On Devuan and Debian

As you probably know, I’m all heart and soul into FreeBSD. I simply love it and cannot accept anything (apart from other BSDs) that would remind me of true Unix. However, I’m also quite naive and in my naivety I again decided to look for Unix in GNU/Linux. I pondered over Gentoo, Arch Linux and CRUX, though for various reasons I realized that neither of them is a distribution I could just quickly install and get on with my life (Python coding, hardware tinkering, etc.). Then, I remembered there is Devuan, which I’ve already used some time ago and was quite satisfied with it.

One could say Devuan was founded by Debian developers who went rogue. That’s the common misconception, unfortunately. The truth is a lot less dramatic. There was a disagreement on the default init system in Debian “Jessie” and a lot of experienced Debian developers decided to exert their right to fork. “Jessie” was the first Debian version to use systemd as the init and supervisor suite. I believe it was far too early as an init system requires years to mature. Roughly half of Debian developers shared such opinion. Devuan, like former Stable “Wheezy”, uses sysVinit and tries to adhere to the Unix ideology by utilizing basic Debian utilities for system tasks. As the notion of Taco Bell Programming states, almost everything can be done with just those basic tools.

Much has been said about systemd already. Despite its presumably obvious merits, it is still buggy and the bugs tend to be quite severe. More so, it doesn’t really solve any problems other than the ones it creates. I wouldn’t want to run something as dangerous on my mission critical machines. As simple as that. It has nothing to do with trust and everything to do with perceived software stability.

Devuan, being a spiritual successor of “Wheezy”, free from the Debian reign may now grow in extremely interesting directions. As sysVinit allows the use of additional supervisors, I think a lot can be done with such flexibility (s6 anyone?). Also, it might be prime time to clean up the GNU mess of numerous overlapping tools and remove the ones that are outdated, vulnerable or lack upstream maintenance.


5 thoughts on “On Devuan and Debian

  1. “There was a disagreement on the default init system in Debian “Jessie” and a lot of experienced Debian developers decided to exert their right to fork.”

    Are those the mythical “500 active developers” that Devuan claims to have? 🙂
    Or how many “experienced Debian developers” went to Devuan? 0?


    • That you would have to ask on one of Devuan’s mailing lists or IRC channels. I think the 500 is just a rough estimate, though judging from the amount of work that has already been done, I think they know their stuff :).


  2. That 500 number was the number of accounts on their gitlab instance back at the time. Hardly any of these 500 people has a commit or any other recorded activity to their name. Calling them developers is a bit of a stretch.

    More realistically there is one packager active in Devuan (if $dayjob permits), there are some mystical VUA entities supposedly around (you only ever see Jaromil and nextime actually doing anything), and about 100 idiots discussing their pet conspiracy theories on why systemd won. Development is basically non-existent: It takes month to fix anything.

    There is no working security infrastructure (a repo for that is there, but it hardly ever gets updated). There is lots of trouble getting packages synced from Debian (e.g. samba being uninstallable for month at a time). The web presence is fucked up most of the time, too. Https is apparently too hard for veteran unix admins.

    I would strongly recommend using anything else for non-experimental Linux testing. That is really sad, as I had high hopes for Devuan at first:-/


    • Thank you kindly for your input. Yes, I do remember reading similar news at some point, too. I agree the 500 is definitely a stretch, though I talked to some Debian developers, who actually aid both projects. They’re more and more reluctant to fix Debian issues, many of which are tied to systemd. While surely still in its infancy, I think Devuan is an interesting project to follow. Even if the only reasons are consistency with Debian’s sysVinit history and the possibility for it to grow in new directions :).


  3. “Https is apparently too hard for veteran unix admins.”

    at any rate, if youre using a free cert authority and they change their plan, if youre growing faster than most distros, if you hit a rate limit for the cert authority and are at their mercy to sort it out, youre going to have the competence of every dev involved questioned over one person involved in charge of certs and an authority devuan has no control over? that sounds reasonable.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s