Slashdot Mirror
Why FreeBSD
An anonymous reader writes "The FreeBSD operating system is the unknown giant among free operating systems. Starting out from the 386BSD project, it is an extremely fast UNIX-like operating system mostly for the Intel chip and its clones. In many ways, FreeBSD has always been the operating system that GNU/Linux-based operating systems should have been. It runs on out-of-date Intel machines and 64-bit AMD chips, and it serves terabytes of files a day on some of the largest file servers on earth."
Simple, choice is good. As muck as I like Linux, I'm glad to see that there are viable, open alternative OS's.
I think I think, therefore I think I am.
Don't ask why, ask why not.
"FreeBSD has always been the operating system that GNU/Linux-based operating systems should have been."
Can it get anymore flaimbaitish than this. Ironicaly enought it comes from I.B.M developer works.
P.S: Opinions are like assholes, everyone has one.
Exactly how is this news?
;)
I've know that FreeBSD was much better than Linux for ages
Joking aside, FreeBSD is a bit hard to install and get working if you're using it as a workstation OS...
I've been using it for 4 years now and it still took most of my free time in a period of 2 weeks to get it installed properly on my newly bought laptop (with all the details and little stuff, that is)
Of course when I was done, it was very much worth it. I don't think any system is as robust and stable as FreeBSD.
A huge "Thank You" to the developers!
Let me make an analogy to a newspaper. Not everything you get in a newspaper is "news". For instance, there's the "Life" section of USA Today or the "Living" or "Features" section of a typical local daily newspaper, which typically runs plenty of articles other than news. Even the "Business" section (called "Money" in USA Today) usually has some articles other than news.
freebsd is dying as is macos x and windows. The future is Linux. The future is Free.
FreeBSD is free'er than Linux, or more accurately the BSD license is free'er than the GPL. That said, the less free GPL's restrictions are meant to be benevolent for certain users.
Mac OS X's share is growing wildly. For some it is replacing Linux as their general purpose unix. Now some people have more specialized needs and Linux may be a better choice but many folks using Linux just need a general purpose unix box and are not into the politics and Mac OS X combines unix, a consumer GUI, FOS software, and off-the-shelf retail software very nicely.
One could also argue that Linux is what FreeBSD should have been, and cite the huge number of supercomputers using Linux, or the success of Linux on the mainframe. However, it would be nice if the poster realized that it's a pissing contest and both operating systems are impressive and have their uses, benefits, and drawbacks. Neither is what one "should have been". They both have their own, very different methodologies, so let's leave it at that.
Not that it's news anyways...
Systemd: the PulseAudio of init systems
The differences between Linux and BSD are minor; anybody who thinks they matter needs to have their head examined, and the kind of (implicit) Linux bashing represented by the article is pointless.
R.
I've admin'd most every flavor of Unix at some point in my life and I really really like how FreeBSD is managed, from development to the ports tree.
Now that there is a push to support binary updates, my last major complaint has been addressed.
Anyone who has ever been stuck in the perl dependancy hell will absolutely love the ports tree - I really don't understand why there hasn't been more adoption of that concept in Linux.
Also, I am suprised that Linux is the platform of choice for all of these appliances that companies are pumping out, like wireless routers, security devices, etc, when the BSD license is so much more attractive to business.
The major stumbling block that FreeBSD has left is their development team. It seems like the way things are organized really creates a lot of opportunity for personality clashes.
Jerry
http://www.cyvin.org/
Starting out from the 386BSD project, it is an extremely fast UNIX-like operating system mostly for the Intel chip and its clones.
This sounds like FreeBSD performs vastly better than any OS in the world. And how much faster is exteremly compared to Linux or Windows? Twice the speed? Four times?
Neither FreeBSD, nor OpenBSD scale as well on large SMP systems as Linux. Period. OpenBSD may have more security features and FreeBSD may have its own strong point, but scalability sure as heck isn't one of them.
Can we please refrain from commentary that compares interesting-if-pointless slashdot postings to a regime responsible for events of genocide and blood-letting on an unimaginable scale? While I agree that the article on IBM's website is very rah-rah-rah, I think it hardly compares to a speech by Hitler or Goebbels.
I mean, this is slashdot, after all. Let's try not to take it TOO seriously...
