[
https://issues.apache.org/jira/browse/YARN-3505?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14535497#comment-14535497
]
Junping Du commented on YARN-3505:
----------------------------------
bq. However, I think we don't need this differentiation in LogAggregationReport
as in case of FAILED case, diagnosticMessage and failureMessages store the same
message which is duplicated. Isn't it?
Just sync offline with [~xgong], the additional challenge here is for RUNNING
LogAggregationState, we cannot 100% sure the log aggregation on this node is
successful or failed (LRS need aggregate log for several times, we always make
it RUNNING until final state). The proposed solution here is we may add a new
state here called RUNNING_WITH_FAILURE for this case, so we differentiate the
message is normal diagnostic one or failure messages. The extra bonus is user
can know something wrong happens to LogAggregation earlier but not get confused
with Log Aggregation already get stopped.
> Node's Log Aggregation Report with SUCCEED should not cached in RMApps
> ----------------------------------------------------------------------
>
> Key: YARN-3505
> URL: https://issues.apache.org/jira/browse/YARN-3505
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: log-aggregation
> Affects Versions: 2.8.0
> Reporter: Junping Du
> Assignee: Xuan Gong
> Priority: Critical
> Attachments: YARN-3505.1.patch, YARN-3505.2.patch,
> YARN-3505.2.rebase.patch
>
>
> Per discussions in YARN-1402, we shouldn't cache all node's log aggregation
> reports in RMApps for always, especially for those finished with SUCCEED.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
