Debianfork

Here you can discuss every aspect of Debian. Note: not for support requests!
Post Reply
Message
Author
User avatar
oswaldkelso
Posts: 1325
Joined: 2005-07-26 23:20
Location: UK
Been thanked: 3 times

Debianfork

#1 Post by oswaldkelso »

Free Software Matters
Ash init durbatulûk, ash init gimbatul,
Ash init thrakatulûk agh burzum-ishi krimpatul.

User avatar
golinux
Posts: 1582
Joined: 2010-12-09 00:56
Location: not a 'buntard!

Re: Debianfork

#2 Post by golinux »

Yes! The Debian bureaucracy is getting tied up in knots. Not hopeful about the outcome - probably some tar 'n feathering and expulsions in the end to consolidate the coup. I think projects like this are going to be the only escape from systemd.
May the FORK be with you!

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Debianfork

#3 Post by keithpeter »

We are excluded from voting on the issue: only few of us have the time and patience to interact with Debian on a voluntary basis.

Now we do what we can, hoping our concerns will be heard by those who can cast a vote about it.
So who would actually be doing the *rather significant* amount of extremely unglamourous grunt work that a fork would require?

I suspect this is a stalking horse. I hope we can keep Debian whole and work out how to allow a (real) choice of init systems even if only for server work loads (or Xorg + window manager on top of a server set up).

User avatar
Gyokuro
Posts: 45
Joined: 2013-03-06 19:33

Re: Debianfork

#4 Post by Gyokuro »

Well I'm not posting that much but I'm watching this systemd dicussion since the beginning and in my opinion it will be another short-lived useless fork in case it's a fork and not a debian flavor with another default init system (and how many people play every day with their init system?? server one time: syslog-ng install, enable in systemd remote login => never touch again such a system (in my case) ). However I'm certain that the same people will cry in case a different init system is the default one - maybe they succed but in case not it will hurt Debian in general. My impression is that some people can not stand systemd for some reasons, lost the vote and therefore try to reach via this step their goals - maybe it's wrong but in such case a manager decision is needed. In my opinion it's ok and needed that Linux distributions get a modern init system which take advantages of the features which are provided by the kernel. Solaris got SMF and not much people cried due to of problems fixing sysv was always a nightmare - BSD has it's old bsd init system and parallel starting is still not in (checked the last time a FreeBSD 10) and the other alternatives not anywhere were systemd already is. It's new and have to mature that's all and for how many years is the sysv-init system in production and Linux distributions should advance and there will be always a minority which will cry about something.

User avatar
edbarx
Posts: 5410
Joined: 2007-07-18 06:19
Location: 35° 50 N, 14 º 35 E

Re: Debianfork

#5 Post by edbarx »

The outcry against systemd is not out of sheer software conservatism, as you boldly try to put it, but the slap in the face of the successful and tried *nix tradition of creating simple programs that do one task right. The latter is why GNU/Linux was successful. Trying to imitate other OSs, will render GNU/Linux unappealing to the majority, as they will lose the interest in using something different.

This is like claiming Italian Cuisine is delicious, so let us discard our own cuisine. Why shouldn't GNU/Linux have its own cuisine? :shock:
Debian == { > 30, 000 packages }; Debian != systemd
The worst infection of all, is a false sense of security!
It is hard to get away from CLI tools.

User avatar
golinux
Posts: 1582
Joined: 2010-12-09 00:56
Location: not a 'buntard!

Re: Debianfork

#6 Post by golinux »

Gyokuro wrote:It's new and have to mature that's all . . .
That's it in a nutshell. Please don't minimize the risks of setting on this course. Who knows just what it will be when it 'matures'. By then, way too late for a do-over. The 'coup' will have succeeded. Debian as we have known it will be but a memory . . .
May the FORK be with you!

User avatar
dasein
Posts: 7775
Joined: 2011-03-04 01:06
Location: Terra Incantationum

Re: Debianfork

#7 Post by dasein »

Gyokuro wrote:It's new and have to mature that's all . . .
Fact-check time...

The monstrosity known as systemd has been in development for four years. The very fact that it's still pre-alpha-quality crap after that much effort really ought to give folks pause.

Once it's finished (if it ever is), and has demonstrated itself to be as robust as sysv, then let the fanbois push for adoption.

User avatar
Gyokuro
Posts: 45
Joined: 2013-03-06 19:33

Re: Debianfork

#8 Post by Gyokuro »

It's new and needs time but why should not try something out different where more developers are working on it - it's another part of the system which is somehow under the umbrella of kernel.org people. Development time says nothing about the maturity of an application and sysv as it seems to be stable is stable because it's a simple but unflexible and somebody can say somehow limiting. However watching all posts over the time I can not find some good posts or a technical explaination why sysv-init is better as it's successors. In my previous post I was mentioning Solaris and SMF and that was a good step and nobody looks back anymore - Apple with launchd developed a system which is perfect integrated within the system. Systemd could be the same due to it's developed somehow under the umbrella of kernel.org folks and get better integrated in the core of a future Linux system and maybe it's time to throw some old pieces and parts overboard. However as somebody who likes Debian in general I hope that there a solution will be found and Debian evolves further but as in real life sometimes you have to say goodbye to move on.

