Hello everyone,
I recently set up a drop-in systemd unit file that adds an “on failure” command to all user-level services on a home server. Specifically the on failure command sends a notification to ntfy.sh detailing the service that has failed. E.g. if my miniflux container fails, it will send a notification such as Le service miniflux.service a échoué. Vérifier : journalctl --user -u miniflux.service
However, particularly on server startup, I have been getting some weird failure notifications for services I cannot find.
Any ideas on what I can do to troubleshoot this? An example is provided bellow.
Title:⚠️ ❌ 303a3d11aa5781aaca14faa1e36fbe4465466f74e49821fee252d01227f8fdfb-73342cf5eacdadbd.service a échoué
Body:Le service 303a3d11aa5781aaca14faa1e36fbe4465466f74e49821fee252d01227f8fdfb-73342cf5eacdadbd.service a échoué. Vérifier : journalctl --user -u 303a3d11aa5781aaca14faa1e36fbe4465466f74e49821fee252d01227f8fdfb-73342cf5eacdadbd.service
dustbin@thelemmy.club 3 weeks ago
The naming convention looks suspiciously like an autogenerated (dynamic) unit; one typically sees these from /etc/fstab being parsed at boot by systemd-fstab-generator into /run/systemd/.
Try perusing www.freedesktop.org/…/systemd.generator.html and /run/systemd/ on your system, see if anything catches your eye…
kaki@sh.itjust.works 3 weeks ago
OP mentionned containers, and Podman/Quadlet generates such services for health-checking a container service, so it might be that.
Whooping_Seal@sh.itjust.works 2 weeks ago
I havent had the change to reboot with the parameters mentioned below, but I think it most definitely is the health check (which I have manually configured for everything but syncthing).
dustbin@thelemmy.club 3 weeks ago
It’s probably most effective at this point to add
systemd.log_level=debug systemd.log_target=kmsgto the kernel boot options to ferret out what’s creating it then go from there