long_running_cluster: Force logrotate at 06:25

The scheduled logrotate cronjob at 06:25 doesn't actually get run like I assumed.  This is because the cron.daily job doesn't use -f.

Signed-off-by: David Galloway <dgallowa@redhat.com>

diff --git a/roles/long_running_cluster/tasks/logrotate.yml b/roles/long_running_cluster/tasks/logrotate.yml
index bf66a6eb3ee912fe92b13babfb5ca3093d06d10a..27607cde67a631d0e3ff2bdd723b0a4e975fad75 100644 (file)
--- a/roles/long_running_cluster/tasks/logrotate.yml
+++ b/roles/long_running_cluster/tasks/logrotate.yml
@@ -2,8 +2,7 @@
 # Because of the high debug level enabled for LRC daemons, the root drives
 # fill up rather quickly.  The drives fill up before the daily logrotate can
 # run so we rotate every 6 hours and keep 3 days worth.  This can be adjusted
-# as needed.  Logrotate already runs daily at 06:25 so there are 3 additional
-# scheduled cronjobs to rotate 4x daily.
+# as needed.
 
 - name: "Write custom ceph logrotate config"
   template:
@@ -14,6 +13,6 @@
   cron:
     name: "Logrotate ceph logs every 6 hours"
     minute: "25"
-    hour: "0,12,18"
+    hour: "0,6,12,18"
     job: "/usr/sbin/logrotate -f /etc/logrotate.d/ceph-common"
     user: root