You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Bibin A Chundatt (JIRA)" <ji...@apache.org> on 2016/11/07 15:56:58 UTC
[jira] [Comment Edited] (YARN-5845) Skip aclUpdated event publish
to timelineserver or recovery
[ https://issues.apache.org/jira/browse/YARN-5845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15644530#comment-15644530 ]
Bibin A Chundatt edited comment on YARN-5845 at 11/7/16 3:55 PM:
-----------------------------------------------------------------
Attaching patch based on solution 1
MInor change in comment from last time instead of startTime should be *submitTime*. Apologies incase created any confusion
{code}
// call RMAppManager to submit application directly
rmAppManager.submitApplication(submissionContext,
System.currentTimeMillis(), user);
{code}
Application start {{submit Time}} and {{aclUpdate}} time are set to same.
The same appSubmit time will be used to aclUpdate also.
In recovery flow the {{acl update time}} will also same as {{submit Time}} and no new entries.
was (Author: bibinchundatt):
Attaching patch based on solution 1
MInor comment from last time instead of startTime should be *submitTime*
{code}
// call RMAppManager to submit application directly
rmAppManager.submitApplication(submissionContext,
System.currentTimeMillis(), user);
{code}
Application start {{submit Time}} and {{aclUpdate}} time are set to same.
The same appSubmit time will be used to aclUpdate also.
In recovery flow the {{acl update time}} will also same as {{submit Time}} and no new entries.
> Skip aclUpdated event publish to timelineserver or recovery
> -----------------------------------------------------------
>
> Key: YARN-5845
> URL: https://issues.apache.org/jira/browse/YARN-5845
> Project: Hadoop YARN
> Issue Type: Bug
> Reporter: Bibin A Chundatt
> Assignee: Bibin A Chundatt
> Priority: Critical
> Attachments: YARN-5845.0001.patch
>
>
> Currently ACL update event is send to timeline server even on recovery {{RMAppManager#createAndPopulateNewRMApp}}.
> For 10K completed application when RM is restarted 10K ACL updated event is added to timelinesever causing unnecessary over loading of system
> {code}
> String appViewACLs = submissionContext.getAMContainerSpec()
> .getApplicationACLs().get(ApplicationAccessType.VIEW_APP);
> rmContext.getSystemMetricsPublisher().appACLsUpdated(
> application, appViewACLs, System.currentTimeMillis());
> {code}
> *Events on each RM restart*
> {noformat}
> "events": [{
> "timestamp": 1478520292543,
> "eventtype": "YARN_APPLICATION_ACLS_UPDATED",
> "eventinfo": {}
> }, {
> "timestamp": 1478519600537,
> "eventtype": "YARN_APPLICATION_ACLS_UPDATED",
> "eventinfo": {}
> }, {
> "timestamp": 1478519557101,
> "eventtype": "YARN_APPLICATION_ACLS_UPDATED",
> "eventinfo": {}
> },
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org