Slashdot Mirror


FreeBSD 8.1 Released

hsn and other readers pointed out that FreeBSD 8.1 has been released. "This is the second release from the 8-STABLE branch which improves on the functionality of FreeBSD 8.0 and introduces some new features. Some of the highlights: zfsloader added; zpool version of ZFS subsystem updated to version 14; NFSv4 ACL support in UFS and ZFS; support added to cp(1), find(1), getfacl(1), mv(1), and setfacl(1) utilities; UltraSPARC IV/IV+, SPARC64 V support; SMP support in PowerPC G5; BIND 9.6.2-P2..." ... and much more. See the release notes summary and the details.

27 of 46 comments (clear)

  1. Re:I Love FreeBSD by xoundmind · · Score: 2, Funny

    Why? The kids are not alright.

  2. Re:I Love FreeBSD by amentajo · · Score: 1

    I don't know what you're talking about. I see one on my screen, and it even says "frist psot" in it.

  3. That's nice, but... by CAIMLAS · · Score: 3, Interesting

    It's nice to see they've got zfsloader in there by default, now. It was otherwise a huge pain to get ZFS to be booted from - you basically had to build your own installer and set up everything manually. Quite the time consuming task.

    Unfortunately, I don't see any mention of these changes:

    * "improved stability for ZFS". Sure, it supports pool version 14! What the fuck does that mean, really, when "bare minimum 4GB RAM" was a requirement for 8.0 to get it even remotely stable (some tuning required)? I don't care if it runs for months without locking the system. It's still locking the system.
    * "decreased memory use for ZFS". It's not even doing deduplication in 8.0 RELEASE yet using 3GB of RAM at an idle load is not unheard of.
    * Why so quiet on the USB front? Nice to see they got ralink devices added, but that does little for the fact that USB is almost completely unreliable in 8.x. Just take a look at the USB mailing list - problem after problem that's the same (USB has many, many timing/boot/detection issues in 8.x), with the seeming consensus being "we don't care, it works for me".

    FreeBSD needs to fix those things or forever be relegated to amateur hour. Seems "quality things that work" gets relegated to "superior design". That's all fine and good, but if you've got to rape an ape just to get the damn thing to work as designed due to implementation flaws, it's essentially worthless.

    --
    ~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
    1. Re:That's nice, but... by mat128 · · Score: 1

      I agree, and besides all this stuff they need to provide an easy way to install it using a PXE server. They have a bugged bootloader (BTX?) since 7.2 so I still have to use 7.1's loader to deploy any new versions. On top of that, I have to uncompress/untar/cpio a bunch of archive just to get access to the actual installer config file... and do the whole process again to get a disk image bootable from PXE.

    2. Re:That's nice, but... by Freultwah · · Score: 1

      What the fuck does that mean, really, when "bare minimum 4GB RAM" was a requirement for 8.0 to get it even remotely stable (some tuning required)?

      I really do not know what kind of orifice you are pulling it from. I set 8.0 up with ZFS on a machine with 3 gigs of RAM, did not tune anything and there were zilch problems with neither stability nor speed. Yes, it is "works for me", but hell, it does work for me. In many configurations and with hardware over three years old. And single core processors.

      * "decreased memory use for ZFS". It's not even doing deduplication in 8.0 RELEASE yet using 3GB of RAM at an idle load is not unheard of.

      For some odd reason then it's been working for me in my home server with 2 gigs of RAM (single core, DDR), no tuning and with extra to spare for daemons and no problems with speed that could be attributed to the file system. It even worked with a single gigabyte, but I must admit that it was a lot on the slow side. In another machine I have 4 gigs of RAM, no tuning, prefetch enabled and I've yet to see memory usage top three gigs under load. Maybe there's not enough load. With prefetch off, memory usage is not even worth mentioning. I think I have been deploying some other FreeBSD.

    3. Re:That's nice, but... by CAIMLAS · · Score: 1

      In the last 3 months, I've had 3 opensolaris related FreeBSD system crashes on multiple systems. I've had two OOM related system locks. Yes, this is on 8.0 RELEASE, on very common 3-year-old to current hardware from SuperMicro, with controllers, etc. specifically picked due to their "good support" under FreeBSD. It's the case both with and without special "tuning", and the system load is light (ie multicore systems with a load average around/under .5). The hardware has all been verified to be error-free.

      Sure, maybe it's me - maybe I'm not massaging it correctly. However, I've never seen such systemic issues in a non-beta OS before.

      --
      ~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
    4. Re:That's nice, but... by bdrewery · · Score: 1

      Deduplication requires even more RAM / CACHE than you could imagine. Don't hold your breath for it.

