OpenSolaris Governing Board Dissolves Itself

mysidia writes "Last month, it was mentioned that the OpenSolaris governing board issued an ultimatum to Oracle. It turns out that Oracle continued to ignore requests to appoint a liaison after the governing board's demands. This morning, the board unanimously passed a resolution to dissolve itself. Source code changes are no longer available, and it would appear that OpenSolaris and community involvement in the development of Solaris have been killed as rumored. We recently discussed a 'Spork' of OpenSolaris called Illumos. Perhaps now, this will have a chance at becoming a true fork."

Comments

What momentum may that fork have?

[alfredos]

With so much core OS work going to Linux and most of the remainder going to *BSD, which also has already ZFS well underway... What do theyhave to attract devs?

Re:What momentum may that fork have?

[Anonymous Coward]

A lot of core Solaris developers are already working on Illumos. This can become a great project, even better than Solaris itself. I expect to see many (Open) Solaris "users" move to support this project instead of supporting Oracle's one.

Re:What momentum may that fork have?

[mewsenews]

What do theyhave to attract devs?

They can probably attract curiosity seekers wondering what the living hell a "spork" is in a development context

Re:What momentum may that fork have?

[Anonymous Coward]

Sorry if you missed it but many of the OpenSolaris "devs" were either directly employed by Sun or companies with close ties to SUN. There was never any real grassroots development of OpenSolaris, and despite all the hype about what "may" happen post Oracle any further development is going exactly where it went before, nowhere.

Re:What momentum may that fork have?

[bsdaemonaut]

Nothing, but if we always did the sensible thing we'd miss out on much of the good software that we have today, such as Linux. There was a time that when it offered very little. It happened to be in the right place at the right time and today we get to enjoy what came about because of it. In regards to OpenSolaris, honestly the whole thing makes me a little sad. I realize commercial Solaris is still around, but it seems like every year we have less choices. I don't know about you, but I don't feel like that's a good thing.

Re:What momentum may that fork have?

[Kjella]

I'd say slim at this point. It doesn't help that all CDDL code is incompatiable with all GPL code so they got plenty wheels that need reinventing unless there's a BSD library for it. Yes, I know you can say the same about the GPL but there's not nearly as much that Linux would want. ZFS and DTrace would be two big ones though, but hopefully the concepts can be incorporated in Linux even if the code can not. Then again, if it'd been GPL then Linux would probably have scavenged all of it already, I guess that was the point of making it GPL-incompatible...

Re:What momentum may that fork have?

[Anonymous Coward]

Sure freebsd might have zfs support, but opensolaris' zfs implementation is THE implementation all other implementations compare themselves against.

Re:What momentum may that fork have?

[guruevi]

ZFS-stable? Seriously ZFS is one of the only points remaining besides the magnificent Fault Management Architecture, very good network stack, zones and dtrace.

Nexenta is doing a good job keeping the current Solaris distro's going but whenever Linux or BSD comes with a decent implementation of the most recent ZFS branch, there is probably going to be a massive conversion.

Re:What momentum may that fork have?

[Abcd1234]

Nothing, but if we always did the sensible thing we'd miss out on much of the good software that we have today, such as Linux. There was a time that when it offered very little.

Oh come on, that's revisionist history at best. When it was first released, it offered an alternative to Minix, and was one of the few protected-mode-capable Unix clones available for x86. As it progressed, it offered the first kernel (sorry Hurd) for a GNU-based OS.

Linux *always* had a niche to fill. I can't see how the same is true of OpenSolaris.

Re:What momentum may that fork have?

[diegocg]

A open and free version of solaris. Oracle is going to delay the release of the source code to make their propietary distro more attractive, but at some time they will release it. Illumos will offer a free version of that, and many people (including oracle customers) will want to use that. It bet it will be popular in the "solaris community". Also, there are companies like Nexenta which can try to develop new features. It won't be as nice as opensolaris was, but it's not the end of the world either. If I use solaris some day, I'm pretty sure it will be illumos.

Re:What momentum may that fork have?

[samkass]

I realize commercial Solaris is still around, but it seems like every year we have less choices. I don't know about you, but I don't feel like that's a good thing.

I kind of felt the same way in the late 90's when BeOS was dying and the MacOS's future looked bleak. Linux had extremely weak driver support, and OS/2 had finally given up the ghost. It looked like Windows might become the only survivor of the 90's. But today there is a new diaspora of OS distributions and platforms. These things ebb and flow. My advice is to not worry so much about choice in general and just try to find something you like and contribute to it.

Re:What momentum may that fork have?

[jd]

Why can't we have a knife! I wanna knife! >sniff<

(One of the big benefits of OpenSolaris is that there's a hell of a lot of commercial software for Solaris that hasn't been - and may never be - ported to Linux. This would matter less if the ABI/IBCS module had been maintained, as Linux could then run Solaris binaries natively.)

Re:What momentum may that fork have?

[Anonymous Coward]

whenever Linux or BSD comes with a decent implementation of the most recent ZFS branch, there is probably going to be a massive conversion.

If there is a decent implementation of ZFS in the Linux or the BSD kernel, Oracle will sue the hell out of them.

Re:What momentum may that fork have?

[metamatic]

This is exactly why I never even bothered downloading OpenSolaris. If they had put it under GPL I'd have been all over it.

Re:What momentum may that fork have?

[amorsen]

There's basically Windows, traditional Unix with X11 (with only Linux, *BSD and AIX left), and Unix + different UI (Mac OS X). Plus some embedded systems and some mainframe-like systems, but Unix is eating both.

We aren't exactly witnessing a Cambrian explosion.

Re:What momentum may that fork have?

[lennier1]

... there's not nearly as much that Linux would want. ZFS and DTrace would be two big ones though, ...

True. There's tons of Linux server admins who wouldn't mind having a proven ZFS implementation available to them.

Re:What momentum may that fork have?

[cgenman]

Don't forget iOS and Android as real platforms, and Chrome as a potential one. The phone boom has given us a splattering of new platforms, reminiscent of the server OS boom during the dot-com days.

Re:What momentum may that fork have?

[Anonymous Coward]

Don't forget iOS and Android as real platforms

Minus Android, when Oracle is done chewing that one up with its patent suits, with pieces of Dalvik stuck to its greedy chin.

Re:What momentum may that fork have?

[samkass]

Unix + different UI (Mac OS X)

and iOS, Android, Blackberry, Windows Mobile 7, WebOS, ChromeOS, Playstation/XBox custom OSes, and a few others. And clumping all "Linux" into one (from Ubuntu to Red Hat's Enterprise) is a bit of over-generalization.

In short, if you like playing around with new and interesting programmable systems, it may not be a Cambrian explosion, but we've certainly come out of the temporary bottleneck of the late 90's.

