Slashdot Mirror
The Future of OpenSolaris
jjrff writes "Phoronix has a little piece about the future (or lack thereof) of OpenSolaris. It appears based on the current support lifecycle, OpenSolaris may be going away. There is a fun thread (read: mild flameage) on a ZFS list about it."
Nothing about this says OpenSolaris is going away. Just support for older versions
"Can somebody show me something good to come from the Oracle-Sun deal? Anything?"
Cutlery related anecdotes for the next 10 billion years
Even if Oracle ditches Opensolaris all together, shouldn't the community keep going and shouldn't third party companies fill the hole left in the market with regards to support?
Or is this a question of reality not working out as the theory? Does that mean that, in a similar vein, Monty was right (and Eben was wrong) ranting and going to the EU about the fate of MySQL in the hands of Oracle?
(I don't know. I don't mean to imply anything. Just asking, sincerely.)
Some of my favourite people are from th US; Vonnegut, Chomsky, Bill Hicks.
A/UX
IRIX
Unicos
Xenix
Ultrix
OSF/1
soon: OpenSolaris
and if Larry Ellison has a bad dream: Solaris
1% APY, No fees, Online Bank https://captl1.co/2uIErYq Don't let your $$$ sit in a no-interest acct.
A bloody good version of Unix it is too. 64 bit from the start back in the early 90s when PC manufacturers and Microsoft were still wetting their trousers about moving to proper 32 bit.
The alpha CPU - what a missed opportunity. Perhaps in some ideal world in an alternate reality people woke up to what a dogs dinner x86 is and the alpha chip had as much development effort put into it. I wonder what apps would be possible on a 2010 alpha chip that is still pie in the sky for x86?
However, it didn't detect ...
Of course it didn't. It's not a desktop OS, even if it does have a purdy interface. Go check the hardware compatibility list - it's pretty friggin' small.
When I put together my home file-server, I made damn sure to check the HCL before purchasing any hardware. Even after doing that, I still had an issue with the on-board LAN chipset - had to compile a different set of drivers in order to stop it from dropping the connection every 5 minutes. OpenSolaris is a great server OS, but it's just silly to expect it to be compatible with some random laptop.
Only mild flamage? You see this is why I prefer Linux!
As you probably are aware of, there are TONS of mission critical servers out there running Debian, CentOS, FreeBSD, OpenBSD and other "there is no company you can blame and/or sue" operating systems, just as well as they run PostgreSQL or MySQL without support contracts for their mission critical databases.
For many companies that's not a problem because they have competent server admin staff and the community support is often way better than what you'd get for money.
An unsupported "debian-testing-style" OpenSolaris would make a lot of sense for both Sun/Oracle and many users. If you want support and someone to blame, just pay for Solaris instead. This model is already proven to work great: Fedora vs RHEL (vs CentOS), openSUSE vs SUSE Linux Enterprise, PostgreSQL vs EnterpriseDB, and so on.
My other account has a 3-digit UID.
I, like others, tested OpenSolaris and the 2 main problems that i saw where, 1) lack of support for fancy/new hardware. 2) not so many native programs as Linux/BSD. I think that OpenSolaris will live forever, but not as a OS, its bests features (ZFS, others) will be incorporated in linux/free bsd/ others)
postgres?
The article doesn't quite say it, and it doesn't have the smoking gun of "We're canning OpenSolaris", but that end of life page for OpenSolaris looks pretty damn final to me and there is little room for interpretation.
I wouldn't be surprised if Open Solaris went the journey. The whole point of it was to arrest the slide of Solaris in the face of Linux, in particular, and so that Sun could tell everyone that Solaris was open and just like Linux. Unfortunately, OpenSolaris has contributed little, if anything, to Solaris. There's no community of developers apart from those Sun sanctioned and things like Solaris's driver support is still a long way behind where Linux is. Development still hasn't been opened and there is no public repository development model. Sun, or Oracle now, is bankrolling it with none of the cost savings you would expect from such a project.
One can only hope that Oracle won't follow the same 'strategy' that Sun have followed for the past ten years, because it got Sun into trouble and it'll cost Oracle rather a lot of money if they get it wrong. However, they look as if they're doing swift about-turns on that and a statement of their future intent is clear when you go to www.sun.com - it redirects straight to www.oracle.com.
"but it's just silly to expect it to be compatible with some random laptop"
I don't think so. Server farms will go with proper supported Solaris and yes they will check the HCL first. The freebie option is for other people who want to try out solaris and who will have all sorts of random desktop and laptop configurations. If opensolaris doesn't support much hardware then who exactly is it aimed at?
