“The FreeBSD operating system is finally through it’s buggy 5.x series and into the more reliable 6.x series. Most of the problems of the old days – kernel panics on multi-CPU machines, AMD64 troubles galore, and shaky network drivers – are gone. FreeBSD still isn’t perfect, but at least with 6.0-RELEASE it’s more stable and functional than it has been in the recent past.”
I downloaded it about 4 days ago, I still have not made the CD’s and installed it yet. Looking like it’s already gonna be better than the 5.x series though… at least from what I’ve read at freebsd.org, here, and other places.
/2 cents
Everything seems much faster, especially once you rebuild world & kernel with optimizations.
I suppose FreeBSD will manage to catch up to linux on of these days, both in terms of speed and stability. But it isn’t likely. At best the FreeBSD kernel is equal to Linux 2.4
I suppose FreeBSD will manage to catch up to linux on of these days, both in terms of speed and stability. But it isn’t likely. At best the FreeBSD kernel is equal to Linux 2.4
I don’t know how you can drive such conclusions. I have tested both Gentoo and FreeBSD 6.0 beta 4 with Oracle in binary compatibility. FreeBSD outperformed Linux in terms of speed and load handling. I don’t know if some one could show a benchmark, but anyway, there’s a significant difference in terms of performance between Linux and FreeBSD.
I have no reason to go back to Linux now.
That’s cute, but it’s obvious you’re making things up.
How could you make the freebsd using the linux compatible’s layer run faster than native linux?
the linux’s interrupt overhead more light than the freebsd’s kernel,you can search this the mailist in the freebsd.
if you say it run on the compattible layer faster than native linux,please give you benchmark
The freebsd run the datebase only using the linuxthreads,but this threads in the linux is 8x slower than the new nptl in the linux,how could you say it can run faster using old linux’s thread than native linux with new nptl?
If you faster,please give you benchmark.If not ,you tell a lie
Can you justify any of your claims?
No, he can’t. He’s just a Linuxtard troll. I know him, because he lives one block away from me and goes to LAN parties with me.
Well, that’s a new one:) “Linux Is Poo” is describing someone “troll”?
???
What’s next in line. Flying dogs?
p.s. You can mod me down without any guilt, but I really couldn’t resist this one:)
FreeBSD 6.0 is now my default OS. It’s blazing fast. the 5.x series were already very good, but this one is a major release.
November has been a good month for all of the BSDs, Open, Free, and Net. I wish them continued success and support. Unfortunately this means we will have to wait about 6 months now before the next round of updates.
“There are reports of FreeBSD starting to benchmark much faster than Linux for certain filesystem tasks on SMP systems.”
“The FreeBSD approach to SMP is a better one than that taken by Linux”
“BSDs in general are also much more corporate friendly than Linux, which, if used in a product by a company, puts some serious restrictions on what the company can do and still comply with the GPL.”
“Linux took an approach that was fairly easy and non-disruptive, but ultimately has limitations that we feel will limit its scalability for general computing,”
http://www.internetnews.com/dev-news/article.php/3561526
In other news:
One of the new features in FreeBSD 6.0 is a multithreaded file system,
which greatly improves data access times for local disks, RAID configurations,
network file systems, and SANs. Recent performance benchmarks show that
FreeBSD 6.0 outperforms Linux in raw data throughput.
http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/st…
“There are reports of FreeBSD starting to benchmark much faster than Linux for certain filesystem tasks on SMP systems.”
Bring on the benchmarks. If this is true, it should be in FreeBSD’s interest to get that out there.
“BSDs in general are also much more corporate friendly than Linux
This is so funny. How about taking a look at what actually takes place in the real world? Both in terms of contributions to the kernel and actual deployment. Almost all the people who work on the Linux kernel these days have that as part of their job description, and the Linux eco system has actually SPAWNED a number of corporations (Red Hat alone has almost 1k employees! not bad for a corporation relying entirely on GPL code), in addition to attracting behemoths like IBM and Intel.
“Linux took an approach that was fairly easy and non-disruptive, but ultimately has limitations that we feel will limit its scalability for general computing,”
How is it possible to put forth such a claim when THE PROOF IS ALREADY IN THE PUDDING?? And why aren’t they publishing any benchmarks that demonstrates that they are even in the ballpark of Linux’ SMP performance?
Btw, immediately when I saw that quote, I KNEW it came from Scott Long. That guy just makes stuff up.
Also, Linux’ approach was non-disruptive because of the Linux kernel development philosophy. Basically, most work should and is done in small incremental steps and each step needs to be a step in the right direction. This makes it much easier for people to review the work, and it also makes it easier for the people doing to work to keep up with current kernels. The Linux kernel people have demonstrated again and again that most work can be completed in this fashion.
In FreeBSD, they prefer to do major work in the perifery on private Perforce repositories and then pull things into the main cvs repository whole hog. No wonder they experience disruptions. It also makes it alot harder for people to do a proper review of all the work.
Remember, because of the SMP work that went into 5.x, 4.x performed significantly better on the common case, UP machines. That kind of regression would NEVER be accepted in the Linux kernel.