I work at a large internet organization that runs thousands of FreeBSD systems. When we need 64-bit though, we switch to Linux because it has a stable 64-bit distribution and FreeBSD does not. I've gone through all the kudo's about FreeBSD being stable, but are you using release 5? and are you using 64-bit? (and don't even get me started about threading support.)
Look, it's a giant troll posting thinly disguised as a news article!
BSD is great, but it's not the only game in town. Suggesting that it is what Linux should have been is nothing more than troll bait.
If you're looking for something close to the style of FreeBSD, but with the new and freshness of Linux, try Gentoo.
Great idea, sir! Spend 6+ hours compiling something. Hooray!
Then again, you could have just installed FreeBSD and saved yourself 5.5 hours.
Free of Flash! Free of Flash!
*pats drsquare on the head*
Yeah. Obviously all linux programmers and moreover linux users are morons. In contrast, the freebsd guys are as gods, and if they can't figure out a problem, they turn to you, the most brilliant entity on the planet.
No shut down your computer and take your medicine, then you're off to bed, it's almost quarter past six.
there was some argument that Microsoft should have stayed with a windowing manager on top of core OS paradigm as they previously had and beefed up DOS to be something like Unix. Fortunately, saner minds prevailed.
No, they didn't. That's exactly what Windows is; you can boot NT into a command line environment and run it completely without a GUI. It would have been insane if they hadn't done that.
Use of a text interface and system fiddling is inevitable. Not so with Windows.
That's total bullshit. There are millions of Linux systems that don't even have a command line and do everything graphically (e.g., Linksys routers). And desktop Linux installations usually come with a full complement of graphical administration tools.
On the other hand, a lot of Windows system management involves going to the command line. It's just that many people give up at that point and just reinstall Windows.
Linux only scales well on said 'large SMP systems' where there has been a tremendous amount of hand-holding by the vendors of said hardware.
I'm having good experiences running NetBSD on a quad CPU server here, but you didn't mention NetBSD...
And I know that isn't necessarily a 'large SMP system.'
Frankly, who *cares* what proprietary vendors are able to twist Linux into doing on their specific hardware? They could do the same thing with any OS they focused on.
You must be out of your mind: the Windows administration GUI is a usability nightmare. If it were easy to use, people wouldn't have to study so much to become Microsoft certified engineers.
...not so fast there sparky, a common misconception.
You are right in thinking that the true "guts" of the kernel is mach, however, it's only really used for the very very low level stuff and message passing, the rest of the system is provided by a BSD server for mach that takes care of 90% of the system duties. What apple have created is a bit of a bastard child of a microkernel and a monolithic kernel.
I am NaN
Yes but there is the licenses issue. BSD style licenses vs the GPL. At least for companies to use the OS with there products.
Just as a factual matter, Linux and GPL software have recieved about 10^6 more corporate support than BSD-licenced software in recent years. The GPL has proven to be a very corporate-friendly license because it allows copyright holders to share their code without giving away the 'exploitation rights'.
Plus, I think you could argue that the big exception (Apple), was driven more by technical reasons than licensing ones. They started with an 1980s BSD-based OS, so FreeBSD code was a better fit. If OS X was a clean slate, who knows?
Whenever I hear the word 'Innovation', I reach for my pistol.
Wow I gotta stop believing all that Linux FUD I've been reading. So Linux can't, hasn't or done any of that? Talk about tooting your own horn at the expense of something else.
My karma is not a Chameleon.
I see no need to pit them against each other, as they both provide freedom and excellence to the user.
FreeBSD has a bit of an identity crisis, they sorta see themselves as "Linux Junior", with a chip on their shoulders. Which is why every single pro-BSD article is basically a comparison to Linux.
If you look at how Linux has been positioned and marketed, they've never felt the need to "eat their own" and convert FreeBSD users. At least not in the last 10 years.
Linux has always been positioned for "world domination" -- first they convinced UNIX/RISC systems to convert, then scientific systems, then embedded systems, and now they are working on Windows systems. Whether it's Java/J2EE, or Oracle, or ERP, or StarOffice, Linux finds a way into a market. These are all new customers.
Meanwhile FreeBSDers sit back and look at the load-average on their sendmail servers and then wonder why the world isn't knocking on their door . Rather than define themselves somehow, they respond by nipping at Linux's heels.
Whenever I hear the word 'Innovation', I reach for my pistol.
Why isn't BSD as well known as Linux if it's supposedly so much better? The answer's pretty obvious to me...
"A truly wise man realizes he knows nothing."
"2 weeks to get installed" can you tell me specifically what advantages you have in a laptop that justify 2 weeks of work ?
even assuming zero maintenance for bsd (ha !) and 30 mins every 2 week for windows (windows update + norton) it is hard to see the advantage, particularly given the wealth of windows apps and ease of communicating with others
The wealth that is in FreeBSD and other UNIX-like OS' is the userland tools and scripting. If you learn the UNIX way, you might find that there are lots of things that you can do really quickly if you are creative enough. Windows is great for certain tasks, but the UNIX-like OS' give you power to apply your creativity seemingly without bounds.
I've just come out of a contract where I was competing with other contractors who had been working on this large project for years. I came in 2 years after it and they started and in a matter of *DAYS* I quickly rose to the top and my methods were chosen to take the lead because I had provided in days many points off a list which other contractors had been trying to provide or claimed were impossible or infeasible. It all came down to the fact that these contractors were using Windows and off-the-shelf applications for this particular industry. I on the other hand was using UNIX-like OS' in addition to Windows and some off-the-shelf applications (using some amount of Windows was unavoidable due to the nature of the project). Through UNIX scripting I very quickly delivered the final chosen solution (in months), what the other contractors had failed to deliver in years.
The problem with off-the-shelf is that you are bound to the constraints of someone elses generic application, whereas employing scripting of the hundreds of text based UNIX tools and applications gave me the flexibility to extract what the customer needed. Gluing some parts together with some C code was also very simple where I could not see a scripting method, because the "glue" could be so small, since the userland apps covered almost all requirements. What's more, the data flowed through processes gracefully until completion. As opposed to the Windows way which attempted to open whole data files at once and often multiple times, causing system crashes on machines which had to have far great memory and CPU resources.
The irony was that I was beating optimized binaries with scripting! All because of the efficiency of pipes. I could avoid having the same or derived data opened multiple times and being held open. Disk, memory and CPU did not have to be thrashed. It just flowed efficiently. Data was getting read once, processed and then results written.
When a project goes on for months or years, system setup of 2 weeks is an investment.
PS, it took me about 2 weeks to set up Windows XP on my laptop. This includes removing all the crap that comes with them nowdays, patching, hardening, installing apps, patching, testing, imaging and putting together a quick full restore DVD of my efforts. Next... the same for FreeBSD on another partition...
When the other guy talks of 2 week installs, I know what he means. Installing an OS is not just a matter of inserting a CD and answering a few superficial questions about yourself, when the task at hand is to get some real work done. I also realise that an initial install and on-going maintenance are two very different beasts.
Here is the kicker though... it takes me about 5 to 15 minutes to install any of OpenBSD, NetBSD or FreeBSD, including X. Then perhaps another 15 minutes to install application packages and a *COPY* of the configuration files for my favourite Windowing environment. I can have a BSD up and ready to go in half an hour. A system which I feel has potential only limited by my creativity to problem solving. What's more, I know at least with OpenBSD that I can package my additions and changes into a tarball for inclusion at install time. Taking future installs of OpenBSD back to about 5 minutes. FIVE MINUTES.
I don't know how easy this is to do with NetBSD or FreeBSD, as I've never tried.
BSD communities have a reputation of being condescending and elitist. This is mostly undeservedt
Through most of the 90s, the BSDers prided themselves with their elitism and their old school Unix cliquishness. I read the lists -- "If you aren't smart enough to figure it out, maybe you should go run Linux, haha." was a common sentiment.
Then, in about 1998, they realized that Linux was being deployed 100:1 over BSD and they had a collective "Oh shit!" moment. They then started a positive effort to be more open and helpful to new and less clueful users, but at that point it may have been too late.
I still read comments like "Well, if you aren't reading such-n-such mailing lists, you shouldn't be running FreeBSD 5", which implies that BSDers still see their world as being pretty small. Consider how many support outlets there are for Linux, and that only a miniscule portion of users read the development lists or come anywhere close to flamethrower-range of the main developers.
Whenever I hear the word 'Innovation', I reach for my pistol.
