[
https://issues.apache.org/jira/browse/HADOOP-3740?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12629085#action_12629085
]
Alejandro Abdelnur commented on HADOOP-3740:
--------------------------------------------
Given the current {{TaskScheduler}} API:
* It requires to duplicate all the scheduler code, specifically the
{{assignTask}} method just to do a check before calling the
{{JobInProgress.obtainNew*Task}}.
With the patch approach the throttling of tasks based on 'license availability'
can be used for different scheduler implementations.
* The {{TaskScheduler}} does not get notified when the task ends (due to
failure or success).
This is not possible to do as this is done in the
{{JobInProgress.complete.updateTaskStatus()}} method.
Maybe there should be something like a {{TaskListener}} with {{start}}
{{update}} {{finish}} methods and the {{start}} call should be able to veto the
task.
Making the {{JobInProgress}} pluggable is a temporary solution, more
specifically to be able to solve the second bullet.
> Make JobInProgress pluggable
> ----------------------------
>
> Key: HADOOP-3740
> URL: https://issues.apache.org/jira/browse/HADOOP-3740
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Environment: all
> Reporter: Alejandro Abdelnur
> Assignee: Alejandro Abdelnur
> Attachments: patch3740.txt
>
>
> By allowing a pluggable JobInProgess it will be possible for provide
> implementations that can do a sophisticated task provisioning to the
> JobTracker.
> For example, by providing alternate implementations of the
> {{obtainNewMapTask}}, {{obtainNewReduceTask}} and {{updateTaskInProgress}} it
> would be possible to implement a license server that allows to throttle use
> of external resources (ie webservices, databases) so at any given time there
> are not more than N tasks using a given resource. For this a task could be
> tagged with the names of external resources and the license server would keep
> track of the tasks running per tag, if the counter reaches zero then the
> {{obtainNew*Task}} method could return NULL instead of a task.
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
