from cStringIO import StringIO
from teuthology.orchestra import run
+from teuthology.misc import reconnect, get_first_mon, wait_until_healthy
log = logging.getLogger(__name__)
log.info("Failed to stop ceph mds service")
remote.run(args=['sudo', 'systemctl', 'start', mds_name])
time.sleep(4)
+
+ # reboot all nodes and verify the systemd units restart
+ # workunit that runs would fail if any of the systemd unit doesnt start
+ ctx.cluster.run(args='sudo reboot', wait=False, check_status=False)
+ # avoid immediate reconnect
+ time.sleep(120)
+ reconnect(ctx, 480) # reconnect all nodes
+ # for debug info
+ ctx.cluster.run(args=['sudo', 'ps', '-eaf', run.Raw('|'),
+ 'grep', 'ceph'])
+ # wait for HEALTH_OK
+ mon = get_first_mon(ctx, config)
+ (mon_remote,) = ctx.cluster.only(mon).remotes.iterkeys()
+ wait_until_healthy(ctx, mon_remote)
yield