So far as I can tell, zfs is the only piece of opensolaris that's exciting enough to make anyone want to install if if they'd otherwise want to install a linux distribution. With that in mind, could someone post an authoritative update on the supposedly intractable licensing issues that prevent ZFS from being incorporated into the linux kernel? Is it still hopeless?
If opensolaris doesn't support much hardware then who exactly is it aimed at?
Small business users, companies like Nexenta which produce their own server hardware/software products, and tech-savvy individuals looking for a home-server solution.
It's not exactly a huge market, but it is a niche (niches?) that needs to be filled. OpenSolaris is currently the best solution for projects such as mine. The ability to build a redundant array with automatic data corruption detection and a simple yet powerful snapshot functionality is what sold me on it. Nothing else on the market can do that, and the solutions which come close would have cost a lot more.
So long as it's FreeBSD 8 and not 7, a properly tuned and setup ZFS install is a breeze to put together. It took me maybe 20 minutes to kernel tune mine (i386 chipset and less than the recommended RAM in it at the time) and it's got good stability on a 1.7TB raidz unit. YMMV, but I wouldn't stick 7 back on another box again. I've no comment on FUSE.
Me failed English...
FreeBSD over Linux. If my comments seem odd, this may explain...
Errrrrr, survival and preventing Sun from going bust, just off the top of my head?
"Hi boss, yeah I'd like to use OpenSolaris. .. No, we specifically can't get support for it from Sun, I mean Oracle. .. Yeah if it breaks we are totally on our own. ... Ok so I guess we're not using OpenSolaris then."
That's not really any different from Fedora, yet businesses still use Redhat.
Uh, what? Redhat is RHEL, for which support is available. Fedora is RHEL beta, and is unsupported. Solaris is supported, OpenSolaris is unsupported. So in fact, it is entirely different.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
You see exactly the same on desktops except there Windwos runs on almost any desktop, OS X only comes on its supported hardware and Linux is fail.
Linux is fail?
I have an 8 gig USB stick that boots into a fully-installed version of Ubuntu. I have used it on at least 6 completely different desktop computers, 4 laptops, and 2 netbooks. Each time I plug it into a different computer it boots and detects the new hardware without a problem. Out of all of those systems, the only exception has been an Asus EEPC on which the wireless card wasn't detected.
I'm not sure how you can consider that "fail". I've yet to see anyone do something comparable with windows.
That's not to say that linux isn't without it's problems - I still use windows as my primary OS - but it certainly does run on just about anything you can throw at it. The only thing keeping me from switching to it permanently is a problem with ATI graphics cards - mine works great ... but causes a memory leak which forces me to reboot every few days (annoying, but not an issue for users who actually like to turn off their computers).
It's just as unfair when people expect to plug in any USB printer or gadget in Linux and have it work, but the world isn't fair.
All of my USB gadgets work just fine on Linux. It even detects my old wifi usb dongle, which windows doesn't.
Can somebody show me something good to come from the Oracle-Sun deal? Anything?
Errrrrr, survival and preventing Sun from going bust, just off the top of my head?
Is that really good? I just met someone who now works for Oracle; they worked for a company acquired by Sun prior to the merger. Sun fired all their best SEs because they made good salaries, while there are people all over Sun (or at least, were) making big bucks for doing nothing. UltraSPARC has its uses, but mostly it is an also-ran. Solaris' claim to fame is ZFS. Under Oracle, Solaris is doomed to either fail (as it was heading towards anyway, due to Linux's increasing dominance) or to become an Oracle RDBMS engine, which is much the same thing.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
The kernel is fairly compatible with the BSDs and OpenBSD at least has superb 100% free support for most documented and many undocumented devices. Especially for laptops. When I looked into OpenSolaris it seemed that they were trying to follow BSD driver development.
However, let's face it. Nobody cares about Solaris enough to spend time porting drivers.
It has a horrible base system, and the alternative is using GNU tools. If it had been more consistent and attractive to Unix fans, like BSD derivatives are, it would still have a decent following. As it is, even Linux beats it in elegance and internal consistency.
A pair of features are coveted by some people, but as vocal as they are, they just want to transplant them to their favorite system. They don't care if the donor dies. Especially if he has an incompatible blood type.
However, it didn't detect: the wifi adaptor the ethernet adaptor the sound
If there's one thing that would make the computer industry move ahead faster, it would be more standards. Why on earth can't simple mundane things like ethernet, sound, etc interfaces come with some sort of descriptors or standards which allow at least basic functionalities to be found more easily by an OS? Couldn't chipmakers, driver and OS writers try to save some work for themselves and talk? Every new OS version has to re-create, re-test, etc every driver for every device on the planet. The mere discussion of standards seems to have been killed by the whole 'de facto' notion, which is basically quitting. Even if we exclude MS, there enough active people now to have some debate over some driver and chip detection standards. VMware, linux, xbsd, the livecd scene, motherboard, device, and chipmakers, etc.
