[
https://issues.apache.org/jira/browse/OOZIE-2897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16216072#comment-16216072
]
Rohini Palaniswamy commented on OOZIE-2897:
-------------------------------------------
bq. but eventually I would hope we could drop those
That never happens in practice as it would involve everyone fixing and
retesting their workflow.xmls. Especially in this case, it will be a total
waste of time for users as it is not going to make any difference. We had the
same situation in mapreduce->tez transition. We support all mapreduce settings
(even though there are equivalent tez settings) and have no plans of ever
dropping support for the mapreduce settings.
bq. Doesn't it seem weird to have users set
oozie.launcher.mapreduce.<some-property> instead of oozie.<some-property>?
Even if we add new Oozie AM specific properties later, for consistency we
should name it as oozie.launcher.<some-property> as it is still for the
launcher AM.
> LauncherAM should support ACLs
> ------------------------------
>
> Key: OOZIE-2897
> URL: https://issues.apache.org/jira/browse/OOZIE-2897
> Project: Oozie
> Issue Type: Sub-task
> Affects Versions: 5.0.0
> Reporter: Peter Bacsko
> Assignee: Peter Bacsko
> Priority: Blocker
>
> In MapReduce, you can define ACL-related properties:
> {noformat}
> mapreduce.job.acl-view-job
> mapreduce.job.acl-modify-job
> {noformat}
> {{acl-view-job}} defines a list of users/groups who can retrieve the job
> statistics. {{acl-modify-job}} defines a list of users/groups who can kill a
> job or adjust the priority of it.
> Docs:
> https://hadoop.apache.org/docs/r1.2.1/mapred_tutorial.html#Job+Authorization
> In YARN, we can provide backward compatibility for these properties. Example
> code:
> {code}
> Map<ApplicationAccessType, String> acls = new HashMap<ApplicationAccessType,
> String>();
> acls.put(ApplicationAccessType.MODIFY_APP, "*");
> acls.put(ApplicationAccessType.VIEW_APP, "*");
> amContainer.setApplicationACLs(acls);
> {code}
> This has to be done before application submission. We have to do what
> YARNRunner.java does:
> https://github.com/apache/hadoop/blob/3721cfe1fbd98c5b6aa46aefdfcf62276c28c4a4/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-jobclient/src/main/java/org/apache/hadoop/mapred/YARNRunner.java#L502-L507
>
> We should also ensure the compatibility with the properties of:
> mapreduce.job.acl-view-job
> mapreduce.job.acl-modfy-job
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
