[07:46:29] <dcaro>	 thanks for that! try to record them in the task, for the next time...
[14:57:56] <andrewbogott>	 taavi or dhinus, https://gerrit.wikimedia.org/r/c/operations/puppet/+/1172656 look ok?
[14:58:47] <dhinus>	 andrewbogott: did it start immediately after the upgrade to epoxy?
[14:58:57] <dhinus>	 I don't love it but I think it's ok...
[14:59:08] <andrewbogott>	 yes, it did
[14:59:19] <andrewbogott>	 how are you with cron intervals? That one is copypasta and looks weird to me
[14:59:40] <dhinus>	 I used to be good at it, but I think I'm losing my kung-fu by lack of practice :D
[14:59:58] <andrewbogott>	 systemd intervals are the same as cron intervals right?
[15:00:30] <taavi>	 no
[15:00:36] <taavi>	 systemd-analyze is your friend:
[15:00:36] <taavi>	 $ systemd-analyze calendar "*-*-* 15:0/21:00"
[15:00:36] <taavi>	   Original form: *-*-* 15:0/21:00
[15:00:36] <taavi>	 Normalized form: *-*-* 15:00/21:00
[15:00:36] <taavi>	     Next elapse: Sat 2025-07-26 15:00:00 EEST
[15:00:36] <taavi>	        (in UTC): Sat 2025-07-26 12:00:00 UTC
[15:00:37] <taavi>	        From now: 20h left
[15:01:17] <dhinus>	 TIL
[15:01:28] <andrewbogott>	 ok, that explains why I was so confused! And also... I hate that they invented a new equally terrible way to do that
[15:01:32] <andrewbogott>	 thanks taavi 
[15:01:43] <taavi>	 andrewbogott: you need to remove `0/` from the minute, you want to run it on the Nth minute and not on every Nth minute starting from :0
[15:01:58] <andrewbogott>	 right, ok!
[15:02:18] <andrewbogott>	 patch updated
[15:02:40] <andrewbogott>	 (I was hoping the pcc would show me the timestamp string but of course it does not)
[15:04:18] <taavi>	 pcc does show the `OnCalendar=*-*-* 15:0/21:00` line, are you looking for something else?
[15:04:25] <andrewbogott>	 ?
[15:04:27] * andrewbogott looks again
[15:05:40] <andrewbogott>	 you're looking at the full json catalog?
[15:05:56] <taavi>	 "full diff" is what I usually look at
[15:05:57] <andrewbogott>	 oh, 'full diff'. 
[15:06:15] <andrewbogott>	 has that...always been there? I'm used to seeing that output on the front diff page
[15:06:20] <andrewbogott>	 Maybe it was split into two places
[15:07:33] <taavi>	 yes
[15:07:45] <taavi>	 the front page shows a diff for changed things, for new things you need the full diff
[15:08:42] <andrewbogott>	 I'm learning all kinds of things today
[15:10:10] <andrewbogott>	 [{'start': 'OnCalendar', 'interval': '*-*-* 15:21:00'}]  and [{'start': 'OnCalendar', 'interval': '*-*-* 15:6:00'}]
[15:10:12] <andrewbogott>	 that seems good
[15:11:40] <andrewbogott>	 can I get a +1?
[16:17:42] <dhinus>	 andrewbogott: do you know what's the CI failure on https://gerrit.wikimedia.org/r/c/operations/puppet/+/1172656? I don't remember seeing that "fail fast" before
[16:18:25] <dhinus>	 ah I see the error now, it's in the CI output
[16:18:41] <andrewbogott>	 I assume (without evidence) that our puppet 5 infra is decaying for lack of use :)
[16:18:54] <dhinus>	 puppet7 is also "warning"
[16:19:06] <dhinus>	 " Experimental build failed."
[16:19:16] <dhinus>	 CRITICAL: Unexpected error running run_host: [Errno 2] No such file or directory: '/var/lib/catalog-differ/puppet/yaml/production7/yaml/facts/cloudnet1006.eqiad.wmnet.yaml.tmp'
[16:19:35] <andrewbogott>	 yeah, so no fact list for puppet 5
[16:19:43] <andrewbogott>	 is 7 warning? I don't see that, I'm looking at https://puppet-compiler.wmflabs.org/output/1172656/4573/
[16:20:44] <dhinus>	 yes no error there, interesting
[16:21:14] <dhinus>	 but I see this in gerrit https://usercontent.irccloud-cdn.com/file/eJHEJvRL/Screenshot%202025-07-25%20at%2018.21.03.png
[16:21:39] <andrewbogott>	 huh
[16:22:19] <dhinus>	 and the "CRITICAL" error above in https://integration.wikimedia.org/ci/job/operations-puppet-catalog-compiler-test/7087/console
[16:23:13] <dhinus>	 ok your link is a different build
[16:23:27] <andrewbogott>	 It's not the first time I've seen puppet 5 fail in the pcc, I'm used to ignoring it. 
[16:23:45] <andrewbogott>	 Maybe it's telling us something but I think most likely it's telling us that the puppet 5 runner doesn't have facts for the host
[16:24:51] <dhinus>	 I'm very confused by the output in gerrit (screenshot above), but maybe you're right it's just puppet5 being flaky
[16:25:18] <dhinus>	 I think that it can be ignored
[16:25:39] <dhinus>	 I +1d it
[16:26:43] <andrewbogott>	 thanks!
[16:28:58] <dhinus>	 can puppet be re-enabled on cloudcontrol2xxx?
[16:41:24] <dhinus>	 I'm off, good weekends!