Looks like I'm gonna exclude most of our #FCOS VMs from daily Proxmox #Backup Server.

Those FCOS VMs tasks are running containers. All we need is just backup quadlet files and /etc/default. For most of container's data, I think we can use application level replication. If it's not available, I rather backup those container's volumes.

Why?

Because full backup of vm disk images is too expensive. Not talking about storage price, but about extraneous load on infrastructures (cpu, storage i/o).

Fedora CoreOS Stable v43.20260316.3.1 has been released few hours ago with 49 packages upgraded, including kernel upgrade from 6.18.13 to 6.19.7.

https://www.fedoraproject.org/coreos

#fcos #fedora #linux

The container optimized OS

A minimal OS with automatic updates. Scalable and secure.

UniFi OS Server installs just fine on Fedora CoreOS when doing "rpm-ostree usroverlay" before starting the installer.

This will allow it to create symlinks in /usr/bin which are not needed and don't need to be persisted. Without temporarily enabling usroverlay the installer will leave UniFi OS Server in a broken state.

Coreos is a great match for UniFi OS because the base system will upgrade automatically while application upgrades are managed by UniFi.

#UniFi #fcos #fedora #coreos

Aaaand it won't. Of course it wants to write in places where it's not even sane to write to on a mutable OS.

#UniFi #fcos #fedora #coreos

"UniFi OS Server" is the successor to #UniFi Network Application.

Is actually not a OS but supposedly a collection of podman containers and scripts that fuck around in the base OS as it pleases. 👀

Will this work in a dedicated #fcos VM? I'll find out.

Fedora CoreOS Release Notes

See the major changes and components version for latest Fedora CoreOS releases.

Major learning curve to use #fedora #coreos is how to create ignition file. #FCOS installation use that for installation configuration. There's no menu or wizard like other #linux distros.

Why? Because one of major goal of #FCOS is to enable easy mass deployment of similar configuration.

After you learn and solved that problem, remaining else is just systemd and #podman #quadlet (most #fcos deployment is to run containerized applications)

@zrail I've managed some #fedora #coreos to run dozens of #podman container. Works nicely.

I can depend zincati to auto-update os, and podman-auto-update systemd service to update container isafely. #fcos update is atomic update.

No need to doing manual major distro upgrade because zincati take care of that.

Yes, u should monitor memory usage. You can run #promotheus & grafana container to monitor it. Also there is prometheus-podman-exporter to monitor podman container with prometheus easily.

@aku dibilang distro-hopping jg gak om.. itu2 aja kan.. kalo skrg ya intinya workstation #fedora, server ya #rhel clone kayak #almalinux atau #rockylinux, kombinasi dikit sama #fcos (fedora coreos). Ada juga opensuse tumbleweed. Tp itu semua kan bisa dibilang satu aliran

I figured out how to make a Fedora CoreOS template for XCP-ng that supports unattended installs by supplying an ignition file through Xen Orchestra's cloud config templates!

https://lethedata.com/blog/xo_fcos_template.html

#fedora #fcos #xcpng #xo

Fedora CoreOS XCP-ng XO Template - LetheData

Create a Fedora CoreOS template for Xen Orchestra with ignition support through cloud config templates.