You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Till Rohrmann (JIRA)" <ji...@apache.org> on 2018/10/04 07:30:00 UTC
[jira] [Commented] (FLINK-10459) Flink cluster does not work
properly in kubernetes
[ https://issues.apache.org/jira/browse/FLINK-10459?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16637888#comment-16637888 ]
Till Rohrmann commented on FLINK-10459:
---------------------------------------
Hi [~gobozov], this problem was caused by FLINK-8696 which removed a command line parameter from the start up scripts. I've opened a PR to fix the Flink docker images: https://github.com/docker-flink/docker-flink/pull/49. The updated images should be available soon (assuming that you are using the Flink images from Docker Hub). Could you verify whether this is still the issue? If not, then please close the image.
For further reference: https://github.com/docker-flink/docker-flink/issues/50
> Flink cluster does not work properly in kubernetes
> --------------------------------------------------
>
> Key: FLINK-10459
> URL: https://issues.apache.org/jira/browse/FLINK-10459
> Project: Flink
> Issue Type: Bug
> Components: Kubernetes
> Affects Versions: 1.6.1, 1.5.4
> Reporter: Georgii Gobozov
> Priority: Major
> Attachments: image-2018-09-28-10-30-27-649.png, jobmanger.log, taskmanager.log
>
>
> Trying to run flink 1.6.1/1.5.4 cluster in Kubernetes, using default kubernetes deployment from flink docs. Jobmanager and taskmanagers starts successfully but in web UI I see zero taskmanagers available. Flink had issue when jobmanger was not able to connect to itself via k8s service name, but this seems different issue.
> Jobmanger config shows *cluster* as *jobmanager.rpc.address*
> !image-2018-09-28-10-30-27-649.png|width=733,height=398!
> Jobmanger and taskmanager logs attached
>
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)