Constant Kafka CR reconciliation loop

[MandMaric]

Hi everyone,

So we have recently upgraded Strimzi to 0.45 on 4 clusters. On 2 of them it works completely fine. On the other 2 we have the issue, that the Kafka CR is in a constant reconciliation loop. It always shows that it has been modified. But the only thing that changed between the modifications is the timestamp of the warning on the Kafka CR, which tells us that we need to upgrade to KRaft soon. Also the generation of the Kafka CR is the same after each reconciliation loop.

We have this Zookeeper/KRaft warning on all 4 clusters, but only on 2 of them it get reconciled almost every second.
We have checked the Openshift versions on all envs, the CRDs and they are the same.

Has anyone had a similar issue before?

[im-konge]

You can hit this issue if you don't replace the CRDs correctly, so you are using just kubectl apply instead of kubectl replace for the CRDs.
But for any other debug, you should share for example logs from the COs or your configuration.