Re:What momentum may that fork have?

[ulzeraj]

I'm not a big fan of Solaris or even OpenSolaris, but recently I've been involved in a project to implement a storage system for the medium-small size company I work for. The storage would hold virtual machine disks over a 1g dedicated switch connection to Xen virtual machines that would be available via iSCSI. Naturally, ZFS comes in mind with the hability to just snapshot the machine and keep multiple backups of it without having to turn off the virtual machine. We chosed FreeBSD-8.0 implementation of ZFS with istgt as the iscsi target software. FreeBSD ZFS seemed pretty ok, but the iscsi connection was deadly slow with 4mb/s write speeds. I thought the reason was the ZIL thingie with sync write operations. I did some research here, asked a question there and a guy from #zfs @ freenode told me to use OpenSolaris. I dont really know if I was doing something wrong with FreeBSD+istgt. All I know is that the OpenSolaris COMSTAR is really, really amazing. I got it working pretty fast with no hassle and no problems looking for documentation. It is also very fast and my tests with disk speed reach the maximum speed allowed by the network. I have 12 virtual machine disks there (both hvm and paravirt) with daily snapshots and live migration between dom0's. I love FreeBSD but I think OpenSolaris is still years ahead with this storage thingie.

Re:What momentum may that fork have?

[bsdaemonaut]

FreeBSD is porting ZFS, not reimplementing it, so it is the very same version that is used in OpenSolaris. Admittedly FreeBSD is not current with the newest ZFS implementation, they are currently using v15 I believe, but they are working to get there.

Re:What momentum may that fork have?

[Kismet]

If the parent's view is wrong (and maybe it is--I don't know), this is almost certainly not because of "revisionist history."

Revisionist history happens, legitimately, when historians review the best sources available and arrive at a conclusion about a story that does not entirely agree with how it has been told in the past. When history is "revised" in this way, it is because the old stories have been based on sources that are less reliable, incomplete, contradictory, or of lesser quality.

Something similar happens in science, when new evidence or understanding comes to light and theories are changed or improved.

For some reason, when it comes to history, "revisionist" somehow always implies "wrong."

Re:What momentum may that fork have?

[rayvd]

Linux *always* had a niche to fill. I can't see how the same is true of OpenSolaris.

ZFS.

And don't say BSD... the BSD port of ZFS is way behind OpenSolaris' feature and performance-wise.

Re:What momentum may that fork have?

[Darinbob]

It's a larger gene pool. That's important by itself.

For open source operating systems, I found OpenSolaris codes that I looked at much easier to read and understand in many places that the terse and cryptic BSD and Linux ones.

Re:What momentum may that fork have?

[Abcd1234]

ZFS.

That's not a niche, that's a technology, and one easily co-opted (the BSD port will improve eventually) or superceded (btrfs).

Re:What momentum may that fork have?

[CAIMLAS]

Well, it's not so much any one thing, but a combination of things make it attractive to me as an administrator.

First, FreeBSD's ZFS may be "well underway", but it's showing no signs of being usable any time soon. Let it suffice to say that anyone paying attention or using FreeBSD ZFS for much more than one or two small servers is likely to agree that their implementation is not "enterprise ready" as they so arrogantly claim.

Second, I'm not so stupid as to fool myself into thinking ports on BSD is a sustainable administrative tool. Nexenta, and I believe Illuminos, use apt.

FreeBSD appears to be in decline as a project. I can't speak for developer activity, but I can say that their ability to actually ship code that works has become diminished since 7.1 or so. Entire subsystems have not worked for quite some time, yet they keep shipping it and saying "it'll be fixed in a couple years" (referring to USB and AHCI). Quite a few drivers have also had regressions.

In addition to ZFS, the Solaris kernel has dtrace, zones, and BrandZ.

Linux will NEVER get ZFS support in mainline.

ZFS on Solaris/OpenSolaris/Nexenta is usable today. Not only does it "have" it, but you're able to trivially export an iSCSI device, use deduplication, and (not 100% sure on this one, just read about it having been added to Solaris in April) do differential filesystem snapshots. FreeBSD's implementation has none of this, giving it little more appeal than current btrfs on mdraid (and in some ways, less).

Re:What momentum may that fork have?

[CAIMLAS]

I may be mistaken, but haven't many of those Sun/Solaris developers kicked off the Oracle train, moving to companies like Nexenta where they can continue working on The Next Big Thing?

I'm not sure on the numbers but I know at least several of the important ones have.

Re:What momentum may that fork have?

[Anonymous Coward]

I'm the poster of the above. So stating the bleeding obvious while posting as AC gets modded +5 while doing similar posts for a couple years with an account got jack nothing? HAHAHA Slashdot sooooo sucks donkey balls.

Re:What momentum may that fork have?

[c6gunner]

That's true for the moment - ZFS is the only reason I use OpenSolaris - but it won't last for long. As soon as the BTRFS folks develop a RAIDZ equivalent, I plan on switching back to linux.

Also, as has already been pointed out, ZFS is available on BSD. With OpenSolaris development effectively stopped I've seriously considered switching, but figured I might as well wait a while and see where BTRFS is in a year or so. However, if I needed to build a new ZFS based file server today, I'd definitely go with BSD.

Re:What momentum may that fork have?

[ChunderDownunder]

Symbian... Nokia are still (perhaps barely) alive and shipping.

There's an effort, "Wild Ducks" to port it to 'generic' (ARM) hardware, i.e. Beagleboard. It might make a nice tablet OS alternative to Android, for those who prefer writing software in Qt, if someone would port it to the touchbook.

Re:What momentum may that fork have?

[Jane Q. Public]

No, that's NOT what's commonly meant by "revisionist history". The commonly accepted meaning is to re-write the history books to tell something other than the true story. As such, it never happens legitimately.

Technically, the phrase "revisionist history" could mean several different things. But as actually used, it doesn't.

Re:What momentum may that fork have?

[jeremyp]

iOS is Mac OS X with another different UI. Chrome is a browser and Android is Linux with another different UI.

Re:What momentum may that fork have?

[Pieroxy]

The point is that they worked on OpenSolaris not because it was their passion, but because it was their job. They've moved on.

Re:What momentum may that fork have?

[SanityInAnarchy]

Maybe. They might also take it as a good thing and use it in their "Oracle Unbreakable Linux".

Re:What momentum may that fork have?

[Anonymous Coward]

traditional Unix with X11 (with only Linux, *BSD and AIX left)

Oh, c'mon. Solaris isn't dead *yet*.

We've still got some work to do. ;-)

Re:What momentum may that fork have?

