check_resource_systemd might break under certain conditions
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
tripleo |
Invalid
|
Medium
|
Unassigned |
Bug Description
During upgrade we called "check_
Interestingly enough this broke at least once for me (likely due to a systemd bug). In short, "systemctl is-active memcached" seems to not work
correctly under some conditions. For example:
[root@overcloud
● memcached.service - Memcached
Loaded: loaded (/usr/lib/
Active: active (running) since Fri 2016-09-23 21:46:36 UTC; 8s ago
Main PID: 25285 (memcached)
CGroup: /system.
└─25285 /usr/bin/memcached -u memcached -p 11211 -m 7383 -c 8192 -l 172.16.2.9 -U 11211 -t 4 >> /var/log/
Sep 23 21:46:36 overcloud-
Sep 23 21:46:36 overcloud-
[root@overcloud
active
[root@overcloud
[root@overcloud
unknown
This breaks at least check_resource_
ERROR - memcached not found to be systemd managed.
For now I set the importance to medium, because I think I can work around it.
I have seen this issue with the mongod service as well. I wonder if this has to do with systemd and pacemaker interaction (i.e. when pacemaker adds a drop-in file) and systemd gets confused?