[
https://issues.apache.org/jira/browse/MAPREDUCE-1818?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12871970#action_12871970
]
Scott Chen commented on MAPREDUCE-1818:
---------------------------------------
Capacity scheduler should have a configuration to specify which queue to use.
I don't know the exact key.
I think we should allow any kind of job configuration.
And raid should not know about the details of the names and configuration keys.
We can configure a parameter like raid.jobconf.location in each policy in
raid.xml
This raid.jobconf.location specify a path of a conf file for the raid job.
For each policy, we can load the corresponding configuration file and passing
them to the jobconf.
This way we can support any parameters.
And we don't have to create extra dependency between raid and other contrib
projects.
What do you think?
> RaidNode should specify a pool name incase the cluster is using FairScheduler
> -----------------------------------------------------------------------------
>
> Key: MAPREDUCE-1818
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1818
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: contrib/raid
> Affects Versions: 0.20.1
> Reporter: Ramkumar Vadali
> Priority: Minor
>
> contrib/fairscheduler (FairScheduler) supports scheduling based on pools. The
> RaidNode should specify a pool name based on configuration to make use of
> pools.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