[TheRaven64]

It's also worth noting that a significant part of the ZFS 'port' for FreeBSD is a solaris-compatibility kernel module, which implements the kernel APIs that ZFS and DTrace need from Solaris. This means that each update to ZFS is less effort. In theory, they just need to grab the code and compile. In practice, they need to make a few tweaks each time, but the more mature the compatibility layer gets, the fewer tweaks are needed.

Re:What momentum may that fork have?

[TheRaven64]

How? The ZFS implementation in the FreeBSD tree is from Solaris, released by Oracle, and distributed in accordance with the terms of the license granted by Sun / Oracle.

Re:What momentum may that fork have?

[Galactic Dominator]

[quote]First, FreeBSD's ZFS may be "well underway", but it's showing no signs of being usable any time soon. Let it suffice to say that anyone paying attention or using FreeBSD ZFS for much more than one or two small servers is likely to agree that their implementation is not "enterprise ready" as they so arrogantly claim.[/quote]

Hmm, it's been usable for many of us for quite some time. Not to say there haven't been problems with it, but less the equivalent point in OpenSolaris history since FreeBSD is able to import well tested code and patches for issues the first gen ZFS had in OpenSolaris.

[quote]Second, I'm not so stupid as to fool myself into thinking ports on BSD is a sustainable administrative tool. Nexenta, and I believe Illuminos, use apt.[/quote]

Again, many of us have used ports with great success for a long time. Personally, I find it easier than working with deb packages, and far easier than working with rpm's.

[quote]FreeBSD appears to be in decline as a project. I can't speak for developer activity, but I can say that their ability to actually ship code that works has become diminished since 7.1 or so. Entire subsystems have not worked for quite some time, yet they keep shipping it and saying "it'll be fixed in a couple years" (referring to USB and AHCI). Quite a few drivers have also had regressions.[/quote]

Wow this is like a broken record. The are no problems with either AHCI or USB for the vast majority of users. There were some changes to USB in 8.0, eg switched to being based off of libusb, device renaming, etc. AHCI and some the other related controller drivers were new in 8.0 as well, There were a few corner cases that didn't work well, but it's seen widespread usage with a great deal of success. The updates included in 8.1 address the corner cases and some further performance improvements. It's hard knowing what your issue is since you speak in generalities, but I assure you there are far more happy FreeBSD users than people like you.

[quote]In addition to ZFS, the Solaris kernel has dtrace, zones, and BrandZ.[/quote]

It's true OpenSolaris's zone are more advanced than FreeBSD jails, but for most uses the differences negligible. BSD Jails also integrate nicely with ZFS and FreeBSD's GEOM layer. OpenSolaris also has XEN dom0 support, I'm surprised you didn't take the opportunity to bag on FreeBSD's lack of it.

[quote]ZFS on Solaris/OpenSolaris/Nexenta is usable today. Not only does it "have" it, but you're able to trivially export an iSCSI device, use deduplication, and (not 100% sure on this one, just read about it having been added to Solaris in April) do differential filesystem snapshots. FreeBSD's implementation has none of this, giving it little more appeal than current btrfs on mdraid (and in some ways, less).[/quote]

It's trivial to export ZVOL on FreeBSD as iscsi targets. Granted it's not quite as nice as OpenSolaris since FreeBSD doesn't have an iscsi target in it's base system but there is a great one in ports and with a two minute wrapper script you have exactly the same functionality.

People talk of deduplication like it's some sort of magic bullet, but I think most of those people have no idea the overhead that imposes on the file system. Once they discover the resources necesscary to run it effectively, much of the enthusiasim fades away. It's a really a very select usage were deduplication would be economically feasible to implement.

Doesn't seem like you understand much about ZFS. Differential snapshot are integral to ZFS, you can't have ZFS with getting it so yes FreeBSD does have differential ZFS snapshots. Perhaps you're misidentifying this? http://netmgt.blogspot.com/2010/03/zfs-snapshot-differences.html Interesting, they have integrated diff into ZFS now. Trivial to do without the integration, if that's actually a make or break feature you need a new sys admin.

While you're getting some learning here, you s

Re:What momentum may that fork have?

[Kismet]

Yes, I am quite aware of how the term is "commonly" used, which is precisely the precondition for my complaint. Otherwise, why complain at all? I'm not sure what your argument is.

If we distinguish "junk science" from actual "science," why not "junk revisionism" or "negationism" from legitimate "revisionist history?"

Since the vocabulary needed for talking about a worthwhile and valid way of reexamining history has been overloaded to mean the same thing as its false and invalid counterfeit, legitimate revisionism suffers.

The problem is compounded every time someone pulls out the fallacy, "That's revisionist history!"

Imagine if politically motivated rubbish and honest research required the equivocal term, "science". One of those disciplines, the one more difficult to justify to the layman, might suffer as a result. Whenever we wanted to debunk something, we'd just hurl the epithet, "science", at it.

That's what I'm complaining about. It's just a crazy, petty Stallman-esque neurosis I have.

Re:What momentum may that fork have?

[Achromatic1978]

Nokia are still (perhaps barely) alive and shipping.

LMAO. "Perhaps barely alive". Uhh, okay, if Q2 2010 profits of US$870M on sales of US$12.7B is barely alive, sure, I guess.

Re:What momentum may that fork have?

[mysidia]

The FreeBSD "port" is more than just behind. It is missing important functionality that was even in the version they ported. It also has some bugs and does not perform that well.

Performance is the big thing. One of the advantages of ZFS is supposed to be its improved performance, in particular it is faster than UFS on server hardware of a large enough scale. But on BSD, ZFS is slow.

Issues abound with things like 'zfs recv' that "just work" on Solaris.

I am sure the BSD devs will do something to work out all these issues eventually, if ZFS support is to continue to be developed; however, it could take a long time for BSD to catch up.

ZFS is complicated enough and different enough from other filesystems that it is not likely to be able to simply be 'ported' or grafted onto BSD or any other OS and work well, without many years of work.

Meanwhile ZFS is stable on Solaris, performs remarkably when setup correctly compared to equivalent config on BSD, and works great.

Re:What momentum may that fork have?

[Anonymous Coward]

Well that's a load of bull if I ever heard one. Your entire post needs a giant [citation needed].

I use FreeBSD ZFS at $work and it works great. Combine it with HAST and CARP and you have a pair of redundant application servers with instant data replication and failover. All on FreeBSD, all very, very usable.

The FreeBSD ports system is awesome IMO, I never understood why some people gets all worked up over it. There are advantages and disadvantages in all package management systems, but I've found FreeBSD ports (with Portmaster being my tool of preference) to be flexible and nice to work with. Maybe it's been a while since you used FreeBSD ? If so, try out Portmaster :)

Saying that the FreeBSD project as a whole is in decline is pretty strange considering the size of their "Quarterly Status Report" filled with all kinds of exciting things. See for example the latest one: http://www.freebsd.org/news/status/ - or the one before that.

