[
https://issues.apache.org/jira/browse/FLINK-33605?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17788402#comment-17788402
]
Martijn Visser commented on FLINK-33605:
----------------------------------------
[~oxwangfeng521] I'm sorry but there's only one screenshot which has a partial
error in there "Could not upload job files". There's no reproducer for the
behavior. The description doesn't provide me with the necessary information as
well
> flink1.17 fd leak
> -----------------
>
> Key: FLINK-33605
> URL: https://issues.apache.org/jira/browse/FLINK-33605
> Project: Flink
> Issue Type: Bug
> Affects Versions: 1.15.0, 1.17.1
> Reporter: feng wang
> Priority: Major
> Attachments: image-2023-11-21-19-02-32-929.png
>
>
> Environmental information:
> Flink 1.17.1;
>
> executed the command in the Flink background, lsof - n | grep deleted, and
> found that a historical process held a handle to a temporary data file. The
> historical data file did not released successfully, resulting in disk space
> not being released；
> !https://dts-szv.clouddragon.huawei.com/v1/nfs/downLoadFile?filePath=20231110P0J0003UW30_image-CORS-1699580879639.png!
>
> Flink SQL model：
> error msg：Could not upload job files，As follows：
> !https://dts-szv.clouddragon.huawei.com/v1/nfs/downLoadFile?filePath=20231110P0J000414VM_image-1699581171355.jpeg!
>
> when disk space is full;， The operation is as follows
> cd /proc/33425/fd/
> ls -l |grep blobStorage |grep deleted | awk '\{print $9}' |grep xargs
> truncate -s 0
> 33425 is taskmanager pid；
> and disk space is release
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
