Forgot your password?
typodupeerror
Operating Systems Upgrades BSD

FreeBSD 8.0 Released 235

Posted by timothy
from the to-be-thankful dept.
An anonymous reader writes "The FreeBSD Release Engineering Team is pleased to announce the availability of FreeBSD 8 stable release. Some of the highlights: Xen DomU support, network stack virtualization, stack-smashing protection, TTY layer rewrite, much improved ZFS v13, a new USB stack, multicast updates including IGMPv3, vimage — a new virtualization container, Fedora 10 Linux binary compatibility to run Linux software such as Flash 10 and others, trusted BSD MAC (Mandatory Access Control), and rewritten NFS client/server introducing NFSv4. Inclusion of improved device mmap() extensions will allow the technical implementation of a 64-bit Nvidia display driver for the x86-64 platform. The GNOME desktop environment has been upgraded to 2.26.3, KDE to 4.3.1, and Firefox to 3.5.5. There is also an in-depth look at the new features and major architectural changes in FreeBSD 8.0, including a screenshot tour, upgrade instructions are posted here. You can grab the latest version from FreeBSD from the mirrors (main ftp server) or via BitTorrent. Please consider making a donation and help us to spread the word by tweeting and blogging about the drive and release."
This discussion has been archived. No new comments can be posted.

FreeBSD 8.0 Released

Comments Filter:
  • Jumping the gun... (Score:5, Interesting)

    by cperciva (102828) on Thursday November 26, 2009 @02:29PM (#30239052) Homepage

    Technically, 8.0-RELEASE has not yet been announced. Judging by the links in the submission, it looks like the "anonymous reader" is whoever owns cyberciti.biz, and he decided to submit the story early in order to drive traffic to his site.

  • FreeBSD rocks :) (Score:5, Interesting)

    by clang_jangle (975789) on Thursday November 26, 2009 @02:30PM (#30239060) Journal
    I was intending to install RC3 on a new desktop machine a few days ago, but got the error message "this version not available on this server". So I went to the options screen and changed it to 8.0-RELEASE just on a hunch and happily it was there and installed without a hitch. Definitely several good performance improvements over 7.2, especially when copying large amounts of data from a USB disk. So far this seems like a nice, solid release and I look forward to migrating my servers to it (after a month or so, just to be sure).
  • by maztuhblastah (745586) on Thursday November 26, 2009 @04:17PM (#30239726) Journal

    Well said. I agree with everything except one bit:

    So, I guess, the real question is why you'd use Linux over FreeBSD?

    Laptops. Power usage.

    FreeBSD isn't (AFAIK) tickless. Furthermore, a lot of my laptop's power saving features (SATA power saving, FB compression) aren't supported at all. My WiFi card is, but I'm not sure if the power-saving stuff is supported for that either.

    With Linux, all of the above features are supported. As soon as FreeBSD gains support for those, I'm switching.

  • by Anonymous Coward on Thursday November 26, 2009 @04:47PM (#30239902)

    The solaris implementation of DTrace is also significantly better than the one found in FreeBSD. If these are the features you are interested in, you really should be using OpenSolaris, not FreeBSD.

  • by OeLeWaPpErKe (412765) on Thursday November 26, 2009 @05:47PM (#30240312) Homepage

    A more complete feature list :

