Weekly Edition
Return to the Distributions page
| |||||||||||||
|
| |||||||||||||
NetBSD 2.0 released
NetBSD 2.0 is out. The list of improvements in this release is quite
large; see the
announcement for the details.
(Log in to post comments)
NetBSD 2.0 Posted Dec 11, 2004 1:21 UTC (Sat) by jordi (subscriber, #14325) [Link] Oh well. There goes the last major Free Software project that hadn't suffered versionitis.
NetBSD 2.0 Posted Dec 11, 2004 13:14 UTC (Sat) by leonbrooks (guest, #1494) [Link] Is that the same as saying that it's dying?<grins, ducks, runs>
NetBSD 2.0 Posted Dec 11, 2004 15:17 UTC (Sat) by indygreen (guest, #26119) [Link] I sincerly hope that you don't mean it's a dying project becuase that just isn't the case. NetBSD is alive and well.
NetBSD 2.0 Posted Dec 14, 2004 1:55 UTC (Tue) by alq666 (guest, #11220) [Link] That (.*)bsd is dying is a joke that must be as old as bsd itself.
first impression good Posted Dec 11, 2004 16:03 UTC (Sat) by rcbixler (guest, #11917) [Link] I've installed it and it looks like a winner to me. Right away, I noticea couple of annoyances from my old NetBSD 1.6.2 installation are gone. CD burning no longer locks up my system and, when using a KVM to switch to another system and to switch back to NetBSD, I am no longer required to reboot to regain use of the console keyboard. Also, it's nice to have the additional security that the "no execute stack" provides.
NetBSD 2.0 released Posted Dec 11, 2004 19:13 UTC (Sat) by tzafrir (subscriber, #11501) [Link] They still use XFree (4.4.0) .
Isn't Xorg more distro-friendly? What about portability?
NetBSD 2.0 released Posted Dec 13, 2004 20:31 UTC (Mon) by TwoTimeGrime (guest, #11688) [Link] It might take time to transition. For example, Debian has not yet made the switch as they are trying to work out some issues before the migration takes place. The same might be true of NetBSD.
NetBSD 2.0 released Posted Dec 11, 2004 20:10 UTC (Sat) by sbergman27 (subscriber, #10767) [Link] I have a good feeling for the strengths of FreeBSD and OpenBSD. Where does NetBSD fit in? If I'm running on a platform that is supported by FreeBSD and OpenBSD, what relative strengths does NetBSD have. Or is it mainly for platforms not supported by other BSDs? It's supported platform list is truly amazing.
NetBSD 2.0 released Posted Dec 11, 2004 21:31 UTC (Sat) by Ross (subscriber, #4065) [Link] NetBSD is simple to install and maintain and it is also small. It is moreconservative in the types of changes which are made and I would expect very little breakage between updates. If you are looking for security OpenBSD is the way to go. If you are looking for fancy desktop or server features then FreeBSD is fine. But if you want something simple that just works... especially on older or more obscure hardware it's hard to beat NetBSD.
NetBSD 2.0 released Posted Dec 11, 2004 20:14 UTC (Sat) by Zarathustra (guest, #26443) [Link] No private namespaces? No decent distributed fs? Still stuck with root?
*rolls eyes*
NetBSD 2.0 released Posted Dec 11, 2004 21:53 UTC (Sat) by pjdc (subscriber, #6906) [Link] NetBSD is about love, not technology.
NetBSD 2.0 released Posted Dec 11, 2004 23:19 UTC (Sat) by chant (subscriber, #20286) [Link] This is not Slashdot.These comments are generally intended to provide informative insight in to the articles. I do not subscribe to this site to read your personal opinion.
Thank-you for please considering this in the future.
NetBSD 2.0 released Posted Dec 12, 2004 0:40 UTC (Sun) by sbergman27 (subscriber, #10767) [Link] pdjc does not specify what the "love" is for. Taken together with Ross' reply to my query above, I would guess he means a love of simplicity, as opposed to adding technology simply because it is considered "cool" or "buzz word compliant".
I'm not sure that it is really best to say this, but honestly, I don't subscribe to this site to read posts from self designated posting police. This is a low volume site and it is really *so* easy to just skip posts we don't find interesting.
-Steve Bergman
NetBSD 2.0 released Posted Dec 13, 2004 20:36 UTC (Mon) by TwoTimeGrime (guest, #11688) [Link] Not every OS is focused on technical one-upmanship. Use whatever OS fits your requirements and that you are happy using.
