qa/cephadm: wait between starting and restarting osd in test_daemon_restart test

The test before the pause test ends with `ceph orch daemon restart osd.0`
which may temporarily cause a CEPHADM_FAILED_DAEMON warning for that
OSD. It seems to be failing with "status=219/CGROUP", logging nothing
else of note, and coming up totally fine when restarted by systemd
a second later (see the linked tracker for more info/logs). This
sleep is me thinking this may be timing related and that if we
separate the start from the restart by a few more seconds it might
avoid the error.

Fixes: https://tracker.ceph.com/issues/69526
Signed-off-by: Adam King <adking@redhat.com>

diff --git a/qa/tasks/cephadm_cases/test_cli.py b/qa/tasks/cephadm_cases/test_cli.py
index ca40a8cdb0f8870b3ca943aea16c9c6285299ec9..faa6fe7c33eaf89b5ee1e13253c926b21d8f0158 100644 (file)
--- a/qa/tasks/cephadm_cases/test_cli.py
+++ b/qa/tasks/cephadm_cases/test_cli.py
@@ -60,6 +60,8 @@ class TestCephadmCLI(MgrTestCase):
         time.sleep(5)
         self._orch_cmd('daemon', 'start', 'osd.0')
         self.wait_for_health_clear(120)
+        # this sleep is to try and address https://tracker.ceph.com/issues/69526
+        time.sleep(5)
         self._orch_cmd('daemon', 'restart', 'osd.0')
 
     def test_device_ls_wide(self):