Why OpenSolaris Failed To Build a Community 280
xtaski writes "Ted Ts'o, one of the earliest Linux developers, points out some serious flaws in OpenSolaris. There is a severe lack of developers, for one. Apparently, after 3 years, the OpenSolaris 'developer community' is still struggling to get the proper tools for developers to develop! Ted also points out some other flaws which make it clear just how disconnected the executives at Sun are from what's really going on in their 'open source communities.' He notes, 'It was never ... Sun's intention to try to promote a kernel engineering community, or at least, it was certainly not a high priority for them to do so.'"
Re:OpenSolaris fails to build community b/c it suc (Score:5, Interesting)
Hmmm - Linux Fan Boys speak out.... (Score:5, Interesting)
There have been a lot of great development work on OpenSolaris in both the x86/x64 and SPARC worlds.
OpenSolaris (much like it's big brother Solaris) does have a list of valid / tested hardware platforms that work out of the box without issue.
If your specific hardware isn't listed and it's fairly well mainstream, document what didn't work, submit it, and it will more than likely get fixed.
I've used OpenSolaris on IBM/Lenovo thinkpads, IBM xServer hardware, SuperMicro / Intel hardware, homebrew systems with rarely an issue.
I've enjoyed the support of the OpenSolaris community as a whole, and found them to be as resourceful as any *inux / bsd community.
It all depends on what you like / want.
For me, gaining the ability to work with Solaris during development cycles to help in some small way guide / assist with the efforts is worthwhile.
Gnu/Solaris (Score:4, Interesting)
Too easy to replace Solaris... (Score:3, Interesting)
Yowsa (Score:1, Interesting)
-Wouldn't let the opensolaris board call the project opensolaris. Probably a legal quagmire of their own creation. The consequences of that lead to this resignation. http://mail.opensolaris.org/pipermail/ogb-discuss/2008-February/004488.html
-There's this gem, most of which I don't pretend to understand. The punchline is on the bottom. http://cryptnet.net/mirrors/texts/kissedagirl.html
-There's this gem, where even Ian Murdock links in suggesting the difficulty is happening above his level. http://ianskerrett.wordpress.com/2008/02/22/a-solution-for-suns-os-community-problems/#comment-17418
As much potential as Sun continues to exhibit, they still can't turn it into anything.
Re:I though it was just because it sucks (Score:3, Interesting)
Re:For those too lazy too read the article: (Score:5, Interesting)
I didn't come to Sun because I like the Kool-Aid, I came by acquisition. I haven't decided yet whether or not this whole "we love Open Source" thing Jonathan keeps plugging is real or a charade. I'm optimistic, but we'll see.
On better days, I like to think that the people way up at the helm really "get it" and are just waiting for the rest of the ship to slowly (slowly!) turn. On not-so-good days, I start to wonder if maybe someone's trying to pull a fast one.
There are lots and lots of people here who really and truly believe that Linux is just an upgrade path to Solaris. In other words... Once people start running Linux on Sun hardware, they'll "want more", and "step into the big league" with Solaris. It's kind of sad, when it's not irritating.
Anyhow... I could bitch for a while, but I won't.
Re:OpenSolaris fails to build community b/c it suc (Score:4, Interesting)
Re:mirror (Score:4, Interesting)
But yeah, I was pretty impressed that my 1 GHz Pentium III with only 512 megs of memory running 2.6.16 linux was able to not only survive, but recover from a slashdotting without needing to reboot. If I had only checked earlier to make sure that wp-cache really was enabled, but as the old saying goes, "no one expects the Spanish Inquisition!"
Re:I though it was just because it sucks (Score:3, Interesting)
Yes, it's a hog, and I say this as someone with a couple of dozen SUNW^H^H^H^HJAVA shares.
I have some old Sun boxen, one of which (Sun Blade 100, 512MB, 500MHz) I used to play at 64-bit big-endian RISCs running Solaris. All of my other boxes are Linux including an old Ultra 1 with splack.
Nevada build 85 only took about 2+ hours to install from DVD on my Sun Blade 100. The previous one I tried (78?) took 3.5. The GUI is unusably slow, and the disk thrashes like mad. I use it headless over my network normally, so I don't mind too much. There are suspicious java processes running all the time, too. I sometimes run SETI@Home on it and it gets just shy of 330MFLOPS, which isn't bad considering its age.
