Page MenuHomePhabricator

maps-wma1 instance unresponsive (second time in 3 days)
Closed, ResolvedPublic

Description

I am unable to log into the instance maps-wma1.maps.eqiad.wmflabs (also reachable at http://wma.wmflabs.org) currently. My public key is not accepted (leading me to speculate that maybe a filesystem is not mounted?)
Could anyone please take a look? This happened on Saturday as well and I "fixed" it with a reboot, but I fear there might be a pattern behind this and I don;t wan t my instance to die every few days.

Event Timeline

dschwen raised the priority of this task from to Needs Triage.
dschwen updated the task description. (Show Details)
dschwen subscribed.

I can't seem to ssh into this even as root :|

A restart didn't do anything either.

Hmm, I wonder / suspect if puppet ever ran on this instance in a while? https://tools.wmflabs.org/nagf/?project=maps is all empty.

We can try with @Andrew or @coren's key when they're back from their vacation since it's been there longer than mine or chase's, but I also recommend rebuilding the instance side by side...

Andrew claimed this task.

My key worked. The puppet cron was clearly disabled -- I started a puppet run and things seem reasonable now. Puppet still throws a bunch of errors, but those should probably be handled by whoever set up the conf in the first place.

(Also, obligatory: don't disable puppet! It makes us waste time like we just did.)

Thanks Andrew. I was not aware that I disable puppet. I'll check out what I did there.