!PSmBFWNKoXmlQBzUQf:helsinki-systems.de

Stage 1 systemd

83 Members
systemd in NixOs's stage 1, replacing the current bash tooling https://github.com/NixOS/nixpkgs/projects/5126 Servers

Load older messages


SenderMessageTime
27 Nov 2022
@me:linj.tech@me:linj.tech
In reply to @me:linj.tech
oneshot does not change this
well, oneshot does make sure my service is before cryptsetup-pre.target. I confirm it with sleep 10
23:20:40
@me:linj.tech@me:linj.techthanks23:20:53
@me:linj.tech@me:linj.tech

It's weird that my pre-device-command.service and cryptsetup-pre.target stops after finishing running nixos activation script.

from man systemd.special:

this target (cryptsetup-pre.target) is particularly useful to ensure that a service is shut down only after all encrypted block devices are fully stopped.

23:27:01
@me:linj.tech@me:linj.techguess our initrd is not the same as what systemd's manual says23:28:06
@elvishjerricco:matrix.org@elvishjerricco:matrix.orgHow is that not in line with the manual?23:28:25
@me:linj.tech@me:linj.tech I think what the manual says is that cryptsetup-pre.target should be stopped when the system is shutdown 23:29:44
@me:linj.tech@me:linj.techbut now, it is stopped after stage 223:29:54
@elvishjerricco:matrix.org@elvishjerricco:matrix.orgThat's not quite accurate23:30:22
@elvishjerricco:matrix.org@elvishjerricco:matrix.org Stage 1 has two big systemd transactions. First is initrd.target, and after that there's a service called initrd-cleanup.service that does systemctl isolate initrd-switch-root.target, which is the second transaction. This causes everything else to be stopped 23:31:49
@elvishjerricco:matrix.org@elvishjerricco:matrix.orgSo that second transaction is in some ways very similar to system shutdown23:32:08
@elvishjerricco:matrix.org@elvishjerricco:matrix.org Any services in initrd that aren't supposed to be stopped by this (e.g. plymouth) need to either be wanted by initrd-switch-root.target, or have IgnoreOnIsolate=true 23:33:37
@arianvp:matrix.orgArian
In reply to @me:linj.tech
my keyfile has a pre-defined key in it
Then don't use /dev/urandom
23:34:18
@arianvp:matrix.orgArianBut the keyfile :)23:35:03
@me:linj.tech@me:linj.techinteresting, will try that later23:35:07
@elvishjerricco:matrix.org@elvishjerricco:matrix.org
In reply to @me:linj.tech
emm, what I want is somewhat weird: generate a keyfile on the fly for luks without storing that keyfile in initrd
I am curious what exactly you mean by this. If it's not randomly generated, then where is this key file you're actually trying to use?
23:36:07
@me:linj.tech@me:linj.tech
In reply to @elvishjerricco:matrix.org
I am curious what exactly you mean by this. If it's not randomly generated, then where is this key file you're actually trying to use?
mkdir -p ${luksKeyFileDir}
mount -t ramfs none ${luksKeyFileDir}
echo -n password > ${luksKeyFile}
chown 0:0 ${luksKeyFile}
chmod 0400 ${luksKeyFile}
23:37:14
@elvishjerricco:matrix.org@elvishjerricco:matrix.org so it is stored in the initrd? Just in the form of a script? 23:37:54
@me:linj.tech@me:linj.techI want to run that before unlocking luks23:37:58
@me:linj.tech@me:linj.techyeah23:38:53
30 Nov 2022
@chillermiller3:matrix.org@chillermiller3:matrix.org left the room.05:29:40
1 Dec 2022
@jkarlson:kapsi.fi@jkarlson:kapsi.fi changed their display name from Emil Karlson to Emil Thorsoe.08:20:44
@jkarlson:kapsi.fi@jkarlson:kapsi.fi changed their display name from Emil Thorsoe to Emil Thorsøe.08:25:10
@hexa:lossy.network@hexa:lossy.network changed their display name from hexa to hexa (22.11 now).13:08:44
@hexa:lossy.network@hexa:lossy.network changed their display name from hexa (22.11 now) to hexa.14:38:33
5 Dec 2022
@janne.hess:helsinki-systems.de@janne.hess:helsinki-systems.de ElvishJerricco: did we have stripping enabled by default on 22.05 already? 18:29:44
@elvishjerricco:matrix.org@elvishjerricco:matrix.org Janne Heß: I don't remember how that panned out off the top of my head. I think we enabled stripping by default for initrd but not for shutdown ramfs? 18:31:13
@janne.hess:helsinki-systems.de@janne.hess:helsinki-systems.deYeah that wasn't too great for some reason18:31:28
@janne.hess:helsinki-systems.de@janne.hess:helsinki-systems.deMy Pi didn't boot because it couldn't load dm_mod because of: https://linux-tips.com/t/how-to-strip-linux-kernel-modules/47218:31:40
@janne.hess:helsinki-systems.de@janne.hess:helsinki-systems.de(only took me 2 days to debug this - the kernel even hinted what could be the issue)18:31:55
@elvishjerricco:matrix.org@elvishjerricco:matrix.orgYea we didn't want binutils in the runtime closure of a system. It's not needed at runtime for initrd but it would be for shutdown ramfs18:32:38

Show newer messages


Back to Room ListRoom Version: 6