    5. Re:That's nice, but... by X0563511 · · Score: 1

      That's interesting... we got some new SuperMicro machines to test to sell at work.

      God help you if you try to enable the RAID controller. It won't ever boot again until you wipe the CMOS.

      Perhaps SMC isn't as reliable as you might think they are.

      --
      For large sets, this will be our guide even unto death, for the LORD will work for each type of data it is applied to...
    6. Re:That's nice, but... by eventhorizon5 · · Score: 1

      I work for an ISP and we were running FreeBSD on our critical (and extremely active) recursive DNS servers and also route master machines (edge-level BGP/OSPF management routers, which are at the level right below the Cisco peer routers - those machines run Quagga). The first issues we started having had to do with random system lockups on the primary recursive DNS server (it appeared like the OS would stop the interrupt controller - pressing the power button on the machine would briefly reactivate the interrupts for about a second probably due to it trying to initiate a sleep mode). This happened on FreeBSD 7.1 and above. I went around in the FreeBSD mailing lists and saw a large discussion about it on there, with some die-hard FreeBSD fans fuming at the kernel devs since this kind of issue went completely unnoticed (was fixed in 8.0). Then I assumed that 8.0 would be fine, but we were building new route master machines with 8.0, and Quagga was having massive issues with the kernel (I tried building the latest Quagga code which didn't solve the issue) - so we had the choice of either dropping back to 7.0 for those machines or just jumping ship to Linux (Debian specifically, which is what most of our machines run anyway). We went with Linux (still have a few FreeBSD machines though), and all our problems disappeared. The machines in question were IBM x335 and x336 1U rackmount machines.

      FreeBSD used to be the standard for high-performance networking systems, but they really need to get their act together and actually field-test things before deploying production code. The code isn't simply being used on some random person's toy box, it's being used in datacenters on critical infrastructure. Situations like this will make people immediately jump ship.

      -eventhorizon

      --
      #Secret Windows Source Code, in MS C% - if (uptime >= "24 hours") then bsod() else print "Windows License Violation!"
    7. Re:That's nice, but... by CAIMLAS · · Score: 1

      The machines in question were IBM x335 and x336 1U rackmount machines.

      That is really, really damn interesting. We've got two machines that are very similar (x335) running FreeBSD still - one is at 6.2 (runs our management VLAN services - SNMP trap, log analysis, etc.). It's got a low CPU/RAM utilization, yet it has locked up numerous times - sometimes as often as twice a day, but usually somewhere in the 2-3 month range. I've been unable to trace the problem to anything, though it's my suspicion that it may be the single channel SATA controller. It very well may be, but at this point I'm hesitant to rule out FreeBSD as the (at least partial) culprit on account of what you're saying.

      As far as "production code in data centers" I'd say half the problem in this area is ports. It's a huge burden and responsibility to test everything thoroughly for a sysadmin before deploying it, much more so than it is on Linux due to the lack of a shared user base (eg. "I'm using package version xyz that comes with debian release y" vs. some ungodly combination of versions of everything).

      --
      ~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
  4. ZFS reliability on 32-bit by gravos · · Score: 2, Interesting

    Anyone know if this release improves ZFS stability on 32-bit machines? even with 2GB of ram i still get occasional kernel panics due to it running out of address space.

    1. Re:ZFS reliability on 32-bit by bdrewery · · Score: 1

      I've had those problems a lot. I highly doubt the fragmentation issues have been fixed though. Limiting the ARC can help a lot (via /boot/loader.conf). I gave up and moved to 64 bit and haven't had issues since.

    2. Re:ZFS reliability on 32-bit by Phantom+Gremlin · · Score: 1

      I'm not a ZFS expert, but IIUC it's simply not possible to have stable production quality ZFS on 32 bit machines. This has nothing to do with FreeBSD, the original Solaris codebase just wasn't designed for it. There are places where the code manipulates 64 bit constructs in what it assumes are atomic operations. That assumption is often invalid on a 32 bit machine. Also the code is written assuming that it has a significant amount of VM to play with.

      I don't think it's reasonable to expect the FreeBSD port to fix/workaround fundamental design decisions like that. The very first dictum in ZFS Best Practices is: Run ZFS on a system that runs a 64-bit kernel. How much clearer can they make that statement? N.B., repeated for emphasis: this is the very first dictum in ZFS Best Practices.

      IIUC it's OK to expect ZFS to run with 2 GB of RAM, just not with a 32-bit kernel.

  5. Re:Go FreeBSD! Whooo! by OSDever · · Score: 1

    Hey, speak for yourself.

