Slashdot Mirror
OpenSolaris One Year On
daria42 writes "In June of last year, Sun Microsystems open sourced its flagship operating system Solaris. This article asks the question, where is the OpenSolaris project after one year of operation? It contains views from Sun itself as well as insights from an external contributor to the code." From the article: "Sun is yet to release some aspects of Solaris as open source software, although that process is due for completion by the year's end. Meanwhile, non-Sun programmers have to date offered some 165 code contributions to the OpenSolaris project, said Eagleton. Of those, 70 have been accepted into the project's code base, while another 95 are still in the review process. To allay early community concerns that the process of getting external code contributions accepted was taking too long, Sun has a temporary buddy system whereby external contributors are partnered with Sun employees."
Sun is new to Open Sourcing its proprietary products. Solaris is a good step and a few glitches here and there are likely to be minor youth problems. The important thing is to know whether Sun will find in this experience enough incentive to open source other stuff (Java anyone?)
--
Krazy Kat, George Herriman
Write boring code, not shiny code!
I want to know why these guys switched from Solaris to Linux when Solaris is now free?
Can anyone with first hand knowledge answer my question?
Has Netcraft confirmed it?
It's like sex, except I'm having it!
Why is this bit of "news" listed under Linix-category?
In Soviet Washington the swamp drains you.
There is nothing Open Solaris can do for me that BSD or Linux cannot. What incentive do I have to use this product over its competition?
I've done a few console installs of Solaris 10 on some headless (and ancient) sparc netras. Here are some things that would make my life easier.
Specially from a user point of view, but also for servers and supercomputers, how do Linux and Solaris compare?
I know there isn't an easy answer to this, but a knowledgeable person could shed some light on us.
I had written a patch for ns_ldap.c to fix an obscure bug. After 2 frustrating weeks of dealing with online registration, which resulted in a heated exchange with one of Sun's adminstrators, I simply gave up trying. They've made it too hard to get involved in the project. For any normal open source project I simply download the tarball, run configure, make, make install, and submit patches to the email address of the most convenient maintainer. With OpenSolaris it's like trying to pull teeth. Even building a binary from the source is a major mission that makes building XFree86 look like child's play. And that's a real shame, because I'd like to fix bugs in Solaris as I find them, but I am not going to go out of my way.
If they want more people to try out OpenSolaris they need to reduce the barrier to entry, i.e. an .iso file or even a bootable CD that folks can order for a minimal fee. The ugly, several step manual process they require now is just too painful.
Solaris is already one of the greatest OSes around, becoming Open Source can only pave a way foreward for Sun Microsystems in the area of OpenSource development. This is good news, however, I wonder if JAVA would become open source with it :\
Follow me..
In the mean time, Sun still refuses to give required documentation so OpenBSD can support the so-called "Open" platform from Sun. As Theo explained, some details have impact on a lot of system subroutines everywhere, and checking Linux source code is not enough (would be easy but BSD is not Linux and there are fundamental differences in how the VM works and so on).
Even a journalist asked very clear and precise questions about the so-called "Open" platform from Sun and he never got any answer (and won't ever, probably).
Nothing to see here, move along.
Rather than statically compiling bash, wouldn't the better choice be to have it's dependencies all available in single user mode?
I used to work at Sun, and it's a company with a slow-moving internal culture. Pretty much any organisation that contains 30,000 people will necessarily not be zippy. The lack of speed says nothing about their intentions, though. For example, I've been talking to a number of Sun people over the past several months as they've been choosing a revision control system for OpenSolaris to use, and they've been keenly aware of the benefits of both doing things in an open manner and doing them carefully. They ended up choosing a wonderful revision control tool called Mercurial, but first they spent a few months evaluating the alternatives and, even better, writing up their evaluations and posting them in public. This is a very useful service to the open source community, as few people have time to evaluate tools in such depth, much less write in detail about why they did or not choose any of half a dozen alternatives.
