[
https://issues.apache.org/jira/browse/HADOOP-3479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12603000#action_12603000
]
Hemanth Yamijala commented on HADOOP-3479:
------------------------------------------
Some points in favor of a separate file for queue configuration:
- All parameters related to queues are controlled by administrators. These are
needed only by the Jobtracker for V1 of the Scheduler implementation
- The file needs to be monitored for updates periodically
- The format is more nested (> 1 level of hierarchy)
- In case the scheduler becomes separate from the JT (conceivably sometime in
the future), a separate file helps because we won't have to separate the
configuration items out of hadoop configuration then.
Given all these, I am thinking of implementing the following:
- Define a hadoop.scheduler.config variable in hadoop configuration that will
define the name of the config file for the Scheduler.
- From Jobtracker read this value and create a SchedulerConf class that will
parse this file and populate itself with the configuration parameters.
- The SchedulerConf class will provide APIs to get and set Scheduler related
configuration params, and also implement functionality such as looking for
updates etc.
Please comment if you prefer a different approach.
> Implement configuration items useful for Hadoop resource manager (v1)
> ---------------------------------------------------------------------
>
> Key: HADOOP-3479
> URL: https://issues.apache.org/jira/browse/HADOOP-3479
> Project: Hadoop Core
> Issue Type: New Feature
> Reporter: Hemanth Yamijala
> Assignee: Hemanth Yamijala
>
> HADOOP-3421 lists requirements for a new resource manager for Hadoop.
> Implementation for these will require support for new configuration items in
> Hadoop. This JIRA is to define such configuration, and track it's
> implementation.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
