[
https://issues.apache.org/jira/browse/FLINK-33613?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dian Fu closed FLINK-33613.
---------------------------
Fix Version/s: 1.19.0
1.18.1
1.17.3
Resolution: Fixed
Fixed in:
- master via 977463cce3ea0f88e2f184c30720bf4e8e97fd4a
- release-1.18 via f6d005681e3f5f83ab1074660c4d9878dabb9176
- release-1.17 via c1818f530617af8996f4a74bb064e203186fd98e
> Python UDF Runner process leak in Process Mode
> ----------------------------------------------
>
> Key: FLINK-33613
> URL: https://issues.apache.org/jira/browse/FLINK-33613
> Project: Flink
> Issue Type: Bug
> Components: API / Python
> Affects Versions: 1.17.0
> Reporter: Yu Chen
> Assignee: Dian Fu
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.19.0, 1.18.1, 1.17.3
>
> Attachments: ps-ef.txt, streaming_word_count-1.py
>
>
> While working with PyFlink, we found that in Process Mode, the Python UDF
> process may leak after a failover of the job. It leads to a rising number of
> processes with their threads in the host machine, which eventually results in
> failure to create new threads.
>
> You can try to reproduce it with the attached test task
> `streamin_word_count.py`.
> (Note that the job will continue failover, and you can watch the process
> leaks by `ps -ef` on Taskmanager.
>
> Our test environment:
> * K8S Application Mode
> * 4 Taskmanagers with 12 slots/TM
> * Job's parallelism was set to 48
> The udf process `pyflink.fn_execution.beam.beam_boot` should be consistence
> with slots of TM (12), but we found that there are 180 processes on one
> Taskmanager after several failovers.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
