Find the answer to your Linux question:
Page 1 of 2 1 2 LastLast
Results 1 to 10 of 11
Hi, I have been doing some reading and trying to find out why the union of the GNU userland and NetBSD kernel was abandoned by Debian, Gentoo or any other ...
Enjoy an ad free experience by logging in. Not a member yet? Register.
  1. #1
    Linux Newbie SL6-A1000's Avatar
    Join Date
    May 2011
    Location
    Australia
    Posts
    120

    Debian or any Linux GNU/kNetBSD or OpenBSD


    Hi,
    I have been doing some reading and trying to find out why the union of the GNU userland and NetBSD kernel was abandoned by Debian, Gentoo or any other Linux Distro, yet FreeBSD was maintained.

    The way i see it is FreeBSD has a large user base and a larger backing than probably both NetBSD and OpenBSD combined, at least from what i can tell. Which raises my question of why unite a userland and Large Linux Distribution like Debian with and a BSD kernel/ large BSD distro, where the maximum gains for either is rather minimal.

    For example FreeBSD already as a majority of the software and applications that Debain provides on Linux. It only take one to look at the repositories of both distro's and you soon realize that FreeBSD as around 30,000 packages in its repo's and Debian as about 32,000. This means that Debian is only really providing approximately 2,000 packages and driver support that FreeBSD doesn't offer on its own.

    Then there is the kernel, and obviously there are a few advantages of running a BSD kernel over a Linux kernel, such as stability, native ZFS support, etc... with a GNU userland which is useful if your a Linux user with little understanding of BSD. But when it comes down to it FreeBSD is only offering those few features and a few more CPU architectures to the Debian Base that exists. While Debian is probably only giving back minimal code fixes, etc simply because FreeBSD is such a large distro with already large support.

    When you compare this to say the NetBSD or OpenBSD project, the same benefits exist in terms of native ZFS support, kernel stability, and better security features etc... But the benefit to either NetBSD or OpenBSD and Debian is far greater than that of what FreeBSD offers. Both NetBSD and OpenBSD do not have the huge backing or developer community behind them that FreeBSD has, thus by Debian or any other Linux Distro choosing to use there kernel on top of the GNU userland has far greater benefits than simply a stable kernel with ZFS support. One of the biggest and most obvious benefits is the developers of that large Linux distro backing the developers of persay NetBSD, which means that NetBSD or OpenBSD would gain more support and bug fixes a lot faster, and the huge package support that is under the hood of running a GNU userland. Not to mention in that Linux Distro's case like Debian they are gaining the huge array of CPU architectures support (practically anything!) and BSD kernel support.

    It just seems like there are greater benefits to be had from the union of the other BSD distros and GNU userland at least in NetBSD's case then there are disadvantages, than those that occur with the FreeBSD and GNU union Debian has offered.
    I also think that this is one of the main reasons many think the Debian GNU/kFreeBSD system is considered so ridiculous as it really doesn't offer much more than the two distro's do independently.

    Note: This is an open discussion that can related to any of the BSD's, but i only have one thing it must be closely related to the subject of this topic. As i want your opinions and thoughts!
    I have nothing against FreeBSD, so don't get me wrong, i respect all the BSD distros, this is simply just a discussion

  2. #2
    Just Joined! Randicus's Avatar
    Join Date
    Sep 2011
    Location
    China
    Posts
    74
    At this point, KFreeBSD is an experiment to see if Linux and BSD can be merged. Which is how the few Debian users who have tried it regard it.

    As for the benefits, if there would be any:
    What is Debian GNU/kFreeBSD ?
    Debian GNU/kFreeBSD is an operating system from Debian, but which uses the kernel of FreeBSD instead of the Linux kernel (thus the name). Some day, most applications would exist both in Debian GNU/kFreeBSD and in Debian GNU/Linux.
    It makes sense that Debian would pick FreeBSD. Debian is a stable system, so choosing the most stable BSD system would be their first choice.

    Whether the experiment has any future is another matter, but it is an interesting project.

  3. #3
    Linux Newbie SL6-A1000's Avatar
    Join Date
    May 2011
    Location
    Australia
    Posts
    120
    Quote Originally Posted by Randicus View Post
    At this point, KFreeBSD is an experiment to see if Linux and BSD can be merged. Which is how the few Debian users who have tried it regard it.

