check_haproxy_servers and check_haproxy_queue have blank command defs causing "NRPE: Command x not defined" errors
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Dashboard Charm |
New
|
Undecided
|
Unassigned | ||
OpenStack Keystone Charm |
New
|
Undecided
|
Unassigned | ||
OpenStack Nova Cloud Controller Charm |
New
|
Undecided
|
Unassigned |
Bug Description
On the most recent release of charms, 18.11 on a Xenial/Queens cloud, I'm noticing that I have all units of openstack-
NRPE: Command 'check_<random check name>' not defined
On each unit, I'm seeing checks unrelated to check_haproxy_
When investigating, I see that the error in sysctl is showing:
Nov 21 18:19:26 juju-45c58f-
Nov 21 18:19:26 juju-45c58f-
Nov 21 18:19:26 juju-45c58f-
Nov 21 18:19:26 juju-45c58f-
Nov 21 18:19:26 juju-45c58f-
$ cat /etc/nagios/
# check haproxy_servers
# The following header was added automatically by juju
# Modifying it will affect nagios monitoring and alerting
# servicegroups: juju
command[
$ cat /etc/nagios/
# check haproxy_queue
# The following header was added automatically by juju
# Modifying it will affect nagios monitoring and alerting
# servicegroups: juju
command[
The parsing of config files by the nrpe daemon stops after detecting this error, leaving other defined checks to fail.
In fact, the shell scripts check_haproxy.sh and check_haproxy_ queue_depth. sh are not found in /usr/lib/ nagios/ plugins, which makes me think something changed in charmhelpers to stop landing those files upon nrpe.add-check calls.