The test could fail at times since the mirror daemon periodically
retries the failed state for a mirror enabled file system. This
flips the status reported via asok interface.
Fixes: http://tracker.ceph.com/issues/50020
Signed-off-by: Venky Shankar <vshankar@redhat.com>
(cherry picked from commit
6f4201d15f37ec8e0d24d192b03cf2a272fe6b78)
self.mgr_cluster.mon_manager.raw_cluster_cmd("fs", "mirror", "enable", self.primary_fs_name)
# need safe_while since non-failed status pops up as mirroring is restarted
# internally in mirror daemon.
- with safe_while(sleep=5, tries=10, action='wait for failed state') as proceed:
+ with safe_while(sleep=5, tries=20, action='wait for failed state') as proceed:
while proceed():
try:
# verify via asok