[
https://issues.apache.org/jira/browse/HADOOP-15449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16466119#comment-16466119
]
Karthik Palanisamy commented on HADOOP-15449:
---------------------------------------------
[~arpitagarwal] Yes, it should re-connect. But Zookeeper already expires the
session because of the timeout (no heartbeat been received from ZK client
within session timeout). In this case, Znode lock could have acquired by
another ZKFC controller which eventually failover to us.
> Frequent Namenode Flipover affecting user Jobs.
> -----------------------------------------------
>
> Key: HADOOP-15449
> URL: https://issues.apache.org/jira/browse/HADOOP-15449
> Project: Hadoop Common
> Issue Type: Wish
> Components: common
> Affects Versions: 2.7.4
> Reporter: Karthik Palanisamy
> Assignee: Karthik Palanisamy
> Priority: Critical
> Attachments: HADOOP-15449.patch
>
>
> We observed from several users regarding Namenode flip-over is due to either
> zookeeper disk slowness (higher fsync cost) or network issue. We would need
> to avoid flip-over issue to some extent by increasing HA session timeout,
> ha.zookeeper.session-timeout.ms.
> Default value is 5000 ms, seems very low in any production environment. I
> would suggest 10000 ms as default session timeout.
>
> {code}
> ..
> 2018-05-04 03:54:36,848 INFO zookeeper.ClientCnxn
> (ClientCnxn.java:run(1140)) - Client session timed out, have not heard from
> server in 4689ms for sessionid 0x260e24bac500aa3, closing socket connection
> and attempting reconnect
> 2018-05-04 03:56:49,088 INFO zookeeper.ClientCnxn
> (ClientCnxn.java:run(1140)) - Client session timed out, have not heard from
> server in 3981ms for sessionid 0x360fd152b8700fe, closing socket connection
> and attempting reconnect
> ..
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