    As for the benefits, if there would be any:

    It makes sense that Debian would pick FreeBSD. Debian is a stable system, so choosing the most stable BSD system would be their first choice.

    Whether the experiment has any future is another matter, but it is an interesting project.
    I do not doubt the decision on using FreeBSD when you put it in that regard "an experiment" than yes, it makes perfect sense to use the FreeBSD kernel.

    But from what i can see Debian has put alot of energy and time into this project, so one has to assume that the project is a much longer term deal. You wouldn't add the kfreebsd packages into the main debian linux server, if this wasn't a more long term project. It wouldn't make sense... over providing periodical updates that one just downloads an iso image of the updated system.

    Which is why i started to question it? I mean sure there is always the chance that Debian will resurrect GNU/kNetBSD now that they have an understanding and a basic concept of how to integrate the BSD kernel into the GNU userland. BUT on that same token having two seperate BSD kernels means more developers to maintain those distributions. Which means A) hiring more developers to work on said projects or B) reducing developers in other sections (projects) to work on said projects.

    Which in my opinion seems unlikely, considering that the project is now underway and already past a beta release, bringing OpenBSD or NetBSD under the Debian wing is unlikely especially for the advantages either kernel would provide over FreeBSD.

    For example if they wanted more cpu architectures the structure and development team is already in place for FreeBSD, thus porting the FreeBSD kernel to other architectures would be easier than creating a new section for NetBSD just to get those same CPU support.

  4. #4
    Just Joined! Randicus's Avatar
    Join Date
    Sep 2011
    Location
    China
    Posts
    74
    KFreeBSD is a small part of the Debian project. I do not know how many people are working on it, but it is not enough to drain resources. So no worries there. And I am not sure how many developers are paid, if any. Debian is mainly volunteer.

    The general consensus among Debian users, at least the ones who participate on the two fora, is that the Linux-BSD hybrid is probably a waste of time. Whether that is true or most users have not looked into it enough to know what it is about, is another question. Myself, I think a hybrid would be a good thing if the best of the two could be combined. Can it be done? I would not pretend to know enough about the mechanics to venture a guess. But is it necessary? Do we really need yet another distribution? That is a question that would probably generate a great deal of debate.

    Then there is the question; What if it is successful? Perhaps it would lead to an open source-Unix revolution. The hybrid might make both Linux and BSD obsolete!

    Judging by how far the project has progressed and how it has been received thus far, I say it will be released, but remain a small distro. It will then be a question of how long Debian decide to keep it going. In the end, whether or not the project is abandoned depends on whether or not enough people use it to justify the work of maintaining it.

  5. #5

  6. #6
    drl
    drl is offline
    Linux Engineer drl's Avatar
    Join Date
    Apr 2006
    Location
    Saint Paul, MN, USA / CentOS, Debian, Slackware, {Free, Open, Net}BSD, Solaris
    Posts
    1,283
    Hi.

    Yesterday I downloaded from Debian -- Network install from a minimal CD the iso for kfreebsd-amd64 so I would say that it is released.

    I plan to evaluate it because the kernel has support for zfs, which (apparently) the Debian kernel cannot because of the BSD license for zfs not being allowable in the Debian view.

    The links caravel posted have more details.

    I have not seen anything that suggests that kfreebsd is an experiment, like btrfs seems to be (too early in development, perhaps).

    Best wishes ... cheers, drl
    Welcome - get the most out of the forum by reading forum basics and guidelines: click here.
    90% of questions can be answered by using man pages, Quick Search, Advanced Search, Google search, Wikipedia.
    We look forward to helping you with the challenge of the other 10%.
    ( Mn, 2.6.n, AMD-64 3000+, ASUS A8V Deluxe, 1 GB, SATA + IDE, Matrox G400 AGP )

  7. #7
    Linux Newbie SL6-A1000's Avatar
    Join Date
    May 2011
    Location
    Australia
    Posts
    120
    Quote Originally Posted by Randicus View Post
    KFreeBSD is a small part of the Debian project. I do not know how many people are working on it, but it is not enough to drain resources. So no worries there. And I am not sure how many developers are paid, if any. Debian is mainly volunteer.
    Exactly my point lol, which puts the idea of them supporting another BSD kernel at being highly unlikely.

