[
https://issues.apache.org/jira/browse/OOZIE-1770?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robert Kanter updated OOZIE-1770:
---------------------------------
Attachment: Prelim OYA Scoping Doc 003.pdf
After playing more with the PoC and some internal discussions, I've updated the
Prelim OYA Scoping Doc to v3, which some changes to how the AMs will work.
I've added more details in the document itself, but basically, we can simplify
this further by using a single AM per Action. The AM runs the Action Launcher
code, much like uber mode today, but without the MapReduce stuff, and with full
control over the AM. There won't be a pool of unmanaged AMs or any of that.
Besides simplifying the code (which is always a good thing), this solves some
issues that came up with logging, user expectation, and queues. See the doc
for more details.
We can still eventually do the Workflow AM by moving the Action AMs into
containers launched by the Workflow AM. Though if the user wants the actions
in different queues, we'll still have to have the Workflow AM submit Action
AMs.
> Create Oozie Application Master for YARN
> ----------------------------------------
>
> Key: OOZIE-1770
> URL: https://issues.apache.org/jira/browse/OOZIE-1770
> Project: Oozie
> Issue Type: New Feature
> Reporter: Bowen Zhang
> Assignee: Bowen Zhang
> Attachments: OozieYarnAM.pdf, Prelim OYA Scoping Doc 001.pdf, Prelim
> OYA Scoping Doc 002.pdf, Prelim OYA Scoping Doc 003.pdf,
> oya-rm-screenshot.jpg, oya.patch
>
>
> After the first release of oozie on hadoop 2, it will be good if users can
> set execution engine in oozie conf, be it YARN AM or traditional MR. We can
> target this for post oozie 4.1 release.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
