| 14 Feb 2023 |
 @elvishjerricco:matrix.org | I mean the real answer is to have a Root= option for mount units (all units?) so that the mount syscall occurs in a chroot or something so that all the scenarios are taken care of | 18:46:14 |
| @elvishjerricco:matrix.org | but that sounds very annoying to validate | 18:46:43 |
 @lily:lily.flowers | Idk, that makes way too much sense. Gotta make sure to pile on more and more hacks (/s) | 18:47:22 |
| @elvishjerricco:matrix.org | lol | 18:47:39 |
| @elvishjerricco:matrix.org | yea I wouldn't dare open a PR with that feature, but it does seem right | 18:47:55 |
| @lily:lily.flowers | In reply to @elvishjerricco:matrix.org
yea I wouldn't dare open a PR with that feature, but it does seem right Yeah that's more systemd refactoring than I am certainly comfortable with, even though that seems objectively better. I'll see what upstream says about at least handling bind mounts for now though (I'll probably open it after their 253 release, when they'll be more able to look at it and my other PR) | 19:47:06 |
| 15 Feb 2023 |
| @elvishjerricco:matrix.org | https://github.com/systemd/systemd/pull/26367#issuecomment-1430670880
jfc "we can't do a real fix because it's remotely possible that some people have an utterly insane configuration that would break" | 04:52:58 |
 @gdamjan:spodeli.org | calm down :) | 10:59:40 |
| 16 Feb 2023 |
 @jkarlson:kapsi.fi | hibernation is a nightmare | 07:18:05 |
| @jkarlson:kapsi.fi | filesytems should really be used with nfs or something with hibernation | 07:18:53 |
| @elvishjerricco:matrix.org | Emil Thorsøe: hm? Why nfs? | 07:19:43 |
| @jkarlson:kapsi.fi | it decouples the processes from internal filesystem state | 07:20:03 |
| @jkarlson:kapsi.fi | so filesystem could be effectively unmounted in any way that could cause corruption | 07:20:40 |
| @jkarlson:kapsi.fi | note that such code is not really implemented, maybe autofs | 07:20:57 |
| @jkarlson:kapsi.fi | I guess filesystemd could be written in a way that adds a nfs wrapper or libc | 07:26:34 |
| @jkarlson:kapsi.fi | * I guess filesystems could be written in a way that adds a nfs wrapper or libc | 07:26:45 |
|  Yinfeng changed their profile picture. | 09:57:16 |
| 18 Feb 2023 |
| @lily:lily.flowers | Is systemd-initrd not working with installer ISOs a known problem? It seems that they set root= on the kernel cmdline to the device label which contains the squashfs, but that doesn't play well with the tmpfs / filesystem defined in the fstab (and the fstab entry ends up getting ignored in favor of the cmdline root=, which is obviously wrong in this case) | 03:02:21 |
| @lily:lily.flowers | (to be clear, the NixOS installer tests pass fine in the test framework, but those don't have to do the ISO trickery) | 03:02:53 |
| @elvishjerricco:matrix.org | Yea that's known but not well explored. | 03:02:59 |
| @elvishjerricco:matrix.org | We've known the isos don't work all along but I don't think anyone's really investigated it | 03:03:17 |
| @lily:lily.flowers | Hmm, I may look for a solution at some point then | 03:03:36 |
| @elvishjerricco:matrix.org | That would be wonderful | 03:03:45 |
| @lily:lily.flowers | I'm not sure if that'll be changing the cmdline or if it'll be doing some hackery with systemd (but I imagine it'll be the first one -- we really shouldn't be setting root= with that for systemd-initrd) | 03:04:12 |
| @lily:lily.flowers | Is there an open issue in nixpkgs for it already? | 03:04:18 |
| @elvishjerricco:matrix.org | I don't think so? Check the systemd stage 1 GitHub project in the repo | 03:04:48 |
| @elvishjerricco:matrix.org | There's probably a correct way to use the root= cmdline, but it probably involves making NixOS use that parameter correctly in the first place (which I have a patch for on a branch somewhere) | 03:05:49 |
| @lily:lily.flowers | I really don't think the installer should be setting it for where to find the squashfs, given it's not an actual root fs. But I'll have to look at how the old stage-1 is handling that before determining the best method | 03:07:30 |
| @elvishjerricco:matrix.org | Ah that's a fair point | 03:07:59 |
| @elvishjerricco:matrix.org | Maybe check how fedora's ISO works? I believe fedora has used systemd in stage 1 for a few years now. No idea about their ISO though | 03:08:33 |
