[ 
https://issues.apache.org/jira/browse/FLINK-18186?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17136827#comment-17136827
 ] 

Andrey Zagrebin commented on FLINK-18186:
-----------------------------------------

Thanks for rising the issues [~fly_in_gis] I think they are valid. I have 
additional comments for some of them:


{quote}NodePort of jobmanager-rest-service should be located in 
30000-32767{quote}

This can be changed with `--service-node-port-range` but it is probably not 
nice user experience for the simple how-to in docs. I agree we use some other 
port in default range.
 
{quote}Array of array could not be support for 
{{io.k8s.api.core.v1.Container.args}}{quote}

This was supposed to mean optional job image arguments. Accidentally, it 
appears as array of arrays. I suggest to either move the optional arguments to 
a comment at the end of the same line or drop the brackets [..] for them at all.

> Various updates on Kubernetes standalone document
> -------------------------------------------------
>
>                 Key: FLINK-18186
>                 URL: https://issues.apache.org/jira/browse/FLINK-18186
>             Project: Flink
>          Issue Type: Improvement
>          Components: Deployment / Kubernetes, Documentation
>    Affects Versions: 1.11.0, 1.12.0
>            Reporter: Yang Wang
>            Priority: Critical
>             Fix For: 1.11.0
>
>
> There are following issues in the current Kubernetes standalone document.
>  * K8s job[1] should not have {{replicas}} and {{spec.selector}} for 
> {{batch/v1}}
>  * NodePort of jobmanager-rest-service should be located in 30000-32767[2].
>  * Array of array could not be support for 
> {{io.k8s.api.core.v1.Container.args}}
>  * Do not set query-state(6125) for jobmanager service and deployment. It is 
> only for taskmanager pod.
>  * {{log4j.properties}} in flink config map should be updated to 
> {{log4j-console.properties}} so that the logs could be output to console and 
> log files simultaneously.
>  
> [1]. 
> [https://kubernetes.io/docs/concepts/workloads/controllers/jobs-run-to-completion/]
> [2]. 
> [https://kubernetes.io/docs/concepts/services-networking/service/#nodeport]



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

Reply via email to