To allay early community concerns that the process of getting external code contributions accepted was taking too long,
You're kidding, right? Solaris is one of the most mature operating systems out there. It runs some of the most powerful servers on the face of the planet. It is the core for a number of institutions, especially in the financial sector. I am not over-dramatizing when I say that Solaris runs a hell of a lot of crucial systems that make our lives easier in a lot of different ways.
That being the case, do these people really think that Sun is just going to say, "Oh, I see. You tested it in a limited fashion and we tested it in a limited fashion in the matter of a few months. Okay, we'll release it to the customers who run massive databases and financial applications on our servers because of a few months of limited testing." I would much prefer Sun take a year if need be to make sure that any modifications will be completely compatible with as many of their customers and equipment as possible, particularly the higher-end systems and major corporate environments.
I understand and share a lot of the aggravation that people feel when it comes to the lack of features, particularly device drivers, in Solaris. This is the one of the main reasons wy I think that Solaris has become so niche, particularly on the x86 side of things. If we're talking about modification to a common tool or enhancements to a graphical interface, okay, I don't see why it would take a year. But if Sun needs a year to make sure that a new device driver doesn't crash a SunFire 25K running a clustered Oracle server during end-of-month, transaction processing, then I'll grant Sun that year.
The Overrated mod is for reversing inappropriate, positive mods, not for voicing disagreement with a post.
No, because if usr is not mounted, the libs would have to be available somewhere in / and I would be responsible for maintaining and updating them myself (if I bothered to remember to do it at all). That would just increase the PITA factor and I might as well grumble and use sh at that point.
Slashdot really needs a "Not Funny" or "Old" modifier for these kinds of comments
Very good points. I for one don't want minimally tested extensions to Solaris (or any other O/S) to be on any system that controls my money! I wonder why no one mentions that IBM isn't doing ANYTHING to make AIX open source, nor is HP doing anything with HP-UX. And hell will freeze before MS does anything open-source with Windoze.
pkgadd, BTW, also supports quite a few URL constructs (e.g., pkgadd http://blah/blah). In this form, the other end of the pkgadd has to be a package stream, however, so that limits its usefulness with the DVD contents.
It's been a while since I've done the text install, but finer grain control has been there in the past. I'd be surprised if it was removed. That said, using Jumpstart combined with a profile will also get you finer grained control without having to do it manually for each install. Information on network-based installs and the like is available here and here."Overrated"
My kingdom for some mod points. You've made some great points that NEED to be taken into account.
Sure, Linux is cool and all, but in terms of maturity - Give me Solaris, or give me Death (actually i'd also accept a flavor of BSD, but that's a whole other post).
They already are. Most of the vital libraries in /usr/lib are softlinks back into /lib.
Thanks for the great advice and information.
/sbin/sh and friends are all dynamically linked.
/sbin directory a unix canon? (Then again, Solaris 10 breaks with the past unix canons on smf versus rc.d/init too) What if /usr is down or needs a low-level fsck? I've always been told that that would mean you are totally screwed, unless you have static binaries in /. If I'm letting the package management/patch system update the libs, I don't see the problem with having it update a few kBs of static binaries in sbin while it is at it (unless the whole update is being done piecemeal).
Jumpstart would be great if I was setting up dozens of boxes, but I'm not. Just two or three. I still think pkg-getting off the internet would be the optimal solution if you just want a few boxes (although that would be leeching bandwith depending on how much you are downloading).
Solaris 10 and up doesn't come bundled with *any* statically built binaries anymore. The
That suprises me. Isn't a static
Sun and Apple both ship a proprietary OS based around an "open source" core. Sun's core is OpenSolaris, and Apple's is Darwin. Sun has done a far better job open sourcing their operating system. I do a 3rd party hardware device driver for both MacOSX and OpenSolaris. To compare Apple's to Solaris' "open source" OS
is quite interesting:
- Source code: Darwin: Must sign up for an Apple account to view source, source code for Intel kernel not even available. Solaris: Source code browseable on web, and available to anybody.