    Quote Originally Posted by Randicus View Post
    The general consensus among Debian users, at least the ones who participate on the two fora, is that the Linux-BSD hybrid is probably a waste of time. Whether that is true or most users have not looked into it enough to know what it is about, is another question. Myself, I think a hybrid would be a good thing if the best of the two could be combined. Can it be done? I would not pretend to know enough about the mechanics to venture a guess. But is it necessary? Do we really need yet another distribution? That is a question that would probably generate a great deal of debate.
    I agree with you the fact the only very few people know about it makes all the more harder, to gauge how sucessful it could be and will be. As the only people that would know about it besides the developers are the one that are similar to you and me... Those questions are hard to dispute, but from what we've seen so far one can assume that it is very possible to create a hybrid, Not that i really know myself.

    Although i wouldn't classify Debian GNU/kFreeBSD as a true hybrid Linux-BSD system. I would put Debian GNU/kFreeBSD simply as one of the first steps in creating Linux-BSD hybrid.
    The reason is because GNU is not Linux, GNU is the software and building components (i.e. GCC, binutils, bash shell, glibc and coreutils) of "GNU systems" of which Linux, Hurd and others come under. Thus, all Debian has done so far is to unify those building utilities so that the same GCC compiler etx that is used for Debain Linux can also Compile Debian GNU/kFreeBSD.

    A true hybrid Linux-BSD system would be something similar the the Mac OS kernel, where components of the BSD kernel are fused with components of the Linux kernel and run under the same userland "GNU".

    Quote Originally Posted by Randicus View Post
    Then there is the question; What if it is successful? Perhaps it would lead to an open source-Unix revolution. The hybrid might make both Linux and BSD obsolete!
    I don't think Linux and BSD would become obsolete at the presence of a Hybrid system that truely takes components from both kernel and all. Every system has its strong points and weak points, its strengths and weaknesses would just be different that of BSD and Linux.

    Quote Originally Posted by Randicus View Post
    Judging by how far the project has progressed and how it has been received thus far, I say it will be released, but remain a small distro. It will then be a question of how long Debian decide to keep it going. In the end, whether or not the project is abandoned depends on whether or not enough people use it to justify the work of maintaining it.
    True, but that's a double-edged sword, they have to provide something that both Linux & FreeBSD don't offer independently, and this is what annoys me about this creation, and perhaps this is my agreement with the opensource ideals of Freedom talking. The GNU userland offers very little to the FreeBSD world unless your a linux user who doesn't grasp BSD very well. While the FreeBSD kernel offers benefits to the GNU userland and people that prefer GNU.
    BUT what it doesn't offer is more choice than what is already provided in a large way, in fact if you were to be blunt about it, it actually reduces choice, because it is promoting the larger distros and eliminating the smaller ones, with the union of two larger communities. This would be equivalent in the closed-source world of Apple and Microsoft uniting to create one distribution (on a relative scale obviously ). Sure the concept sounds great, but it totally elimantes any choice.
    Thats part of why my opinion still stands, sure FreeBSD is stable it has alot to offer, but it doesn't need the backing of Debian to become something more. I believe NetBSD and OpenBSD can offer just as much as FreeBSD but have been overlooked. They are the BSD's that would benefit and be benefited greatly.

    How can a smaller system become better if its always overlooked by the more developed system?

  8. #8
    drl
    drl is offline
    Linux Engineer drl's Avatar
    Join Date
    Apr 2006
    Location
    Saint Paul, MN, USA / CentOS, Debian, Slackware, {Free, Open, Net}BSD, Solaris
    Posts
    1,283
    Hi.

    My recollection is that installing packages in FreeBSD means compiling them. I don't remember any packaged binaries being available.

    Using any measure that seems reasonable, does anyone know how similar the kernels are for {Free, Open, Net} BSDs?

    I just searched in the manuals Manual Pages: Index Page and zfs does not appear. If OpenBSD does not support zfs, then it would be of limited use for me, at least for the time being.

    Every few years I install the BSDs, just for practice, but I use them rarely. Does SL6-A1000 use them?

    Interesting philosophical discussion, but I look at primarily the practical (possibly short-term) value of many things, not necessarily long-term.