Saying that the USB or AHCI subsystems hasn't worked for years is also, obviously, very wrong. For example, the AHCI driver works very well for my SATA drives, and it includes SATA hotplug support and whatnot. The USB system has worked fine for everything I've used it for - from 4.x through to 8.1 today.

The only thing I'll grant you is that deduplication and a couple of other nice ZFS features are not yet in FreeBSD, and they would indeed be nice to have. But saying ZFS on FreeBSD is useless because of this fact is a bit harsh don't you think ?

You may be happy with Solaris but please stop spreading FUD, check your facts, and have a nice day.

ps. sorry for AC, forgot my login somewhere, been a while since I last posted.

Re:What momentum may that fork have?

[fotoguzzi]

FreeBSD appears to be in decline as a project.

Has Netcraft confirmed it?

Re:What momentum may that fork have?

[avgapon]

Your FreeBSD bashing seems to be lacking in facts or examples. Or do you just want to attract points from Linux fans?

Re:What momentum may that fork have?

[ChunderDownunder]

yes i was being somewhat tongue in cheek, with the proviso that android and ios are flavors of the month. Still, i think the roadmap for symbian^4 looks promising as a complement to meego.

Re:What momentum may that fork have?

[gmhowell]

traditional Unix with X11 (with only Linux, *BSD and AIX left)

Oh, c'mon. Solaris isn't dead *yet*.

We've still got some work to do. ;-)

Queue "The Dead Parrot Sketch" or "Bring Out Your Dead" from Holy Grail.

Re:What momentum may that fork have?

[TheGratefulNet]

I've been a freebsd user since freebsd 4.x (heavily) and linux user since the 1.2 kernel days. started off using DEC unix and vax/vms. I'm no newcomer to unix and unix-like systems.

what the poster says seems right to me. freebsd had its day but its just not keeping up with linux and even stability is lacking (had my freebsd8.0 system lock up on me and that never happened with freebsd 6.x and 4.x; at least not mainline versions). bsd simply is less travelled and still supports less hardware than linux. and linux has become so good that the argument of 'bsd is more stable' is just not true anymore.

what's the story with bsd and samba? why is it always always slower than linux samba on the same hardware? curious, that.

what's the story with TRIM and SSD on freesbsd? is there any story? linux is at least 'mostly there' (just not turnkey in their distros). bsd will lag, as usual.

I'm about to retire my freebsd8 system once I get all my data off the zfs pool. then I close down that system and re-use the hardware.

the one shining thing that kept me on bsd was zfs; and with my recent zfs hiccup, I'm now willing to leave zfs behind and freebsd along with it.

1 less system type to have to manage, as well.

Re:What momentum may that fork have?

[sjames]

They don't allow knives, belts, or shoelaces when you're on suicide watch.

Re:What momentum may that fork have?

[Anonymous Coward]

Android = Unix + different UI
WebOS = Unix + different UI
Chrome OS = Unix with X11

Blackberry, WM7 and console stuff are definitely their own thing. Not sure where iOS fits.

Re:What momentum may that fork have?

[anilg]

Nexenta, and I believe Illuminos, use apt.

Nexenta uses apt. Illumos as yet is packaging independent. It's chosen by downstream distributions.

Re:What momentum may that fork have?

[LordLimecat]

Andriod IS Linux, as are ChromeOS and WebOS.
Windows Phone 7 is (according to wikipedia) windows CE. Xbox (the first) used the NT kernel; as best as I could tell, Xbox360 uses some kind of windows kernel as well.
iOS is based on Mac OSX, so again, not really a new OS here, just a modified one.

That basically leaves Blackberry OS, the playstation OS, and whatever your "few others" are. As Playstation doesnt really count (you cannot run anything but that software on your PS3-- hypervisors aside-- and you cannot run that software on anything else), and neither does Blackberry (ditto), that really doesnt add much to the list at all.

Re:What momentum may that fork have?

[BitZtream]

If this was modded funny, I'd let it go but um, ZFS in FreeBSD has been considered production ready by most for a while now, and officially so since 8.0 was released.

Ports are used by at least 4 BSD based releases that I know of, it works, of course, theres also packages too if you don't want to build the apps.

You do realize there have been 4 releases since 7.1 right? 7.1 was released in Jan 09 ...

FreeBSD has dtrace, jails and I don't know to what level FBSD has support for BrandZ type features, it was just starting out last I looked a few years ago, I don't use jails for much though so I'm probably just out of date.

Uhm, FreeBSD is fully capable of exporting zvols via iSCSI, I don't think dedup is in yet, and if solaris just got differential filesystem snapshots (I'll have to figure out what that is considering how snapshots work in the first place) so you got something there, I want dedup too, I could drop compression in favor of dedup and get far better performance but otherwise ZFS on FreeBSD with sufficient ram works great. RAM is the key though isn't it.

Shit, I just got trolled didn't I?

Re:What momentum may that fork have?

[BitZtream]

I'm trolling but ... how sad is it that you missed the experience with OpenSolaris, that no doubt would have been useful to you just from learning about a different way of doing things ... and the only reason was because the license wasn't GPL compatible.

Thats pretty sad.

You probably are better off for not trying OpenSolaris at this point, it sucks, but still, you completely ignore things because your license is so restrictive? CDDL doesn't prevent you from using their code, GPL prevents you from using CDDL code. I just can understand how you can be such a tool and not realize it, let me guess, you're a goth too, right?

Re:What momentum may that fork have?

[mulvane]

As of FreeBSD 8.1 with ZFS v14 I found it quite stable and usable enough to move it from my solaris install. the rig is a 16 core xeon with 96GB of RAM and 2 areca 24 port raid controllers with 4GB of cache each. Now, the 52TB of data configured into various raidz's and joined into the zpool all work quite nicely. And as of testing with separate drives, I have found it a VERY fast solution maxing connectivity across the 2 gig-e interfaces I have currently enabled on the machine. When you open your mouth, its a good idea to know WHAT you are talking about.

Re:What momentum may that fork have?

[TangoMargarine]

CDDL doesn't prevent you from using their code, GPL prevents you from using CDDL code.

The CDDL was a modification of the Mozilla license specifically made to be incompatible with the GPL.

Re:What momentum may that fork have?

[Anonymous Coward]

Don't forget Nucleus, VxWorks, QNX, Symbian OS...all RTOS prevalent in embedded devices.

Re:What momentum may that fork have?

[rivaldufus]

Solaris is still around, and probably will be for a few more years. For whatever reason, Oracle loves the idea of running oracle on sparc Solaris. I guess it's rather "enterprisey."

Re:What momentum may that fork have?

[eosp]

If there is a decent implementation of ZFS in the Linux or the BSD kernel, NetApp will sue the hell out of them.

Fixed that for you.

Re:What momentum may that fork have?

[pevans]

One word: btrfs

All the posts here mostly talk about the wonders of zfs. I was enamored of zfs myself when it first came out.

Thing is, even if they got rid of the license problems, the thing is still not very linux kernel-friendly. The way it is, it will only ever exist in user-land/fuse. Yes, it can be made to work. So what?

Btrfs is better, uses zfs ideas and leaverages the linux kernel natively.

  http:// en.wikipedia.org/wiki/Btrfs

Re:What momentum may that fork have?

[TyFoN]

The irony is that BTRFS was started by Oracle.

Re:What momentum may that fork have?

[EEPROMS]

Linux would want. ZFS and DTrace would be two big ones
errr why would I want ZFS and Dtrace when Linux already has better alternatives "now". ZFS is nice but btrfs is now in many cases catching up to ZFS fast and in some new benchmarks surpassing ZFS. Also ZFS has just as many things missing, as there, especially when looking at the enterprise side of things it is far from complete and btrfs now has many of the missing bits in place.

Re:What momentum may that fork have?

[Nyder]

Unix + different UI (Mac OS X)

and iOS, Android, Blackberry, Windows Mobile 7, WebOS, ChromeOS, Playstation/XBox custom OSes, and a few others. And clumping all "Linux" into one (from Ubuntu to Red Hat's Enterprise) is a bit of over-generalization.

In short, if you like playing around with new and interesting programmable systems, it may not be a Cambrian explosion, but we've certainly come out of the temporary bottleneck of the late 90's.

Don't forget AmigaOS!!!

Re:What momentum may that fork have?

[Anonymous Coward]

> let me guess, you're a goth too, right?

Way to call yourself out as a dick and invalidate anything else you said that might have had any worth.

> I'm trolling but...

No kidding.

Re:What momentum may that fork have?

[amorsen]

And clumping all "Linux" into one (from Ubuntu to Red Hat's Enterprise) is a bit of over-generalization.

I really don't think it's an over-generalization. E.g. Sun OS vs. Solaris was a MUCH larger difference than whether the package manager is apt or yum.

As for the rest, LordLimecat said it better than I could.

Re:What momentum may that fork have?

[amorsen]

From a commercial viability view point, Solaris has pretty much been dead since the tech crunch. It took a while for Sun's money to run out, but now it has, and there is no way Oracle can bring Solaris back to profitability without almost completely killing development.

Itanium killed Irix and HP/UX, although the latter is still pretending to be alive.

Re:What momentum may that fork have?

[drinkypoo]

(One of the big benefits of OpenSolaris is that there's a hell of a lot of commercial software for Solaris that hasn't been - and may never be - ported to Linux. This would matter less if the ABI/IBCS module had been maintained, as Linux could then run Solaris binaries natively.)

When I ran Solaris for x86, the available software was pretty limited. I can't see how that wouldn't be even more true of OpenSolaris.

Re:What momentum may that fork have?

[bsdaemonaut]

Have you actually used the version included in 8.x? It's much improved -- I believe it's even considered production ready. I'm not sure how well you can compare the UFS versions between Solaris and FreeBSD so relative performance expectations between FreeBSD's ZFS and UFS seem premature. It's worth noting though that in the current version of FreeBSD, in the benchmarks that I've seen, the ZFS numbers match or exceed UFS+S/UFS+J.

http://www.phoronix.com/scan.php?page=article&item=zfs_ext4_btrfs&num=5

The other response to my above message was also pretty informative.

Re:What momentum may that fork have?

[drinkypoo]

and iOS, Android, Blackberry, Windows Mobile 7, WebOS, ChromeOS, Playstation/XBox custom OSes, and a few others. And clumping all "Linux" into one (from Ubuntu to Red Hat's Enterprise) is a bit of over-generalization.

iOS is based on OSX. Android is Linux. Blackberry actually has its own OS. Windows Mobile 7 is just a shitty rehash of WinCE, which ought to have been mentioned just to bag on it. WebOS runs on Linux. Chrome OS runs on Linux. Playstation has its own OS that nobody cares about, but Xbox has rehashed windows (Win98 in Xbox, WinNT in Xbox 360.) So you're creating more operating systems than actually exist. Besides, you left out Symbian (circling the drain but still with us) and BREW. Wait, those are no more relevant than Blackberry OS, which I suspect is ALSO on its way out. It offers nothing compelling.

Re:What momentum may that fork have?

[tehcyder]

the MacOS's future looked bleak.

The Mac OS had no future. It was killed off and replaced by BSD (running on Intel hardware) with some pretty widgets on top.

Re:What momentum may that fork have?

[ulzeraj]

" Not only does it "have" it, but you're able to trivially export an iSCSI device, " zfs shareiscsi isnt working. You have to set up COMSTAR for that.

Re:What momentum may that fork have?

[CAIMLAS]

And by moved on, do you mean they're working on OpenSOlaris for someone else now, but under a different name? Because that is happening.

Re:What momentum may that fork have?

[Bengie]

I don't know about your company, but my cousin works in a data center with 13,000+ harddrives. They're not allowed to use a file system unless it's been tested and used in enterprise settings for at least 5 years.

I would assume that most enterprise situations require tested and reliable file systems and not some buggy alpha grade FS.

I googled BTRFS within the last month to see how it was doing and I got lots of people talking about stability issues and lost data. Sounds like a great platform to store several petabytes of data you can never recreate.

In the future it may/will be good, but for now ZFS is the only *great* FS

Re:What momentum may that fork have?

[CAIMLAS]

As I was reading this, I had one of our FreeBSD 7.2 machines kernel panic due to a ZFS/opensolaris kernel crash. No, the pool is not approaching full, and yes it is "properly" tuned. This is, unfortunately, hardly an isolated incident, as several ZFS and yes, other non-ZFS FreeBSD machines of ours have stability issues. Several have been converted over to Linux and have ceased having these problems.

Re:What momentum may that fork have?

[CAIMLAS]

Oh, good God, where to start....

If this was modded funny, I'd let it go but um, ZFS in FreeBSD has been considered production ready by most for a while now, and officially so since 8.0 was released.

So, what - they decided to compromise stability elsewhere to improve ZFS stability? Because I've certainly seen a decrease in other subsystems, too.

I've had three ZFS/opensolaris kernel module related crashes in FreeBSD in the last month alone - on different hardware and releases. It crashes in scenarios where the "evil" NTFS would not only be OK, but in the even of an actual failure would be recoverable.

What the FreeBSD developers consider "production ready" and what everyone else considers "production ready" are different things. It's one thing to eat your own dog food, but actually listening to what others are experiencing is actually pretty useful.

Besides, isn't the latest buzz phrase "enterprise ready ZFS" for FreeBSD? Not even close: I can give a half dozen use cases where it falls flat (some of which have been commented on elsewhere, but it shouldn't take long to find more examples).

Here's a hint: "production ready" or "enterprise ready" filesystems do not need tuning for basic stability, only for performance above and beyond typical work loads. ext3 is "enterprise ready"; XFS is "enterprise ready"; NTFS is "enterprise ready". ZFS is not, at least on FreeBSD.

Ports are used by at least 4 BSD based releases that I know of, it works, of course, theres also packages too if you don't want to build the apps.

That's a line of complete shit. In building a system it is not uncommon to run into the following scenarios with fairly commonly used packages:
* The package does not work with what everyone else uses (eg. apache 2)
* The package has some other unmet dependency which is unavailable
* Deviating from the default config results in broken dependencies
* Available ports commonly have security issues several weeks+ old.

I'm sure I could come up with a couple others, if I cared.

Furthermore, the whole

You do realize there have been 4 releases since 7.1 right? 7.1 was released in Jan 09 ...

Re:What momentum may that fork have?

[Jane Q. Public]

I won't argue with that. I misunderstood, and thought you were saying that the meaning of "revisionist" as used was simply to make a revision. But that is the literal meaning, of course. As for "otherwise why complain", that's simple: it appeared as though you were complaining that people were hurling the phrase "revisionist" at any kind of revision, legitimate or not. Which of course isn't the typical case, thus your actual objection.

As I know full well from my research into political history: a lot of the history that "everybody knows" and that they were taught in school in the U.S., has been distorted over time for political ends. It doesn't match the actual historical records. (Never underestimate the political agenda of a school board. For example, look at the boards that choose textbooks in many states, attempting to insert creationism into the science curriculum when it just plain isn't science.) At least I can say with authority that it has turned out that much of the history I was taught in school was distorted, and I was in a pretty standard public school system. Which means at some point it was "revised" by "revisionists". And if we ever hope to get it back on the factual track, then it will have to be "revised" again, to reflect the actual historical record.

Is it really dead?

[Tubal-Cain]

Has Netcraft confirmed it?

Re:Is it really dead?

[Frequency Domain]

It's not dead, Oracle just told them to go fork themselves.

Re:Is it really dead?

[Anonymous Coward]

It's not dead. It's sleeping.

Uses for Opensolaris

[cc-rider-Texas]

I have to use opensolaris on my intel boxes because I have to use gcc-2.95.3 to compile some ancient software I use for research, and not because of some fancy file system or dtrace. I can't help but wonder if other people are in the same boat. Anybody?

Re:Uses for Opensolaris

[Anonymous Coward]

U mad bro?

Re:Uses for Opensolaris

[Anonymous Coward]

I MAD

Re:Uses for Opensolaris

[Anonymous Coward]

I have to use opensolaris on my intel boxes because I have to use gcc-2.95.3 to compile some ancient software I use for research, and not because of some fancy file system or dtrace. I can't help but wonder if other people are in the same boat. Anybody?

Thankfully, no.

Re:Uses for Opensolaris

[Hatta]

Why can't you use gcc-2.95.3 on Linux?

Re:Uses for Opensolaris

[LWATCDR]

And you can not get it to compile under a more modern version?
Nope but ewwwww.

Re:Uses for Opensolaris

[Anonymous Coward]

If your app requires GCC extensions, since you didn't mention trying the Intel CC, and the compilation failure is the result of deprecated features, really your only viable option is to patch the software so it compiles with a modern GCC. You should do that anyway, for the sake of code portability and ongoing maintenance.

Re:Uses for Opensolaris

[cc-rider-Texas]

For my 64 bit machine, both opensolaris and gcc-2.95.3 and the ancient software will install/compile seamlessly, including my T61 laptop. I have installed it on an older 32 bit ubuntu install, but the old software doesn't like linux very much, and will not compile under any other gcc either. Opensolaris works just fine for it though, especially considering that the old software was developed for unix in the early 90's anyway. I just really don't want to get into modifying the gcc configure files that would allow gcc 2.95.3 to be compiled on a newer 64 bit machine with linux on it.

Re:Uses for Opensolaris

[rubycodez]

why not just use the old RedHat 7 or similar that ran that moldy ol' 2.95.3? if you're behind firewall shouldn't be problem as long as you're not too intimate with internet

Re:Uses for Opensolaris

[NuShrike]

Can't you get that through FreeBSD and anything with a Pkgsrc derived system that doesn't peg binaries to OS releases? (aka RPM Hell).

Re:Uses for Opensolaris

[Antique Geekmeister]

Finding a compiler and build system ancient enough to deal with the no longer POSIX compliant unfrastructure for gcc 2.9x compilers is fraught with pain for inexperienced engineers. Rolling your build environments that far back is _awfully_ painful.

Re:Uses for Opensolaris

[grouchomarxist]

Yeah, that's what I was thinking. If it is worth using, it might be worth porting. In the long run some time spent getting it working on newer systems will free you up in a number of ways.

Perhaps write up a number of conformance tests to make sure it works as it did before.

Honey,

[Anonymous Coward]

Does this mean you won't spork me tonight?

Re:Honey,

[Anonymous Coward]

Only if you forck me first...

Meanwhile, in Redwood Shores...

[Anonymous Coward]

The Open Solaris council will no longer be of any concern to us! I've just received word that Emperor Ellison has dissolved the council permanently. The last remnants of Sun have been swept away.

From now on, fear will keep potentially traitorous Solaris users in line. Fear of our software patents - and our new super death-ray powered ELAs!

Re:Meanwhile, in Redwood Shores...

[Anonymous Coward]

Do you think Emperor Larry would stroke my gel sack?

bOrg

[mark72005]

When will we get an Ellison/Borg icon for /. ?

Re:bOrg

[Mongoose Disciple]

Maybe it could be Ellison-as-Hitler instead. Just go ahead and pre-bust the Godwin cherry on all the Oracle stories.

Re:bOrg

[l0ungeb0y]

I see, so Oracle, and Apple = BAD/EVIL
While MS = GOOD/TRUSTWORTHY

Arbitrary logic is arbitrary.

Re:bOrg

[twoears]

Considering that Ellison is Jewish, the Hitler thing could be poignant.

Re:bOrg

[AnonymousClown]

GP:

Don't know why anyone ever trusted Oracle or Ellison. Ill never trust anyone who is too much of a pussy to eat meat.

I see, so Oracle, and Apple = BAD/EVIL While MS = GOOD/TRUSTWORTHY

Arbitrary logic is arbitrary.

Yes, indeed.

Re:bOrg

[Anonymous Coward]

How about an Ellison icon with a gas mask on. Have a link to a comment: 'Oracle, taking care of Sun one department at a time.'