    --
    What is the airspeed of a fully laden swallow?
  6. Re:I Love FreeBSD by Jeremiah+Cornelius · · Score: 2, Funny

    I don't want a joke. I want OS support for the Elan framebuffers in my R4000 Indigo and R4400 Indigo 2.

    --
    "Flyin' in just a sweet place,
    Never been known to fail..."
  7. Re:Who cares? by rivaldufus · · Score: 3, Insightful

    Yeah, Linux will eventually be included on that list for Slashdot, too. Mac fanbois have taken over; the Ubuntists are out.
    Pretty soon, a minor Linux kernel update won't make the front page.

  8. Re:Who cares? by tulare · · Score: 1

    Perhaps, although the number of (non-Ubuntu) Linux machines deployed outside of people's parent's basements exceeds fBSD and Ubuntu combined, which is kind of a built-in interest base. My expectation is that most smart people try to stay abreast of trends in the industry they work in ;)

    --
    political_news.c: warning: comparison is always true due to limited range of data type
  9. Big problem with ethernet adapter by rhomp2002 · · Score: 1

    I had a very common ASUS ethernet adapter. When I tried to install FreeBSD to try it out I found that they had no support for this adapter. It was the one that most of the ASUS MB's used at the time. I see they still, 2 years later, do not support this very common adapter. It is an ATTANSIC L1 ethernet adapter and is very common. Until FreeBSD starts opening themselves to the common hardware that is used on many of the MB's then I don't see why I should jump through hoops to use their O/S. I did try DeskBSD on an older computer I had and it worked fairly well but not any better than Linux did and Linux had a lot more supporting applications than the BSD did. Saw no reason to go to BSD over what I could find elsewhere and with more support. Real shame because the basic structure is great but the details that are missing make all the difference. I find that ARCH Linux and Mint Linux give me more with speedier applications than BSD does and still have the protection that Windows does not offer.

    1. Re:Big problem with ethernet adapter by rhomp2002 · · Score: 1

      Stupido. I have a very common ASUS motherboard, not an ASUS ethernet adapter. Preview is your friend.

    2. Re:Big problem with ethernet adapter by imp · · Score: 1

      I have the ATTANSIC L1 on one of my systems, and it works great on FreeBSD 8.1.

    3. Re:Big problem with ethernet adapter by rhomp2002 · · Score: 1

      Wouldn't even hook up with the Asus MB I had at the time. When I did a little research with the FreeBSD forum I was told that they did not intend to implement it. Same thing with DeskBSD as well. I had the same problem with a couple of Linux distros but when I found a Linux app from ATTANSIC with the code for it the distro creator made up an module to implement it. BSD people wouldn't do that. Consider yourself lucky that your worked. It didn't work with my ASUS MB.

    4. Re:Big problem with ethernet adapter by rhomp2002 · · Score: 1

      I just took a quick scan of the forum at FreeBSD. I last tried to install version 7.1 and when I got the response from the BSD people dropped the idea. I see that there were enough other people who complained that there was an update to version 7.2 that finally included the ATTANSIC/ATHEROS L1 in the hardware that would be accepted by FreeBSD. The initial release of 7.2 would still not accept the L1 but the update to that release made it work. Maybe I will give it another chance now but I no longer have the computer with the ATTANSIC L1. I have updated to a newer MB (Gigabyte UD2). Maybe BSD will accept that one.

  10. Re:No Drivers - No Java - No Good by bdrewery · · Score: 1

    Excuse me, but that troll you just pasted is over ten years old. I've perused my logs and found myself referencing said troll over two hundred times. Thank you.

  11. Re:No Drivers - No Java - No Good by trasz · · Score: 1

    Linux is a nice 'hobby' OS; however, its hardware support sucks for virtually all modern devices. It's 2010, and Linux still has problems with basic functionality like suspend or graphics acceleration support on many machines. ;-> Of course, you're missing the fact that desktop is not the whole world. Lack of 802.11n support for a few devices doesn't really matter when you're building a core router, NAS, or a http server. (Also, your trolling would be better if it didn't contain claims that are easy to verify as false. In particular, the point about ACLs is bad, because it's easy to check that e.g. Linux doesn't support standard NFSv4 ACLs at all. ;-)

  12. Re:Awesome! by trasz · · Score: 1

    According to "man -k L1", Attansic L1 is supported by the age(4) driver.

  13. Re:No Drivers - No Java - No Good by trasz · · Score: 1

    Yeah, especially the "hobby" part, and the "up to ve level of other unix systems", which is false, as explained above. ;->

  14. Re:Who cares? by bonch · · Score: 1

    What the hell are you talking about? Apple-haters have almost completely taken over the comments section to any Apple article. What a lame attempt at karma-whoring on your part.