[
https://issues.apache.org/jira/browse/FLINK-32110?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yunjun Luo updated FLINK-32110:
-------------------------------
Description:
If job use time window in Pyflink thread mode, TM native memory will grow
slowly during the job running until TM can't allocate memory from operate
system.
The leak rate is likely proportional to the number of key.
was:
If job use time window in Pyflink thread mode, TM native memory will grow
slowly during the job running util TM can't allocate memory from operate system.
The leak rate is likely proportional to the number of key.
> TM native memory leak when using time window in Pyflink ThreadMode
> ------------------------------------------------------------------
>
> Key: FLINK-32110
> URL: https://issues.apache.org/jira/browse/FLINK-32110
> Project: Flink
> Issue Type: Bug
> Components: API / Python
> Affects Versions: 1.17.0, 1.16.1
> Reporter: Yunjun Luo
> Priority: Major
>
> If job use time window in Pyflink thread mode, TM native memory will grow
> slowly during the job running until TM can't allocate memory from operate
> system.
> The leak rate is likely proportional to the number of key.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
