Martin Emrich

Have the same issue. In the "standalone" ebs driver, this apparently can be configured with a helm variable "tolerateAllTaints", too: https://github.com/kubernetes-sigs/aws-ebs-csi-driver/blob/master/charts/aws-ebs-csi-driver/templates/node.yaml#L46 I have opened a ticket with AWS support on...

Just to report back: AWS support recommended to use a self-managed installation of the CSI driver for the time being, and requested me to watch this issue on Github.

@idanl21 Indeed, currently the only solution is to remove the AWS-managed Add-In, and install the ebs-csi-controller yourself.

Same/Similar issue here: Updating a StatefulSet (thus replacing the pod on the same node) gave me this: ```` Warning FailedMount 2m34s kubelet Unable to attach or mount volumes: unmounted volumes=[mariadb-data],...

@vdhanan I encountered it when I updated a StatefulSet, replacing the Pod on the same node (in that case, the EBS volume was bound to the same node before).

@gtxu Just happened again, terminationGracePeriodSeconds of the pod was set to 30 in this case.

I am just about to introduce helm here, and am thinking hard on a policy regarding this... And I am more or less set on keeping them always identical. That...

FYI: as a workaround, I specified the `NodePort` in the service and added it to the ports list. But I'd prefer the automatic port assignment by K8s, to avoid conflicts.

Yes, I created the TargetGroupBinding manually... I did not know that Instance Mode ist not supported yet. But with my little workaround, it works fine... Are there any traps in...

/remove-lifecycle rotten