    Best wishes ... cheers, drl
    Welcome - get the most out of the forum by reading forum basics and guidelines: click here.
    90% of questions can be answered by using man pages, Quick Search, Advanced Search, Google search, Wikipedia.
    We look forward to helping you with the challenge of the other 10%.
    ( Mn, 2.6.n, AMD-64 3000+, ASUS A8V Deluxe, 1 GB, SATA + IDE, Matrox G400 AGP )

  9. #9
    Linux Newbie SL6-A1000's Avatar
    Join Date
    May 2011
    Location
    Australia
    Posts
    120
    @drl: To answer your question drl, yes i have run FreeBSD, NetBSD, and Debian GNU/kFreeBSD. I have attempted to run OpenBSD but my system isn't fully support by OpenBSD so i have never got a running system beyond the installation. Debian GNU/kFreeBSD has given me alot of issues especially when it comes to getting the GUI working.

    I tend to prefer NetBSD because i understand the pkgsrc system, it is very easy, and only takes 3 steps to have pkgsrc up and running, and its very similar to the Slackware package manager. I tend to get confused with the FreeBSD package system, i have used PC-BSD as way to try a full-blown FreeBSD system.

    ZFS is supported by all of them, to what to degree i have no idea. I know FreeBSD has the latest ZFS and i know NetBSD supports it because i have used a ZFS filesystem in NetBSD. As i said though i don't know about OpenBSD or during install.

    As for the kernels i think you would find that they do differ, more so than the difference that exist between a Debian Kernel and a Red Hat Kernel. As the way i think of it, the more architectures a system supports the more work as to be implemented into the kernel to support those architectures. So one would assume without really knowing, that NetBSD would build a very generic kernel that supports all 52 CPUs at its most basic level, and would improve performance and errors, bugs etc by making small changes/ fixes to that base kernel. Otherwise, they would have a huge workload of porting and tweaking there improved kernel to each architecture.
    I would take a stab in the dark and make the guess that the kernels wouldn't probably differ so much if it wasn't for the huge CPU architecture support NetBSD has. Then again i don't really know... :S

    Maybe thats part of the reason the GNU/kNetBSD project was abandoned?.... debian didn't want the support for all those architectures, and thus when it come down to it the FreeBSD kernel was "better" for the architectures it did want?
    Last edited by SL6-A1000; 03-17-2012 at 01:57 AM.

  10. #10
    drl
    drl is offline
    Linux Engineer drl's Avatar
    Join Date
    Apr 2006
    Location
    Saint Paul, MN, USA / CentOS, Debian, Slackware, {Free, Open, Net}BSD, Solaris
    Posts
    1,283
    Hi.

    @SL6-A1000:

    As for packages, I found:
    For most ports, precompiled binary packages also exist. This method is very quick as the whole compilation process is avoided, but the user is not able to install a program with customized compile time options.[14]
    -- FreeBSD - Wikipedia, the free encyclopedia
    but my recollection is that I needed to compile perl, for example -- perhaps that was long ago.

    As for the kernel, as far as I know all the Linux distributions obtain the kernel from kernel.org. Many of the posts in forums discuss compiling one's own kernel, and a common first step is to download source from kernel.org.

    Regarding ZFS, in the table at Comparison of BSD operating systems - Wikipedia, the free encyclopedia
    zfs is shown ownly in FreeBSD and GhostBSD, the article being last modified on 9 March 2012.

    However,
    The ZFS filesystem developed by Sun Microsystems was imported in to the NetBSD base system in 2009. Currently, the NetBSD ZFS port is based on ZFS version 13.
    is noted at NetBSD - Wikipedia, the free encyclopedia

    Whereas,
    Will ZFS be added to OpenBSD?
    Not unless someone can convince Oracle to change the license for it to something compatible with OpenBSD policy.
    at 1 - Introduction to OpenBSD

    So I certainly wish all the BSDs success at achieving their goals, but if they don't offer something that appeals to a large number of people, they probably will stay small. That's not necessarily bad, it's just a fact.

    Best wishes ... cheers, drl

    ( edit 1: deleted reference to monolithic kernel )
    Last edited by drl; 03-17-2012 at 04:05 AM.
    Welcome - get the most out of the forum by reading forum basics and guidelines: click here.
    90% of questions can be answered by using man pages, Quick Search, Advanced Search, Google search, Wikipedia.
    We look forward to helping you with the challenge of the other 10%.
    ( Mn, 2.6.n, AMD-64 3000+, ASUS A8V Deluxe, 1 GB, SATA + IDE, Matrox G400 AGP )

Page 1 of 2 1 2 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •