flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xintong Song (Jira)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-13554) ResourceManager should have a timeout on starting new TaskExecutors.
Date Wed, 08 Jan 2020 08:06:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-13554?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17010451#comment-17010451
] 

Xintong Song commented on FLINK-13554:
--------------------------------------

We have confirmed that the release-1.10 blocker FLINK-15456 is actually caused by the problem
described in this ticket.
Since this problem is not introduced in 1.10, I believe it should not be a blocker. But how
do we fix the problem, and whether it needs to be fixed in 1.10 still need to be discussed.
I'm setting this ticket to be release-1.10 critical for now, to avoid overlooking it before
a decision being made.

> ResourceManager should have a timeout on starting new TaskExecutors.
> --------------------------------------------------------------------
>
>                 Key: FLINK-13554
>                 URL: https://issues.apache.org/jira/browse/FLINK-13554
>             Project: Flink
>          Issue Type: Improvement
>          Components: Runtime / Coordination
>    Affects Versions: 1.9.0
>            Reporter: Xintong Song
>            Priority: Critical
>             Fix For: 1.10.0
>
>
> Recently, we encountered a case that one TaskExecutor get stuck during launching on Yarn
(without fail), causing that job cannot recover from continuous failovers.
> The reason the TaskExecutor gets stuck is due to our environment problem. The TaskExecutor
gets stuck somewhere after the ResourceManager starts the TaskExecutor and waiting for the
TaskExecutor to be brought up and register. Later when the slot request timeouts, the job
fails over and requests slots from ResourceManager again, the ResourceManager still see a
TaskExecutor (the stuck one) is being started and will not request new container from Yarn.
Therefore, the job can not recover from failure.
> I think to avoid such unrecoverable status, the ResourceManager need to have a timeout
on starting new TaskExecutor. If the starting of TaskExecutor takes too long, it should just
fail the TaskExecutor and starts a new one.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Mime
View raw message