[
https://issues.apache.org/jira/browse/HDFS-11180?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15702354#comment-15702354
]
Kihwal Lee commented on HDFS-11180:
-----------------------------------
bq. NameNode holds a lock of FSEditLog and requires a lock of MetricsSystemImpl
when registering IPCLoggerChannel metrics.
It looks like this deadlock can happen only when the QJM is used.
Metrics update does not need a precise txid. We could introduce unsynchronized
methods for metrics and perhaps use volatile for txid?
> Intermittent deadlock in NameNode when failover happens.
> --------------------------------------------------------
>
> Key: HDFS-11180
> URL: https://issues.apache.org/jira/browse/HDFS-11180
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 2.6.0
> Reporter: Abhishek Modi
> Labels: high-availability
> Attachments: HDFS-11180.00.patch, HDFS-11180.01.patch, jstack.log
>
>
> It is happening due to metrics getting updated at the same time when failover
> is happening. Please find attached jstack at that point of time.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
