OpenSolaris Or FreeBSD? 405
Norsefire writes "I am in quite a predicament. I decided a while back to branch out and use a new operating system (currently running Debian). After a bit of searching (trying Gentoo, Gobo and Arch along the way), I decided to use something that isn't Linux. Long story short: I narrowed the choices down to OpenSolaris and FreeBSD, but now I'm stuck. OpenSolaris is commercially backed by Sun, has nice enterprise-y tools in the default install, and best of all, a mature implementation of ZFS. FreeBSD is backed by a foundation, has a minimal default install and a rather new (but recently improved in the 8.0 release) implementation of ZFS, however it offers the Ports Collection (I quite like the performance boost due to compiling from source, no matter how small it might be) and a bigger community than OpenSolaris. That is just a minimal mention of the differences. I would be interested to see what the Slashdot community thinks of these two operating systems."
Re:Go the whole hog... (Score:5, Informative)
Ah, but OpenSolaris isn't a clone, it's one of the true heirs to the throne, a direct descendent of the original UNIX lineage.
The *BSD family are now cousins to the original UNIX as all the original code was excised to make the 4.3BSD-lite codebase.
Re:Performance boost? (Score:5, Informative)
For x86, you may get a very slight boost, because binaries in conservative distros/OSes (like FreeBSD) are still typically compiled for i686. Turning SSE and other such stuff on can let gcc generate more optimal code, particularly when floating point is involved.
On x64, it is of course quite meaningless.
In practice, either way, it's not worth the hassle at all.
Re:Performance boost? (Score:2, Informative)
I personally don't care about the little performance gain from the flags. BUT you can get a lot of performance and customization options if you compile it from source because there are many options available for you only if you compile it. A simple example: try installing pidgin from ports, and you will see a bunch of options you probably never saw before! You can disable networks you don't use, enable some underground ones, etc. Now try compiling apache and other server stuff...
It is time consuming, but ports make it really easy for you.
Re:What? (Score:2, Informative)
C'mon ! Parent is funny not a Troll :) Mods try to have some second degree ...
Re:OpenSolaris is more supported (Score:3, Informative)
I have to say that I really love OpenSolaris.It's polished, works out of the box with nvidia, has good Java support (LiveConnect actually works in Firefox) and the admin tools for stuff like zfs, zones, glassfish, fault management, system services etc are really excellent.
The list of software packages is still a bit limited, but at least most important things are there. Blastwave, /contrib and /pending helps a lot.
The thing that really bothers me, however, is the lack up security updates in /release. There have been very few updates to 2009.06, even though Mozilla, for instance, has released Firefox 3.5 updates several times. It's hard to believe that 3.1 beta 3 (which is what's in 2009.06) would be immune to all these security issues found in 3.5 ..?
Debian GNU/kFreeBSD (Score:5, Informative)
You are used to Debian ? Then try Debian GNU/kFreeBSD [debian.org].
The Debian distro on top of a FreeBSD kernel.
Re:Why? (Score:4, Informative)
Surely it depends upon what you mean by "support?"
OpenSolaris is backed by one of the big UNIX developers and is a true, direct lineage UNIX. You can also pay Sun for full enterprise OS support, which could include getting their programmers to fix a particular kernel or core OS bug for you within days.. if you're rich enough to afford the Platinum Support.
And you've bought hardware on their "supported hardware" list.
hm (Score:5, Informative)
Re:FreeVMS (Score:5, Informative)
Instead of FreeVMS which isn't ready for prime time... Get the OpenVMS hobbiest edition, load up SimH and run OpenVMS on a real emulated Vax. For fun you could boot OpenBSD, NetBSD or BSD4.x on the emulated Vax.
As far as Solaris vs. BSD -- I run 'em both here. Solaris mostly on Sparc and BSD on x86. I've done Solaris x86
and it's ok, but it's really fun to set up a jumpstart server and load up some old Sparcs.
I've even got SunOS 4.1.4 up...
