Hello everyone.

Long time user of the Devuan project (Since its inception), new user of this forum.

First of all, huge thank you to everyone involved with the Devuan project and making sure we have a reliable and sane close-as-it-can-be systemd-free Debian based distro.
Devuan has saved the day many times for me, both for embedded devices and setting up servers.
I hope I'm able to contribute back in some way to further advance the adoption of Devuan.

Apologies for the bit of a vague subject line, but I am trying to develop a more generic/automated way to ease conversion of Debian based distros but also to serve as a boiler plate for individuals and businesses who decide to roll their own custom distro for their product.

To name a few distros I've spent the last 7 months investigating their build systems and how they could be transitioned to Devuan are;

* TrueNAS Scale.
* Proxmox VE.
* Proxmox Backup Server.
* Open Media Vault.

Yes, I know, it is a massive undertaking.

The TrueNAS Scale build system, written in Python seems quite flexible in this regard, as it relies on debootstrap and adding their own packages. However it seems some of the packages would have to be modified to exclude systemd specific stuff.

Proxmox VE and Backup Server are... well.. interesting.
They have done a lot of customization and tightly coupled it with systemd sprinkled with some Rust in between.
Without trying to sound grumpy, there are a lot of their default settings and decisions which, unfortunately, causes a lot of frustration once you start tweaking things, especially around ZFS.
Those of you who are more knowledgeable in the art of forking distros and the legals behind, what would be the best plan of action here?
The end goal here is to have a KVM/LXD turnkey host/server distro which aren't suffering from systemd (I frequently fight with Proxmox systems and systemd).

Open Media Vault has been a goto for some time when building a little NAS appliance, this is again one of those situations where systemd has been defaulted to, without regard to users who are not interested in running systemd.
It would require some effort to get rid of the hard systemd dependencies, but it would be doable.

Lastly - init system
Personally I've grown quite happy with a combination of busybox + OpenRC and runit for embedded, but I see s6 is also quite attractive as it appears to be very well written.
What are your guys inputs when it comes to server focused distros where service supervision and management is important while the bootup and shutdown of a system is still fast.
My plan here is to write some supporting tools to better convert systemd units into a desired service system, be it OpenRC, s6 or something else.

Any input or suggestions are very welcome

Thanks in advance
Glenn