- Installable OS: Darwin was never updated from 8.0.1, which was released over a year ago. Solaris: Solaris Express is released at least monthly.
- Project direction: Darwin code appears after a MacOSX release. There is no way to see the source code of an upcoming MacOSX version, there is no way to even know what features will be present aside for signing up for a $500/yr ADC account. You are not allowed to talk about this in public. This is in stark contrast to OpenSolaris, where Sun engineers publically debate virtues of different features, and future directions on their forums/mailing lists, and anybody is welcome to contribute.
In short, OpenSolaris is a real open source project. Darwin is a sham, and would not survive without Apple.
Another option is to utilize and plan for using LiveUpgrade. That allows you to 'clone' your running OS onto other slices, perform upgrades on them, patch them, etc, while the system is running... your only downtime is the time it takes to reboot. If for some reason /usr fails, you can reboot onto your alternate root copy and then do repairs.
Not to mention that it's pretty hard to build a fully static binary because Solaris 10 does not include static analogs of libc and friends.
is it possible to download an iso file of OpenSolaris ? On the solaris website, only the source files and some tools are ready to download. Schillix (which is based on OpenSolaris) is available in iso file.
Solaris isn't OpenSolaris and no one should be running OpenSolaris on anything that mission critical. If you consider OpenSolaris a testbed for future Solaris updates there is no good reaons why it should be unreasonably hard to get your code into OpenSolaris
Some seriously usefull information there. thanks again.
Hey buddy, it's time for you to sit yo' ass down, warm up the chair, and learn how to set up and use JumpStart(TM) to deploy Flash(TM) archives.
Installation lasts a few minutes, literally. Systems are deployed and configured fully automatically, without human intervention.
If you use just the JumpStart(TM) framework, you get to write rulesets selecting (or deselecting) exactly which packages you want (or don't want).
http://docs.sun.com/ is that way.
And BTW, there's a far better and more powerful shell than that bash PoS. It's called tcsh, the Tenex C-shell.
Opensolaris is often referred to as ON and it looks like even the fellow Slashdot Mason couldn't resist the temptation with the title "Opensolaris One Year On". The designation ON appears quiet often on their site such as in "ON Copyright Notice"
"TOI for ON Developers"
and on a dozen other places.
I like to think that the designation 'ON' that appears all over the Opensolaris pages refers to Heliopolis, the CITY OF THE SUN, in fact I'm pretty sure it does.
Check out the relevant Wikipedia article on Heliopolis:
The city's Egyptian name (shown in hieroglyphs, right, transliterated wnw), is often transcribed as Iunu (literally "[place of] pillars"), and was often written in Greek as On, and in biblical Hebrew as Ôn and wen.
(Interestingly enough, firefox showed the Greek and hebrew glyphs as the latin characters 'On' when I previewed this on slashdot)
One thing I know for sure, any endeavor with such a powerful sigil as the Sun can only succeed.
There's also nothing stopping developers from distributing device drivers separately and indeed you can find third party drivers for peripherals like Ethernet cards out on the web. The device driver services, documentation, and stable ABI even made this possible before the OS was open sourced. Properly written drivers can be portable across architectures and can be used across OS versions without recompilation.
That's right, all of the best software engineers in the world are working for Sun and everyone else is on crack ... and Linux is losing to Solaris in such a big way on Wall St. and Sun have been able to keep using their same old development methodologies, don't change what works right?
However back in the real world, the changes people submit to a project they haven't been contributing to generally don't involve re-writting the VM or network stack, they involve the lots of much smaller "annoying itch" type things that a competant maintainer should just be able to quickly read and ACK (or NACK, or munge and ACK). The fact they've only got about one submission every three days, for their entire subset of OS they've released, and have ack'd less than half of those does not bode well for OpenSolaris.
