbstractService: Service org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore
failed in state INITED; cause: java.lang.NullPointerException
> java.lang.NullPointerException
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.constructZkRootNodeACL(ZKRMStateStore.java:200)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.initInternal(ZKRMStateStore.java:236)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.serviceInit(RMStateStore.java:682)
> 	at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.deleteRMStateStore(ResourceManager.java:1447)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.main(ResourceManager.java:1374)
> 16/11/12 13:29:30 INFO event.AsyncDispatcher: AsyncDispatcher is draining to stop, ignoring
any new events.
> 16/11/12 13:29:30 FATAL resourcemanager.ResourceManager: Error starting ResourceManager
> java.lang.NullPointerException
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.constructZkRootNodeACL(ZKRMStateStore.java:200)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.initInternal(ZKRMStateStore.java:236)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.serviceInit(RMStateStore.java:682)
> 	at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.deleteRMStateStore(ResourceManager.java:1447)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.main(ResourceManager.java:1374)
> 16/11/12 13:29:30 INFO resourcemanager.ResourceManager: SHUTDOWN_MSG: 
> {noformat}
> {noformat}
> 16/11/12 16:58:58 INFO resourcemanager.ResourceManager: registered UNIX signal handlers
for [TERM, HUP, INT]
> 16/11/12 16:58:59 INFO recovery.RMStateStoreFactory: Using RMStateStore implementation
- class org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore
> 16/11/12 16:58:59 INFO event.AsyncDispatcher: Registering class org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStoreEventType
for class org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore$ForwardingEventHandler
> 16/11/12 16:58:59 INFO service.AbstractService: Service org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore
failed in state INITED; cause: java.lang.NullPointerException
> java.lang.NullPointerException
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.constructZkRootNodeACL(ZKRMStateStore.java:200)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.initInternal(ZKRMStateStore.java:236)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.serviceInit(RMStateStore.java:682)
> 	at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.removeApplication(ResourceManager.java:1462)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.main(ResourceManager.java:1377)
> 16/11/12 16:58:59 INFO event.AsyncDispatcher: AsyncDispatcher is draining to stop, ignoring
any new events.
> 16/11/12 16:58:59 FATAL resourcemanager.ResourceManager: Error starting ResourceManager
> java.lang.NullPointerException
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.constructZkRootNodeACL(ZKRMStateStore.java:200)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.initInternal(ZKRMStateStore.java:236)
> 	at org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.serviceInit(RMStateStore.java:682)
> 	at org.apache.hadoop.service.AbstractService.init(AbstractService.java:163)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.removeApplication(ResourceManager.java:1462)
> 	at org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.main(ResourceManager.java:1377)
> 16/11/12 16:58:59 INFO resourcemanager.ResourceManager: SHUTDOWN_MSG: 
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org


Mime
View raw message