On the long run, the csi pods suffer the issue presented for the version v2.5.0 of the image k8s.gcr.io/sig-storage/csi-node-driver-registrar as reported in the known issue: https://github.com/kubernetes-csi/node-driver-registrar/issues/188
The version v2.5.1 should fix the problem.
Tested that prevents the restarts on a real k3s cluster for more than half an hour.
On the long run, the csi pods suffer the issue presented for the version v2.5.0 of the image k8s.gcr.io/sig-storage/csi-node-driver-registrar as reported in the known issue: https://github.com/kubernetes-csi/node-driver-registrar/issues/188 The version v2.5.1 should fix the problem. Tested that prevents the restarts on a real k3s cluster for more than half an hour.