how to restart kafka-connect properly

[jack1902]

What is the correct way to restart kafka-connect instances?

I am aware that i can restart the following in the following manner:

• kafka - annotate pods with strimzi.io/manual-rolling-update=true
• zookeeper - annotate pods with strimzi.io/manual-rolling-update=true
• connectors - https://github.com/strimzi/proposals/blob/main/007-restarting-kafka-connect-connectors-and-tasks.md

But i am unsure how to effectively restart the kafka-connect instances as i have previously used kubectl rollout restart deployment <connect-deployment-name> but on occasion have observed the newer pod getting killed off by a reconcile being carried out by the operator.

The main reason i am restarting the connect instances is because of outdated sidecars which need updating.

[scholzj]

Discussed in Slack => this is a missing feature which should be added in a similar way as for Kafka brokers for example.

[jack1902]

Raised: #9142

[scholzj]

Thanks 👍

[jack1902]

Adding a comment here as i always come to this issue whenever i google search this

kubectl annotate strimzipodset <cluster_name>-connect strimzi.io/manual-rolling-update="true"

https://strimzi.io/docs/operators/latest/deploying#proc-manual-rolling-update-strimzipodset-str