Page MenuHomePhabricator

PAWS SSL CRITICAL - Certificate paws.wmcloud.org valid until 2022-05-17 15:59:48 +0000 (expires in 2 days)
Closed, ResolvedPublic

Description

I haven't investigated this yet but am creating this task for tracking purposes.

Assuming that paws is using acme-chief this probably just requires a service restart.

Event Timeline

Mentioned in SAL (#wikimedia-cloud) [2022-05-14T16:16:57Z] <andrewbogott> restarting acme-chief.service on paws-acme-chief-01 for T308383

Thank you! I think the restart did it:

openssl s_client -servername paws.wmcloud.org -connect paws.wmcloud.org:443 2>/dev/null | openssl x509 -noout -dates
notBefore=May 14 15:16:50 2022 GMT
notAfter=Aug 12 15:16:49 2022 GMT

Looks like we're good until August.

This should not be needed, the service has a set of timers that should be triggering the config updates instead, maybe that is not working anymore, see T307333: [tools] toolserver.org cert is expiring in 2 days

yep, that seems to be the case (see that NEXT is n/a):

root@paws-acme-chief-01:~# systemctl list-timers
NEXT                         LEFT          LAST                         PASSED       UNIT                                            ACTIVATES
...
n/a                          n/a           Mon 2021-11-22 23:59:14 UTC  5 months 22 days ago reload-acme-chief-backend.timer                 reload-acme-chief-backend.service
...

It's not clear why (a race condition with the timer + when puppet changed it) but restarting it works, and it should be up again the next time the VM reboots too:

root@paws-acme-chief-01:~# systemctl restart reload-acme-chief-backend.service
...


root@paws-acme-chief-01:~# systemctl list-timers
NEXT                         LEFT          LAST                         PASSED       UNIT                                            ACTIVATES
...
Mon 2022-05-16 10:22:20 UTC  59min left    Mon 2022-05-16 09:22:20 UTC  30s ago      reload-acme-chief-backend.timer                 reload-acme-chief-backend.service