“Bring on the benchmarks. If this is true, it should be in FreeBSD’s interest to get that out there.”
No need benchmark it, just install FreeBSD-6.0 on your dual-boot Linux and see for your self the difference.
“How is it possible to put forth such a claim when THE PROOF IS ALREADY IN THE PUDDING?? And why aren’t they publishing any benchmarks that demonstrates that they are even in the ballpark of Linux’ SMP performance?”
With the enhanced ULE scheduler, FreeBSD’s SMP is much better than Linux, squidclam is hogging to death on Gentoo Linux while FreeBSD-6.0 with ULE sched has no effort at all
“Remember, because of the SMP work that went into 5.x, 4.x performed significantly better on the common case, UP machines. That kind of regression would NEVER be accepted in the Linux kernel.”
I really accept that FreeBSD-4.x is fast, and IMHO, FreeBSD-5.x is the developmental version which put FreeBSD-6.0 on its scalability and performance.
BTW, if linux is _STABLE_ why they’re putting “EXPERIMENTAL” stuff on it? Does it mean that “EXPERIMENTAL” is “STABLE”?
Never ever ever? Try 2.6. There are so many performance regressions going to 2.6 from 2.4 that you’d blow milk out of your nose.
Of course you can’t quantify them. Same as nobody can quantify any of these magical freebsd performance advantages.
They just are because you’re a freebsd homo ching.
when will osnews editors learn to hyperlink?
is the word “gone” descriptive of the page it points to?
is it too difficult to hyperlink “FreeBSD” to the projects page?
is the word “gone” descriptive of the page it points to?
If you read the teaser, then yes. The days of… are gone.
is it too difficult to hyperlink “FreeBSD” to the projects page?
Yes. I’m not going to link every project in every teaser.
> is it too difficult to hyperlink “FreeBSD” to the projects page?
Yes. I’m not going to link every project in every teaser.
tease..
Thom,
You are such a dick.
I don’t really mind a bit. I’ve started to like linking on “secondary words”.
But if you really want to link the most representative expression, that would be “more reliable 6.x series” or “6.0-RELEASE”.
I agree that “gone” is in no way repressentative of the article. It should have said something like “jem report posted a review:”, where the word review was the one was linked.
Unfortunately Thom doesn’t like critisism. Not even when its constructive.
You’re writing an article reviewing the latest release of FreeBSD, and yet you don’t see the need to link to the page of the software you’re reviewing?
I hope you’re not getting paid for this crap.
“You’re writing an article reviewing the latest release of FreeBSD, and yet you don’t see the need to link to the page of the software you’re reviewing?
I hope you’re not getting paid for this crap.”
FreeBSD is my favourite OS. I’m simply a user and thus I just speak for myself, but I really can’t understand all these gratuitous, anonymous attacks towards the OSNews editors (they’re just a handful, including other threads, but still too many in my view).
This trollish behaviour has nothing to do with the BSD community, and *especially* with the FreeBSD community. I know it well, because I’ve been reading BSD forums, mailing lists and chat discussions for quite a while. So, the attitude of these anonymous comments is really surprising.. definitely weird, I’d say.
That’s why I’ve got this strange feeling.
A feeling that, since FreeBSD is gaining momentum, it’s starting to bother some corporate interests gravitating around other OS’s.
To make it short, I’ve got a feeling that somebody’s trying to mess things up between FreeBSD and the press.
Maybe it’s just a silly idea.. but anyway, as a humble OSNews reader, I hope to see people behave more politely and respectfully, especially around almost irrelevant issues.
Sorry for this not-minding-my-own-f**king-business thing, but I had to say it
I’ve been using FreeBSD for one of my desktop OSes (I have a couple of desktops through the house) for a couple of years now.
The things that I love about it are how incredibly easy it is to update to the latest and greatest. Yes, it does require a lengthly make buildworld process, but it still beats how tough it is to go from say FC3 to FC4 without any problems.
The ports system is amazing, though there is some competition now from Gentoo and portage, which has made progress in leaps and bounds since I first used it back in 2002/03 (around then).
The only things that I desktop user might want to see that aren’t really available right now is something similar to using DBUS and HAL on linux. (Of course, I haven’t looked too hard for a FBSD solution, and I would love to hear anyone’s suggestions).
Of course, it is a little more involved to use than something like Ubuntu, but if you’re looking for a machine that you can play with and really get immersed in unix-like systems with a great community and easily updated, and (for the most part) bleeding edge software, then really try it out. You just may be surprised with how well it works.
And of course, for servers it still kicks serious ass as well.
Is this the same mendicant from indecline?
The only thing I really want to see in FreeBSD is something better than portupgrade for upgrading installed apps. Apt is primarily binary-based, but it’s spoiled me as far as upgrading goes. Portupgrade tends to choke when you give it too much to upgrade at once, which is why the Gnome on FreeBSD guys provide their own upgrade script for each new release.
An easier way to mount my iPod would be nice, too, but I don’t use FreeBSD for user-friendliness. For all the complaints, 5.x has been rock-solid and fast as hell for me and I’m looking forward to the improvements of 6.0.
