[09:08:09] just fyi. I left a process running on tools-k8s-worker-nfs-47 gathering some stats of dependencies usage by tools (populates `toolforge_internal_dependencies` metric in prometheus), in case you find it and wonder about it [09:11:59] https://usercontent.irccloud-cdn.com/file/09jMAutW/image.png [10:18:34] I'm creating a new cloudvps project T391742, the cookbook is warning me that the name should not contain underscores [10:18:34] T391742: Request creation of wmgmc_monitoring VPS project - https://phabricator.wikimedia.org/T391742 [10:19:11] but dcaro suggested a name with an underscore in the task... are they ok? I remember dashes also had issues in the past [10:19:34] The cookbook is saying "Project names should generally be limited to alphanumeric characters and dashes, otherwise they might have issues with DNS." [10:19:59] yes, please use a dash instead of an underscore [10:20:23] is the problem with dashes fixed? I think it was with object storage [10:20:44] yes, we no longer use the name as project id [10:21:27] nice, I didn't follow at which stage we were with that change :) [10:29:04] the cookbook wmcs.vps.create_project worked very nicely :) [15:26:38] T392294 is looking for a +1 followed by a quota bump and exposing a flavor in the new zuul3 projects. [15:26:39] T392294: Quota increase for zuul3 project - https://phabricator.wikimedia.org/T392294 [15:28:20] bd808: I can +1 but I'm also on clinic duty so ideally we would need somebody else to +1 so I can apply the change :) [15:45:09] dhinus: heh. seems reasonable. I don't think it is urgent [16:40:09] I went to +1 but was too late. The zuul3 work has needed doing for almost a decade, we must not interrupt the momentum! [16:42:58] It's going to be a project all quarter at least, but yes it is quite long overdue at this point. [16:43:56] taavi or anyone else who understands our go+tofu setup: I was up late last night banging my head against this particular wall: [16:43:57] https://phabricator.wikimedia.org/P75286 [16:44:09] any ideas how to detect that very very simple thing? [16:44:40] Zuul starts and stops have correlation with hunting for a completely different system to replace it, picking one, and then having a course correction that said we would replace the whole gerrit stack. [16:46:47] yeah [16:47:38] I remember being pretty impressed by zuul3 when they introduced it, but I also think the devs are maybe jerks for doing a total tear-down/rewrite with no migration path. [16:53:53] andrewbogott: we are going to contract with the core zuul folks to help us move the jobs, so maybe long term job security? I think Tyler said the guy said something like "I haven't done that migration for anyone in a very long time" though. ;) [16:54:42] andrewbogott: for your other question, I'm not following all of the discussion at https://github.com/gophercloud/gophercloud/issues/939, but I think it is related. [16:55:22] oh! You're right -- I was assuming that I didn't understand how to parse the response but that suggests that it's just silly. [16:56:41] Jenna knows Go so I spent a long time last night with her looking over my shoulder and I kept saying things like "surely I'm not supposed to just do a string match on this error message" [16:57:08] (we're running gophercloud 1.0 so I bet all those 'we fixed it' messages in that forum don't help but I will check) [16:57:11] heh. I love hearing that Jenna is an oracle for you now too! [16:57:25] It's handy [16:57:32] Thank you for googling for me! [16:57:53] yw. I am easily distracted on Friday mornings (and most other times) [16:58:48] waaait they're suggesting that I do the thing that I'm already doing [16:59:05] the very clear and concise "if _, ok := err.(gophercloud.ErrDefault404); ok {" [17:12:22] oooooh, I think my code is correct but the get is returning 400 and not 404 [17:12:30] that makes me feel better for some reason