Nutcases Posted Dec 12, 2004 1:18 UTC (Sun) by ncm (subscriber, #165) [Link] One is tempted to say, but can't, that NetBSD has fewer nutcases involved than FreeBSD, OpenBSD, or DragonflyBSD. However, one can say with confidence that it has different nutcases.The project's response to Fejj's scalability benchmarks was astounding. It took only a couple of weeks' tuning to match Linux 2.6's years of work. My only complaint is that PIPE_BUF is still 512 bytes, as inherited from the PDP-11 days. They elected not to grow it to (e.g.) 4K, as found in modern systems, from what seems to be a combination of founder worship, base superstition, and misplaced fear that some program somewhere might have been coded to depend on PIPE_BUF not being any bigger. (Saner heads say that such a program was broken to begin with, and the sooner you find out the better off everyone is.)
Nutcases Posted Dec 12, 2004 17:07 UTC (Sun) by khim (subscriber, #9252) [Link] The project's response to Fejj's scalability benchmarks was astounding. It took only a couple of weeks' tuning to match Linux 2.6's years of work. Not at all. They only replaced old O(n) algorythms with O(1) ones. As far as actual speed is concerned they are still slower then Linux 2.6. And algorythms are uncopyrightable so I'm pretty sure they just lifted them from Linux 2.6 (or more likely from FreeBSD). Since it was possible to do for ages one should ask "why did they only do this when unfavorable benchmarks were showed?" instead. Most of Linux 2.6 was focused on slightly different type of scalability - not from 1 to 10000 processes but from 1 to 100 CPUs!
Nutcases Posted Dec 12, 2004 17:36 UTC (Sun) by philips (guest, #937) [Link] This guys love to do things right.For example, I used twice NetBSDs implementation of ipv4 stack - it is simple and readable. I'm not talking about benchmarks - I'm talking about "enabling technology." And I generally do not like benchmarks. Linux kernel beats any other kernel on benchmarks - but still have some long-standing funny little annoying problems (*). (*) Even with 2.6 copying large file (large ~ twice bigger than size of memory) tends to render system unusable. Kernel people still insist that it doesn't happen to them. Well, it happens to me since 2.0 kernels. And I still see no progress on issues in 2.6. Never happened to me on *BSD.
Nutcases Posted Dec 12, 2004 18:51 UTC (Sun) by khim (subscriber, #9252) [Link] It does not happen to me as well and I'm dealing with video-files 10-20Gb in size of regular basis. Gentoo (stable+kernel 2.6.9-r9), nothing fancy. Perhaps you should investigate problem sowhat deeper ? It happened for me with PATA, but now I do not have PATA drives anymore - not sure if it's related to your problem or not...
Nutcases Posted Dec 13, 2004 0:26 UTC (Mon) by larryn (guest, #3457) [Link] Sounds like the problem only happens to your box. You didn't say specific what fs being used and what your hardware spec is. Otherwise, how do you think people making movies using Linux? I use both NetBSD and Linux but your argument does not sound right AFAIK.
Nutcases and hdparm Posted Dec 13, 2004 14:11 UTC (Mon) by zblaxell (subscriber, #26385) [Link] If you're using PATA drives, run 'hdparm' on them and make sure unmaskirqand using_dma are both set. If they aren't, the system will basically stop whenever you try to write large files, as your CPU is doing all of the heavy lifting and turns off interrupts in the meantime. Other symptoms you'll see in this case is that your clock will start getting slow (or lose sync if you're using NTP), and lost data on serial ports. This is really common, because unmaskirq=0 is the default and using_dma=0 can be set if the IDE driver sees anything it doesn't like about your chipset, drives, or cables. I have a DVD burner that keeps sneaking into using_dma=0 mode...when this happens, the system basically stops until the current read or write ends, after which I can use hdparm to turn DMA back on again. That said, I do observe that Linux 2.[024] basically grinds to a near halt (but not a complete one) if you have one slow device anywhere (e.g. a hard disk on USB1.1) and you suddenly write a lot of data to it. All that data wanting to escape from RAM keeps the I/O queue and buffer cache full until it's all gone, which could mean that minutes pass before you can do I/O again. Combine this with NFS, NBD, or some variants of loop device, and the system deadlocks. I can't keep 2.6 running with such devices long enough to test whether it has the same problem.
Nutcases Posted Dec 16, 2004 11:25 UTC (Thu) by nix (subscriber, #2304) [Link] Er, by far the best way to wring speed out of something slow is to change the algorithm: and changing things for speed's sake in the absence of profiles is extremely stupid, because human intuition is a very bad guide.
And `just lifting' cool ideas from other OSes is a perfectly good idea: grabbing the code from FreeBSD would be fine, as well. Isn't that what the license is BSD *for*?
|
|
||||||||||||