[
https://issues.apache.org/jira/browse/YARN-8045?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16606394#comment-16606394
]
Shane Kumpf commented on YARN-8045:
-----------------------------------
Thanks for the proposal [~ccondit-target]. Moving the meat of the diagnostics
field to DEBUG makes sense to me and would meet the requirement with minimal
change.
My one concern is how that might impact compatibility. HADOOP-13714 recently
updated the [compatibility
guide|https://github.com/apache/hadoop/blob/trunk/hadoop-common-project/hadoop-common/src/site/markdown/Compatibility.md#log-output],
which includes logs. Given that logs are considered Unstable, I think we are
safe, but there is a note about ensuring existing parsers don't break. Can we
consider the parser requirement in moving this entry to DEBUG?
> Reduce log output from container status calls
> ---------------------------------------------
>
> Key: YARN-8045
> URL: https://issues.apache.org/jira/browse/YARN-8045
> Project: Hadoop YARN
> Issue Type: Improvement
> Reporter: Shane Kumpf
> Priority: Major
>
> Each time a container's status is returned a log entry is produced in the NM
> from {{ContainerManagerImpl}}. The container status includes the diagnostics
> field for the container. If the diagnostics field contains an exception, it
> can appear as if the exception is logged repeatedly every second. The
> diagnostics message can also span many lines, which puts pressure on the logs
> and makes it harder to read.
> For example:
> {code}
> 2018-03-17 22:01:11,632 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.ContainerManagerImpl:
> Getting container-status for container_e01_1521323860653_0001_01_000005
> 2018-03-17 22:01:11,632 INFO
> org.apache.hadoop.yarn.server.nodemanager.containermanager.ContainerManagerImpl:
> Returning ContainerStatus: [ContainerId:
> container_e01_1521323860653_0001_01_000005, ExecutionType: GUARANTEED, State:
> RUNNING, Capability: <memory:1024, vCores:1>, Diagnostics: [2018-03-17
> 22:01:00.675]Exception from container-launch.
> Container id: container_e01_1521323860653_0001_01_000005
> Exit code: -1
> Exception message: <unknown>
> Shell ouput: <unknown>
> [2018-03-17 22:01:00.750]Diagnostic message from attempt :
> [2018-03-17 22:01:00.750]Container exited with a non-zero exit code -1.
> , ExitStatus: -1, IP: null, Host: null, ContainerSubState: SCHEDULED]
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