Build your own energy sources from scratch. http://otherpower.com/
there are TONS of mission critical servers out there running Debian, CentOS, FreeBSD, OpenBSD and other "there is no company you can blame and/or sue" operating systems
Absolutely false inaccurate information, at least for Debian.
As per
http://www.debian.org/consultants/
"811 Debian consultants listed in 64 countries worldwide."
Now, you can hire a consultant whom might actually moonlight as a debian developer, perhaps even the maintainer of something that is critical to you. And, as a private citizen or at least small consulting company, you could sue them when/if they screw up.
On the other hand, if you think you you can sue microsoft, and win, next time exchange falls over, you are in for a BIG surprise.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
FOSS is FREE only if you don't value your time.
*gasp* I value my time but I also value flexibility and independence from vendor whims.
I have an equally naive cliche for you right here: Proprietary software is only cheaper if you are incompetent.
My other account has a 3-digit UID.
FreeBSD - ZFS is no longer in experimental status as of version 8.0. I haven't heard anyone recommend FUSE on Linux. As far as other BSDs go, I know that at least OpenBSD has no plans to include it at this stage.
http://kerneltrap.org/mailarchive/openbsd-misc/2009/1/17/4750954 - But that was over a year ago.
At the moment I'm learning FreeBSD over OpenSolaris because I want ZFS, FreeBSD is fully free and open source, FreeBSD looks to have a wider array of ports, which should be easy to install, even though with the LiveCD of OpenSolaris it boots up straight to X. On a production server or maybe even workstation, I think the choice would be down to FreeBSD versus Solaris, rather than OpenSolaris. Someone correct me if I'm wrong. Solaris does have a lot of nice features though, FMA (fault management architecture - lets you know when something has gone kaput and what to do about it.) And FreeBSD will lag in terms of the version of ZFS it supports. Deduplication looks to be a pretty cool feature - if you copy some data to another part of the HDD, and then you leave it a bit and your hoarding nature kicks in and you don't know whether you can delete it or not - no fear, ZFS will recognize the data as the same, only store it in one place (unless modified of course) and so there is no benefit to deleting the copy other than being a neat freak.
I'm presently wrestling with setting up FreeBSD on wireless. After that I have to get X set up. It would be nice if FreeBSD had version specific handbooks ala PostgreSQL, but they don't. So it's a combination of man pages, handbook, googling, etc to get me where I want to go. It's a bit of a contrast to Ubuntu which I set up on another box in the space of about an hour, including updates. Unmetered FOSS mirrors on ISPs kick ass!
Anyway, I suspect that the user base of FreeBSD will grow by leaps and bounds when people realize the advantages of ZFS and don't want to wait for BTRFS or whatever the results of this meeting might be: http://blogs.sun.com/bonwick/entry/casablanca
If I have seen further it is by stealing the Intellectual Property of giants.
Well, Oracle offered about 3x the prevailing Sun stock price, so the Sun shareholders have done well. At least, well in relative terms--- some probably still lost money, but there was really not much else on the horizon that was looking likely to triple Sun's stock price. Before Oracle came along, the just-over-$3.00 stock was almost mocking its owners with its stock ticker of JAVA, an anachronism from the days that Sun management thought Java would somehow make them rich.
Coincidentally, for public companies, if you make a really good offer to stockholders (something >2x the current stock price usually qualifies), it's usually an offer the buyout target will find hard to refuse. That's the tradeoff you make when you IPO a company and put its ownership in the hands of the stockowning public.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Yeah - same problem - of about nine or ten systems that I've tried it on (up to snv_133), all of them have at least one hardware problem.
eg from my memory
NIC drivers (Broadcom, Even Intel)
W/LAN drivers (Atheros for instance)
Display driver support (not just VESA!)
HW RAID drivers (Compaq, Promise)
AHCI drivers (including NCQ and hot plug support (slated to fix in snv 135)
AMD PhenomII support (fixed now since snv 126)
and I've had issues with the install (eg installation from USB CDROM)
However, saying all this, the journey is worthwhile - some features really are fantastic - especially together:
ZFS + snapshots + dedupe + Virtualbox VMs.
YMMV
Those, who are crying here "OpenSolaris gone", read the fucking article CAREFULLY (never happens on Slashdot, though):
So use letter-by-letter approach if you're unable to see word-by-word or sentence-by-sentence:"Future releases of the Solaris OS will also be based on the OpenSolaris community codebase."
