PV gets stuck "timeout expired waiting for volumes to attach/mount for pod" on Dedicated

Issue

• When we scaled down and up the pods, the pods stopped working. Event logs are below:

FIRSTSEEN   LASTSEEN   COUNT     NAME                   KIND      SUBOBJECT                     REASON       SOURCE                                                        MESSAGE
...
1h        1h        1         mysql-1-094kd         Pod                                                    Scheduled           {default-scheduler }                                          Successfully assigned mysql-1-094kd to <HOST>
1h        1h        4         mysql-1-094kd         Pod                                                    FailedMount         {kubelet <HOST>}   Unable to mount volumes for pod "mysql-1-xxx-xxxx(e2cc2fc6-d85b-11e6-8938-026f2b6a837f)": timeout expired waiting for volumes to attach/mount for pod "mysql-1-094kd"/"xxxxx". list of unattached/unmounted volumes=[mysql-data]
1h        1h        4         mysql-1-094kd         Pod                                                    FailedSync          {kubelet <HOST>}   Error syncing pod, skipping: timeout expired waiting for volumes to attach/mount for pod "mysql-1-094kd"/"xxxx-xxxx". list of unattached/unmounted volumes=[mysql-data]

• I heard from Red Hat that it was is an issue on OCP 3.3 and should not scale down/up quickly. But when I waited 10 seconds, and then scaled them back from 0 to 1. The issue happened again.