You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "leishuiyu (Jira)" <ji...@apache.org> on 2025/04/10 04:01:00 UTC

[jira] [Created] (FLINK-37646) flink jobmanager oom

leishuiyu created FLINK-37646:
---------------------------------

             Summary: flink jobmanager oom
                 Key: FLINK-37646
                 URL: https://issues.apache.org/jira/browse/FLINK-37646
             Project: Flink
          Issue Type: Bug
          Components: Runtime / Task
         Environment:  

flink version 1.16.3 

deployment on k8s 

HA cluster

 

seatunnel 2.3.5

 

 

 
            Reporter: leishuiyu
         Attachments: image-2025-04-10-11-47-21-883.png, image-2025-04-10-11-51-59-574.png, image-2025-04-10-11-56-31-392.png, image-2025-04-10-11-59-43-542.png

 

when use seatunnel ,the batch job submit more time ,the flink jobmanager heap OOM,

I think it may be caused by jobmanager not releasing the temporary files in the blob after the job is finished

I found a lot of temporary files in the jobmanger container showing deleted, but the thread still holds the file handle, causing the java zip class not to be recycled

 

 

the MAT show 

!image-2025-04-10-11-47-21-883.png!

 

 

 

 

!image-2025-04-10-11-56-31-392.png!!image-2025-04-10-11-51-59-574.png!

 

 

 

!image-2025-04-10-11-59-43-542.png!



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