Take a look at the software available on the http://www.openvmshobbyist.com/ [openvmshobbyist.com] site. A ton of VMS languages including C, ADA, Pascal, Macro32... TCP/IP and Clustering.
http://simh.trailing-edge.com/ [trailing-edge.com]
Re:Debian GNU/kFreeBSD (Score:4, Informative)
ATT Bell Labs started Unix (and the C language) (Score:2, Informative)
You have that backwards [slashdot.org].ATT&T Bell Labs invented C, and then used it to write Unix, which was a play on the name of the OS called Multics, which was also AT&T Bell Lab's baby (along with MIT and General Electric.)
Re:Article is trollbait (Score:3, Informative)
> Now who's trolling/flambating?
I was over the top, but...
You said you like to compile your own because of "speed issues." If your speed issues are that sensitive, you don't want Solaris. Solaris on x86 has been known to be a pig for speed in the past. While this has improved in Solaris10, I think you'll find Solaris on x86 to be slightly below par compared to a precompiled Debian kernel for your architecture.
The reasons for running Solaris are not speed related.
> I also like setting compile-time options, applying patches etc. that you can't do with packages
I believe that the whole "locally compiled = better performance" is a load of hogwash since you chew up time compiling, and if you compile in the background, doesn't that affect what's going on in the foreground? This is why some people have a separate "compile machine" for Gentoo. I believe this "wasted" effort and time outweighs whatever potential benefits you get from a local compile.
And someone up there mentioned going from hypothetical versions x.x.3 to x.x.4rc or something, leaving you with the choice of not being arsed to do it and put up with a potential security hole or do another recompile for a minor bugfix.
Your argument for local compilation also assumes that packages go unpatched. That's simply not true. And if you're really impatient or can't find it in the repository, you can build your own packages from source with chkinstall. I have my own little repository called local.debs just for that.
> But there's only one of me so I can't use both.
Sure you can. They're called virtual machines. It's also called multiple boot.
Make some space and try 'em out for yourself.
--
BMO
Re:Go the whole hog... (Score:3, Informative)
The question was about OpenSolaris which is much - but not quite - like Solaris.
Although I suspect you meant your question to be rhetorical it's a good question. HPUX and AIX has never meant to be run on consumer hardware. They are/where tailored to the hardware they where sold for. Solaris and MacOS aren't much different here.
OpenSolaris on the other hand is an open project with ongoing efforts to make it run on thing not sold by Sun or Fujitsu. It has nowhere near the peripheral device support Linux has, but as with other open source OSs that is an ongoing effort. I haven't found a PC I couldn't install OpenSolaris on for a few years now.
Re:Only copyleft is "commie", BSD isn't. (Score:3, Informative)
Many small-time programmers do pick GPL for irrational ideological reasons - "don't let evil corporations steal our code". That was the prevailing culture from the early days of open source software, back when everyone lived in mom's basement and thought money grew on trees. As FLOSS got bigger, a lot of software authors simply didn't give much thought to the GPL-vs-BSD debate, and went with the herd mentality (pun intended). Some bigger players like Qt (now Nokia) also used GPL's restrictiveness to make money, which is perfectly fine as long as you don't claim that restrictively licensed software is somehow more "free" than the permissively licensed / public domain kind. A lot of people also thought GPL would be more effective at "hurting Microsoft" than BSD, which has proven to be completely the opposite - as I predicted. (Google - smart, IBM - dumb.)
I'm not "trying to pin the non-success of SFU" on anyone but the regulators. The FLOSS community doesn't have any obligation to support a particular platform, but it's very telling that they snubbed Interix as much as they did...
So, anyway, I'm just making a long-term prediction of a libertarian-minded counter-movement in open-source software - people like me picking *BSD over Linux / Solaris for ideological reasons. We'll see how that prediction holds out.
No free security updates for OpenSolaris (Score:1, Informative)
Remeber that you don't get any free security updates for OpenSolaris. That means you are stuck with the security problems and bugs until the next release,
Of course you can buy support from Sun.
Re:Only copyleft is "commie", BSD isn't. (Score:4, Informative)
While I agree with much of what you say, it doesn't exactly help your case when you layer your own post with fairly fanciful and stupid assertions, while rebutting the exact same in the GPs post. For one, the BSA aren't Microsoft's enforcers anymore than the RIAA are the Bee Gees' enforcers. They are a group that exists to enforce copyright and software licences, and while I don't agree with much of their policy or their actions in enforcing it, suggesting they are some puppet of Microsoft's is just absurd. Check the BSA membership, it's full of huge industry giants many of them direct competitors of Microsoft's; IBM, Apple, Dell, Adobe, Symantec, RSA, to name just a few. Further, military style raids might be a slight exaggeration, like calling the GPL communist or anti-capitalist for example.