Have a look at portmanager. I haven’t played with it yet, but there’s a lot of positive feedback around it on the various FreeBSD mailing lists.
“I have no reason to go back to Linux now.”
So don’t do it. FLOSS is all about choice and freedom.
It still won’t recognize my Centrino WiFi card and would not recognize my Marvell Yukon Gigabit ethernet (Tecra M3). Many of the applications are not there in packages – Installed Gnome and got 2.10 when 2.12 has been out for some time now.
-D
Many of the applications are not there in packages
This is due to license restrictions. There are commercial applications that will never be downloaded using pkg_add -r
Installed Gnome and got 2.10 when 2.12 has been out for some time now.
You need to update your ports first. Use CVSup or Portupgrade.
On a fresh install, I was hoping that the FreeBSD gnome website was correct, and that I could just pkg_add -r a Gnome 2.12 desktop. It installed 2.10. So now my laptop is chugging away on the upgrade install from ports. I hope to wake up to a nice Gnome 2.12 desktop!
Ports usually get the new software faster than the pkg respiratory, so I’m not sure if they’ve made packages for it yet.
One more thing, on arch gnome 2.12 seemed a bit slow (for example metacity lagging), can you give me a report if this is like this on freebsd too?
Not knowing what kind wifi chip you have, looking into iwi and ipw drivers, these take some special configuring since you have to load an intel firmware.
After that setup my thinkpad T41 detects and loads the wireless automatically at boot.
The wireless on there is Intel PRO/Wireless 2200BG or in the same family. Install the iwi-firmware port.
No current driver for the Marvel PC Express gigabit ethernet chip.
/me is happy with how freebsd runs on his Tecra M3
They need to get over the ports system and start supporting binary upgarde of packages. The packages are incomplete and many of them fail to upgrade properly. Desktop users will not stand for hours of compile time for a 3% performance gain. PC-BSD fills in some of the gap but its KDE based. Gnome users will have to wait …
Its there, its called pkg_add -r
is it true that the amd64-version is unstable?
No. At least officially it is a tier-1 platform. See here: http://www.freebsd.org/platforms/amd64.html
and here: http://www.freebsd.org/doc/en_US.ISO8859-1/articles/committers-guid…
Check out the mailing lists, they can be a good indicator as to what problems people are experiencing as well as the best way to get help and improve support for your specific hardware.
5.3 and 5.4 are stable on my amd64. I havn’t tried any other versions. However, some ports dont work on amd64, when you try and build them you are informed that they do not work. Overall 5.3 upwards has been stable on all my machines (I have 2 heavily loaded webservers, one backup server and one desktop computer (amd64)). I did not try versions below since they were not considered STABLE by the FreeBSD team. 6.0 is considered stable so I guess it should work fine even on amd64
No, it’s false.
My favorite Fiction writer strikes again!
Go jen, go!
“Also, in contrast to most of the 5.x releases, you have to recompile the kernel to get SMP support.”
That’s because FreeBSD 6.x with an SMP kernel is still much slower than FreeBSD 4.x on a single processor machine. AFAIK, DragonFlyBSD handles this performance gap in a better way, although the transition isn’t complete yet.
Therefore, I still prefer FreeBSD 4.x on UP machines.
Is this a logical argument, because it makes no sense.
“Also, in contrast to most of the 5.x releases, you have to recompile the kernel to get SMP support.”
A fact that leads to the conclusion you would use 6.x with out recompiling in smp support. As is the common case when most people don’t have smp boxes.
Adding in overhead for full smp support does cause a slow down when you don’t have a smp box, but a desktop user would never know whitout benchmarks. So please link to a little proof when your wording implies such a drastic performance loss.
Then a random, and admittently not certain, fact about another os which has nothing to do with your argument.
Then your conclusion. You perfer an os that is no longer updated and only suppported for security fixes.
You don’t want a production release on your UP machinces?
This might be a stupid question, but will SMP support boost prefoemmance of a multi core cpu?
Not quite sure, but I would think so
Yes. Without it the os would only use one core, and you would have that level of performance.
Why not recompile a UP kernel for FreeBSD 5.x or 6.x? 5.x appears slower only if you neglect to compile SMP out of the kernel. See this comparison between UP and SMP kernels on a UP machine: http://www.cons.org/tmp/smpkernel.wall.html
Having said that, FireflyBSD is a good project. Firefly’s approach is to pass messages between parts of the system to queue threads as efficiently as possible.
Rather than more message pasing and queueing, FreeBSD’s SMPng tries to improve performance by finer locking. It replaces spinlocks with sleep locks, and introduces preemptable kernel threads. The “softer” sleep locks permit CPUs to run other threads while the locking thread to release. In contrast, spinlocked CPU’s sit idle until the locking thread is able to come back.
