Description of problem:
Delete a Pod which uses a local storage(directory not mountpoint), the Pod's status is "Terminating" after about 20 minutes.


Version-Release number of selected component (if applicable):
openshift v3.7.0-0.126.4
kubernetes v1.7.0+80709908fd

How reproducible:
Always

Steps to Reproduce:
1. On one onde, create 2 directories(/mnt/disks/vol1, /mnt/disks/vol2).
2. Ensure 2 PVs was created for vol1, vol2
3. Create a PVC using one of the PVs created step forward.
4. Create a Pod use PVC created step forward.
5. Read/write from/to Pods.
$ oc exec local-volume-pod-2 -- touch /mnt/local/testfile1
$ oc exec local-volume-pod-2 -- touch /mnt/local/testfiles
6. Delete Pod
$ oc delete pod local-volume-pod-2


Actual results:
The Pod's status is "Terminatiing" after about 20 minutes.

Expected results:
Pod was deleted successfully after a few minitues.

Master Log:

Node Log (of failed PODs):
Sep 20 05:51:40 host-8-241-40 atomic-openshift-node: W0920 05:51:40.751977    1742 util.go:87] Warning: "/var/lib/origin/openshift.local.volumes/pods/063cf03e-9de9-11e7-9400-fa163e501ea4/volumes/kubernetes.io~local-volume/local-pv-b77559f6" is not a mountpoint, deleting
Sep 20 05:51:40 host-8-241-40 atomic-openshift-node: E0920 05:51:40.752027    1742 nestedpendingoperations.go:262] Operation for "\"kubernetes.io/local-volume/063cf03e-9de9-11e7-9400-fa163e501ea4-local-pv-b77559f6\" (\"063cf03e-9de9-11e7-9400-fa163e501ea4\")" failed. No retries permitted until 2017-09-20 05:51:42.752007921 -0400 EDT (durationBeforeRetry 2s). Error: UnmountVolume.TearDown failed for volume "pvol" (UniqueName: "kubernetes.io/local-volume/063cf03e-9de9-11e7-9400-fa163e501ea4-local-pv-b77559f6") pod "063cf03e-9de9-11e7-9400-fa163e501ea4" (UID: "063cf03e-9de9-11e7-9400-fa163e501ea4") : remove /var/lib/origin/openshift.local.volumes/pods/063cf03e-9de9-11e7-9400-fa163e501ea4/volumes/kubernetes.io~local-volume/local-pv-b77559f6: device or resource busy
Sep 20 05:51:42 host-8-241-40 atomic-openshift-node: I0920 05:51:42.756039    1742 reconciler.go:186] operationExecutor.UnmountVolume started for volume "pvol" (UniqueName: "kubernetes.io/local-volume/063cf03e-9de9-11e7-9400-fa163e501ea4-local-pv-b77559f6") pod "063cf03e-9de9-11e7-9400-fa163e501ea4" (UID: "063cf03e-9de9-11e7-9400-fa163e501ea4")
Sep 20 05:51:42 host-8-241-40 atomic-openshift-node: W0920 05:51:42.756188    1742 util.go:87] Warning: "/var/lib/origin/openshift.local.volumes/pods/063cf03e-9de9-11e7-9400-fa163e501ea4/volumes/kubernetes.io~local-volume/local-pv-b77559f6" is not a mountpoint, deleting
Sep 20 05:51:42 host-8-241-40 atomic-openshift-node: E0920 05:51:42.756248    1742 nestedpendingoperations.go:262] Operation for "\"kubernetes.io/local-volume/063cf03e-9de9-11e7-9400-fa163e501ea4-local-pv-b77559f6\" (\"063cf03e-9de9-11e7-9400-fa163e501ea4\")" failed. No retries permitted until 2017-09-20 05:51:46.756216316 -0400 EDT (durationBeforeRetry 4s). Error: UnmountVolume.TearDown failed for volume "pvol" (UniqueName: "kubernetes.io/local-volume/063cf03e-9de9-11e7-9400-fa163e501ea4-local-pv-b77559f6") pod "063cf03e-9de9-11e7-9400-fa163e501ea4" (UID: "063cf03e-9de9-11e7-9400-fa163e501ea4") : remove /var/lib/origin/openshift.local.volumes/pods/063cf03e-9de9-11e7-9400-fa163e501ea4/volumes/kubernetes.io~local-volume/local-pv-b77559f6: device or resource busy
Sep 20 05:51:44 host-8-241-40 atomic-openshift-node: I0920 05:51:44.221594    1742 openstack_instances.go:45] Claiming to support Instances
Sep 20 05:51:45 host-8-241-40 atomic-openshift-node: W0920 05:51:45.665453    1742 helpers.go:771] eviction manager: no observation found for eviction signal allocatableNodeFs.available
Sep 20 05:51:46 host-8-241-40 atomic-openshift-node: I0920 05:51:46.763666    1742 reconciler.go:186] operationExecutor.UnmountVolume started for volume "pvol" (UniqueName: "kubernetes.io/local-volume/063cf03e-9de9-11e7-9400-fa163e501ea4-local-pv-b77559f6") pod "063cf03e-9de9-11e7-9400-fa163e501ea4" (UID: "063cf03e-9de9-11e7-9400-fa163e501ea4")


PV Dump:

PVC Dump:

StorageClass Dump (if StorageClass used by PV/PVC):

Additional info:
$ oc get pod
NAME                                       READY     STATUS        RESTARTS   AGE
local-volume-pod-2                         0/1       Terminating   0          2m
local-volume-provisioner-5lfv2             1/1       Running       1          52m
local-volume-provisioner-bootstrap-4bwgb   0/1       Completed     0          52m
local-volume-provisioner-bvgcp             1/1       Running       0          52m
local-volume-provisioner-ghs53             1/1       Running       0          52m

$ oc get pod
NAME                                       READY     STATUS        RESTARTS   AGE
local-volume-pod-2                         0/1       Terminating   0          23m
local-volume-provisioner-5lfv2             1/1       Running       1          1h
local-volume-provisioner-bootstrap-4bwgb   0/1       Completed     0          1h
local-volume-provisioner-bvgcp             1/1       Running       0          1h
local-volume-provisioner-ghs53             1/1       Running       0          1h