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Debianfork

#9 Post by keithpeter »

Gyokuro wrote:However watching all posts over the time I can not find some good posts or a technical explaination why sysv-init is better as it's successors. In my previous post I was mentioning Solaris and SMF and that was a good step and nobody looks back anymore - Apple with launchd developed a system which is perfect integrated within the system.
Over the next 2 to 10 years or so as the big users move over (Amazon cloud services, Google, Facebook, most ISPs) we will all be able to judge the reliability, security, and attack surface of systemd as it will be running most of the Internet!

Quite a thought isn't it?

tomazzi
Posts: 730
Joined: 2013-08-02 21:33

Re: Debianfork

#10 Post by tomazzi »

I doubt that Google will decide to use systemd - they are using heavily modified (not to say forked) linux kernel, based on version 2.4.x

Even if they decide to use systemd, then it will be havily modified version too...
And in such case, I'm entirely sure that google's systemd will differ at least in that it'll have implemented the most basic protection against segfaults in release builds - unlike that crap from upstream version...
Odi profanum vulgus

timbgo
Posts: 265
Joined: 2013-04-14 12:17

Re: Debianfork

#11 Post by timbgo »

But, dear FOSS Linux users [1], some of you are looking up to the Surveillance Engine as a model to follow, or learn from? Or even Apple Mac?

There is no freedom there. Apple is as surveillance compliant as M$ is. Schmoog OTOH, is the Surveillor. And the freedom that you achieve without privacy is close to zero, nada, zilch! (That is why, since the spyware SELinux there!, gNewSense doesn't get you freedom either, see this msg of mine on gNewSense list or see what it was about in Grsecurity Forums in case censorships moves in.)

I'm with Debian because I had seen some chance that it were salvageable for user's control and those you compare this systemd intruding story in Debian... (and others are lost, along, if Debian goes...)

And those some of you who compare this systemd intruding story in Debian with passive/active total surveillors' OSes that some of us would feel like mental prison having to use them... are maybe afraid to look the truth in the eyes or maybe haven't understood it for some other reason... or something yet other is the matter.

Also: Why, oh why, lots of you here who participated in the quest to find the solution to the systemd steath imposition in Debian, seem not to have noticed that DD Thorsten mirabilos Glaser solved the issue?... But maybe some of you were simply busy elsewhere yet...

And instead of the fork, what should be done is:
steveL wrote: As for http://debianfork.org/ someone should tell those guys they don't need to fork anything; just use mirabilos' prevent-systemd-completely package (which is "Important: yes") and keep integration-testing with whatever pid1 and rc-manager they want. Then swap notes and do the QA work to keep it going as a collective, using debian infra where they want as it's a perfectly valid use of debian.
as you can read here.

See also here. This is, it looks to me currently, I'm really worried, my last ditch effort to contribute to saving Debian. That is what it is about all along since I noticed it in my first post about it on Debian Forums.

[1] I decided that while for me, Linux can not stand alone as the name of that OS, because it's so much more than the Linus' contribution and because of what in my signiture you can clearly see on that guy's character, and that GNU is dead for me as the name of it since Richard Matthew Stallman supports SELinux insanely, see the Emacs page on gnu.org.
Miroslav Rovis
Zagreb, Croatia
http://www.CroatiaFidelis.hr
Anyone can dismiss these: kernel hooks for rootkits
linux capabilities for intrusion?

User avatar
harrycaul
Posts: 105
Joined: 2008-12-23 17:06

Re: Debianfork

#12 Post by harrycaul »

http://distrowatch.com/weekly.php?issue=20141027
Debian, systemd and forks

In the past we have touched on the systemd init technology. The systemd project is an attempt to revolutionize the way Linux starts up and manages services. In theory, systemd allows computers to get up and running faster and the technology offers some new methods for controlling system services. The project doesn't stop there though, systemd has also expanded to add binary logging, a userland text console and several other features. Some people see systemd as a unifying force that will reduce distribution fragmentation while decreasing boot times. Others see systemd as a project that is expanding unnecessarily outside of its scope while reducing flexibility and transparency.

Until the start of 2014 most of the Linux distributions that had adopted systemd as their default init technology had been relatively cutting edge projects. It was natural to see Fedora, always an early adopter of open source software, embrace systemd. Arch Linux, being a cutting edge distribution, also adopted systemd quickly. It was a little surprising to see Debian, a famously conservative distribution, vote to adopt systemd earlier this year. Fans of the Debian project usually prefer tried and true technologies to the latest and greatest. Fans of Debian are often also fans of choice. Debian claims to be the "universal operating system" and its many branches, architectures and ports speak to people who value flexibility. Last week we talked about a proposal to keep choice alive in the Debian project by enabling users to select their preferred init technology when setting up Debian. The proposal suggests systemd may remain the default init system, but administrators should be able to swap out systemd for another init technology.