If you’re curious, read Greg Lehey’s SMPng paper: http://www.lemis.com/grog/SMPng/Singapore/
I found the following email rather informative WRT DragonFly’s threaded messaging system and why it was chosen for DragonFly over the spinlocks of Linux, or the “softer” locks of FreeBSD:
http://leaf.dragonflybsd.org/mailarchive/users/2005-01/msg00194.htm…
>The FreeBSD operating system is finally through it’s buggy 5.x series
> and into the more reliable 6.x series
>it’s more stable and functional than it has been in the recent past.
Really? I think 6.0-RELEASE is not so stable yet.
See “FreeBSD 6.0 Open Issue.”
http://www.freebsd.org/releases/6.0R/todo.html
Possibly 4-STABLE or 5-STABLE will be more stable. (in this time)
Edited 2005-11-08 22:12
I for one have experienced 6 being considerbly more stable than 5.x
Of course this greaty depends on your hardware, I have one dual cpu machine that was hell with 5.x, I haven’t had the time to really test 6 on that system yet…
I will have to agree with the above comment. 6.0-Release is much less buggy than any of the 5.x releases I have tried.
When 5.3 came out and it was still pretty buggy I decided to jump (back) to Archlinux, but with the release of 6.0 I picked up FreeBSD again. None of the annoying bugs of 5.x I experienced were present anymore.
most of the materials i find on the web regarding this are seriously outdated. anyone have any hints or warnings?
It works fine, but I reccomand you to
1) don’t create a partition for freebsd on an extended partition as freebsd won’t find it
2) use GRUB to choose between freebsd and linux. Remember to get the right line for freebsd, and in the freebsd installer, make sure freebsd doesn’t touch the MBR (or else you have to boot a livecd to get grub back in the MBR).
I always wanted to know it FreeBSD had a better performance than Linux. A couple of weeks ago I made my small test. I used a server (lighttpd) since BSD’s are always said to be server oriented. I benchmarked it with httperf. My box is a P4-2.6Ghz-512RAM.
FreeBSD-5.4 (default install)
lighttpd-1.3.16 (using same conf file as in linux)
The same html page for both tests (index.html)
httperf –hog –server localhost –num-conns 10000
Result: 1700 requests/sec
Slackware-10.2 (kernel 2.6.13, everything else default)
lighttpd-1.3.16 (same conf file)
Same HTML page (index.html)
httperf –hog –server localhost –num-conns 10000
Result: 4500 requests/sec
Could someone reproduce a similar test and post results? I didn’t expect that big difference. And I must say I have nothing about BSD’s or claim that Linux is better. But those were my results.
“httperf –hog –server localhost –num-conns 10000 ”
–hog doesn’t make much sense when benchmarking localhost. You should run the tests in a local network.
Linux has had the equivilant to multi-threaded filesystems for years, FreeBSD is seriously behind.
As for benchmarks of FreeBSD 6 being on par or just as fast as Linux, I will believe it when I see them from an independent source.
Linux has had kernel threads for years… yes.<BR>
FreeBSD (and NetBSD too, although not the same) chose a much more elegant implementation based on scheduler activations that the linux guys decided was too much work. FreeBSD 6 has caught up with an important difference; they have the infrastructure to do much better now.
“Then a random, and admittently not certain, fact about another os which has nothing to do with your argument.”
The FreeBSD work on SMP doesn’t pay off for the average user on a UP machine. That’s the connection. On an SMP machine you have to, explicitely, enable the “main feature” of FreeBSD 6.x, namely SMP. So, perhaps, it seems just more stable on UP boxes than 5.x because SMP is disabled by default now.
“Then your conclusion. You perfer an os that is no longer updated and only suppported for security fixes.”
Because it works reliably.
“You don’t want a production release on your UP machinces?”
At least not FreeBSD >4.x
Yes, that is actually the only thing that keeps me from coming back to FreeBSD. USB2.0, at least on my hw, transfer rates were rather low when compared to Linux(IIRC ~5x-8x slower). But I must say that once I was a really happy FreeBSD user. Then came Debian and now Ubuntu which both *just work* better
Oh yeah, and one more thing that keeps me in linux world is APT.
6.0 seems to do USB2 right. I’ve been using it with an external drive, and it is working very well now.
I know I may Google, but it always turns up a bubble gum machine of random noise for this sort of thing, so I’m going to ask here:
Which is the best, security wise:
FreeBSD or OpenBSD?
Now I’m not asking which is the best for any other reason. Please don’t reply with questions, I would appreciate (and I know others would too) some opinions. Thanks!
Free and Open take security seriously by default. However, OpenBSD takes it to another level; FreeBSD does a balance between the two.
If pure security is your motivation; then go OpenBSD. On a side note; check out the mailing lists and documentation for both projects. Both have very good docs and support; however, one might scratch an itch better than another (aka your personal preference).
PS: don’t forget to make sure that the applications you want to use are available on your platform of choice. On the server side; your based should be equally covered, with maybe one or two exceptions.
PS: Hardware may or may not play a role in your choice. It depends on what you have for your system(s).
Just some things to keep in mind when making a choice.
Regards,
bsd user at large
Aside from the great points you’ve made here, I think another thing to remember is the better one groks a system, the more secure s/he can make it.
