[00:51:25] inflatador: hm, I'm not sure why it got stuck, I've restarted it (the healthcheck should've done that automatically...) - if you sign up for toolforge, I can make you a maintainer :) tbh it might be one of those things that should get properly implemented on apt.wm.o properly instead of having a tool scrape it if folks are relying on it [13:10:35] legoktm sure, I'm happy to help out as a maintainer. Will ping you later today [18:58:20] urandom: so what was the fix for aqs102? [18:58:23] *1012 [18:58:26] I see signs of life there [18:58:43] I'm trying to reimage, preserving the partitions [18:58:50] but I think that will probably fail [18:59:04] because disks were replaced, the partitions tables aren't right [18:59:06] but you did manage to get past the failing boot screen? [18:59:10] oh no [18:59:18] sukhe: https://phabricator.wikimedia.org/T215183 [18:59:47] which is basically my theory from yesterday [19:00:24] vaguely recall this one. sorry for all the fun you suffer with hardware <3 [19:01:21] I don't like hardware. [19:01:29] And I'm pretty sure the feeling is mutual. [19:02:23] urandom: T350179 [19:02:23] T350179: Reimage cookbook on new eqiad hosts stuck at PXE booting - https://phabricator.wikimedia.org/T350179 [19:02:37] we debugged it for months -- actual months -- before papaul found the issue [19:02:40] so yeah... [19:03:29] anyway, good luck for your adventures. shout out here if we can help :] [19:03:58] I think this host is going to be a loss [19:04:18] does d-i have any partitioning tools in the shell? [19:04:34] does d-i have any partitioning tools in the shell? <-- see the level of desperation? [19:04:39] ... [19:04:51] but yeah, once you are at the partiotining step, you can drop to a shell [19:05:03] right, but can you manually partition from there? [19:05:15] or simply switch to the alternate screens and run it [19:05:20] urandom: I think so [19:05:29] I don't see fdisk, or parted [19:05:40] or s[fg]disk [19:06:32] try anna-install'ing it? anna-install [19:10:44] swfrench-wmf: thought you might find this interesting RE: foreachwikiindblist ignoring file paths (which we seem to do in Puppet, e.g. maintenance/growthexperiments.pp). https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/1169737/1//COMMIT_MSG [19:18:50] Krinkle: thanks for the pointer! ah, yeah I vaguely recall some discussion around the "looks like a dblist file path is supported, but not really" behavior at one point ... I think that was in the context of considering adding support for it [19:19:43] There is a beta-specific patch that adds "& all-labs" to all these in expanddblist [19:20:06] so that would at glance conflict with the idea that passing a single arg would be a "real" file read as-is. [19:20:14] although the two aren't mutually exclusive [19:20:24] it could do both, but anyway, that's how it works today :)