@gamingonlinux I would have tested it, but this far it's been deeply depending on #systemd which I think it's wrong way to go, as there are multiple init system and not the #unlx
way of doing things, posrting getting nasty as say BSD don't use systemd at all.
My money is on the new pope issuing a bull against #systemd.
@aires I'm already using it! Although not because it's made in rust, purely because I found some Nix tooling really doesn't accommodate doas
I can't remember them all, but I know deploy-rs was the one that "pushed me over the edge", as it would only work with sudo, so I went with the more minimal sudo-rs implementation.
That said, what I've seen mentioned elsewhere which I'm really interested in, is run0
https://mastodon.social/@pid_eins/112353324518585654
Although it doesn't seem to work on NixOS, unless anyone's aware of any workarounds? This is the most recent (Nix oriented) post I've seen.
$ poweroff: /usr/lib64/libcrypto.so.3: version `OPENSSL_3.4.0' not found (required by /usr/lib64/systemd/libsystemd-shared-257.so)
People: hahaha, #systemd.
Me: cool, this just saved me from a boot failure the next day.
"Replace #Kubernetes with systemd" they said, because apparently using a hammer to fix a computer makes total sense.
Why innovate when you can just duct tape your way to 2024, right? #TechGenius #InnovationOrInsanity
https://blog.yaakov.online/replacing-kubernetes-with-systemd/ #TechHumor #Innovation #DuctTape #Systemd #HackerNews #ngated
Which distribution ships a working rootless systemd-nspawn?
Good blog post, basically sums up my current thoughts about the home lab. Kubernetes is nice, but overkill, and Podman+systemd and even further Quadlet can replace k8s really well on a single person scale.
Wish there was a good NGINX-based solution to mimic what ingress-nginx and cert-manager can do on k8s, not that it's that much work to handle a couple domains manually.
https://blog.yaakov.online/replacing-kubernetes-with-systemd/
#podman #systemd #quadlet #k8s
Not sure how long it will last, but I tried installing #emacs with a #systemd system extension which was pretty interesting!
https://www.alcarney.me/blog/2025/installing-emacs-with-systemd-sysext/
configuring nginx to auto-redirect from port 80 to 443 may be useful
if #nginx can automatically do #acme that'd be news to me, you might want to try #caddy for that
personally i use #nixos options to configure nginx and tls/ssl, which achieve this by creating a cert-renewing #systemd service that calls out to #letsencrypt. i assume that's sort of the canonical solution for not generating certs manually - even if you don't want to use nixos you may want to look into how they do it
I sometimes find that, if I log out and immediately log back in, it takes a weirdly long time for #KDEPlasma to start up. Is it because part of Plasma is running as a #systemd service and a new session can't start until the old one has completely terminated? Seems like a bad idea…
@irfan while this points to a specific bug triggering this behaviour it also means, #KDE should extend the process isolation of desktop components and background services as #systemd user services. Ensuring desktop interaction is always possible and snappy and limiting the potential outfall in terms of resource consumption of individual processes by enforcing upper limits.
The groundwork is done and already widely used, it just needs to spread to a few more places and needs some finetuning.
systemd system extensions for Fedora image based systems via @jordimoncompanys https://lobste.rs/s/p63gea #systemd
https://travier.github.io/fedora-sysexts/
On today's episode of knit-picking #SystemD;
Obviously, we'd need all the #ZFS storage to be fully set-up, running, and /dev/zvol/ populated, 1st. Only then, can we start our #iSCSI target service.
How do YOU do it? (please reply, *splaining)
The first two options might get evaluated to the same result by the systemd "magic" - either option should make sure the storage is ready for services, right?
Tried to set up #systemd in a devcontainer using a feature:
This is not possible, as the entrypoint gets wrapped with a `/bin/sh`
So the only way so far has to override the Containerfile, set the entrypoint and use overrideCommand: false:
https://github.com/orgs/community/discussions/68954#discussioncomment-11812864
Thinking out loud: is there a way to get systemd to pass a D-Bus system bus socket FD to a service when it’s bus activated (i.e. via `LISTEN_FDS`)? Would mean the service could sandbox AF_UNIX socket connectivity (if it only needed that to connect to the bus, which I guess is true for some bus daemons). Downside ottomh: system bus connection policy would be bypassed (though it allows all connections by default) and the auth would still need to be done by the service.