[08:01:21] greetings [08:04:38] Morning [08:18:21] FYI re: reimaging cloudvirts, I am reimaging them with the new role to get netbox updated with the vlan interfaces, FWIW [08:18:43] so yes easier if reimage happens when transitioning virt roles from insetup [10:14:04] ack [12:57:07] the name is not confusing at all... https://phabricator.wikimedia.org/T436892 [13:41:47] at last it isn't 'labs' [13:44:22] toolserver labs? [13:45:10] lol [14:34:21] It's weird that systemd just doesn't provide any facility for removing units, right? Isn't that a thing that people want to do sometimes? [15:00:48] I guess you can just delete them? (though there's things like templates and such that make it harder to interpret what "remove unit" means) [15:01:04] you can mask them I guess for that [15:01:51] I always wind up spending 15 minutes chasing down loose ends when if could provide a 'delete' command that removed all the parts and fixed dependencies... [15:02:10] not a big deal, just weird that their design model assumes people will only ever add units and never remove them [15:02:16] oh yep, deps can be tricky [15:45:59] thilp: hey, if you need any reviews for the logs, just ping me here, the fact that there's 4 MRs attached to the task and each has it's own review process makes it impossible for me to know if/which MRs need reviews [15:46:15] (by just looking at where the task column) [15:46:26] sounds good! [15:47:02] Honestly I think we could have done 1 MR with 4 commits, and then squash at merge [15:47:44] if people are fine reviewing commit by commit (if they so choose), I’d do that next time [15:55:13] it's usually easier to review the changes one by one (if they are big enough), and can be merged independently [15:55:49] (as always, this does not mean "always" split your changes in as many MRs as possible) [15:56:44] as a rule of the thumb, I use >400 lines of reviewable changes, usually start thinking on how to split if possible [16:12:50] reviewing changes one by one works commit by commit too, but I agree with you on the benefits of merging asap [16:13:52] testing too (we don't really have a 'paved' way to toolforge_depoy a specific commit from an MR) [16:14:06] ack [17:02:54] * dcaro off [22:25:24] something noisy is happening with wmf-pt-kill@s4.service on clouddb1024 but I'm not going to get into it before dinner (and possibly not after dinner either) [23:51:47] anybody have history on this page and what it was intended to hold? https://wikitech.wikimedia.org/wiki/Wikitech:Cloud_Services_Cross_Site_Policy -- apparently it is mentioned in the WMCS TOU but was never really fleshed out