[
https://issues.apache.org/jira/browse/FLINK-27650?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17537773#comment-17537773
]
Thomas Weise commented on FLINK-27650:
--------------------------------------
That's as per
[https://github.com/apache/flink-kubernetes-operator/blob/main/flink-kubernetes-operator/src/main/java/org/apache/flink/kubernetes/operator/utils/FlinkUtils.java#L83]
Maybe it is better to just override?
This isn't a "blocker" though.
> First environment variable of top level pod template is lost
> ------------------------------------------------------------
>
> Key: FLINK-27650
> URL: https://issues.apache.org/jira/browse/FLINK-27650
> Project: Flink
> Issue Type: Bug
> Components: Kubernetes Operator
> Affects Versions: kubernetes-operator-0.1.0
> Reporter: Simon Paradis
> Priority: Major
> Fix For: kubernetes-operator-1.0.0
>
> Attachments: flink-27650.yaml
>
>
> I am using the Flink operator image *apache/flink-kubernetes-operator:0.1.0*
> to deploy Flink 1.14.4 job. The deployment manifest makes use of pod template
> feature to inject environment variable to control structured JSON logging.
> I noticed the first defined environment variable is never injected into the
> JobManager nor TaskManager pods. The work around is to define a dummy env.
> var.
> Here's the manifest template. This gets processed by a tool that will first
> expand ${ENV_VAR} reference with values provided by our CI pipeline. We
> should not have to create the FLINK_COORDINATES_DUMMY env var.
>
>
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