Also, don't forget, that they're pushing an entire new FS/block device design (layers, who needs layers) ... and have re-written the network stack for Solaris 11 (to be competitive with FreeBSD/Linux). So if you think change is bad, you really need to migrate from Solaris.
ustr: Managed string API with ave. 44% overhead over strdup(), for 0-20B
Something I also never knew about was that if any developer makes a change that causes Solaris to run any slower, that change is immediately rejected until the code can be reworked. I love my linux, but it is definately taking much longer to boot these days - totally in contrast to my experience with Solaris from 2.6 till now.
Well, they had the docs. They had the code. But dtrace and some nasty bug lopped it all off. It would have been better to lop off that specific breaking feature, and not a whole architecture that didnt have the release. If you can keep enough of sbus to run a U2 and the sbus hardware, the u1/ss5/ss20 should be a (relative) piece of cake (notwithstanding dtrace, the way to HCL out anything before a Ultra2).
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
Wow. The clue meter is reading zero.
One of the main reasons behind OpenSolaris is for people outside of Sun to make drivers, tools, utilities, etc. that can be included into future versions of (non-open) Solaris. Because of the nature of the systems that run on Solaris, it's critical for Sun to make sure that changes, regardless of how benign it might seem, have no impact on any kind of potential, mission-critical appliction. What one indeveloper might think is a great driver or enhancement for their partcular use could possibly break something else. Considering how much is riding on Sun -- not the least of which is rebuilding Sun's own reputation -- they have every reason in the world to make sure that everything that "everyone else on crack" submits will have zero impact on anything else in Solaris.
Sheesh. Get off your damned high horse.
You're right. Some of the best software engineers in the world are working for Sun. Sun's been hammered in the market place because it didn't react well to the dot-com bust. No layoffs, no significant restructuring or refocusing for years. Everyone cut margins and moved to commodity hardware, except Sun. (Note - Times have changed, see Galaxy and Coolthreads lines) The fact that Linux is "great" is precisely because it adheres to many of the design principals of a Unix-type Operating System and because it's free -- as in freedom, not beer. I also happen to think the BSDs are great, and that they've contributed a lot to the F/OSS community. But if you can't see all the amazing things coming out of the Open Solaris project that I just wish was in Linux, then there's no use even discussing it with you.
Sun has it's problems, but the engineers there are not one of them, and frankly you seem to be implying that they re somehow trying to impede progress or shut developers out of Open Solaris. On the contrary, they're working with Open Solaris contributors every single day. Let's not confuse high standards with a lack of enthusiasm for community building.
By your logic, Microsoft operating systems must be technically superior to Linux, because of both profit margin and marketshare right?
I would love to see OS X and Solaris 10 merge into an insanly great OS. They would seem to compliment each other. I love OS X. It has a wonderful interface but, IMHO, it is lacking on the UNIX side. Darwinports and the Fink project have done much to fix this, but it is not perfect. Solaris is wonderfully advanced with features like predictive self healing and tools like dtrace. But Solaris is really lacking on the user interface side (last time I tried Solaris, the Desktop was a modified version of Gnome 2.6).
Sun and Apple seem fairly similar in that each have their own unique OS and Hardware that is more or less designed for the OS. Platform specific features applying to both isn't that far-fetched.
Just a thought...
Slow Down Cowboy! It's been 1 hour, 47 minutes since you last successfully posted a comment
All Linux distros do this (that I know of), since bash is often used early in the boot procedure way before
I wasn't trying to say that, I think people wouldn't have missed my point if I'd said the same thing in person ... Ahh well, communication via. text sucks sometimes.
Personally I think that the license and the time of release have impeaded adoption of OpenSolaris, and has certainly imeded people from contributing (everyone I know has been told that legally they can't even look at it, as they'd become tainted for any Linux work -- but then a lot of people I know are employed doing Linux work). None of that is any Sun engineers fault though.
I wasn't trying to say that no good engineers are at Sun, merely that not _all_ of the good ones are there. I wasn't trying to say that you don't need to exercise caution when accepting patches, merely that FreeBSD/Linux/etc. engineers already do this as do Red Hat/SuSe/Ubuntu ... and they process a lot more than 70 changes a year from "the community".