But what happens if Debian not only adopts systemd, but also insists on making systemd a dependency, one which cannot be easily removed? Some people will adapt, learn the new init technology and continue on using Debian. Others have expressed an interest in migrating to more conservative projects. Another group has suggested Debian should be forked if systemd becomes a fixed part of the Debian GNU/Linux operating system. The Veteran Unix Admins (VUA) group has declared that they like Debian, want to continue using Debian and they hope systemd will remain a removable option. However, if systemd is here to stay, the VUA has declared their intention to fork Debian.

I had a chance to exchange e-mails with one of the VUA members and here is what they had to say about Debian, systemd and a potential fork of the project.

DW: Would this be a complete fork of the distribution with entirely separate repositories? Or would you mostly use Debian's existing repositories and just fork the low level components relating to start-up/init?
VUA: We are still evaluating the possibilities and analyzing dependencies of systemd and related components (d-bus for instance) to understand the span of this fork, which we still hope won't be necessary.

The majority of us involved in this early phase see it as a fork taking distance from Debian repositories, cutting down on packages maintained and eventually optimizing the governance model thanks to the fact we will start smaller.
DW: How many contributors does this potential fork have?
VUA: VUA is a group counting little less than 1,000 subscribers on private fora, obviously the debianfork.org initiative doesn't involve all of us. We are receiving an overwhelming quantity of e-mails with offers to help, about half of them are by people that seem to know well the issues at stake and to be well capable to give a hand.
DW: Popular projects like Debian need a lot of storage space, code repositories, mirrors, etc. How can people help with providing hardware/mirrors/donations? What resources do you have already?
VUA: Many of us are senior IT professionals and also considering our targets are sysadmins, infrastructure shouldn't be a problem. Of course we will not start big, but grow as needed. Having said that, this would be a challenging adventure for many reasons and we still hope it won't be necessary.
DW: Your website mentions the people involved in this potential fork do not have time to get involved with Debian's governance. What sort of governing body will the fork have?
VUA: It will be a governing body that puts the benefits of the users first, not the mystification of a "doacracy" delivering all the power to the package maintainers.

Originally, Debian was created as a universal operating system for the users. The Free Software movement itself is there to defend users rights. Sgryphon explains it well in this thread.

We will likely reproduce the governing body of Debian to follow its original mandate, with the advantage of starting small and more focused, hopefully with less pressure from the interest of commercial developers.
DW: If Debian is forked, how long do you think it will take before we see a stable release of the fork?
VUA: Those that could be in need of a fork are in such a situation because they need well known and tested technology on production servers and cannot face such a big overhaul as a dist-upgrade to a systemd based Debian 8, also considering the quantity of customization on top of sysvinit that in many cases has to be dealt with. It is foreseeable that the majority of people in such a situation will likely rely on the long term support of Debian 7 even if a sysvinit updated fork is available.

The stable release of the fork should be in place at least as an installer alternative to Debian 8 and available in the same period: it should be an environment familiar to sysvinit Debian users who are doing a fresh install.
DW: Thank you and good luck.

User avatar
buntunub
Posts: 591
Joined: 2011-02-11 05:23

Re: Debianfork

#13 Post by buntunub »

Yep. I saw that today too. DebianFork is off to the races. Wonderful! Looks like all those looking for a home from "the Distro that is no longer true Debian" now have a place to go. And it certainly sounds like he wants a return to Debians roots. I think all true Debian hard cores should put their support behind this worthy effort to save our beloved Debian.

User avatar
edbarx
Posts: 5410
Joined: 2007-07-18 06:19
Location: 35° 50 N, 14 º 35 E

Re: Debianfork

#14 Post by edbarx »

Like many 'antiquated' GNU/Linux greybeards, I am with those considering a fork of the Debian distribution. Since Debian is a huge repertoire of software, it makes sense to pinpoint those areas that are the source of incompatibilities, one of which is the infamous abomination of systemd.

I contacted debianfork that I can code although I am not formally qualified to do so. They published my email. Let us start working, I am eager to help in making Debian respect CHOICE.

Code: Select all

progress != software_lockins
Long Life CHOICE! :mrgreen:
Debian == { > 30, 000 packages }; Debian != systemd
The worst infection of all, is a false sense of security!
It is hard to get away from CLI tools.

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Debianfork

#15 Post by keithpeter »

tomazzi wrote:I doubt that Google will decide to use systemd - they are using heavily modified (not to say forked) linux kernel, based on version 2.4.x
Very interesting: any references? I did try googling but just got generic pages about a 'customised Linux'

@all: no idea which poll alternative to pick.

mmix
Posts: 205
Joined: 2012-07-14 00:08

Re: Debianfork

#16 Post by mmix »

when debianfork available, i wish it comes with quantum-shell
it is systemd-gnome-free GUI.
https://quantum-os.github.io/

Image

User avatar
keithpeter
Posts: 502
Joined: 2009-06-14 08:06
Location: 5230n 0155w

Re: Debianfork

#17 Post by keithpeter »

mmix wrote:when debianfork available, i wish it comes with quantum-shell
it is systemd-gnome-free GUI.
https://quantum-os.github.io/

Image
Are you sure that a qt5 based UI will be able to not use logind? Especially when Wayland arrives?

Looks nice though.

Post Reply