I remember when Sun started the Java desktop thing. It is a port of GNOME, but they insisted on writing a lot of replacement applets in Java for it. Why have XMMS for audio when you can write a lame mp3 player in Java, for example? *sigh*
Re:For those too lazy too read the article: (Score:5, Interesting)
So, for example, the Thumper [sun.com] is one of Andy's creations. It's pretty hard to beat the storage density you get for the price. Put a mess of those under a Lustre filesystem, and people start to take notice of Sun as a player in HPC. The recent TACC Ranger [utexas.edu] system is all Sun gear: storage, compute, and network (with sun-built Magnum [sun.com] switches). The OS? Linux.
There's more interesting stuff coming down the pike, and from my perspective, it seems that there's a shift toward making money on volume rather than margins. In other words, somewhat less awesome, but more of it.
I dunno. I don't profess to have much more special knowledge than anyone outside of the upper echelons. I'm hopeful, though. I read somewhere that many of the big Solaris egos were hired away by teh google. Hopefully they keep going. They can have our kool-aid-drunk sales and marketing people too.
Re:Not welcoming your Scott McNealy overlord? (Score:1, Interesting)
IBM is at every technical conference I've been to in the last four years. OLS? ISO? POSIX? GCC Summit? LSB? IBM people write for LWN, even though IBM published a ton of linux-specific documents on their own site. IBM people are active on all major FOSS lists, including kernel, gcc, glibc, they donated Eclipse, etc etc.
Intel is about the same.
Sun gets maybe 20% on this stuff. For the most part, they give the appearance of not caring to involve the community even on the software (openoffice, see build divergence, java, now mysql?) that they "own." Sun doesn't even test gcc on Solaris. Let alone actually try to fix bugs in native support.
Jeeze. And I don't even work for IBM.
Re:Gnu/Solaris (Score:3, Interesting)
GPL was a reactionary statement to the BSD licenses' freedoms being taken advantage of by corporate powers. The license OpenSolaris is under now ( CDDL ) is much more balanced. It keeps the code under a viral clause ( change it? share it! ) but it doesn't spread to the surrounding code like the GPL does ( link to it? share it! )
The GPL illustrated some fine points about giving users too much freedom, and now it's time to lay it to rest with Marx and pure Smith economics as political tools that were found to be unworkable except in a modified form.
That Linux is successful and also GPL is an anomaly rather than a general rule, and even in that case it serves to illustrate that the hardware manufacturer's unwillingness to write GPL drivers is holding Linux back. Were Linux under a more permissive license, with the market share it has now, hardware mfrs would be rushing to write drivers for it
Re:For those too lazy too read the article: (Score:5, Interesting)
Funny. It has been exactly the opposite for us.
We're running a bunch of xfires (14 boxes total, 4100, 4200, 4150) here
and initially started out with solaris because the wise guys said it's faster,
more stable, oh and no least you get that shiny "platinum support" badge...
Yea it was all that and the zfs hype, what could possibly go wrong?
Nothing much to be honest. We fell in love with the hardware immediately
and the machines hummed along without too much trouble. Postgres performs
well, java performs well, and ZFS snapshots are a blessing.
Despite all that superficial happyness we switched most of the hosts to linux
(and aim for 100% linux) after a few months. We still love ZFS (and can't wait
for a linux equivalent) but that alone couldn't justify sticking to solaris for us.
What broke it for us is the userland with all its subtle differences
to linux, or in other words: the learning curve. This may sound strange when
talking about a UNIX OS but as a linux shop we're spoiled by the GNU toolchain,
by dead-simple package management and all the little everyday things that just
work a tiny little bit different under solaris.
I'm not saying the linux-UI is better (actually, it is in many
places, but that's not the point here), it's just that we all grew
up with linux, so the solaris CLI "felt like a really old version of linux"
(to paraphrase a coworker) from the start.
We didn't slack, mind you. We tried hard to make that feeling stop. We read the
books and collected bigadmin bookmarks like trophies. We changed the default-shell
to bash and installed the GNU tools to keep our sanity but otherwise did our best
to treat solaris with respect and resisted the urge to dress it up to look more
like linux.
It didn't work out.
I could rant for days about the many little things that drove us away but
I'll try to focus on a few of the most significant points here:
1. Package Management (the lack thereof)
Pkg-add is bad joke when you're used to apt-get and emerge. JumpStart feels
like an insult when you're spoiled by FAI. I can only guess how Sun expects
us to keep our multiple solaris boxes in sync. Maybe they sell that as
one of their many enterprise service?