Captcha: Slander

Re:bOrg

[Anonymous Coward]

There are no shortage of kapos these days, sadly. So his ancestry is irrelevant.

If Oracle starts losing money, maybe we'll see a Ellison/Downfall parody on You Tube.

Re:bOrg

[Anonymous Coward]

When he allows it?

Re:bOrg

[interval1066]

After the Jobs one.

Re:bOrg

[Anonymous Coward]

have you never seen the Gates/Borg icon for /. ?

Re:bOrg

[Anonymous Coward]

Hitler's Jewish ancestry

You could be on to somethng!

Re:bOrg

[martin-boundary]

Shouldn't that be more of a Ellison/Leisure Suit Larry icon?

Re:bOrg

[zigfreed]

Any company who's definition of 'compete' is 'sue into the ground' is evil. I.e. MS vs. Salesforce.com , the recent Oracle vs. Google, and Apple... well Google doesn't show any recent 'Apple sues' that's actually related to competition. Looks like just the Gates and the Ellison.

Rotten

[dandart]

I think Oracle are jolly rotten. I wish the very best to the Illumos team the very best of luck, and the continuation of development.
This is pivotal, as ZFS is just spiffing although its license is less than optimal.
I'd join in if I could program in C.

Good day!

Oracle shooting his own foot.

[mr exploiter]

This move by Oracle reminds me yahoo in the .com era. Open Solaris was not a revenue source but it was important as the means to get developers interested in Solaris. I don't think there will be much development or support for solaris from the open source community from now on.

Re:Oracle shooting his own foot.

[Anonymous Coward]

This move by Oracle reminds me yahoo in the .com era. Open Solaris was not a revenue source but it was important as the means to get developers interested in Solaris. I don't think there will be much development or support for solaris from the open source community from now on.

Maybe, but OTOH Oracle was very profitable without much contribution from the open source community, while Sun despite having open sourced tens of millions of lines of code had a long sequence of black zeros and red zeros at the bottom of the balance sheet.

Re:Oracle shooting his own foot.

[khb]

Except that there is absolutely no evidence that it got any incremental developers "interested in Solaris". Thinking it would in the future is probably wishful thinking.

Feel good ranting amongst ourselves isn't evidence. Yes, we like Open Source, otherwise we wouldn't be /. regulars.

Developers could, and did, get interested in Solaris from Solaris Express before it was killed to make OpenSolaris the "one way forward".

I haven't been with Sun for some years, and was never with Oracle. But the people there did have metrics, and while people may say unkind things about Larry Ellison, it is a mistake to think him (and his staff) stupid.

Nor are they purely motivated by $$/quarter, or they never would have invested in Sun.

That isn't to say profits aren't important to them; but Schwartz converse ("we lose money every quarter, but we'll make it up in future revenue") was suicidal and the eventual result is now clear even to the most optimistic.

Turning Solaris into an open project might well have worked better earlier (think SunOS 4.1.x ... could have gone BSD ... since it started as BSD ;>). The reasons for not doing that are indubitably financial (ATT paid Sun many $$$ directly and indirectly) and it's hard to go back (even if Management had thought that a wise course).

It clearly didn't have the desired results, and if Oracle Management believes they can get effective change faster by narrowing focus, one can see why they are interested in trying.

Re:Oracle shooting his own foot.

[drinkypoo]

Except that there is absolutely no evidence that it got any incremental developers "interested in Solaris". Thinking it would in the future is probably wishful thinking.

What it does is keep systems administrators interested in Solaris. If I can only run Solaris at home on Sun hardware, I won't bother, because it is not power efficient by modern standards unless you have new, expensive machines. If I run Linux at home, I want to run Linux at work, so that the skills I develop at home will carry over directly, and vice versa.

At least, it did that to SOME people... I couldn't be arsed about it because of the deliberately chosen incompatible license and the horribly crap hardware support. Sadly, Solaris for x86 had the same problem back when it was a shipping product. Due to that failure, I installed a whole bunch of Linux at a formerly all-Sun shop for which I worked. They had people sitting at SS1s and SS2s and I came in and installed P233MMX systems with 20" viewsonic monitors for less than half the price of buying used sparcstations with blurry old Sony monitors. Sun has never put in the time to have enough hardware support on x86 to make their software credible there.

OpenSolaris Governing Board Dissolves Itself

[Anonymous Coward]

Wicked Witch was not available for comment.

Only a good thing

[MonsterTrimble]

I think this is a good thing for everyone. Hear me out, I'm not an idiot, or at least I'm trying not to be!

Solaris, unlike the other big two open source operating systems - Linux & BSD - has always had the problem of the double edged sword because it always had to serve Sun, which in turn supported it. Now with Oracle punting Solaris to the curb, the community can really see what it's made of. Whether Solaris is an also-ran or if the community can really go full tilt and go in directions that it couldn't because of it's omnipresent master is entirely up to the people who support it. The future is bright for smaller form factors, as well as servers (again) so now it's time to see if they are big time players or another name to the pile of failed OSes.

I wish them well.

Re:Only a good thing

[bsdaemonaut]

It's worth noting that Oracle is not kicking Solaris to the curb, they are kicking OpenSolaris to the curb.

Re:Only a good thing

[interval1066]

"...has always had the problem of the double edged sword because it always had to serve Sun..."

Not exactly sure what you mean by "serve", and even so, not sure how that was a problem. Of course OpenSolaris needed to compatible with Sun products, it was a SUN effort, why shouldn't it run the Sun library of products? How exactly was that a problem? OpenSol suffered from a slight lack of interest, really. And I mean slight. Lots of folks liked the idea of an OSS Solaris, I'm just exactly sure there was a real need for it. Of course some are going to disagree with me, but honestly, for all the wonder and mystery of OpenSolaris; the universe will shed a single tear and move on. The world wasn't exactly beating Sun's door down for another flavor of Unix. But having some kind of problem? Not any more so than the problems the other Unoids have.

Re:Only a good thing

[erice]

It's worth noting that Oracle is not kicking Solaris to the curb, they are kicking OpenSolaris to the curb.

It's really the same. OpenSolaris wasn't really about being open because they thought it was right. It was an attempt to push the bulk of OS development onto the community. Internal development ground almost to a halt. The last major release of Solaris was Solaris 10, in 2005. At least from what I see from the publicly available patches, Oracle isn't even keeping up with security patches anymore. Since I haven't heard about Oracle beefing up internal Solaris development, I can only conclude that killing OpenSolaris means they are "done" with Solaris.

So they took their toys and went home

[BitZtream]

So what? OpenSolaris was a bad joke anyway.

In the true spirit of OSS, they packed up their bags, stuck out their tongues and said 'fuck it, we're done dealing with you guys, we're going home' ... and thats perfectly within their rights.

