[
https://issues.apache.org/jira/browse/FLINK-26377?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17498393#comment-17498393
]
Aitozi edited comment on FLINK-26377 at 2/26/22, 7:01 AM:
----------------------------------------------------------
I have the same sense to extract common reconciler interface. We can choose or
create the target reconciler based on the FlinkDeployment. I volunteer to do
this refactor.
One more further question I want to discuss: Do we need to introduce the extra
filed like {{mode}} to reflect the mode of FlinkDeployment, for example:
{{JOB}} and {{SESSION}}. MayBe {{Standalone}} mode on the way. By this, we can
directly know the mode of a FlinkApp
instead of deciding by the null or not null of {{JobSpec}}.
was (Author: aitozi):
I have the same sense to extract common reconciler interface. We can choose or
create the target reconciler based on the FlinkDeployment. I volunteer to do
this refactor.
One more further question I want to discuss: Do we need to introduce the extra
filed like {{mode}} to reflect the mode of FlinkDeployment, for example:
{{JOB}} and {{SESSION}}. MayBe {{Standalone}} mode on the way. By this, we can
directly know the mode of a FlinkApp
instead of deciding by the null or not null of {{JobSpec}}.
> Extract Reconciler interface
> ----------------------------
>
> Key: FLINK-26377
> URL: https://issues.apache.org/jira/browse/FLINK-26377
> Project: Flink
> Issue Type: Sub-task
> Components: Kubernetes Operator
> Reporter: Gyula Fora
> Priority: Major
>
> We should extract a common interface for the different reconciler classes
> (Job and Session for now) and create the reconciler instance on the fly based
> on the FlinkDeployment.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
