[13:25:29] <Krinkle>	 Puppet question - Should I bother with ensure=>absent for `systemd::timer::job` or is this a managed resource where e.g. puppet knows to remove untracked stuff on the server?
[13:25:56] <Krinkle>	 I looked in the job() function for some (to me, obvious) indications of it being managed but didn't see anything stand out.
[13:26:30] <legoktm>	 you have to ensure=>absent
[13:27:52] <Krinkle>	 ok :) Is that something you'd know by reviewing the code or is there a trick or shortlist way to find out, like if you came across a resource type you hadn't worked with before
[13:47:53] <jbond>	 Krinkle: its not something that can be easily ascertain.  Its safer to assume that you have to explicitly set `ensure => absent` 
[13:50:11] <claime>	 Krinkle: And not remove the attributes, or the type will cry
[13:50:30] <claime>	 (well it's *some* of the attributes, but I can't remember which exactly)
[13:50:39] <jbond>	 in genral only core resources have the ability to automaticly purge unmanaged resources and for them you would see use of the resource type e.g. {'user': purge => true } 
[13:50:55] <Krinkle>	 ack, I used to do that but then I recall more experienced folks with more minimalist styles but maybe that was too long ago or not actually better. I'll stick to simple absence for now.
[13:51:08] * Krinkle jumps back into a burrow labelled "JavaScript & PHP" – https://en.wikipedia.org/wiki/Burrow#/media/File:Pr%C3%A4riehund_P1010308.JPG
[13:51:16] * Krinkle is annoyed that "burrow" is not "mole hole" in English.
[13:51:16] <jbond>	 there asr some caveates to this for example apt has a purge_unmanaged_source parameter
[13:51:24] <jbond>	 but in genral best to assume you need to be explicit
[13:51:30] <claime>	 I usually make a stacked patchset with the ensure => absent, and a patch on top with the actual removal of the resource
[13:51:46] <claime>	 Krinkle: mhole
[13:52:07] <Krinkle>	 careful, you might trigger a 15-min Austin Powers sketch
[13:52:16] <claime>	 :')
[15:14:05] <apergos>	 whack-a-mhole?
[16:27:11] * James_F hugs legoktm for doing the first Shellbox deploy for four months.
[16:27:46] <legoktm>	 :D
[16:27:49] <legoktm>	 it's...too stable?
[16:31:15] <James_F>	 legoktm: Ideally services would get refreshed for minor updates along the way, though
[16:31:35] <James_F>	 legoktm: `rg 'version: "2021'` in deployment-charts makes me sad, for instance.
[16:44:35] <duesen>	 Hi all! https://phabricator.wikimedia.org/T329366 seems stuck, what can I do to un-stick it? How do we determine a good set of wikis to enable the parsoid cache warming jobs on? Should I just do group0?
[17:09:10] <duesen>	 ...I got impatient: https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/912929