In order to debug some operations it was necessary to redeploy the kfdef (kfdef). The general process we followed is to delete the kfdef so that the operator detects a deletion of the kfdef and then delete all kubelow resources on the cluster. Apply some changes in the kfdef and then deploy it again by creating the kfdef custom resource, the operator detects a new kfdef instance and then deploy the resources again on the cluster. We faced the issue which was supposed to be resolved in version 1.2 of kfctl as pr was already merged. But now we faced this kind of deletion directly after the deletion of the kfdef. We don't have monitoring and logging applied at the moment but i can assure to you that in the logs of the operator there were delete events for profiles and namespaces. Is this an expected behavior? Is this a new known bug?
NB: We did the deployment several times and the deletion of the profiles with namespaces only happened once.
In order to debug some operations it was necessary to redeploy the kfdef (kfdef). The general process we followed is to delete the kfdef so that the operator detects a deletion of the kfdef and then delete all kubelow resources on the cluster. Apply some changes in the kfdef and then deploy it again by creating the kfdef custom resource, the operator detects a new kfdef instance and then deploy the resources again on the cluster. We faced the issue which was supposed to be resolved in version 1.2 of kfctl as pr was already merged. But now we faced this kind of deletion directly after the deletion of the kfdef. We don't have monitoring and logging applied at the moment but i can assure to you that in the logs of the operator there were delete events for profiles and namespaces. Is this an expected behavior? Is this a new known bug?
NB: We did the deployment several times and the deletion of the profiles with namespaces only happened once.