[
https://issues.apache.org/jira/browse/FLINK-10641?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16659003#comment-16659003
]
Jiayi Liao edited comment on FLINK-10641 at 10/22/18 2:14 PM:
--------------------------------------------------------------
I think it's not a flink problem, it seems that yarn runs slow when allocating
last container.
was (Author: wind_ljy):
I think this is a bug, and I'll check it later.
> Slow when applying new containers
> ---------------------------------
>
> Key: FLINK-10641
> URL: https://issues.apache.org/jira/browse/FLINK-10641
> Project: Flink
> Issue Type: Bug
> Components: YARN
> Affects Versions: 1.6.1
> Reporter: Jiayi Liao
> Assignee: Jiayi Liao
> Priority: Major
>
> When it comes to the last reuqested container, the container is received and
> returned over and over again like this:
> 14:36:19,486 INFO org.apache.flink.yarn.YarnResourceManager - Received new
> container: container_1535124617388_1936_01_000929 - Remaining pending
> container requests: 0
> 14:36:19,486 INFO org.apache.flink.yarn.YarnResourceManager - Returning
> excess container container_1535124617388_1936_01_000929.
> But the truth is that the program still needs a container, it should not be a
> "excess".
> Sometimes it will last several minutes, which is out of our expectations.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
