[
https://issues.apache.org/jira/browse/FLINK-16406?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17050890#comment-17050890
]
Xintong Song commented on FLINK-16406:
--------------------------------------
A summary of bad cases collected so far:
* [~Jamalarm] reported in FLINK-16142. He had both problems of memory leak
and insufficient default size. He has not mention how large metaspace size
fixes his problem.
* [~blablabla123] also reported in FLINK-16142 about the insufficient default
size, fixed by increasing to 256 MB.
* [~nielsbasjes] reported in a [user ML
thread|http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Kubernetes-java-lang-OutOfMemoryError-Metaspace-td33285.html]
and in FLINK-16142. From what he described that the problem occurs with
repeatedly executing a job, it sounds like a memory leak problem to me. We
would need to investigate more into this.
* John Smith reported in another [user ML
thread|http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/MaxMetaspace-default-may-be-to-low-td33049.html]
about insufficient default size, fixed by increasing to 256 MB. I believe he
is also the one who opened FLINK-16278 ([~javadevmtl]).
> Increase default value for JVM Metaspace to minimise its OutOfMemoryError
> -------------------------------------------------------------------------
>
> Key: FLINK-16406
> URL: https://issues.apache.org/jira/browse/FLINK-16406
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Configuration, Runtime / Task
> Affects Versions: 1.10.0
> Reporter: Andrey Zagrebin
> Assignee: Andrey Zagrebin
> Priority: Critical
> Labels: pull-request-available
> Fix For: 1.10.1, 1.11.0
>
> Time Spent: 10m
> Remaining Estimate: 0h
>
> With FLIP-49
> ([FLINK-13980|https://issues.apache.org/jira/browse/FLINK-13980]), we
> introduced a limit for JVM Metaspace
> ('taskmanager.memory.jvm-metaspace.size') when TM JVM process is started. It
> caused '_OutOfMemoryError: Metaspace_' for some use cases after upgrading to
> the latest 1.10 version. In some cases, a real class loading leak has been
> discovered, like in
> [FLINK-16142|https://issues.apache.org/jira/browse/FLINK-16142]. Some users
> had to increase the default value to accommodate for their use cases (mostly
> from 96Mb to 256Mb).
> While this limit was introduced to properly plan Flink resources, especially
> for container environment, and to detect class loading leaks, the user
> experience should be as smooth as possible. One way is provide good
> documentation for this change
> ([FLINK-16278|https://issues.apache.org/jira/browse/FLINK-16278]).
> Another question is the sanity of the default value. It is still arguable
> what the default value should be (currently 96Mb). In general, the size
> depends on the use case (job user code, how many jobs are deployed in the
> cluster etc).
> This issue tries to tackle this problem by firstly increasing it to 256Mb and
> overall default process size to 1728Mb in flink-conf.yaml to have no impact
> on default sizes of other memory components. We also want to poll which
> Metaspace setting resolved the _OutOfMemoryError_. Please, if you encountered
> this problem, report here any relevant specifics of your job and your
> Metaspace size if there was no class loading leak.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
