[
https://issues.apache.org/jira/browse/FLINK-33288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17776124#comment-17776124
]
Xin Chen edited comment on FLINK-33288 at 10/17/23 10:14 AM:
-------------------------------------------------------------
yeah，I see now. [~Zhanghao Chen] I have incorporated your modifications and
verified that after the task is executed normally, the directory can be
deleted. However, in abnormal scenarios, such as when the task is running on
Yarn and I kill it, both the directory and subdirectory blob will still exist.
Perhaps the community does not consider abnormal scenarios, as we have
previously added a hook to clear the appid directory under '/user/hadoop/.
flink', It includes jar packages uploaded by Flink when submitting tasks.
But this hook involves Hadoop and it also needs to be modified. In the hook
code, it needs to provide us with feedback about the status of the task. I
think it may be due to the interaction between multiple communities, and the
Flink community did not actively consider clearing in this abnormal situation.
was (Author: JIRAUSER298666):
yeah，I see now. [~Zhanghao Chen] I have incorporated your modifications and
verified that after the task is executed normally, the directory can be
deleted. However, in abnormal scenarios, such as when the task is running on
Yarn and I kill it, both the directory and subdirectory blob will still exist.
Perhaps the community does not consider abnormal scenarios, as we have
previously added a hook to clear the appid directory under/user/hadoop/. flink,
It includes jar packages uploaded by Flink when submitting tasks.
But this hook involves Hadoop and it also needs to be modified. In the hook
code, it needs to provide us with feedback about the status of the task. I
think it may be due to the interaction between multiple communities, and the
Flink community did not actively consider clearing in this abnormal situation.
> Empty directory residue with appid name in HA(highly-available) related
> directory of hdfs, not cleaned
> ------------------------------------------------------------------------------------------------------
>
> Key: FLINK-33288
> URL: https://issues.apache.org/jira/browse/FLINK-33288
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Configuration
> Affects Versions: 1.16.2, 1.17.1
> Reporter: Xin Chen
> Priority: Major
> Attachments: image-2023-10-17-16-43-07-859.png, screenshot-1.png,
> screenshot-2.png
>
>
> When I submitted a large number of tasks in Flink-on-Yarn mode and
> successfully executed, I unexpectedly found a large number of empty
> directories left in the directory related to 'high availability.storageDir'
> on hdfs, with appids as shown below. I believe this must be cleared! However,
> after verification in the environments of 1.16.2 and 1.17.1, it was proven
> that neither of them solved this problem.
> my flink-conf.yaml about 'high availability.storageDir':
> {code:java}
> high-availability.storageDir: hdfs://hdfsHACluster/flink/recovery
> {code}
> !screenshot-1.png!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
