-
Task
-
Resolution: Done
-
Undefined
-
None
-
None
Implement all agreed changes in RollingUpdateST.
- testRecoveryDuringZookeeperRollingUpdate
- check that only 1 zk pod is in pending state even after second reconciliation
- testRecoveryDuringKafkaRollingUpdate
- check that only 1 Kafka pod is in pending state even after second reconciliation
- testKafkaAndZookeeperScaleUpScaleDown
- enable this test to kraft if possible
- scale kafka only to 5 rep. (to save time/resources)
- After Kafka scale create new topic with replicas on newly added brokers (cannot easilly scaledown after that - remove topics should handle it)
- testZookeeperScaleUpScaleDown
- Create new topcis after ZK scale operation and attach clients to it to see if the ZK works correctly
- testBrokerConfigurationChangeTriggerRollingUpdate
- see how this triggering by change of config is covered in diff. test suits.
- testManualKafkaConfigMapChangeDontTriggerRollingUpdate
- remove this test
- testExternalLoggingChangeTriggerRollingUpdate
- rename test according to what is happening in steps.
- consider migration to logging test suite
- testClusterOperatorFinishAllRollingUpdates
- Try to dig form the history why we have there check for pending state of Kafka pods and why we delete CO twice
- testMetricsChange
- move to metrics?
- deploy Kafka with disabled metrics and then enable it to see it will trigger rolling update