If you want to seriously compare two open-source Unix-like systems, the only instrinsic difference is the kernel. Arguing that one system is better because of the default configuration of network services, the package system, the organization of the rc scripts, and so on, is a red herring, because there is no reason you can't take all of the userspace from one system and run it on top of the kernel from the other -- and there are projects which do this.
In that light, these benchmarks are the most enlightening comparison I have seen to date. Some BSD users have attacked the methodology, but none of them has gone on to do alternative tests of their own, and the author has been very conscientious about addressing some of the criticism. The bottom line is that FreeBSD is, whichever version you choose, at best equal to Linux in low-level kernel performance, and usually slower.
When you also take into account the greater ease of use of most common Linux distributions, broader hardware support, greater availability of commercial software (yes, you may be able to run it under FreeBSD's Linux emulation layer, but the vendor is unlikely to officially support that, which matters to large corporations), and better scalability, it really isn't suprising that most people considering a free Unix-like operating system choose some distribution of Linux.
Undoubtedly for a long time, perhaps until the 2.4 kernel came out, FreeBSD probably was superior, and had a well-deserved reputation as a better choice for serious usage. For some purposes (there are some routing benchmarks that FreeBSD people always bring up, which I can't find right now) it may still be. But through some combination of the AT&T lawsuit, media coverage, and pure chance (licensing may also have played a part), the commercial support and developer mindshare swung decisively to the Linux kernel, and today it is clearly the best choice for most uses. We can wonder what would have happened if FreeBSD had won out instead -- the resulting kernel might very well be better than either Linux or FreeBSD is today -- but that doesn't change the facts about which is the better choice today.
"(Man) tries to live his own life as if he were telling a story. But you have to choose: live or tell." --Sartre
Are you kidding me? So *BSD is made of PHD holding engineers who are absolutely the most mature people ever. While Linux developers are blathering idiots who wouldn't know '==' from '='?
In comparison, yes.
Linux community = larger, therefore we have more "fanboys". With such a broad culture of people you're bound to run into a few you don't particularly like. I have been using Linux for 5 years. I'm not insanely anti-Microsoft, no I don't like their business but I don't go around telling everyone they are idiots for using their stuff. I really don't care what people use. I use my computers to get things done, I can do all those things in Linux so I choose to.
Yeah, so what? I like a small community as opposed to a large one, mainly because it doesn't attract band-wagon jumpers-on, who then naturally become part of the "It's the greatest thing since sliced bread!" crowd.
So I guess I kinda break up your statistic there huh? Probably the other thousands of people who just use Linux to get things done might as well.
Not really. You fit into them rather nicely.
You act like people just walked into developing Linux. "Jee you know what, I have no clue about anything computers but I think I'll make a kernel today!".
Um... That was almost the approach Linus took. He didn't really study existing kernels or anything. He just went for it. In fact, it started out as a terminal emulator and evolved from there. Not exactly screaming great engineering.
Post when you actually have something other then anecdotal evidence to try and make you look intelligent. Oh wait this is Slashdot, thats asking too much.
Highly suggest you do the same. As you said, that is far too much to ask (okay, I added the "far").
Free of Flash! Free of Flash!
"Each problem that I solved became a rule, which served afterwards to solve other problems." -- Rene Descartes.
/etc/sysconfig from RedHat systems. The equivalent on Debian is /etc/defaults, which is mentioned nowhere in the manual page, or *any* manual page, and Debian takes more care with their manual pages than most distributions. Why should I have to figure this out? This is one trivial example, but these kind of glitches in integration multiplied many-fold engender a degree of inconfidence in the system. When I read something in a *BSD manual page, I can usually go to the bank with it. OpenBSD takes great pride in their manual pages. I liken this to confidence with a hammer. If you know you can trust your hands and the tool, you can swing hard and drive your nails in. If you think you've got a slightly bent nail, a slightly rotted board, or shakey hands, you have to tap the screw in. Linux always makes me tap nails, and *BSD lets me drive them in like a pro.
This, in my opinion, is what *BSD has over Linux. The *BSD's have more stable API's, better documentation, and well-established procedures for system management, etc. Some of the quirks have to do with not invalidating old-knowledge, not breaking things that work, not causing radical upheavals. Linux, of course, has no problem doing these things on a routine basis.
