[
https://issues.apache.org/jira/browse/YARN-7607?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16278026#comment-16278026
]
Weiwei Yang commented on YARN-7607:
-----------------------------------
{{ContainerImpl#addDiagnostics}} extracts each message string and add timestamp
as prefix, if a line break {{\n}} was added as a separate message, e.g
{code}
container.addDiagnostics(exitEvent.getDiagnosticInfo(), "\n");
{code}
this will cause the duplicated trailing timestamp. The fix is simple, the line
break should be part of the message instead of a separate argument.
> Remove the trailing duplicated timestamp in container diagnostics message
> -------------------------------------------------------------------------
>
> Key: YARN-7607
> URL: https://issues.apache.org/jira/browse/YARN-7607
> Project: Hadoop YARN
> Issue Type: Bug
> Components: nodemanager
> Affects Versions: 2.9.0
> Reporter: Weiwei Yang
> Assignee: Weiwei Yang
> Priority: Minor
> Labels: log
> Attachments: YARN-7607.001.patch
>
>
> There are some container diagnostic messages currently malformed, like below
> ###
> 017-12-05 11:43:21,319 INFO mapreduce.Job: map 28% reduce 0%
> 2017-12-05 11:43:22,345 INFO mapreduce.Job: Task Id :
> attempt_1512384455800_0003_m_000012_0, Status : FAILED
> \[2017-12-05 11:43:21.265\]Container Killed to make room for Guaranteed
> Container{color:red}\[2017-12-05 11:43:21.265\] {color}
> \[2017-12-05 11:43:21.265\]Container is killed before being launched.
> ###
> such logs are presented both in console and RM UI, we need to remove the
> duplicated trailing timestamp from the log message. This is due to the
> mis-use of {{addDiagnostics}} function in these places.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
