[
https://issues.apache.org/jira/browse/FLINK-14963?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Piotr Nowojski updated FLINK-14963:
-----------------------------------
Component/s: (was: Runtime / Checkpointing)
> Massive Duplicated Useless Oss Message in jobmanager.log
> --------------------------------------------------------
>
> Key: FLINK-14963
> URL: https://issues.apache.org/jira/browse/FLINK-14963
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / FileSystem
> Affects Versions: 1.9.1
> Reporter: Eric Lee
> Priority: Major
> Attachments: 7120B204-1B34-46D5-A9ED-D5BAB2D76F26.png
>
>
> When I was using AliCloud Oss as the external file system for flink
> checkpoints, there were massive oss related logs which are useless because
> checkpoints can still complete successfully.
> After I debugged into checkpoint coordinator and file system, I found the
> root cause, the wrong use of oss sdk to check whether the target exists in
> hadoop fs module.
> (Attached image is the screenshot of jobmanager.log)
> ----
> Solutions:
> I have submitted a [pull request
> |https://github.com/aliyun/aliyun-oss-java-sdk/pull/264]in aliyun-oss-client
> repository and v3.8.0 has released. We can alter the oss client version with
> [3.8.0|https://github.com/apache/flink/blob/master/flink-filesystems/flink-oss-fs-hadoop/pom.xml#L31].
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