Sometimes it’s not the software that makes it secure* it’s the sysadmin and the way s/he implements her/his “solution.” (ugh, I just used a sales term!) This applies greatly to UNIX and the BSDs in particular, for the most part you don’t have a WIMP env. by default to hold your hand, it’s up to you to secure your own system.
*(though this helps a great deal, especially with stability and bug prevention measures in place when the software is written and maintained)
“Which is the best, security wise:
FreeBSD or OpenBSD? ”
This is not the place to ask.
FreeBSD versus Linux? Here is my opinion on the matter: you have two great operating systems, both are CLI on steroids enabled (unlike Solaris), both give you full control, both are blazzingly fast and stable, both are free and open. The only real difference is that both are based on completely different kernels. But if you ask me, thats the best part because I like having choices. I don’t want everything to be Linux. I don’t want everything to be BSD. FreeBSD has made Linux better and Linux has made FreeBSD better. Whenever I use Linux, I cry for certain features BSD has. But also when I use FreeBSD, I cry for the features Linux has and BSD is missing. So I ask everyone to stop this stupid bickering and give some praise for the FreeBSD team for doing such a wonderful job on the 6.0 release. It is a wonderful little OS and in the long run will only help make Linux better just as Linux has helped FreeBSD become better. Deploy both solutions and enjoy.
One of the best opinions I’ve seen posted on OSNews in a long time…
WOW! This is the best comment I have ever seen on OSNEWS and on other geek news site.
For some reason, this view is not shared by all (see the thread on Nexenta). But I couldn’t agree more. Up another point to you! Thanks for being so forthright.
You saved me from having to say it. Anyone that comes into a story like this and says that operating system X is better than operating system Y because of reason Z should have their account revoked.
Richard
I don’t know, a lot of people don’t bother with Z.
Heh…finally someone gets it. As member of the OSS community I applaud you. I myself am a Gentoo user and active Gentoo Developer. Why…not because I would advocate the use of Gentoo over any other option but because it is what works best for me, it allows me to do what I need to do in a manner that I find easy and stable (and no quips on this please). The whole point of the OSS movement is freedom of choice, the kind of freedom that can only come when everyone can beneift from the hard work of others, Linux distros, BSDs, and the myriad other hobbiest OSs out there. I say Good Job! to the FreeBSD team…and hear is to more (and better) releases to come. I raise my glass to you!
why does everyone say freebsd is better than linux for servers? if something is slower than linux, has less hardware compatibility/drivers (eg server class RAID disk controllers, network), is more difficult to maintain, .. then it must be better right?
the ports system is NOT robust. if your services depend on it then you are risking them breaking during an upgrade. if you pre-compile them elsewhere or test elsewhere why not just use package systems like rpm/deb? even the kernel can be safely upgraded for security fixes – try that with the BSD kernel.
let’s go with the myth (since no-one has yet confirmed it with numbers) that linux is less stable/breaks more often than BSD. but if you fold in the time it takes to find fixes, and then the mechanism to update the servers – you find that linux is back up much quicker. and since most dev/rpm are transactions they can be easily undone. unlike undoing recompiling kernels to an older snapshot of the source.
BSD are used in inductry purely for the licensing reasons – of that is not an issue – then the industry uses linux based systems.
OK, I sentence you to six re-readings of the “I Hate Zealots” post above. Then print it out and hang it over your workstation, please.
why does everyone say freebsd is better than linux for servers? if something is slower than linux, has less hardware compatibility/drivers (eg server class RAID disk controllers, network), is more difficult to maintain, .. then it must be better right?
I am not “everyone”, but in my experience (and I started out as a staunch Linux man), almost every time I used Linux for a mission-critical server, I was bitten sooner or later. For certain things, FreeBSD might well be a little slower, and might have less broad hardware support, BUT, for the things it does support, it does so rock-solid. FreeBSD is *conservative*. That’s the whole point. I don’t plan on running an insanely broad array of hardware devices on a system–just the ones I have tested and work to my satisfaction.
Even the supposedly “unstable” 5.x series in FreeBSD caused me less problems than the Linux 2.6 kernel series. I don’t have anything against Linux, but my experiences were simple fact. Maybe a few other people had similar experiences and that is why they use FreeBSD.
the ports system is NOT robust. if your services depend on it then you are risking them breaking during an upgrade. if you pre-compile them elsewhere or test elsewhere why not just use package systems like rpm/deb? even the kernel can be safely upgraded for security fixes – try that with the BSD kernel.
Anything can be a problem if it breaks halfway through update. Are you trying to tell me that a binary update that depends on 15 other dependencies can be guaranteed even if there is a network outage halfway through updating the libraries? Anyways, source updates are only a part of FreeBSD’s software management suite. With pkg_add, and portupgrade/portdowngrade there is no constraint to do source compiling; you can force it to download binaries every time.
let’s go with the myth (since no-one has yet confirmed it with numbers) that linux is less stable/breaks more often than BSD. but if you fold in the time it takes to find fixes, and then the mechanism to update the servers – you find that linux is back up much quicker. and since most dev/rpm are transactions they can be easily undone. unlike undoing recompiling kernels to an older snapshot of the source.
