You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After upgrading from 19.04.1 to 19.07 on a Rancher RKE deployed cluster that has both linux and windows nodes, trident is trying to start a csi pod on the windows node. Obviously this won't work so the pod sits forever at "ContainerCreating."
Should just need to edit the deployment yaml to only try to start on linux nodes to avoid this for future customers.
Here's the get po output for the stuck pod. tst-k8s-wm-01 is our windows node.
kubectl get po -n trident
NAME READY STATUS RESTARTS AGE
trident-csi-867d54588b-6xsqv 4/4 Running 0 13m
trident-csi-rhlp9 2/2 Running 0 13m
trident-csi-xqlz5 0/2 ContainerCreating 0 13m
Thanks. We'll need to add a node selector to anything that schedules a container, like our Deployments and DaemonSets. At least this: beta.kubernetes.io/os=linux.
After upgrading from 19.04.1 to 19.07 on a Rancher RKE deployed cluster that has both linux and windows nodes, trident is trying to start a csi pod on the windows node. Obviously this won't work so the pod sits forever at "ContainerCreating."
Should just need to edit the deployment yaml to only try to start on linux nodes to avoid this for future customers.
Here's the get po output for the stuck pod. tst-k8s-wm-01 is our windows node.
The text was updated successfully, but these errors were encountered: