When all pods end up in Pending state => for example because of badly configured label in rack awareness, the KafkaRoller is not able to recover from it (after CR is fixed) and instead always fails with the following errors: https://pastebin.com/n7nKi9z7