[ https://issues.apache.org/jira/browse/HDFS-5832?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13881927#comment-13881927
]
Rakesh R commented on HDFS-5832:
--------------------------------
Thanks Uma for the interest.
Looks like similar, but I could see still there is a chance of reverse locking in the code.
I'm seeing the following code flow like,
- In one side, when running SafeModeMonitor thread, it is invoking safeMode#canLeave() by
acquiring 'safeMode.this' lock and here it is asking for the DataNodeManger#getNumLiveDataNodes
by trying to acquire 'datanodeMap' lock.
- On the other side, DN heartbeat comes and DatanodeManager#handleHeartbeat() has acquired
'datanodeMap' lock and calling namesystem#isInSafeMode(), which inturn trying to acquire 'safeModeInfo.this'
lock
> Deadeadlock found in NN between SafeMode#canLeave and DatanodeManager#handleHeartbeat
> -------------------------------------------------------------------------------------
>
> Key: HDFS-5832
> URL: https://issues.apache.org/jira/browse/HDFS-5832
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: namenode
> Affects Versions: 3.0.0
> Reporter: Rakesh R
> Assignee: Rakesh R
> Priority: Blocker
> Attachments: HDFS-5832.patch, jcarder_nn_deadlock.gif
>
>
> Found the deadlock during the Namenode startup. Attached jcarder report which shows the
cycles about the deadlock situation.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
|