    -> ZFS
    -> DTrace (if anyone tells you SystemTap is equivalent, take them out and shoot them in the head. Twice. Carefully make sure they're dead. Shoot a few more times to make sure)
    -> jails (if anyone tells you chroot or even vserver is equivalent, see above)
    -> pf, and even other options (anyone tells you iptables ... you know what to do)
    -> faster routing code (all juniper routers run a modified version of fbsd, for good reason)
    -> actual coherent system (ever worked on a busybox linux system ?)
    -> drivers are not so plentiful, but they are better quality. Rare hardware simply doesn't work, instead of sort-of working and then crashing your system every now and then
    -> binary compatibility, not just with linux, but SCO and System V
    -> sensible network defaults
    -> default filesystem supports snapshots. It's not impossible with linux, but still this is nice
    -> netgraph : there is no such thing as a way to mangle packets that fbsd doesn't support. Plus : easy to add stuff to. Fast (if not perfect, but faster than doing stuff in linux user space), and ultimate flexibility
    -> application-level firewall (meaning applications can push firewall rules into the kernel that only apply to that one application. Helpful for stability and ddos mitigation)
    -> coherent, existing organisation, polite even (try getting something done via lkml. I hear it's been known to cause nervous breakdowns. (jokingly) Even murder [wikipedia.org])

    So, personally, I'd agree with "better quality". Especially for network servers.

    If you want to use it on a laptop ... better look elsewhere. It will run, though.

  • Re:Awesome! (Score:5, Interesting)

    by laffer1 (701823) <luke@GAUSSfoolis ... m minus math_god> on Thursday November 26, 2009 @08:02PM (#30241360) Homepage Journal

    That depends on your definition of BSD. Some people look at the userland and the large amount of BSD code in Mac OS X and call that BSD. I'd say there are more than 0.01% of users that are on mac os.

    I started a project to make a desktop friendly BSD operating system called MidnightBSD. There's also PC-BSD and the now defunct DesktopBSD. The new problem is that Linux folks have grown inpatient with the linux on the desktop idea. They want it now and feel that supporting other operating systems in their FOSS work is slowing linux down. A few projects have really done some serious changes to their software to make it function poorly (or not at all) on other OSes including *BSDs. Sometimes it's a lack of people to make reasonable updates to the kernels for various things like "new" video interfaces. Even things like X.org have done shifts that make hardware acceleration a real pain in the butt on BSD platforms. I've been shunned many times for trying to provide patches both for MidnightBSD and previously FreeBSD to other projects.

    The FreeBSD project has had trouble getting patches upstreamed for things like GCC and binutils in the past. In general, I think many GNU projects are starting to get grumpy with respect to *BSD patches. There's a backlash with BSD developers trying to write alternatives that are under the BSD license because we must to survive. Also, you get into situations like Apple buying cups and switching to LLVM because of fear of the GPLv3. Perhaps fear is not the right word.

    The open source community is not one big happy group but a series of factions that don't get along. It's a shame really.

  • by TheRaven64 (641858) on Friday November 27, 2009 @09:34AM (#30245274) Journal

    Sorry if this comes across as a flame, but as the guy who wrote FreeBSD Update, perhaps you can answer a couple of questions:

    Firstly, why is it so slow? I can cvsup and recompile the tree and install in less time that it takes freebsd-update upgrade to run; the two install steps then take even longer. If I run systat -iostat, I see it hitting the disk incredibly hard. Couldn't it just compare the last modification date of most of the files with the time of the last upgrade? Possibly this has been fixed, as updating from 8-RC3 to 8-RELEASE seems to be going a lot faster than going from RC-2 to RC-3 (and has actually finished while typing this post. Previous updates took a good 3-4 hours; longer than compiling all of LLVM + Clang on that machine).

    Secondly, is it possible to make the merge marginally less stupid? Every time I run freebsd-update it prompts me to okay changes to files that I've never edited (and have no idea what it is even for), where the only change is the comment at the top telling me the FreeBSD version the file came from. I then have to say 'yes' a few dozen times, and at twice while doing that I've managed to type yes to okay one change which actually was important and left my system unbootable. One was a change to one of the init scripts in the bit that mounted the default filesystems; it tried to merge the two versions (even though the file was not one that I had modified, or that users ever should touch) and ended up with complete nonsense so the shell aborted the script while trying to mount filesystems. I only fixed that one by booting to single user mode, deleting the script and grabbing a copy from cvs.

"Out of register space (ugh)" -- vi

Working...