[
https://issues.apache.org/jira/browse/HBASE-1018?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12650285#action_12650285
]
stack commented on HBASE-1018:
------------------------------
Patch looks great.
I suppose you don't need to up the protocol version since its just carrying a
ServerLoad instance still -- its just that ServerLoad now carries more stuff.
You're using mbeans to figure JVM heap size. Why not
Runtime.getRuntime().maxMemory() and totalMemory()? Seems like you need to
bring in less 'stuff'.
> Regionservers should report detailed health to master; master should flag
> troubled regionservers in UI
> ------------------------------------------------------------------------------------------------------
>
> Key: HBASE-1018
> URL: https://issues.apache.org/jira/browse/HBASE-1018
> Project: Hadoop HBase
> Issue Type: Improvement
> Components: master, regionserver
> Reporter: Andrew Purtell
> Assignee: Andrew Purtell
> Fix For: 0.19.0
>
> Attachments: 1018.patch
>
>
> Regionservers should report detailed health to master. The master should flag
> troubled regionservers in the UI.
> The concern at the moment is primarily heap. Regionservers should report
> used, committed, and max heap metrics in the periodic report. The master
> should flag in the regionserver list on /master.jsp those regionservers where
> available heap is below a configurable threshold.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
