| 11 Jan 2023 |
 @andreas.schraegle:helsinki-systems.de | In reply to @andreas.schraegle:helsinki-systems.de
iirc the decision was made not to just pull in all that (meaning kernel modules + tooling), because not that many people use it.
in the old initrd, it's just enabled by default. by default == there's probably not even a way to disable it | 14:34:18 |
 Mic92 | * Andreas Schrägle: but even enabling swraid seems not to fix. one has to write configuration | 14:34:33 |
| Mic92 | Andreas Schrägle: https://github.com/NixOS/nixpkgs/blob/2ff336107f66ce6c3f89eb36d0b7492b1b9d3c2a/nixos/tests/systemd-initrd-swraid.nix#L23 | 14:35:00 |
| Mic92 | This test has to add configuration to work | 14:35:19 |
| @andreas.schraegle:helsinki-systems.de | ah. interesting. the old initrd probably just does mdadm --scan or something like that? | 14:35:37 |
 @elvishjerricco:matrix.org | Mic92: I would look at the foobarCommands options and try to see what might be setting those | 14:36:42 |
| @elvishjerricco:matrix.org | if that's not the problem then something is much more confusing than I would expect :P | 14:37:00 |
| @andreas.schraegle:helsinki-systems.de | ah, looking at nixos/modules/tasks/swraid.nix, it seems like there's some udev rule stuff happening | 14:37:34 |
| Mic92 | Yeah. I think the udev rules are missing | 14:37:49 |
| @elvishjerricco:matrix.org | interesting... | 14:37:59 |
| Mic92 | Janne Heß: you merged this into the swraid module, do you know why udev is disabled? | 14:38:40 |
| @elvishjerricco:matrix.org | oh wow yea it just wholesale cps those udev rules lol | 14:38:44 |
| @elvishjerricco:matrix.org | Mic92: I think we just decided not to include any foobarCommands on principle. | 14:41:33 |
| Mic92 | ElvishJerricco: ok, but including udev rules is fine, no? | 14:41:52 |
| @elvishjerricco:matrix.org | eh, the way that particular thing works is iffy | 14:42:31 |
| Mic92 | what is the current way? | 14:42:45 |
| @elvishjerricco:matrix.org | we don't have a phase where random commands can be run | 14:42:47 |
| @elvishjerricco:matrix.org | the scripted initrd does | 14:42:59 |
| @elvishjerricco:matrix.org | (i.e. the extra-utils derivation) | 14:43:07 |
| Mic92 | Can I now just include a package that gets merged somehow? | 14:43:09 |
| @andreas.schraegle:helsinki-systems.de | shouldn't boot.initrd.services.udev.packages = lib.mkIf cfg.enable [ pkgs.mdadm ]; do that? | 14:43:16 |
| @elvishjerricco:matrix.org | In reply to @andreas.schraegle:helsinki-systems.de
shouldn't boot.initrd.services.udev.packages = lib.mkIf cfg.enable [ pkgs.mdadm ]; do that? Oh, yes | 14:43:26 |
| Mic92 | will test | 14:44:00 |
| @andreas.schraegle:helsinki-systems.de | nixos/modules/tasks/swraid.nix already does that, so it should already work without a config. idk why the test adds one. | 14:50:29 |
| @elvishjerricco:matrix.org |
the test
What test?
| 14:50:56 |
| @andreas.schraegle:helsinki-systems.de | In reply to @joerg:thalheim.io
Andreas Schrägle: https://github.com/NixOS/nixpkgs/blob/2ff336107f66ce6c3f89eb36d0b7492b1b9d3c2a/nixos/tests/systemd-initrd-swraid.nix#L23 . | 14:51:11 |
| Mic92 | also seems to be not enough from I can see | 14:51:19 |
| @elvishjerricco:matrix.org | ah | 14:51:23 |
| Mic92 | I see a lot of lines like this: /nix/store/maw4dvj4zyn7l4gn7l4ic77w6mir8yr8-systemd-stage-1-252.1/lib/udev/rules.d/50-udev-default.rules:85 Unknown group 'disk', ignoring | 14:51:58 |
| Mic92 | Is this normal? | 14:52:02 |
