[
https://issues.apache.org/jira/browse/MAPREDUCE-3703?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13190250#comment-13190250
]
Hitesh Shah commented on MAPREDUCE-3703:
----------------------------------------
Would it be better to just provide a count of nodes in the jmx output and have
additional webservice apis ( with appropriate filter/pagination params to
reduce the size of returned data ) for querying the RM for all the
active/unhealthy/rebooted/... nodes?
I am aware that the jobtracker does emit this information today but on a large
cluster, the amount of data being sent back in the json output could be
potentially huge ( and a performance issue if any rogue layer pings the jmx
servlet too many times)?
> ResourceManager should provide node lists in JMX output
> -------------------------------------------------------
>
> Key: MAPREDUCE-3703
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3703
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 0.23.1
> Reporter: Eric Payne
> Assignee: Eric Payne
> Priority: Critical
>
> In 0.20.*, the JMX UI for the JobTracker (http://<JobTrackerHost>:50030/jmx)
> showed lists of Live and BlackListed Nodes under the JobTrackerInfo section.
> In 0.23, the ResourceManager JMX UI shows the number of active,
> decommissioned, lost, unhealthy, and rebooted nodes under the ClusterMetrics
> section, but does not give the list of nodes.
> At least the list of active nodes is needed in JSON format.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
