[
https://issues.apache.org/jira/browse/FLINK-40092?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18096324#comment-18096324
]
Yixing Zhang commented on FLINK-40092:
--------------------------------------
+1 on this. We have seen a case when cutover happened and blue was deleted. But
then before the next reconciliation cycle, green became unhealthy, which was
blocking the state to transition to ACTIVE_GREEN.
> BlueGreen transition should finalize immediately after deleting previous
> deployment
> -----------------------------------------------------------------------------------
>
> Key: FLINK-40092
> URL: https://issues.apache.org/jira/browse/FLINK-40092
> Project: Flink
> Issue Type: Bug
> Components: Kubernetes Operator
> Affects Versions: kubernetes-operator-1.15.0
> Reporter: Calico Shop
> Priority: Major
> Labels: pull-request-available
> Fix For: kubernetes-operator-1.16.0
>
>
> During a BlueGreen deployment transition, once the previous FlinkDeployment
> is successfully deleted, the operator currently waits for another
> reconciliation loop before finalizing the FlinkBlueGreenDeployment into the
> next ACTIVE_* state.
> This extra reconciliation is unnecessary because the old deployment has
> already been removed and the transition can be finalized immediately. The
> operator should call the existing finalization logic in the same
> reconciliation that successfully deletes the previous deployment, while
> preserving retry behavior when deletion is not acknowledged.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
