qa/tasks/ceph: retry several times to tell mons ot stop logging health

If we have any sort of failure injection, one attempt is not enough.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 662dbf9c6c6567a5eaad8d5c9fe7f936b939b423)

diff --git a/qa/tasks/ceph.py b/qa/tasks/ceph.py
index 7f0976536959d107a81e5b2f7649e8535b6c169d..95d094e4178f6fcd4cc6441c77c669170c9c3cba 100644 (file)
--- a/qa/tasks/ceph.py
+++ b/qa/tasks/ceph.py
@@ -1938,11 +1938,13 @@ def task(ctx, config):
             # a bunch of scary messages unrelated to our actual run.
             firstmon = teuthology.get_first_mon(ctx, config, config['cluster'])
             (mon0_remote,) = ctx.cluster.only(firstmon).remotes.keys()
+            # try this several times, since tell to mons is lossy.
             mon0_remote.run(
                 args=[
                     'sudo',
                     'ceph',
                     '--cluster', config['cluster'],
+                    '--mon-client-directed-command-retry', '5',
                     'tell',
                     'mon.*',
                     'injectargs',