2. Google doesn't work well for solaris
Not really something we can blame on solaris or Sun but time after time we
were astonished as to how hard it is to find useful help for specific solaris
problems via google. Howto's and Tutorials about all things solaris are generally
very sparse. Due to this "learning by doing" doesn't work as well for solaris
as it does for linux.
3. The sun website SUCKS
Sure there is a lot of documentation, if you can find it in the pile
of rubble that sun calls a website. But even the stuff we found was
not always helpful. Sun documentation tends to be very verbose while
still often glossing over important details. Sun docs often feel like
they expect you to print them out and put them under your pillow.
We don't work that way. We're spoiled by straighforward howtos,
examples, stuff that gets us going fast. We're the impatient
youngsters.
Well, this got longer than I intended. I'll close with saying that
we'll keep buying sun hardware. The xfire series is the best piece of kit
(at a very competitive price) that I have ever seen and it runs linux
like a champ.
But linux as an "upgrade path to solaris"? Ha. Good joke.
In my world solaris has it's place on big iron and in areas where
the last bit of performance really matters. For everyone else the
natural choice is what they're familar with. And who grows up with
solaris these days?
Re:For those too lazy too read the article: (Score:4, Interesting)
As a cross-platform sysadmin with a decade-plus of experience, dude. Just... dude. My "Ineptitude" sense is tingling.
I mean, how much messing around on the command line do you really need to do on a daily basis? If the answer is "A Lot" - then I can sort of see Linux taking the advantage. Kinda. Yet, most, if not all, command line wrangling on modern servers is to munge shell scripts, and that's mostly been supplanted by python and ruby these days, anyhow. You've never even heard of JumpStart? "pkgadd -d" is too hard for you? You can't find anything on Solaris administration by googling the problem keywords with "sun Solaris administration" appended after? Really? Wow.
Not to be mean or anything, but seriously. If you can't handle Solaris, you shouldn't be in the sysadmin game. Only OpenBSD us easier for Unix wonks to tinker with. Stuff like AIX and NonStop would =break= you into a quivering pile of goop. Based on your post, I have doubts you'd be able to hack the more challenging Linux distros, too, like Slack or Debian.
There are good reasons to go with Linux on Sun hardware - pure speed, cross-platform compatibility (esp. with LAMP stack stuff), the need to tinker with the kernel to meet project objectives, or stripping down the OS to a bare minimum for performance or security advantages, re-purposing old hardware with an up-to-date OS that demands fewer resources. Then there are the reasons you gave.
I suppose if you really wanted to re-orient your entire computing platform around the needs of the sysadmins to play with the shell rather than the needs of the project to utilize its very expensive hardware to its fullest with modern OS features not yet available on Linux, your reasons are valid. Stupid, short-sighted, luddite and likely to get you fired at any other Unix shop of any size, but valid. I guess.
Re:Not welcoming your Scott McNealy overlord? (Score:3, Interesting)
It is very obvious, however, that it is not a clause that gives IBM any more rights than others, as you claimed. Were IBM to sell me a consumer product, they would be bound by the same rules as anyone else doing so.
Re:For those too lazy too read the article: (Score:3, Interesting)
Most of the younger staff here take a few months to get used to Solaris, it's the little things, netstat doesn't work as they expect (Why can't I just see the PID/binary name), ls doesn't have the switches they expect and to perfectly honest, while pkgtools is familiar to anyone who's used Slack etc kids today are spoiled with apt and they're right to like it, it "Just works"
We can't over modify the system as they need to be able to provide advice to clients with production Solaris environments so they have to learn for now , but I guarantee that all those shiny GNUisms will be available on Solaris in the medium term because Linux has given people an expectation of usability from the CLI that Solaris doesn't have @ the moment.
Was Linux born in 3 years? (Score:1, Interesting)
There were lots of people messing around with floppies and definately not the crowd that it has today. I'm trying to think back, was it 1991 or 1992 when the 0.99 was about? And by 1995 or 1996, was it so much better? Or the community?
The proper time frame to judge whether or not OpenSolaris will be able to build a successful community is somewhere in the 5 to 8 year time frame.
It is still too early to decide if OpenSolaris has failed or not to build a community.