Debian 6.0 Released In GNU/Linux, FreeBSD Flavors

itwbennett writes "After two years of work, the Debian Project has announced the release of Debian 6.0. 'There are many goodies in Debian 6.0 GNU/Linux, not the least of which is the new completely free-as-in-freedom Linux kernel, which no longer contains firmware modules that Debian developers found troublesome,' says blogger Brian Proffitt. And in addition to Debian GNU/Linux, Debian GNU/kFreeBSD is introduced as a technology preview. 'Debian GNU/kFreeBSD will port both a 32- and 64-bit PC version of the FreeBSD kernel into the Debian userspace, making them the first Debian release without a Linux kernel,' says Proffitt. 'The Debian Project is serious about the technology preview label, though: these FreeBSD-based versions will have limited advanced desktop features.' The release notes and installation manual have been posted, and installation images may be downloaded right now via bittorrent, jigdo, or HTTP."

Comments

NetBsd kernel...what's the advantage?

[aaaaaaargh!]

I used to run NetBSD on an old PP Mac booted from a zip drive in the nineties. It was running great but since then I haven't looked at it again. I know that the 3 free BSDs (open-, free- and net-) are security audited and support old hardware very well. But I wonder what advantages the kernel itself brings. So my potentially stupid questionis:

What's the advantage of running Debian with a BSD kernel instead of linux?

Re:NetBsd kernel...what's the advantage?

[bk2204]

You can find some of the reasons here. Among them are ZFS, jails, and pf. I've used Debian GNU/kFreeBSD in the past and found pf significantly easier to use than iptables and tc.

Re:NetBsd kernel...what's the advantage?

[rtfa-troll]

... it has always befuddled me as to why they didn't 'just use' BSD instead.

To be honest, believing exactly the argument you gave companies mostly they did until the last few years. However, you never knew about out because they didn't publish the code. The reason for this is that there is no need to and if they do release their code, their competitors can use it, so their lawyers advise them against. After a few years they either get so wildly successful (JunOS / OSX / Microsoft TCP/IP stack) that they keep their own completely proprietary branch and never help anyone else or they get abandoned (IPSO / AlchemOS / BSDi / SunOS / etc. etc.)

The thing is, that the because of the effects of copyleft, the Linux people cooperate and release code and so, even though the resources put into Linux are much less, there is less duplication and so more is achieved. This has become much more visible recently with Android and other successes and means that corporate types have begun to see copyleft as a platform which makes limited cooperation with potential competitors possible and safe.

If you are choosing a system for your own platform, this becomes a good reason to choose an AGPLv3 base as much as possible and, if you have any proprietry code, layer that separately on top. Your work on the commodity underlying components can be safely released and will move forward with the rest of the community. Whatever investment you put in will be preserved instead of becoming obsolete.

Re:NetBsd kernel...what's the advantage?

[Morth]

Are they using glibc or the freebsd one? Because one of the developer advantages of the BSDs are that kernel and libc are more in sync. Ie. there's no system calls in libc that are not in the kernel, and vice versa.

Re:NetBsd kernel...what's the advantage?

[Joe Jay Bee]

One of the things I appreciate most about it is the proper OSS sound support, with mixing that actually works out of the box without having to deal with shit like PulseAudio or the clusterfuck that is Alsa.

Re:NetBsd kernel...what's the advantage?

[TheRaven64]

Working, in-kernel, low-latency, sound mixing is another one. No messing around with portaudio or ALSA, sound Just Works(tm).

Re:NetBsd kernel...what's the advantage?

[TheRaven64]

So why is ZFS the only feature ever mentioned? I don't care about ZFS, only servers want that

Indeed. There are absolutely no use cases for ZFS that aren't on the server. For example, no average user would want easy-to-use snapshots - being able to easily revert a file from an earlier version is a server-only feature. They definitely wouldn't want to be able to do simple incremental backups just by streaming the disk changes with something like zfs send / receive - only server users care about that. Data integrity is probably an enterprise feature too - no one on a desktop wants checksums in their data, because data loss only matters to enterprise users.

Re:NetBsd kernel...what's the advantage?

[TheRaven64]

If you really want snapshots, LVM is probably easier

A comment like that implies that you have never worked with ZFS. Snapshots are simple, are constant time, and don't have a performance overhead, and don't require any preconfiguration. A fairly typical ZFS setup has a cron job that takes snapshots every hour and preserves hourly ones for a while, daily ones for a bit longer, and then monthly ones, so you can always go back and find the file that you deleted by mistake.

and works on all filesystems

So does ZFS, if you decide to run another filesystem in a ZVOL (but why would you? We're talking about end users.

The trouble is to come up with a good usecases for snapshots on the desktop, especially since a good backup solution is still needed in case of hardware failure

zfs send / receive makes backups trivial to a NAS. Just take a snapshot, and you can then send all of the deltas between that snapshot and the previous one to a remote drive, over SSH (for example). Then, on the NAS, you also have the snapshots, as well as the latest version, so you can revert to earlier versions easily.

