[
https://issues.apache.org/jira/browse/HADOOP-7455?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13078606#comment-13078606
]
Justin Joseph commented on HADOOP-7455:
---------------------------------------
@Aaron : When an Active Namenode lose the connection to Zookeeper ensemble, it
cannot remain in the Active state. One option is to transition to Standby
state, but it is unnecessary, the same Namenode may be asked to become Active
once the connection with Zookeeper is established again. The second option, the
better one, is to move to an idle state - neutral state..
Moving from Active to Standby involves many complexities, at least in Backup
Node based HA. First the Active state needs to be cleaned up; it should
transition to Standby; it should connect to Active Namenode & then trigger a
check pointing. This is purely an overhead considering the fact that the same
Namenode may be asked to become Active again soon.
Neutral state ensures that a Namenode does not act hastily when it loses the
connection to Zookeeper ensemble. Rather it just 'wait & watch'.
> Introduce HA Service Protocol Interface
> ---------------------------------------
>
> Key: HADOOP-7455
> URL: https://issues.apache.org/jira/browse/HADOOP-7455
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: util
> Reporter: Suresh Srinivas
> Assignee: Suresh Srinivas
> Attachments: HADOOP-7455.1.patch, HADOOP-7455.2.patch,
> HADOOP-7455.3.patch, HADOOP-7455.4.patch, HDFS-7454.patch
>
>
> This jira introduces a protocol interface to be implemented by services that
> provide HA functionality.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
