storageClass set, PVC pending about helm HOT 5 CLOSED

Thank you for your feedback! I managed to get around this by setting the glusterfs storage (named slow) as default, and then issuing the original command without setting the storageclass. This is however just a workaround, someone having multiple storageclasses might experience the same.

Btw. the statefulsets had always been created during the previous attempts (when setting the storagecalss name), but they had always stayed in a pending state, and then had been deleted (because of the --atomic switch)

from helm.

Could you query the Statefulset kubevious-mysql like this:

kubectl get statefulset kubevious-mysql -n kubevious -o yaml

is there any chance you did the deployment once without setting the class and the persistent volume got created with the default storage class?

from helm.

This is a new new deployment, so there is no statefulset to query. Also there is no existing pv. I have a glusterfs cluster with dynamic provisioning, so the pvc with a named storageclass should be sufficient to create the pv (?). Doesn't the meassage no persistent volumes available for this claim and no storage class is set indicate, that the -set mysql.storageClass=slow parameter is not being passed?

from helm.

This is a new new deployment, so there is no statefulset to query. Also there is no existing pv. I have a glusterfs cluster with dynamic provisioning, so the pvc with a named storageclass should be sufficient to create the pv (?). Doesn't the meassage no persistent volumes available for this claim and no storage class is set indicate, that the -set mysql.storageClass=slow parameter is not being passed?

There should be StatefulSet created, because the data-kubevious-mysql-0 PVC is created because of the volumeClaimTemplates setting. See here:
https://github.com/kubevious/helm/blob/master/kubernetes/templates/mysql/statefulset.yaml#L80-L91

For debugging you issue we first need to make sure that the mysql.storageClass=slow setting was indeed applied to the StatefulSet.

from helm.

@leftbrake, sometimes such issues are easier to debug in a Slack channel. Feel free to join if we don't find a solution quickly here.

from helm.