Pod using RBD backend storage does not unmount/mount properly

Issue

• When we deployed pods, new pods cannot launch on new node.

• It looks like pods failed to unmount/mount:

# oc get event -w| grep FailedMount
2018-08-14 21:08:26 +0000 UTC   2018-08-14 21:01:36 +0000 UTC   4         mytest-13-7z2kj.154adbfb84345dd5       Pod                                                              Warning   FailedMount                   kubelet, elastic
HOSTNAME   Unable to mount volumes for pod "mytest-13-7z2kj_kfryklun-test(f1e4b6e9-a004-11e8-a5e9-fa163ed573e6)": timeout expired waiting for volumes to attach/mount for pod "kfryklun-test"/"mytest-13-7z2kj". list of unattached/unmounted volumes=[volume-g8s0j]
2018-08-14 21:10:43 +0000 UTC   2018-08-14 21:10:43 +0000 UTC   1         mytest-13-7z2kj.154adc7b03ade53f       Pod                                                              Warning   FailedMount                   kubelet, elastic
HOSTNAME   Unable to mount volumes for pod "mytest-13-7z2kj_kfryklun-test(f1e4b6e9-a004-11e8-a5e9-fa163ed573e6)": timeout expired waiting for volumes to attach/mount for pod "kfryklun-test"/"mytest-13-7z2kj". list of unattached/unmounted volumes=[volume-g8s0j default-token-hk9xq]

• We are using RBD for backend storage.