Whatever. You are responding with another myth. And anyway, all I do to roll back to my previous kernel is boot to kernel.OLD, or another named version, and then rename the kernel images. It’s that easy to roll back. I don’t pretend to know whether FreeBSD is “better” than Linux. I just know that it makes my life a little easier.
BSD are used in inductry (sic) purely for the licensing reasons – of that is not an issue – then the industry uses linux based systems.
Thank you for speaking for me, who happens to be the decision-maker on OS platform for my employers. We chose FreeBSD for the reasons I described above and more. Even though we did suffer a small (small) performance decrease in some areas. The trade-off was well worth it, with hundreds of machines deployed in clients’ offices that never cause us to break a sweat (in fact, I would estimate that supporting those machines remotely doesn’t even take up 1% of my time). Yes, the license did help, but was nowhere near the deciding factor.
And interestingly, it looks like our stuff will run faster on FreeBSD 6, so we might not even have a speed disadvantage re: Linux anymore. I’m NOT saying it will be that way for everyone or every application. I’m sure there will be people for whom Linux is faster or is the better choice for some other reason. So why can’t we all just…get along? Why the need for all the animosity?
Are you trying to tell me that a binary update that depends on 15 other dependencies can be guaranteed even if there is a network outage halfway through updating the libraries?
Yes. In every sane package manager, all needed packages are downloaded at first and then upgraded.
If you don’t know this, I doubt you have ever used linux for something serious (as you are trying to convince us)…
Yes. In every sane package manager, all needed packages are downloaded at first and then upgraded.
Yes, you are right, and that was the one part of my comment I didn’t stop to think about. The problem still exists, though, if there is a system failure in the middle of the post-download upgrade. Yes, FreeBSD ports doesn’t *by default* download everything before upgrading all at once, but it is quite easy to do that if you want. Also, with portupgrade, you can use the –backup-packages directive to roll back to any point within the upgrade chain. We could go back and forth on this argument, but really the pros/cons of each are within a small range when you get down to it.
If you don’t know this, I doubt you have ever used linux for something serious (as you are trying to convince us)…
I could say the same about your FreeBSD knowlege. Of course I have used FreeBSD more for my serious stuff, since every time I was *starting* to get into something serious with Linux, I would run into certain showstoppers that took longer to solve on Linux than on FreeBSD. Note: I’m NOT saying they couldn’t be solved, nor am I saying that it is this way for everyone, but just that for the things I was working on, which usually involve a lot of heavy PostgreSQL stuff, as well as PostGIS, R statistical language, or other database-related things, FreeBSD either didn’t have a problem, or let me deal with the problem much faster, and get on with designing my systems. So I gladly traded a bit of bleeding edge performance for stability and predictability.
For an example of the kinds of problems Postgres people were having with kernel 2.6, just check out http://www.powerpostgresql.com/Downloads/terabytes_osc2005.pdf (page 6).
I was only replying to you comment about binary upgrades in linux. Notice that I have neither compared FreeBSD ports to linux binary packages, nor stated anything about my FreeBSD knowledge or experience…
For me, both Linux and FreeBSD are excelent systems and I only wonder why there are so many people (I don’t mean you) that need to convince everyone around that their favourite system is the best and the other system is piece of shit (poo
.
One thing that BSD is still lacking is a proper audio framework – ALSA on Linux comes to mind. I would like to run software that records and plays back audio at low-latency, but that is not an option on any of the BSDs, just as it isn’t on Solaris.
For this, Linux is the only option appart from the two commercial and mainstream biggies. If anybody know of a solution to this, please let me know.
And yes, I do know that neither BSD nor Solaris was originally intended as desktop operating systems for multimedia use.
This this guy against FreeBSD or what?
It’s not even a reivew!
He didnt justify any of his tests. He didnt even say what Machines he is testing on.
I think he is just giving false info.
I like the new icon
I know we’re not on Slashdot, but I ask everyone to print out the parent post, put it in a nice golden frame and hang it above your desk.
PRAY TELL BROTHER
[++++] insightful
P.S. dammit this was supposed to be an aswer to this thread…
http://www.osnews.com/read_thread.php?news_id=12583&comment_id=5809…
Edited 2005-11-09 02:57
too bad i can’t tell what the parent is because osnews doesn’t make threads obvious.
The benchmarks that they talk about, where done by Matt Olander(offmyserver.com), they’ve done these benchmarks in order to test scalability/performance on their SMP servers for various customers, of course among them Yahoo. These benchmarks should be posted fairly soon, and they were carried out without any tweaks/optimizations on either Linux or FreeBSD. It seems Yahoo has been interested in abandoning the MySQL servers running RHEL 4 due to guess what, stability problems. They’ve just stopped using some of their Solaris/Oracle systems in favor of RHEL4/MySQL. And they didn’t consider FreeBSD 5.X for databse usage series because of it’s publicized instability and it’s sluggishness running MySQL, although their web and mail servers run entirely on FreeBSD. It seems that with the release of MySQL 5.X, some people at Yahoo decided to optimize MySQL for the new threading model on FreeBSD in order to move away from the idea of adopting RHEL4 widely, since then whole situation has changed, and Yahoo is now interested on running all of their servers on FreeBSD systems, with people inside Yahoo claiming that it’s not only more stable but also quite faster. This would end the Solaris/Oracle era and stopping the RHEL/MySQL era prematurely.
