qa: allow client mount to reset fully

... without this there is a sutle race where it takes a
bit of time after a hard reset of a client mount causing
further checks to fail as the (still up) client is still
connected to the MDS.

Fixes: http://tracker.ceph.com/issues/42213
Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit dcbdc726a694713b9ebd01954e173b0d4ec4af0f)

diff --git a/qa/tasks/cephfs/test_client_recovery.py b/qa/tasks/cephfs/test_client_recovery.py
index 747669d88ec885bbd4a54b453031bbb9e9334579..5d2cd708f82b55870e751bb437e1ce20edc3baa7 100644 (file)
--- a/qa/tasks/cephfs/test_client_recovery.py
+++ b/qa/tasks/cephfs/test_client_recovery.py
@@ -187,6 +187,9 @@ class TestClientRecovery(CephFSTestCase):
         # The mount goes away while the MDS is offline
         self.mount_a.kill()
 
+        # wait for it to die
+        time.sleep(5)
+
         self.fs.mds_restart()
 
         # Enter reconnect phase