[09:23:34] that's wrong ^ btw. It's not effie, it's me that's oncall. We 've swapped 2 months ago, I had even set up the overrides, they magically vanished [09:23:42] let's see if I can re set it up [09:29:07] supposedly, just {{done}} [09:29:09] let's see [09:31:15] perfect :-( [09:31:48] Amir1: fyi, you are not alone [09:32:19] !oncall-now [09:32:19] Oncall now for team SRE, rotation business_hours: [09:32:19] A.mir1 [09:33:06] !oncall-now [09:33:06] Oncall now for team SRE, rotation business_hours: [09:33:06] A.mir1 [09:33:09] pffff [09:34:03] !oncall-now [09:34:04] Oncall now for team SRE, rotation business_hours: [09:34:04] a.kosiaris, A.mir1 [09:34:08] aaaah, success! [09:47:19] Thanks akosiaris ! [10:16:07] I didn't get if akosiaris is on-call, can somebody tell me? [10:16:12] * elukey runs away [10:18:28] elukey: you can get the answer instantly as long as you ask 3 times [10:19:40] instantly* (*conditions apply) [10:22:40] elukey: I had switched with effie since july [10:22:52] and had gone ahead and setup the overrides since then [10:23:09] it was a mutual exchange of oncall shifts [10:23:39] and well today I went in, but found no override for effie. I did find one for me, telling me I was overriding myself [10:24:01] I have no idea what on earth, I am gonna pretend if was my fault, although it doesn't add up. [10:24:25] akosiaris: I admire the fact that you answered to me in a very precise way without any swear words. [10:24:48] :D [10:27:13] did anyone tried any vscode plugin for gerrit? like https://marketplace.visualstudio.com/items?itemName=SanderRonde.vscode--gerrit ? [10:29:01] Yeah, I tried that one, didn't find it very useful but I may be holding it wrong [10:29:29] elukey: :D [16:10:44] I'm reimaging a host and getting stuck at the partitioning stage; mdadm is trying to rebuild a raid and it appears that the rebuild will take days or weeks. The thing is, I don't /want/ to rebuild the raid, I just want to wipe it and start over. [16:10:47] Has anyone encountered this? [16:16:52] partman recipe is [16:16:56] - partman/standard.cfg [16:16:57] - partman/raid10-4dev.cfg [16:17:03] which as I understand it should be giving me a fresh start [16:31:22] I don't know if it helps, but in the past I had to "wipe" with the same installer the software raid because, despite the disk being logically formatted, the installer wanted to be smart and "recover" from found metadata [16:31:53] I don't remember if I had to write some 0s at the start, in adition from the regular removal or something [16:33:40] I can't figure out how to wipe, anytime I dry to remove the raid it tells me that it's in use [16:33:50] But yes, wiping it is surely the right thing to do [16:34:12] # mdadm --stop /dev/md0 [16:34:12] mdadm: Cannot get exclusive access to /dev/md0:Perhaps a running process, mounted [16:34:52] jynus: do you remember what you did to clear it? [16:35:26] I think dd zero and kept for some time until it worked, but it was a long time ago, so I don't remember [16:35:35] either that or manually partition something else [16:36:10] hm, I dd'd the partition table and that didn't seem to change anything [16:36:13] * andrewbogott googles further [16:36:27] it is one of those things that you try stuff until it works, despite "what should be working" (manual removal) didn't [16:36:33] * andrewbogott nods [16:36:42] I've already tried a dozen things :( [16:58:19] * andrewbogott dd's everything, yet the raid persists [17:07:23] maybe `shred` would work? You may have to hack the preseed file to get the installer to drop into interactive mode [17:07:48] and I guess that assumes `shred` exists in the installer, which is probably a bad assumption [17:13:30] dd should work if I'm hitting the right place, but the metadata must be hiding someplace else [17:13:42] Yeah, basically none of the tools that I want are present in this busybox installer though. [17:14:14] * andrewbogott is going to eat a sandwich and hope that the sandwich has the answers [17:15:44] only smarties have the answer [17:16:14] I'm in the US, wrong kind of smarties :( [17:19:57] maybe copy the `shred` binary into the directory that hosts TFTP from the installer servers? Of course that all depends on how crazy you wanna get [17:30:09] oh, it's moving past that step now! Maybe the dds fixed things but rebuilding took a few minutes [17:30:36] hmmm that being the bookworm installer. We'll see if bullseye can do the same [18:18:11] andrewbogott: did you succeed, happy to help with ideas. [18:18:42] going to work around it I think. It only installs with bookworm and I can maybe/sort of make that work [18:22:27] maybe you have to delete the superblock and could use mdadm itself for that. mdadm --zero-superblock [18:23:32] at least that's where the metadata should be [19:24:10] FWiW, I was messing around with preseed/partman in my homelab this weekend and I discovered that neither Ubuntu and the Debian Cloud Team use preseed/partman. Debian Cloud Team uses https://fai-project.org/fai-guide/ , maybe something to look into at some point? [19:24:41] neither Ubuntu **NOR** the Debian Cloud Team, that is [19:31:09] We don't have any plans at present inflatador, but I used FAI, many years ago, and really liked it [19:32:08] one nice feature of FAI is that it uses golden images, so you don't have to wait for packages to install, the files are just copied via nfs if I recall correctly [19:41:09] jhathaway that's cool. I came from RHEL-world so I mostly know kickstart (and packer). I might play around with https://salsa.debian.org/cloud-team/debian-cloud-images/-/tree/master if I get bored next weekend [19:43:13] definitely worth looking at, I would love to find some time to take a look at it as well [21:42:58] helmfile.d/admin_ng/helmfile.yaml