Skip to content

Run Komodo periphery agent on Alpine Linux

Status

No official solution.

Official releases

You can't. Alpine Linux uses OpenRC instead of systemd and compiles with musl libc instead of glibc.

The official Komodo Periphery binary is compiled with glibc and the official setup file and ansible roles use systemd for managing it as a service.

Community resolutions

jarjarthidev recompiled Komodo Periphery with musl and modified the setup-periphery.py to add support for generating OpenRC init files. The repository contains a GitHub workflow that rebuilds directly from the upstream using Alpine Linux.

It is available from jarjarthidev/komodo-periphery-alpine: Modified setup-periphery.py with OpenRC support for Alpine Linux

If going the Proxmox Virtual Environment route for Komodo Core like I did, a setup will end up with at least two docker hosts. An example can be found at 16 - Komodo Setup - Norbert Csicsay - Homelab.

The reason:

The community-scripts.org install Komodo as Add-on into an existing LXC container. This could be Debian, but I ended up using Alpine Linux. The Add-on automatically installs Docker if missing and adds a Docker Compose file with the following services:

The "solution":

If you install via Komodo | Proxmox VE Helper Scripts on Alpine, keep it a bare docker instance only running the Komodo stack. Think of it as a throw-away docker management machine.

Type update_komodo in the console or re-run the install script to update.

Run and manage all other containers on dedicated docker hosts on systemd supporting distros like Debian or Arch Linux with Komodo Periphery as a service.

That way you manage all your docker hosts centrally from your Git-backed Komodo repositories without any of them knowing about (read: not being altered for) being managed, which makes Komodo such a great alternative to solutions like Dockge.