I have run Debian. The boot(7) web-page mentions that
Security is a big point for OpenBSD, but I would consider the documentation and the ease of administration as being bigger points. OpenBSD has a very minimalist approach, which translates to being very simple to learn and run. It also has the policy that any commit to CVS that changes the user-visible behaviour of any part of the base system must also include an update to the man pages. FreeBSD doesn't do this - something that stung me when they changed the interface to Project Evil between 5.3 and 5.4.
I am TheRaven on Soylent News
Its certainly not the clear-cut decision it was two years ago though, when I would have said Free on x86 and Net on anything else.
I am TheRaven on Soylent News
I think the answer is that the GPL code has benefited from the "fair play" users of the code for so long that it has improved a lot and even become the defacto standard. So these companies prefer to use this code and still screw the GPL license rather than legally use BSD-style code. It just shows the success of the GPL license - companies like Cisco, Linksys etc just have to be educated.
What is the sound of one hand clapping?
cat
I am not going to get into which OS "is better" because actual performance is not the issue here. If I had to rate what I saw, FreeBSD (4.1x) worked okay for the hardware it was put on, although it probably would have worked better on a "stock install" than the kludged clusterfuck that we deal with now.
The background is this: a few years ago, the small company I worked for had two admins who were FreeBSD fanatics. They pressured the IT department to use FreeBSD because it was free, their Windows infrastructure was taxed, and they had just bought a whole lot of new hardware. The pressured FreeBSD over Redhat, and made an impressive demo. So the company started going to FreeBSD. The admins, who had impressive mod skills, "tuned and tweaked" FreeBSD to work under the specific loads of the various server functions.
This would have been a good situation to be in, but then one of them got lazy, and updates got further and further behind. The other quit. The lazy one got fired. The other admins didn't know FreeBSD and barely knew Linux. Both of them eventually quit, too. I don't blame FreeBSD for the personnel problems, but this is leading to the main problem.
The company searched for someone with FreeBSD experience. The few people they found were not the kind of people they were looking for (inexperienced, would not pass clearance, had poor work records), and now they were stuck with a rapidly aging system that wasn't supported by anyone who had a clue. The new admins they hired tried to match the previous admin's skills, but were spending so much time diagnosing crashes, they didn't have time to learn new FreeBSD skills via online sources, which are sparse, confused, unorganized, and unsupportive (don't flame me on this, because this is pretty much the opinion of the whole company). And finding corporate-level supported software and hardware to run on FreeBSD was next to impossible ("We don't support FreeBSD for our fiber channel cards," says a SAN company desperate for our business, "but we hear some guy in the Netherlands had a flaky beta driver that can see things as long as the partitions are less than 256 GB." then the Sourceforge project hasn't been updated since 2002, doesn't work on our kernel version, and the guy's website is 404...)
So they decided to go with Redhat Linux. It just works. It worked faster than FreeBSD. It had an easy-to understand packaging and script-driven administration system, corporate support, and better yet: they could find LOTS people skilled in Redhat Linux in resumes. I was a particular gem because when the hired me I was an RHCT and had experience with OpenBSD and FreeBSD experience to boot. My first project was "Get us off FreeBSD!!!" by direct order. Yes, you could argue this is not a FreeBSD issue at all, but some management of people issue, and you would be right, and that is my exact point.
If FreeBSD had a sensible corporate base, a well-thought out directory structure (I have boot scripts in /etc and /usr/local/etc... and have you ever had to diagnose which one broke?), better hardware/software vendor support, and a huge skills base, maybe with some certs... THEN we will see true competition with Linux in the corporate sector. Redhat is the type of company businesses want. They understand the support language Redhat speaks. And maybe I'll see stats that the Redhat kernel is bloated, runs 20% slower the what FreeBSD does on Apache pulls, or some fanatic going on about, "Oh yeah? What about PORTS, dumbass???" But you know what? If FreeBSD wants to be taken out of the hobbyist corner and shine in the corporate arena... it's got a lot of marketing work to do.
The Ports system is far superior to the rpm system. It actually tracks dependancies, and has a system to grab them for you. You are way off base on that statement.
It is not superior to RPM, nor DEB for that matter. Anyone who has built or managed a large number of servers with disperate services on them should be able to recognise why.