I wasn't trying to say Solaris doesn't still have benifits over Linux, merely that it's not all benifits.
ustr: Managed string API with ave. 44% overhead over strdup(), for 0-20B
yup, about 0% of internet sites are using it, with 10% growth per year. and their average uptime is 100%.
So - basically, you want Solaris to be Linux. Sorry, d00d, not gonna happen. Sounds like you started with linux, and now anything that's not linux sux.
Shutting down free speech with violence isn't fighting fascism. It IS fascism!
There's http://www.gnusolaris.org/gswikiNexenta
On the Windows marketshare, last I checked, Linux's growth rate was quite robust compared even to that of Windows. Not that the Linux fanboys should be complacent or anything, but the Windows and Linux fanboys might have reason to be worried...
I'd rather Sun focus on hardware I actually want to run, not crap I find in the back of my closet.
Well, that's why I'm not asking about sun4c's and the VME machines.
Heck, their competitors have done well with - IBM's AIX 5.1 supports their equivalent of Sun's IPC/IPX's, and
throws in features that at least give the oldest IBM 70xx's the "swan song" release. That means that they gave older
machines at least some of the newer features despite being relatively undocumented and dropped from support next
release.
(TEN... YEARS... that's back to 486 days in the Intel world)
That's back to when PREP and Microchannel CHRP machines were beginning to be supported, not in the "ancient but
supported in a relatively recent release" of AIX. If you havent guessed, that's where I moved to escape the Sun
HCL game.
It's more or less to warn people that if you're going to go Sun, get the longest support contract possible (long enough
to warrant an extension of hardware support beyond their normal cycle) and to make sure that no hardware can be HCL'ed
out until the end of the contract.
They didn't just sort of stop supporting the stuff, they dropped 32-bit support entirely from the S10 kernel.
That means the sun4m, sun4c, and early edition of ultrasparc machines are impossible to use. It also means
no expensive engineer time spent trying to deal with 32/64 bit issues in existing code.
Or they could just put all of Build 22 up(the last sun4m release) and put a huge warning that this is a release
for the 32bit end, with the explicit part that some parts do not work despite being a buildable, runnable system.
Just put all of it in there as a build of its own (Unless they could just release Solaris 9 + ZX & S24 code - same
terms and intent that it is intended for 32bit sun4's), which could be built from source(from a sun4u if they
have to add a hurdle), but does not have dtrace(hardware limitation), and may not have certain features related to KCF.
That's enough of a compromise for those who have that hardware, and wouldnt mind just being able to give their Quad
Ross SS/20 a relatively secure (having sadmind bugs just for full ZX acceleration isnt my idea of security) setup that
lets them run their Sparcstation as intended (with Solaris)
To bmc, the "it could be done, but he didnt bow to the gods" keeper of the code:
It's not that biblical of a task to do some the above or the suggestion of documentation down below, bmc - the legal bits could
be the largest of any challenges. It's not like I didnt hear that you didnt like how I was asking it in some irc channel. Saw the
list reply and the irc messages that sort of gave away who you were talking to.
Or run OpenBSD on them. OpenBSD loves the 4m.
It's ironically the only thing that runs reliably with one of the buggiest sun4m's - the SS5/170. Look up turbosparc
cache bug, that should point you towards the bugs that the machine had that were fixed only in OpenBSD for a long while.
The only catch is that ZX and S24 cards (decent cards for the task) are dog slow to run given the complete lack of
documentation (which is what the people behind OpenBSD ask to have) on anything regarding the chip's use beyond a
slow 2d frame buffer. If one ends up running OpenBSD, at least dont make their developers worse off for the
hardware documentation.
If anything could be released that wasnt a full release of OpenSolaris/sun4, would be the documentation to drive
about every sbus/afx card that has no documentation or incomplete documentation. Then it would be a nonissue to deal
with 32bit machines.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.