[
https://issues.apache.org/jira/browse/HDFS-10373?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Nauroth resolved HDFS-10373.
----------------------------------
Resolution: Invalid
Hello [~piaoyu zhang]. This doesn't look like a bug. If ZKFC cannot contact
its peer NameNode for a successful health check RPC, then an HA failover is the
expected behavior. This looks like an operational problem in this environment
that needs further investigation. "Connection reset by peer" means the remote
end (the NameNode) closed out the socket before sending the expected response
data. I recommend looking at the NameNode logs to see if anything unusual
happened during the timeframe of the HA failover. If you need further
assistance, then consider sending an email to u...@hadoop.apache.org. I hope
this helps.
> HDFS ZKFC HealthMonitor Throw a Exception Cause AutoFailOver
> ------------------------------------------------------------
>
> Key: HDFS-10373
> URL: https://issues.apache.org/jira/browse/HDFS-10373
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: auto-failover
> Affects Versions: 2.2.0
> Environment: CentOS6.5 Hadoop-2.2.0
> Reporter: zhangyubiao
> Attachments: screenshot-1.png, 屏幕快照_2016-05-06_上午10.17.22.png
>
>
> HDFS ZKFC HealthMonitor Throw a Exception
> 2016-05-05 02:00:59,475 WARN org.apache.hadoop.ha.HealthMonitor:
> Transport-level exception trying to monitor health of NameNode at
> XXX-XXX-XXX-hadoop.jd.local/172.22.17
> 1.XX:8021: Failed on local exception: java.io.IOException: Connection reset
> by peer; Host Details : local host is:
> "XXX-XXX-XXX-hadoop.jd.local/172.22.171.XX"; destinat
> ion host is: XXX-XXX-XXX-hadoop.jd.local":8021;
> Cause HA AutoFailOver
--
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
