[
https://issues.apache.org/jira/browse/HBASE-15318?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15163158#comment-15163158
]
Matteo Bertozzi commented on HBASE-15318:
-----------------------------------------
I think [~zhangyongxyz] is talking about the rs znode used to add the server in
the DeadServer list and not the assignment znode.
I'm changing ServerManager to add more information like RS VersionNumber,
reduce the number of Maps we have around and avoid the
array.contains()/array.remove() instead of set.contains()/remove() we have
around serverManager and balancer. there were discussion about using heartbeat
to identify if a server is alive but at the moment we are not focused on that.
so, you can go on and implement it. I assume is not invasive and you are going
to replace just how we add the rs to the dead server list
> Zk-less region server state management
> --------------------------------------
>
> Key: HBASE-15318
> URL: https://issues.apache.org/jira/browse/HBASE-15318
> Project: HBase
> Issue Type: Improvement
> Reporter: Yong Zhang
> Assignee: Yong Zhang
>
> Current region server state is managed via znode created by region server,
> master just listen these nodes. but we find many user network issue is not
> network disconnected but package lost, which is hard to capture because
> connection between region server and zk is fine.
> This jira goal is region server state is managed by master without shared
> info in zk, via enhancement heartbeat from region server to master.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