But one point in particular I'd like to address is your assertions on the Interix system. Firstly, I think it's absurd to suggest that Interix was "created solely for the reason of destroying UNIX". Where's your proof? What leads you to this conclusion? Or does providing compatibility now (much like a huge number of other projects, like Wine) automatically entail an objective of destroying the target platform? Unix (and Unix-like) systems have always played and continue to play a major role in computing, and this is a good thing, surely some degree of compatibility with these systems at the API level is a good thing? This is a large part of what Interix does, it provides a POSIX implementation on Windows as well as a Unix-like environment for development and productivity. So you have the POSIX API, Csh/Korn shells, a large set of Unix utilities, compiler, libraries and headers, and a lot more. The idea is to provide a Unix environment on Windows for migration, compatibility and development.
Cygwin I suspect wasn't "fixed" by Microsoft for several reasons. One would be that Interix/Cygwin began development around the same time, another would be whether the developers would be receptive to development efforts by Microsoft, another might be legal concerns and all the usual licensing crap, but perhaps most of all, the way they accomplish their functionality is very different. Cygwin provides a POSIX implementation and Unix-like environment _ON TOP_ of the Win32 API. This is done through a DLL (cygwin1.dll) which translates POSIX calls into Win32 calls which in turn call into the NT Native API. Interix by contrast does not use Win32 at all, but runs directly on top of the POSIX subsystem, thus, Interix apps go POSIX Subsystem -> NT Native API. Of course, you still have to use the Win32 API as that's what the Windows OS is primarily built on, but the POSIX subsystem runs alongside it and Interix on top of it. This is indeed the point of the NT Native API and much of the NT design; the Native API is (as the name implies) the base API for the NT OS and environment subsystems run on top of it providing an API for client applications. The Windows API is one such subsystem and the one that 99% of people use, POSIX is another, Win16 is another (I think?), and in the past there has been a (fairly crippled) OS/2 subsystem, and possibly others.
This affords some unique functionality for Interix in that it can do things at the API level that the Win32 API doesn't really support, simple example: fork(). The Win32 API to my knowledge has no real fork() equivalent, however, this is supported by the POSIX subsystem. The reason is that the Native API does support fork() but does not expose it through Win32 (but does through POSIX). Clearly, the Cygwin developers have worked around this, although how they've done it I'm not sure. Perhaps they translate fork() calls to loose Win32 equivalents? Or they call directly into the Native API (possible, but strongly discouraged)? Whatever, my point is the implementations of these two environments are very different, and I suspect they offer varying functionality as well as differing in actual POSIX implementation. I gather there's quite a nice Interix community, and Microsoft has put a
Re:Go the whole hog... (Score:3, Informative)
Mac OS X Server is licensed to run in a virtual environment when the host is running on Apple hardware. Actual support from virtualization software is lacking, and limited to Parallels and VMware for OS X only.
Solaris is supported and certified to run on xVM, VMware ESX and VirtualBox on SLE 11; this means that Sun will provide support for running it on these virtual environments. Solaris is supported by VMware's virtualization products as well, and can run as a Xen guest.
Re:ATT Bell Labs started Unix (and the C language) (Score:1, Informative)
Comment removed (Score:3, Informative)
Re:Debian GNU/kFreeBSD (Score:1, Informative)
How about suggesting something that is not in a very highly and unstable alpha condition? Have you ever used it? I have, and it's not ready for prime time.
Re:FreeBSD ZFS kernel panics? (Score:1, Informative)
A fellow has asked the freebsd-stable mailing list that exact question, and still to date there have been 0 responses.
http://koitsu.wordpress.com/2009/10/08/freebsd-and-zfs-is-it-truly-stable/
Apparently people are seeing better stability in general with FreeBSD 8.0 (which just came out), but there are some claims of serious performance degradation as a result of ARC thrashing. Patches are being tested in -CURRENT for this problem.