Apache Flink

Hey there, folks.

I have some Flink deployments running with the k8s operator. The restore configuration is `savepoint`. Sometimes when a job gets updated, instead of loading from the latest savepoint, it loads from the oldest. Have you seen this?

that sounds strange, could you please open a JIRA with the operator logs / CR?

I found it <https://issues.apache.org/jira/browse/FLINK-32890?jql=project%20%3D%20FLINK%20AND%20fixVersion%20%3D%20kubernetes-operator-1.7.0|here>. I think it’s baking for 1.7

We will release the 1.6.1 version containing this fix soon with <@U03JBQ34Z41>

Hey there. Sorry about bringing it again. We got good time after the upgrade but it seems this still happens sporadically. Have anyone seen this?

<https://issues.apache.org/jira/browse/FLINK-32890?jql=project%20%3D%20FLINK%20AND%20fixVersion%20%3D%20kubernetes-operator-1.7.0|Original Issue >

<@U03N7A9DTQR> please open a jira ticket with operator logs and all details you can provide 

Then we can investigate the problem further :)