It checks for required dependancies and installs them as required, something even CPAN can do, and it can even (usually) uninstall them (with varing levels of success), that is more or less it.
What it does not do is check for conflicts between libraries or the dependancies of other applications (meaning it's possible to fuck up one application, by installing another because it may overwrite an existing, older, installed library).
Systems like 'ports' do not verify package integrity, nor do they it support using a previons minor revion of the same application (often a requirement when the 'latest' version of whatever application or library your using breaks a feature you've been relying on, or is simply not a release you've had time to test in your test environment).
An even bigger problem related to the reliance of ports system, one of the most time consuming, is the process of upgrades. Upgrading between newer versions of FreeBSD is a mine field, awash with the potential for screwups, because there is no system in place to handle this task elegantly. Upgrading between disperate versions of the same branch (e.g. 4.6 to 4.11) will often cause serious problems you'll have to sort out manually at the console, upgrading from systems that are not in the same release (from 3.x or 4.x to 5.x) will often take up a good chunk of an afternoon to sort out the resulting mess. Upgrading a DEB or RPM system which is fully packaged managed (kernel and all), even between quite disperate releases is far more straight forward (more along the lines of 'apt-get update; apt-get dist-upgrade', or in the case of a Red Hat system, pop in the CD and click the 'Upgrade' button).
Another feature that makes more feature rich package management systems more desirable is the reduced risk. RPM is staggeringly powerful with regard to the sorts of operations it is able to carry out with packages (pretty much every feature you could ever want) and it is also trivial to build packages for - as is also true of DEB. This massively reduces the propensity for mistakes - sure you can write individual custom bash install scripts for packages in ports, but that is not a robust approach when you think about how many packages your likley to use.
There are two practical reasons why people typically dislike RPM:
1) It's refused to do what they wanted because a package was trying to be installed did not have it's appropriate dependancies also installed, or it conflicted in some way.
In these circumstances most users opt to 'force' it to install because they are frustrated at installing dependancies and sub-dependancies, and then they wonder why their system does not work as they expected, and so declare the RPM system to be at fault.
Arguably, it is at least partly at fault, but it was only being accurate by alerting them to problems that they would not otherwise have known about, which is pretty hard to critisize.
2) This is compounded by the fact that they are usually exposed to it in the form of 'Red Hat', or a similar distrobution, which does NOT feature something like 'apt-get'. 'apt-get' is in fact package management system agnostic, at least technically, and their have been RPM based commercial distrobutions which have shipped with it. It's absolutly more associated with Debian, and I think it's a mistake for Red Hat and other vendors not to include it in their RPM based distros because of the frequency of the scenario above, I suspect the cost of maintaining apt repositories (man hours, infrastructure and running costs such as bandwith) are key reasons.
Disclaimer:
I use ports on an almost daily basis, along with De
Well, the X-windowing system was implemented on unix in the mid 80's at MIT. Than came Stevie and Billy boy and "innovated" it on their boxes harvesting the fame. (They did it actually legally because Xerox Lab in Palo Alto had old fashioned non-profit, "infornmation want to be free" attitude - that's the reason why they never ever got rich, not even as famous as they should be) /. mob do not even check who were the ancestors.
:) So there's a clear advantage of diversity. I hope there will be no AUTHORYTY to decide which Linux is kosher or not. In that case I will go with an apocryph distro >:->
Linux of course borrowed from the unix source and of course the Linux fanatix
Isaac Newton said that he could see far ahead while standing on the shoulders of giants.
It seems to be so that more and more Linux fans do not look back anymore, so hear the message of the day: UNIXES DO HAVE X WINDOWING SYSTEM!
Solaris and Irix are besides using X too are a lot more stable than Linux is. I'm administering unix systems runnig for years - not days or at best months as Linuxes. Installing Solaris and Irix is not more complicated than Linux.
Indeed, the trouble with *BSD is that they are pain to install and they are proud of it.
Imho Linux, especially RH, became more and more similar to the evil empire fighting it: All these fancy, worthless features, unneccessary complicated solutions (e.g. crontab, init) aimed to lock-in users to a specific distro making it more or less incompatible with all the others. (Started by Mandrake using their "own" libraries breaking programs and mastered by RH bullying other distros and alienating users)
Luckily Linux is GPL OSS so Ubuntu, Gentoo may continue in the right direction