That means RedHat/Fedora model. Clear now?
http://www.sun.com/software/solaris/lifecycle.xml
And then only if your vendor is competent.
My interpretation of part of the point of OpenSolaris - Sun were using it as a testing ground for concepts that would make it over to Solaris, e.g. ZFS. AFAIK it's not exactly Solaris beta, but it is at least somewhat analogous. Both Fedora and OpenSolaris are unsupported, RHEL and Solaris are supported. I don't think you'd use either Fedora or OpenSolaris if you fear cutting edge stuff breaking on you, and if you fear that sort of thing enough to want something more stable, you might also buy support contracts, which would be less costly to provide because less stuff would break (hence RHEL and Solaris and their support contracts). But the utility of OpenSolaris/Fedora in business - someone in IT could get their feet wet for free with either Fedora or OpenSolaris and then make a case to management that they want to go with RHEL or Solaris after they have done proof of concept.
I could be wrong though.
If I have seen further it is by stealing the Intellectual Property of giants.
Nobody cares about open solaris. Nobody in their right mind would have chosen it as a platform.
I'm not surprised that IBM is the last company, AIX the last proprietary unix platform. Power the last proprietary hardware platform...
HP & Itanium? Laughable... And Linux on x86 has eaten the rest.
IBM 'get' services in the way the rest never have. They get that it's the bloody hardware which matters. This is why power is hitting 5GHz. The OS is just there to make it work. You want the fastest, lowest latency, highest throughput. You use IBM. You just want it to work and are on a budget? Linux.
The 'executives' of the rest of the companies clearly didn't know or care what their customers want, or what their business really is.
Deleted
And virtualisation is a big deal. Who cares what hardware the OS supports, so long as it can run under a hypervisor, which supports your actual hardware?
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
Not the same thing. I've been dicking around with such images since BartPE first came out. It can be useful for certain things, but it's nothing like having a complete OS running from flash.
There's really no good excuse for an x86 Unix to not have robust ETHERNET support.
This sort of nonsense is what kept Solaris x86 on the sidelines to begin with.
Not supporting the desktop frills is one thing, but ethernet is pretty fundemental to a Unix.
A Pirate and a Puritan look the same on a balance sheet.
Sun was going to get acquired. The only alternative to Oracle was probably a deal with IBM. You can speculate if IBM would have been a better owner, but IMHO they'd have had many of the same corporate priorities: making money and cutting losses on things they couldn't make money on. If Sun had focused on these things earlier, rather than doing crazy stuff like spending $1B on mySQL, they might have had a chance surviving on their own.
On the contrary... if you want zfs, and stability under load, there's nothing wrong with it.
I wouldn't disagree, but there are usually alternatives. I would personally never make a choice of OS on the availability of zfs (though I have no doubt the filesystem is quite nice). And stability under load can be had with any Linux or BSD with an appropriate configuration. So yes, there's nothing wrong with OpenSolaris, but then there's nothing so right with it that we can't afford to do without it.
I know there's a resistance among the various *Solaris/SunOS communities about the other UNIX variants, but it's time for these people to get over themselves. The world has moved on, and the alternatives are plenty mature enough to cope, and whining about them just makes these people look like craniorectal adolescents.
ZFS is available now. Running numerous places including my own home network.
btrfs still has a ton of "EXPERIMENTAL." "DON'T USE FOR ANYTHING IMPORTANT" warnings everywhere.
That right there clinched it.
And some other Solaris tech. Heck, they should GPL the whole thing and get a group of engineers to port the juiciest morsels of it to Linux. That way, Solaris going away would be much less of a loss, and Linus would be a happy man (he said, half-jokingly, he wants Solaris to die :-).
Anyone has been read the thing at Oracle? They say they will release their Solaris on top of the codebase of OpenSolaris. In other words, commercial Solaris from Oracle is a same as an OpenSolaris++. Oracle will add some proprietary features to Solaris that will be enabled only on their exclusive hardware.
In other words, this is Fedora/RHEL or OpenSuSE/SLES model. You like generic OS and you like to piss with it yourself, wasting a load of time, then go ahed, get OpenSolaris and GA support, if you like. But if you want advanced stuff and you have no time to waste a time for the cheap mess, then get Oracle Solaris for pay, get Oracle hardware and that's it.
So that's basically a message. Which is very good: it will actually push generic OpenSolaris to be up to date and financed by Oracle.
Oracle is the primary driving force behind btrfs, which is a ZFS-like next generation FS for Linux. Now that they own ZFS, continuing development of btrfs makes far less sense.