[ 
https://issues.apache.org/jira/browse/FLINK-16293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17046733#comment-17046733
 ] 

Andrey Zagrebin edited comment on FLINK-16293 at 2/27/20 3:26 PM:
------------------------------------------------------------------

[~nielsbasjes] Thanks for looking into this!

I would say it rather belongs to docker [doc 
section|https://ci.apache.org/projects/flink/flink-docs-release-1.10/ops/deployment/docker.html]
 because it is about how to build a custom docker image. The docker image can 
be used for many types of deployment, not only native Kubernetes. Now we have 
some confusion in our examples, different places in docs and Flink repo about 
docker and Kubernetes related topics because of historical reasons. Therefore, 
it is a bit difficult to integrate properly the plugin guide into the existing 
docs. Right now, we are preparing a new FLIP to address this docker topic 
misalignments and figure out a unified clear approach.

The common [docker 
entry|https://github.com/apache/flink-docker/blob/master/docker-entrypoint.sh] 
already has some scripts to setup plugins. In this particular case of native 
Kubernetes integration, the entry point gets overridden and plugin script gets 
lost. Therefore, the new idea is to prepare scripts/examples for this kind of 
standard actions and make it part of base docker image to call in inherited 
Dockerfile, like you do in your example. The plan is also to rewrite the 
mentioned general docker image docs with this new approach in mind including 
general example of plugin installation, hadoop etc. All other places (e.g. 
Kubernetes integration) should refer to this as a separate if user has to 
create a custom docker image.

Taking this into account, the addition to the docs, introduced in the current 
PR, will have to be completely replaced anyways by a reference to the main 
docker image guide with an example of a simple script call in custom Dockerfile 
or entry point.

What do you think?


was (Author: azagrebin):
[~nielsbasjes] Thanks for looking into this!

I would say it rather belongs to this [doc 
section|https://ci.apache.org/projects/flink/flink-docs-release-1.10/ops/deployment/docker.html]
 because it is about how to build a custom docker image. The docker image can 
be used for many types of deployment, not only native Kubernetes. Now we have 
some confusion in our examples, different places in docs and Flink repo about 
docker and Kubernetes related topics because of historical reasons. Therefore, 
it is a bit difficult to integrate properly the plugin guide into the existing 
docs. Right now, we are preparing a new FLIP to address this docker topic 
misalignments and figure out a unified clear approach.

The common [docker 
entry|https://github.com/apache/flink-docker/blob/master/docker-entrypoint.sh] 
already has some scripts to setup plugins. In this particular case of native 
Kubernetes integration, the entry point gets overridden and plugin script gets 
lost. Therefore, the new idea is to prepare scripts/examples for this kind of 
standard actions and make it part of base docker image to call in inherited 
Dockerfile, like you do in your example. The plan is also to rewrite the 
mentioned general docker image docs with this new approach in mind including 
general example of plugin installation, hadoop etc. All other places (e.g. 
Kubernetes integration) should refer to this as a separate if user has to 
create a custom docker image.

Taking this into account, the addition to the docs, introduced in the current 
PR, will have to be completely replaced anyways by a reference to the main 
docker image guide with an example of a simple script call in custom Dockerfile 
or entry point.

What do you think?

> Document using plugins in Kubernetes
> ------------------------------------
>
>                 Key: FLINK-16293
>                 URL: https://issues.apache.org/jira/browse/FLINK-16293
>             Project: Flink
>          Issue Type: Improvement
>          Components: Deployment / Kubernetes, FileSystems
>    Affects Versions: 1.10.0
>            Reporter: Niels Basjes
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> It took me some time to figure out how to enable plugins when running Flink 
> on Kubernetes.
> So I'm writing some documentation to save other people trying the same a lot 
> of time.



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

Reply via email to