[
https://issues.apache.org/jira/browse/YARN-2946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14244050#comment-14244050
]
Rohith commented on YARN-2946:
------------------------------
Another deadlock detected in different flow after fixing previous deadlock :-(
Basically, by convention all the locks should acquire from
*StateMachine.doTransition() -> zkRMStateStore.class* or directly
zkRMStateStore.class , but in {{RMStateStore#isFencedState()}} method locking
order is reversed i.e *zkRMStateStore.class -> StateMachine.doTransition()*
{noformat}
Found one Java-level deadlock:
=============================
"org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore$VerifyActiveStatusThread":
waiting to lock monitor 0x0000000000e55698 (object 0x00000000c0272cb0, a
org.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine),
which is held by "AsyncDispatcher event handler"
"AsyncDispatcher event handler":
waiting to lock monitor 0x00000000013adcf8 (object 0x00000000c0272b10, a
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore),
which is held by
"org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore$VerifyActiveStatusThread"
Java stack information for the threads listed above:
===================================================
"org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore$VerifyActiveStatusThread":
at
org.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine.getCurrentState(StateMachineFactory.java:442)
- waiting to lock <0x00000000c0272cb0> (a
org.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.isFencedState(RMStateStore.java:693)
- locked <0x00000000c0272b10> (a
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore$VerifyActiveStatusThread.run(ZKRMStateStore.java:1020)
"AsyncDispatcher event handler":
at java.lang.Object.wait(Native Method)
- waiting on <0x00000000c0272b10> (a
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore$ZKAction.runWithCheck(ZKRMStateStore.java:1043)
- locked <0x00000000c0272b10> (a
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore$ZKAction.runWithRetries(ZKRMStateStore.java:1070)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.doMultiWithRetries(ZKRMStateStore.java:906)
- locked <0x00000000c0272b10> (a
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.doMultiWithRetries(ZKRMStateStore.java:920)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.createWithRetries(ZKRMStateStore.java:929)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore.storeApplicationStateInternal(ZKRMStateStore.java:608)
- locked <0x00000000c0272b10> (a
org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore$StoreAppTransition.transition(RMStateStore.java:146)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore$StoreAppTransition.transition(RMStateStore.java:131)
at
org.apache.hadoop.yarn.state.StateMachineFactory$SingleInternalArc.doTransition(StateMachineFactory.java:362)
at
org.apache.hadoop.yarn.state.StateMachineFactory.doTransition(StateMachineFactory.java:302)
at
org.apache.hadoop.yarn.state.StateMachineFactory.access$300(StateMachineFactory.java:46)
at
org.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine.doTransition(StateMachineFactory.java:448)
- locked <0x00000000c0272cb0> (a
org.apache.hadoop.yarn.state.StateMachineFactory$InternalStateMachine)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore.handleStoreEvent(RMStateStore.java:699)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore$ForwardingEventHandler.handle(RMStateStore.java:754)
at
org.apache.hadoop.yarn.server.resourcemanager.recovery.RMStateStore$ForwardingEventHandler.handle(RMStateStore.java:749)
at
org.apache.hadoop.yarn.event.AsyncDispatcher.dispatch(AsyncDispatcher.java:173)
at
org.apache.hadoop.yarn.event.AsyncDispatcher$1.run(AsyncDispatcher.java:106)
at java.lang.Thread.run(Thread.java:745)
Found 1 deadlock.
{noformat}
> Deadlock in ZKRMStateStore
> --------------------------
>
> Key: YARN-2946
> URL: https://issues.apache.org/jira/browse/YARN-2946
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.7.0
> Reporter: Rohith
> Assignee: Rohith
> Priority: Blocker
> Attachments: 0001-YARN-2946.patch, 0002-YARN-2946.patch,
> TestYARN2946.java
>
>
> Found one deadlock in ZKRMStateStore.
> # Initial stage zkClient is null because of zk disconnected event.
> # When ZKRMstatestore#runWithCheck() wait(zkSessionTimeout) for zkClient to
> re establish zookeeper connection either via synconnected or expired event,
> it is highly possible that any other thred can obtain lock on
> {{ZKRMStateStore.this}} from state machine transition events. This cause
> Deadlock in ZKRMStateStore.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