Good job Debian team

[JackieBrown]

This looks like a solid release. I only use stable for as long as it takes for the new queue to start start dumping back in Sid but I appreciate the hard work that has gone into this.

And the new artwork really rocks. I was shocked to see plymouth working out of the box with my nvidia card. The consistency from grub to kde launch is really stunning and makes the whole bootup feel seamless.

I love it!

[no known priors]

I'm a Ubuntu user, but I know where it comes from. Debian has been the dream operating system of mine for ages. Easy to install thousands of packages, stable, safe, etc. The only trouble is, when I first tried to install it in 2007, I couldn't get it to work with my wireless card. Ubuntu just worked. I'm going to guess that it wouldn't work now either; my wireless card is one of those Intel ones with the locked up firmware so that I don't start spamming the airwaves... (If I recall correctly the software is ipw2200, or similar.)

Anyway, one thing I note from the press release, is that it is still including OpenOffice.org 3.2.1. I wonder when they'll get LibreOffice (Ubuntu will get it in the 11.4 release).

Great job Debian!

Re:I love it!

[mmj638]

it is still including OpenOffice.org 3.2.1. I wonder when they'll get LibreOffice

Debian's OpenOffice uses patches from Go-OO (now merging with LibreOffice) anyway, so in some ways it is already more similar to LibreOffice than to stock OpenOffice.org. It opens .docx documents very well, for example.

This is also true of Ubuntu's, and generally other distros' OpenOffice packages.

LibreOffice itself came into existence too late for an actual LibreOffice version to make it into Debian 6.0.

I expect it will be a smooth and uncontroversial transition from Debian 6.0's Go-OO enhanced OpenOffice to Debian 7.0's LibreOffice. I'm guessing they'll no longer use the OpenOffice.org branding for that particular variation of it though.

Re:I love it!

[clang_jangle]

That's the whole reason Ubuntu exists though, for people who lack the time or ability or whatever to just install and configure Debian (which exists for people who lack the time or ability or whatever to just build gentoo, which exists for people who lack time or ability or whatever to just create their own private distro from scratch, which of course is for people who lack time or ability or whatever to just go ahead and code their own custom OS using vi or emacs...).

Some CPU microarchitectures dropped from Debian...

[ornia]

It's interesting to note, that while Debian has traditionally supported more CPU microarchitectures than any other mainstream GNU+Linux distribution out there, they have decided to officially stop supporting multiple microarchitectures with the release of Squeeze. The dropped architectures are alpha, hppa, and arm, the latter of which is replaced by the new "Embedded" ABI of ARM, which Debian calls armel.

Although kfreebsd-i386 and kfreebsd-amd64 have been added, these are not true new CPU microarchitectures in and of themselves, as they are compiled to standard x86 and x86_64 respectively, but obviously with the fairly radical change of not using Linux at all with a different GNU libc requiring all packages to be recompiled. This is the same situation as we have traditionally seen in the never-officially-released hurd-i386 port of Debian (which makes sense to call Debian GNU I suppose, as the Hurd kernel is part of the GNU project already) which seems to be missing so far with Debian 6.0 so far, pending a decision to potentially drop it as well.

All in all, amazing work by all in the Debian project. It remains an incredibly impressive feat that such a project can have no corporate oversight or ownership yet maintain such an impressively influential, relevant, and useful place in the operating system ecosystem. Even with dropping a couple of architectures, Debian still supports more computer types than most people even know exists, and continues to provide package updates that many many other operating systems base their repositories from. Also wonderful to see the website be updated!!

Re:First release without a linux kernel??

[guppysap13]

GNU/Hurd has never been released in Stable. It is available in Unstable, but isn't complete enough for them to upgrade to Testing and Stable yet. GNU/kFreebsd however, is now an official Stable release with Squeeze (6.0).

Re:Yeah!

[icebraining]

Is this not one of the fastest Debian release cycles?

Not really. It's been two years since Lenny, which was two years after Etch.
http://en.wikipedia.org/wiki/Debian#Release_history

Re:yay. two more variants that nobody will want.

[johnw]

Can you provide an actual example of Debian fans complaining in the way you indicate, or is it all in your imagination?

Debian tends to be the way it is because Debian users (and builders) like it that way. Of course they do end up being rather smug as well, but complaints about those who choose to use lesser distributions are notably absent.

Re:Works well, but significantly higher requiremen

[DieByWire]

Squeeze has significantly higher minimal install requirements than Lenny, to the point it wouldn't fit on my Dockstar or my Dt360.

I'm running squeeze on a dockstar right now by booting from a USB stick. Some smart people made it easy for the rest of us.

Re:yay. two more variants that nobody will want.

[ffreeloader]

What a load of bs.

I started with Debian as total Linux noob back when Woody was the official release. I've stayed because Debian stable is so stable, and because the APT system is about as good as installers get. I've never had to wonder whether something wasn't working because it was buggy, or because I lacked the requisite knowledge to configure it correctly. That alone made learning Linux much, much easier and far more straightforward. I'd used a couple of other distros before I heard of Debian, but even simple things in the gui didn't work on them because of bugs and I got very frustrated with them. I never knew if any problem I ran across was a bug or because I'd done something stupid. With Debian I could know with a high degree of certainty that the problems I encountered were my own stupidity, not someone elses.

