Neutron netns check CRIT/CRITICAL mismatch

[sudeephb]

The included check_netns.sh script for checking neutron gateway netns issues uses "CRIT" to report critical errors, but check_status_file.py expects "CRITICAL".

The former script is here, and writes its output to a file when called by cron, so the "exit $STATE_CRIT" never reaches the check, which can only read the contents of the log file:

https://git.launchpad.net/charm-nrpe/tree/files/plugins/check_netns.sh

An error state will be recorded in the log file as e.g. "CRIT: [...] aren't responding".

But the script used by the check, check_status_file.py, onlny looks for "CRITICAL" in the output, so doesn't match "CRIT":

https://git.launchpad.net/charm-nrpe/tree/files/plugins/check_status_file.py

---

Imported from Launchpad using lp2gh.

• date created: 2023-01-20T05:54:44Z

• owner: barryprice

• assignee: None

• the launchpad url

[sudeephb]

(by barryprice)
Marking this as also affecting the neutron-gateway charm, as that's where I came across it.

If desired, this could be worked around in that charm's code with something like:

•    check_cmd='check_status_file.py -f /var/lib/nagios/netns-check.txt'
  

•    check_cmd='check_status_file.py -c CRIT -f /var/lib/nagios/netns-check.txt'
  

https://opendev.org/openstack/charm-neutron-gateway/src/branch/master/hooks/neutron_hooks.py#L347

Happy to propose that fix if desired.

[sudeephb]

(by barryprice)
Actually coming around to the idea that this may be a bad check, I've found it in CRIT state on other clouds which appear to be working fine.

Have marked the linked MP as WIP while we investigate further, but it may be that a better fix is to either remove or update check_netns.sh rather than to trust its current output.

[sudeephb]

(by barryprice)
I don't think Launchpad allows blocking one bug against another, but I believe the linked bug here should be addressed before we decide what to do with this one:

https://bugs.launchpad.net/charm-nrpe/+bug/2003641