[
https://issues.apache.org/jira/browse/FLINK-17976?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17129965#comment-17129965
]
Yang Wang edited comment on FLINK-17976 at 6/10/20, 5:00 AM:
-------------------------------------------------------------
For question #1, it is probably a bug or improvement of
{{ActiveResourceManager}}. Currently, we decrease the pending workers after
allocated, not TaskManager registered. So both in K8s and Yarn, we may come
into this issue. cc [~xintongsong], i think you could give more suggestions
here.
For question #2, make the pending pods/containers could be released when we do
not need any pods/containers(e.g. Flink job failed/canceled). I have created a
new ticket FLINK-18228 to track this.
was (Author: fly_in_gis):
For question #1, it is probably a bug or improvement of
{{ActiveResourceManager}}. Currently, we decrease the pending workers after
allocated, not TaskManager registered. So both in K8s and Yarn, we may come
into this issue. cc [~xintongsong], i think you could give more suggestions
here.
For question #2, make the pending pods/containers could be released when we do
not need any pods/containers(e.g. Flink job failed/canceled). I will create a
new ticket to track this.
> Test native K8s integration
> ---------------------------
>
> Key: FLINK-17976
> URL: https://issues.apache.org/jira/browse/FLINK-17976
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Coordination
> Affects Versions: 1.11.0
> Reporter: Till Rohrmann
> Assignee: Robert Metzger
> Priority: Critical
> Labels: pull-request-available, release-testing
> Fix For: 1.11.0
>
> Attachments: enough_tm_wait_5min.txt
>
>
> Test Flink's native K8s integration:
> * session mode
> * application mode
> * custom Flink image
> * custom configuration and log properties
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
