[15:47:16] andrewbogott: catching up with the patch, and test you did [15:47:32] are you seing we could merge that also right now? [15:49:13] Yep, let's merge it. [15:49:40] bleh, the clinic duty rotation script is failing because the etherpad export endpoint is apparently failing [15:52:53] transient or did they block something ? [15:53:14] it's a 504 but i'm getting it constantly, so not sure [15:54:33] maybe ask collab services, if they are not too busy right now [15:55:02] andrewbogott: are you ok with https://gerrit.wikimedia.org/r/c/operations/puppet/+/1371719 too? [15:57:08] have we really not noticed it being broken for 6 years? [15:57:27] volans: yes, +1 [15:57:57] I mean, maybe it's actually working even with the colon? I haven't tested it to be honest [15:58:24] i know at least :: has some special meaning in arrays in apt configs, i wonder if : is some operator as well [16:01:51] doesn't seem so from man page and "the internet" [16:02:10] while :: is the scope separator and a trailing :: has a meaning [16:02:39] re etherpad, filed a task https://phabricator.wikimedia.org/T445565 [16:02:46] and i guess i'll do the steps manually then [16:03:08] :'( [16:04:08] taavi: Max got back to me about IRC. He's MadMaxWP, but currently intermittently online as he's been using irccloud.com free tier which disconnects after 2h idle. [16:04:33] I told him to check out https://wikilounge-users.wmcloud.org/ and see if it would work for him. [16:05:07] yep, /var/lib/apt/periodic/autoclean-stamp not present on any VM I tested (all tools ones for exmaple), so has not been runing [16:26:02] Raymond_Ndibe: if you still have that lima-kilo build up from yesterday, can you tell me if the 'envvars are set inside jobs' test is passing for you? I updated dependencies for envvars-admission and it seems to not work anymore (although it also doesn't crash) [18:16:40] https://gitlab.wikimedia.org/repos/cloud/toolforge/toolforge-deploy/-/merge_requests/1533 [18:17:07] that's live (in dry-run mode) in toolsbeta now. once someone stamps that i'll put it in dry-run mode in tools, then look at turning dry run mode off in both places [18:18:38] Raymond_Ndibe: disregard for now, I think I found a workaround [18:47:25] hm, the toolforge production chart has a toolsbeta-harbor reference for some reason [18:49:47] which chart? [18:55:01] dhinus: the new kube-container-updater one, https://gitlab.wikimedia.org/repos/cloud/toolforge/kube-container-updater [18:57:47] hm, the values.yaml in the repo needs to point to tools-harbor and ci rewrites it to toolsbeta-harbor when necessary, and not the other way? [18:58:27] I always get lost myself with that flow :) maybe we should revisit it and/or document it better [18:59:11] I'm also not sure we're doing it in a consistent way for all the other components [18:59:53] let's try that then https://gitlab.wikimedia.org/repos/cloud/toolforge/kube-container-updater/-/merge_requests/5 [19:03:03] that looks correct based on https://gitlab.wikimedia.org/repos/cloud/cicd/gitlab-ci/-/blob/main/toolforge-cd/only_helm_publish.yaml?ref_type=heads#L62 [19:03:11] but there are so many moving parts that I'm not 100% sure [19:04:15] also I would probably not use a real repo name as a placeholder, but I'm sure it was because [19:04:16] ahhh i was looking at that line earlier but confused the source and target of the replacement i guess