Since it’s hard to benchmark stability, I guess we could refer ourselves to netcraft to see how many of the top 50 servers are running what OS, and you’ll see that among the top 25 there is one form or another of a BSD system(some are even BSDOS). Evenmore, not only Yahoo uses FreeBSD widely, giants such as Pair Networks and Verio base most of their hosting on FreeBSD, enterprise friendliness has nothing to do about it,and let’s not forget that router giant Juniper Networks based their JunOS entirely on the FreeBSD kernel, I can’t recall any such example on the linux camp, maybe Linksys with their crappy AP’s but that’s all.
On the side of SMP support, due to the nature of the interrupt driven model adopted by Linux, which is a 1:1 implementation, performance on UP or SMP systems is not heavily affected. On the other hand the M:N implementation for SMP on FreeBSD impacts negatively on UP systems when running a SMP kernel. That’s why it’s turned off by default. Performance-wise FreeBSD 6.0 is very close to FreeBSD4.X on UP systems, not quite there yet but very close. Due to some locking still present in several subsystems.
I think It’s nothing to do with the juniper using the freebsd for their router for freebsd’s perfomance,and it can’t prove the freebsd has superiority.Because the cisco’s cws-1’s router much more power than the biggest juniper’s freebsd router but it’ os is only a micro kernel os writting by cisco itself.
We all know the micro kernel os’s performance is much more worse than the monomothic kernel.so cisco using the os proved that the freebsd in the juniper only because it didn’t want to reinvent the wheel.
we all know the freebsd in the juniper only control the route list,it’s been choosed not for it tcp/ip stack,but it’s free as peer and juniper don’t want to release the source like GPL’s license.nothing else not to choose linux
Since it’s hard to benchmark stability, I guess we could refer ourselves to netcraft to see how many of the top 50 servers are running what OS
I can see that this is VERY popular argument among FreeBSD users. So again, read http://uptime.netcraft.com/up/accuracy.html#whichos
FreeBSD is among few OSes that report uptime remotely. Linux, Solaris, AIX, HP-UX etc. does not (or cycle back to zero after 497 days).
Top 9 uptimes are all FreeBSD, followed by BSD/OS in 10th place.
http://uptime.netcraft.com/up/today/top.avg.html
I never liked the FreeBSD 5.x because it appeared to be a departure from their long-running reputation for the most robust free OS.
I’m running Debian 3.1 and FreeBSD 4.11. I don’t have any plans of ever installing FreeBSD 5.x but I’m very interested in trying 6.x after a couple more minor releases are out.
Why nobody reads the FAQ of those uptimes? Linux, and some versions of FreeBS are NOT included for technical reasons !
http://uptime.netcraft.com/up/accuracy.html#whichos
does it finally support USB2 to a decent level?
i like the freebsd system, however, the hardware support honestly sucks
if it doesnt support a common piece of hardware such as that mass storage, i dont know what to say.
netbsd doesnt either which was annoying becuase i want to use a well engineered bsd system and grew tired of the linux distros. but it seriously cant support some rather common hardware.
FreeBSD 6.0 runs exceptually good on my amd64 box.Neither did i have any serious problems with the 5.x series other than it didn’t flawlessly support x86_64.
KDE is to my (perceived) experience on FreeBSD 6.0 snappier than on SuSE 10.
Once properly configured FreeBSD 6.0 can match the experience you get with any of the mainstream linuxes. It goes without saying that things may vary on your mileage.
kudos to the devs and crew
The crucial fact for my upgrade to FreeBSD 6.0 is whether my WLAN card is going to work. Is the Project Evil wrapper expected to work on amd64 release with a 32bit WinXP binary driver?
Is a precompiled package of OpenOffice.org (preferably 2.0) available for the amd64 release? I remember at 5.3 I had to enable Linux binary compatibility and to install i386 Linux version…
You want to run a 64-bit kernel using a 32-bit driver? And you can’t see why that will be a problem?
Either run the 32-bit version of FreeBSD using the 32-bit driver, or wait for the 64-bit driver to be released and run a 64-bit kernel.
You can mix-and-match the bit-ness for applications running on top of the kernel. But everything inside the kernel must be the same bit-ness.
There is some way to get sun java 1.5 to work in Frrebsd ?
I can’t remember the link, but if you check Google for FreeBSD and java you should get some links that will tell you how to compile it. I’ve done it for 1.4 some months ago, the patches needed for 1.5 were available, just not complete and the recommendation was to use 1.4.
Hummm… cd /usr/ports/java/jdk15 && make install clean?? xDD
only days after its release, the reviewer claims v6 is more reliable. oh please.
what a f–king joke.
gimme win2000 or win2003 any day.