Debian was a breath of fresh air compared to all the bugs in other distros and Windows. I've played with Ubuntu a few times, but always abandoned it because it's not gotten any better over the years. It's always buggy, buggy, buggy. If I wanted a buggy OS I would have stayed with Windows. And, I find fewer bugs and newer software in the vast majority of cases in Debian testing and unstable than I do in Ubuntu.

Re:yay. two more variants that nobody will want.

[rl117]

You should be aware that Debian is not allowed to use the trademark "Firefox" and also have the ability to apply patches such as security fixes(1). It's not called "Iceweasel" out of anything but necessity. You think this is a Debian-specific issue? Well, no, it's actually a major problem for all other distributors as well(2).

1
2

So the links are 5 years old, but the issues surrounding the trademarks haven't changed or gone away. Distributions shipping "Firefox" have abrogated their ability (and responsibility) to be able to apply changes and security updates to the software without the explicit concent of Mozilla Corporation.

Not exactly free software when it comes on those terms, is it?

Regarding the kernel, I assume you're referring to the non-free firmware removal. Maybe you haven't been fully informed that the non-free firmware was actually removed from the upstream kernel sources as well. As a result, the Debian kernels are far from "non-standard", they are standard!

Regards,
Roger

[FFS Slashdot, it's 2011 and you still can't handle UTF-8!]

Most RC-free release in a long time

[gringer]

I'm amazed that they stuck this release freeze out long enough to get the RC bugs for the testing release down to what looks like the lowest since the graph began tracking testing in 2004 -- I would like to believe that this means squeeze will end up being the most stable/reliable release so far.

Now that the release is done and the freeze is over, an upgrade of the Linux kernel (from 2.6.32 to 2.6.37) in unstable should be soon to follow. Also, Firefox (probably 3.5.9 -> 4) and LibreOffice (OOO 3.2.1 -> LO 3.3).

Re:Some CPU microarchitectures dropped from Debian

[hedwards]

I'm guessing that there are more developers interested in maintaining the m68k port than the Alpha port. Or at least that's how that typically goes. Unless you've got a strange OS like NetBSD which is obsessed with running on absolutely every possible architecture from mainframes to wrist watches, some platforms tend to not have enough people with the hardware and interest to keep updating the branch.

Really excited, then...

[CAIMLAS]

I got really excited when I read this at first, but then I realized it's probably going to have many of the same bugs that the FreeBSD kernel has surrounding the various subsystems (jails) and drivers (recent Intel ethernet crashing, USB, etc. that still don't work for the better part of a year), as well as crippling limitations as it regards adaptability on filesystems (ext*, NTFS, NFS - all limiting) and the like.

i wonder if they managed to get ZFS to work fully with the userland utilities written? That would be the biggest point that might pull me over to give it a go.

Re:Compelled by FSF diff than by church or gov't?

[perpenso]

Lets say your company adopts a BSD project and a GPL project..... The difference is that on the GPL you can only keep the code in house if you do not want to give out improvements. BSD? Rip, sell, earn money, do not contribute back. Or... at the least you can do that. Which does not help the community.

Well for a limited definition of "helping the community". GPL'd projects have directly benefitted by incorporating BSD code and indirectly benefitted from the knowledge and experience of UC Berkeley's pioneering work in developing a *truly* free and open UNIX implementation. Apple, Sun, SGI and other have advanced the world of computing using BSD based code *and* they have also given back in various ways *including* giving code to the community. For example Apple HFS, which I believe has been incorporated into FreeBSD. And some have argued that some GPL folks take without giving back. One key BSD developer writes:
"GPL fans said the great problem we would face is that companies would take our BSD code, modify it, and not give back. Nope -- the great problem we face is that people would wrap the GPL around our code, and lock us out in the same way that these supposed companies would lock us out. Just like the Linux community, we have many companies giving us code back, all the time. But once the code is GPL'd, we cannot get it back. Ironic."
http://kerneltrap.org/OpenBSD/Stealing_Versus_Sharing_Code

Basically I'm saying that the meme that GPL gives back and BSD does not is false. Things are far more complicated than that.

Re:The other way around...

[godrik]

I actually see two main things coming out of the freebsd kernel on debian.

1/ having a really good kernel without the stupid port system.
I know that sounds like a troll. But I really elieve linux is a crappy kernel. It is supposed to be monolithic so everything got thrown in the kernel. And now, we realized it is not going to work, so we start using micro kernel types techniques such as network manager, udev, hal... That's not the way to go with a monolithic kernel.

On the other hand freebsd has an awful packing system in my opinion. I need to install weird packages all the time and I don't want to spend so much time compiling everything. I think debian really rocks at having a lot of packages that are overall well compiled with appropriate dependencies. I expect a lot out of debian/freebsd

2/ using a different kernel is likely to activate different code path. That's a great thing for debugging purpose. As parent said, that will help to find GNU dependent code and probably linux dependent assumption. That's a good thing for make our tools more reliable.

Debian: here is an attaboy from me!