[ 
https://issues.apache.org/jira/browse/YARN-11955?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Konstantin Bereznyakov updated YARN-11955:
------------------------------------------
    Description: 
FileSystemApplicationHistoryStore and RMApplicationHistoryWriter log 
non-critical exceptions at ERROR level, inconsistent with similar handling 
elsewhere in the same files.

Affected Code

1. 
[FileSystemApplicationHistoryStore.java#L212|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-applicationhistoryservice/src/main/java/org/apache/hadoop/yarn/server/applicationhistoryservice/FileSystemApplicationHistoryStore.java#L212]:
LOG.error("History information of application " + appId
+ " is not included into the result due to the exception", e);

2. 
[RMApplicationHistoryWriter.java#L201|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ahs/RMApplicationHistoryWriter.java#L201]:
LOG.error("Error when storing the start data of container "
+ wcsEvent.getContainerId());

Inconsistency

The same file (FileSystemApplicationHistoryStore.java) already uses WARN for 
similar non-critical conditions:
 - LOG.warn("Start information is missing for application " + appId)
 - 
[FileSystemApplicationHistoryStore.java#L183|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-applicationhistoryservice/src/main/java/org/apache/hadoop/yarn/server/applicationhistoryservice/FileSystemApplicationHistoryStore.java#L183]:
 LOG.warn("Finish information is missing for application " + appId)

Impact
 - Unnecessary ERROR-level log noise
 - Inconsistent logging behavior within the same component
 - False alerts in monitoring systems tracking ERROR volume

  was:
FileSystemApplicationHistoryStore and RMApplicationHistoryWriter log 
non-critical exceptions at ERROR level, inconsistent with similar handling 
elsewhere in the same files.

Affected Code

1. 
[FileSystemApplicationHistoryStore.java#L212|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-applicationhistoryservice/src/main/java/org/apache/hadoop/yarn/server/applicationhistoryservice/FileSystemApplicationHistoryStore.java#L180]:
LOG.error("History information of application " + appId
+ " is not included into the result due to the exception", e);

2. 
[RMApplicationHistoryWriter.java#L201|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ahs/RMApplicationHistoryWriter.java#L201]:
LOG.error("Error when storing the start data of container "
+ wcsEvent.getContainerId());

Inconsistency

The same file (FileSystemApplicationHistoryStore.java) already uses WARN for 
similar non-critical conditions:
 - LOG.warn("Start information is missing for application " + appId)
 - 
[FileSystemApplicationHistoryStore.java#L183|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-applicationhistoryservice/src/main/java/org/apache/hadoop/yarn/server/applicationhistoryservice/FileSystemApplicationHistoryStore.java#L183]:
 LOG.warn("Finish information is missing for application " + appId)

Impact
 - Unnecessary ERROR-level log noise
 - Inconsistent logging behavior within the same component
 - False alerts in monitoring systems tracking ERROR volume


> Application history store logs non-critical failures at ERROR level
> -------------------------------------------------------------------
>
>                 Key: YARN-11955
>                 URL: https://issues.apache.org/jira/browse/YARN-11955
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: Konstantin Bereznyakov
>            Priority: Major
>
> FileSystemApplicationHistoryStore and RMApplicationHistoryWriter log 
> non-critical exceptions at ERROR level, inconsistent with similar handling 
> elsewhere in the same files.
> Affected Code
> 1. 
> [FileSystemApplicationHistoryStore.java#L212|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-applicationhistoryservice/src/main/java/org/apache/hadoop/yarn/server/applicationhistoryservice/FileSystemApplicationHistoryStore.java#L212]:
> LOG.error("History information of application " + appId
> + " is not included into the result due to the exception", e);
> 2. 
> [RMApplicationHistoryWriter.java#L201|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/ahs/RMApplicationHistoryWriter.java#L201]:
> LOG.error("Error when storing the start data of container "
> + wcsEvent.getContainerId());
> Inconsistency
> The same file (FileSystemApplicationHistoryStore.java) already uses WARN for 
> similar non-critical conditions:
>  - LOG.warn("Start information is missing for application " + appId)
>  - 
> [FileSystemApplicationHistoryStore.java#L183|https://github.com/apache/hadoop/blob/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-applicationhistoryservice/src/main/java/org/apache/hadoop/yarn/server/applicationhistoryservice/FileSystemApplicationHistoryStore.java#L183]:
>  LOG.warn("Finish information is missing for application " + appId)
> Impact
>  - Unnecessary ERROR-level log noise
>  - Inconsistent logging behavior within the same component
>  - False alerts in monitoring systems tracking ERROR volume



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to