Oh well, I can see that trying to have a discussion here based on facts as opposed to wishful thinking is a no go. My bad.
You cannot claim superior quality over 5.x only one week running. I never had any major issues with 5.3. Howerver, 6.0 doesn’t boot. md5sums match, burned at slow speed, several workstations: boot errors. So?
If you are already running 5.3, why are you downloading CDs? I have never understood .ISOs when it came to upgrading FreeBSD; there are more elegant ways of doing it, in less time. CVSuping the source tree, for example, or doing a binary upgrade with /stand/sysinstall would download only what needs to be upgraded, not 650MB worth of data that you may already have installed. Again, this is just me; I had assumed ISOs were for Linux/Windows users who were accustomed to software coming on CDs.
Again, this is just me; I had assumed ISOs were for Linux/Windows users who were accustomed to software coming on CDs.
For me it comes on CD’s to install it on other PC’s that don’t have a base install yet. The rest is handled through APT and (sometimes) Synaptic.
At least for myself, I transitioned to 6 as soon as it was 6.0-CURRENT, that and through the BETAs all worked better for me than experiences with 5.
The way I felt about it was, each new release of 5 made the previous 5.x seem unusable and for me 6 has just continued that trend.
My two students installed the 6.0,they all crashed for some reasons.can you say it more stable than 5.x only released so short days?
I guess so, since 1) they tested it (I haven’t yet) and 2) 6.0 contains many fixes to bugs in 5.x
…they could get it working properly with USB keyboards and mice. I got tired of unplugging my mouse and keyboard and replugging it at every boot. I love the ports system, but until they solve the USB problem it is unusable to me. Please don’t tell me to buy a ps2 keyboard and mouse…or recompile the kernel. It is just not worth that much trouble….and yes I have read the handbook through at least 3 times. Most things about it I like, and the documentation is the best for any system, but the USB support is spotty.
USB works fine. The only issue is it doesn’t support USB keyboards at the boot prompt. But after the kernel loads, I have USB keyboard and mouse, and both work perfectly.
Not having USB working at the boot prompt is a big issue. I have Dell computers that have no PS/2 ports. USB is my only port for keyboards, mice, etc.
So basically, that Dell wouldn’t be a good computer to run FreeBSD 6.0? That sucks, because it’s a really nice machine. BTX design, hyper-threading (although some say that’s bad), and other goodies.
As much as I loved FreeBSD in the past, they really need to just fix there stuff and make sure that everything works (i.e. take a note from the OpenBSD people).
Depends on the BIOS. Does the BIOS support Legacy Keyboards, or PS/2 emulation of USE ports, or however the BIOS maker wants to call that feature?
If it does, then you can run FreeBSD on there without any issues. When it boots, the BIOS will emulate a PS/2 keyboard port, the installer/boot loader will see and use a PS/2 keyboard, and once the kernel loads, it will switch to using the USB keyboard directly.
If the BIOS does not support any kind of Legacy USB support or PS/2 emulation, then you’re out of luck.
Why don’t you enable legacy keyboard support in your bios like everybody else who has this issue? Thanks for flying FreeBSD. Google is your friend.
I have already done this. I keep this option on by default at all times. I know my hardware quite well. I have an AMD 64 system, so I don’t know if this problem exists with i386.
is FreeBSD’s GCC patched for buffer over flow attacks and similiar attacks?is there anything code auditing like in OpenBSD?can we say that FreeBSD is as secure as like OpenBSD?if it is not,then why?why the developers don’t think about security too?
I’ve been in love with freebsd from back in the 3.x days. But, the 5.x branch gave me a foul taste in my mouth due to its lack of stability. I have had servers crash randomly and I’ve even gone to the extreme of changing all the hardware in a system (including cpu, mobo, and psu) and the only thing left unchanged was an image of the OS, and I still got random crashes that I couldn’t trace.
That was the last straw with 5.x and I rolled back about 5 servers to 4.11, which is where I’m at now. Hopefully, 6.0 will restore FreeBSD’s legendary stability and uptime. I’m off downloading…
until they have the most important apps I use in ports.
Just one example (out of about ~6), I replaced mod_fastcgi with mod_fcgid on all my Debian Apache2 servers because it is more reliable. mod_fastcgi has memory corruption issues on Apache2 (but Apache 1.3 is OK).
Latest mod_fcgid is 1.06 ( Apr 27th 2005 ).
Latest in FreeBSD ports is 0.80 and the description still states “The bad news is that it does not currently support suEXEC,[…]”
There are many other server-oriented apps that I need, and are more currently up-to-date in Debian Stable than FreeBSD.
Also, FreeBSD needs to provide 100% support for EXT3 filesystem (one that doesn’t cause problems when mounted read/write). EXT3 support is necessary for VPS hosting companies to offer FreeBSD with web-based virtual partition management (create/delete/resize virtual partitions using web browser).
References:
mod_fcgid – http://fastcgi.coremail.cn
mod_fcgid port – http://www.freshports.org/www/mod_fcgid/