NixOS systemd | 611 Members | |
| NixOS ❤️ systemd | 172 Servers |
| Sender | Message | Time |
|---|---|---|
| 21 Feb 2025 | ||
x-systemd.required-by= will make it so that the mount is not required by loca-fs.target, but instead just by the value of the option | 06:01:41 | |
| meaning if it fails, it won't send the system to emergency mode | 06:01:54 | |
but of course a required-by relationship isn't an ordering relationship, hence why I also mentioned x-systemd.before= | 06:02:38 | |
ElvishJerricco: ah, RequiresMountsFor is what i wanted, and does work | 06:09:18 | |
but only if its in unitConfig, doh | 06:09:23 | |
service.d/overrides.conf:18: Unknown key name 'RequiresMountsFor' in section 'Service', ignoring. | 06:09:34 | |
cleverca22: Sure, though that requires adding nofail to the file system if you want it to not cause emergency mode | 06:10:05 | |
| yep, i already added nofail, and that caused an issue of the service starting without the disk | 06:12:43 | |
nofail plus RequiresMountsFor results in no emergency mode, and the service doesnt start | 06:13:01 | |
Feb 21 06:10:16 backend1 systemd[1]: ipfs.service: Job ipfs.service/start failed with result 'dependency'. | 06:13:32 | |
| thats perfect | 06:13:35 | |
yea, that's nearly the same thing as simply adding x-systemd.required-by=foo.service,x-systemd.before=foo.service, with the only difference being that my way doesn't mount the file system if foo.service isn't scheduled at boot. | 06:15:29 | |
| under normal conditions, it should mount at boot, and start the service at boot | 06:19:59 | |
| yea, the difference is negligible | 06:20:12 | |
the problem, is that somebody put the fileSystems config in a shared moduleand then deployed it to a dev machine in digital ocean a machine lacking the DO volume.... | 06:20:26 | |
| it would only matter if you wanted to mask the service for some reason | 06:20:27 | |
| so the machine instantly fell over | 06:20:32 | |
| and nobody bothered fixing it, for 5 months | 06:20:39 | |
| It'll mount at normally boot with either way | 06:21:28 | |
but the x-systemd. way would not mount if you masked the service for some reason | 06:21:54 | |
| * It'll mount normally at boot with either way | 06:22:02 | |
| ah | 06:22:13 | |
| 24 Feb 2025 | ||
| 08:06:23 | ||
| Do I need to worry about these logs in systemd-initrd?
| 10:02:26 | |
we are including libnss_files so I am confused. Also ld-linux-aarch64.so.1 missing sounds bad? | 10:02:47 | |
| oh wait our ld-linux is just a stab that says “this doesn’t work” anyway | 10:11:18 | |
| ah this is just glibc being glibc | 10:32:55 | |
| This happens because glibc is effectively an implicit part of RPATH. So all the glibc libs are not found in the literal RPATH of an executable, but are found by the dynamic loader because it just knows to look there | 20:57:53 | |
I just found out that systemd-fstab has a —generate-fstab flag that uses MountPoints= to synthesize an fstab file | 22:07:34 | |
* I just found out that systemd-repart has a —generate-fstab flag that uses MountPoints= to synthesize an fstab file | 22:07:42 | |