It should be noted however, since they were about the only ones using OpenSolaris, no one is really going to notice they are gone.

Using OpenSolaris is roughly the same as running Darwin instead of OS X. Roughly, not really the same, but both are pretty much pathetic bases of the what the person actually WANTS to run, which is the full version. Anyone who would consider running OpenSolaris will just pay the tiny little fee and run the real deal which has slightly (not much mind you) more spit and polish on it.

Re:So they took their toys and went home

[bsdaemonaut]

I personally never noticed much difference between the two, though I'm not sure that I can claim the kind of familiarity to be 100% on that. OpenSolaris was a full and useable OS, OpenDarwin was barely more than a kernel. Since you used to be able to get single licenses of Solaris for free I'm not sure the comparison works well, but I get your point.

Re:So they took their toys and went home

[Anonymous Coward]

Until oracle got their grubby hands on it, even the full version was free for most use.

Re:So they took their toys and went home

[SIGBUS]

Actually, OpenSolaris was way ahead of Solaris 10 (as it was originally conceived as the development version of the yet-to-be-released Solaris 11). It has a more advanced version of ZFS, the COMSTAR storage framework, a package management framework, and other goodies not in Solaris 10.

Unfortunately, Oracle's corporate culture of radio silence was incompatible with Sun's open development model - and the very first thing they did when they took over was shut off the distro snapshots. I'm surprised it took them this long to quit updating the Mercurial repository.

The Achilles heel of OpenSolaris was that Sun chose a half-assed approach to opening the code; there are still lots of closed binaries, which means more work for the Illumos effort. They at least have an unencumbered C library now, but a lot more needs to be done.

Re:So they took their toys and went home

[TheRaven64]

OpenDarwin was barely more than a kernel

OpenDarwin was a lot more than a kernel. Unfortunately, it was not a complete kernel. Some quite important parts, such as most of the sound subsystem, were not released.

Re:So they took their toys and went home

[gmhowell]

Sound subsystems oughta be easy to replace; haven't they developed eleventyhundred of them for Linux?

Re:So they took their toys and went home

[NormalVisual]

a package management framework

...that broke sparse zones...

Re:So they took their toys and went home

[Xtifr]

Yes, but Sun deliberately chose a license that was incompatible with Linux. In their efforts to prevent Linux devs from "stealing" their precious ZFS, they cut themselves off from access to all those nice Linux drivers that might have actually made their system into a general purpose one.

Of course, as I understand it, Linux and the Solaris kernel are incompatible enough that porting drivers either way is going to be only slightly less complex than a full rewrite from scratch anyway. So maybe it's all for the best.

Re:So they took their toys and went home

[TheRaven64]

More or less, yes. There are two issues. One is that IOKit has a completely different (although, actually quite nice) design from any other *NIX kernel. This means that porting drivers is a bit of effort. The other is that the userland interfaces for sound on OS X are completely different, so if you port the subsystem from somewhere else you'll end up with something that doesn't look or behave like OS X from a developer perspective.

The same problem applies to the video interfaces. If you want 3D support, you either need to reimplement Quartz, add a whole new set of drivers to X.org, or port DRI from another platform. Porting DRI is the least effort (although still far from trivial), but then you've basically got a system that looks like FreeBSD, but has a poor VM subsystem (improved from 10.4 and earlier, when it was embarrassingly bad - now it's just bad), a terrible pthread implementation, and Mach ports.

You'd be better off writing a kernel module that provided Mach ports to FreeBSD, finishing the Launchd port, and ditching the rest of Darwin. About the only nice things in the open source part of Darwin are Launchd and libdispatch, both of which already work on FreeBSD.

Re:So they took their toys and went home

[bsdaemonaut]

Admittedly I did exaggerate, but as you seemed to agree, there is a noticeable difference between the completeness of the two.

Re:So they took their toys and went home

[TheRaven64]

OpenSolaris also has some serious additions. I've recently been doing some work for a company that is interested in using OpenSolaris as the base for a HPC platform. All of the locale stuff is basically missing. The kernel is there, although a few drivers are proprietary, but a lot of the userland stuff is completely missing. Some things are totally undocumented. For example, there is no documentation of the Solaris C++ ABI (actually, ABIs), even internally. The only documentation is the code, and not all of the code is released. I could reverse engineer it, but they're probably just going to use the Itanium C++ ABI, which pretty much everyone else uses (even Darwin, although they use an ARM-insprired compression format for the unwind tables as well).

Question here.

[Beelzebud]

I know that a portion of the Open Solaris code is closed source. Will this hamper any effort to actually fork it? I'm asking from a place of ignorance, as I just know that "some code" is closed, but I'm not sure if it's the code that would be essential to actually forking the project.

Re:Question here.

[WebMink]

Check out the Illumos announcement. Slides 18 and 19 in the deck about that. The Illumos people have made a bootable system with closed bits of libc (including full locale support) replaced, replacements for the most critical closed source utilities and replacements for some drivers. Still to do:

• NFS/CIFS lock manager
• Full kcf module/daemon (crypto framework)
• Trusted Extensions (labeld)
• Many more drivers

That's plenty of work but there are people willing and able to get it done and they have a bootable system to evolve. The real question is when someone will kick off a full distro around it (since Illumos is purely a kernel).

Re:Question here.

[CAIMLAS]

Illumos is currently working on replacing the few parts of the kernel which are still "closed source"; I believe they're fairly trivial things, mainly, from what I recall reading. Trivial, at least, compared to the Important Parts, like dtrace and zfs.

Re:Question here.

[Beelzebud]

Thanks.

And nobody cared....

[Fished]

Oracle didn't care, because Oracle has said that they are no longer interested in having an open development model for Solaris. In fact, the fact that Oracle doesn't care is why they're dissolving in the first place. Solaris users don't care because, let's face it... does anybody actually use OpenSolaris? I work for a huge Solaris shop, and we use stock, Sun-supported Solaris. If we wanted an Open Source operating system, we'd use Linux. We use Solaris for huge database servers that are too big to run Linux (mostly Oracle DB.)

So, that leaves OpenSolaris developers. Look, this is the risk you take when you work on a project dominated by one company, especially when you have a license like the CDDL. I feel bad that you're in this position, but it was kind of predictable, and I really think you're missing the boat with Illumos. You're unlikely to get enough interest to ever make a go of it with Oracle being disinterested. Go work on making Linux better instead!

Re:And nobody cared....

[StuartHankins]

Too big even for RHEL? Help me put this into perspective, what kinds of sizes are we talking here?

Theoretically I understood RHEL Advanced Server was capable of an unlimited number of CPU's and memory ( http://www.redhat.com/rhel/compare/ ).

Perhaps you're reaching some contention at high loads or with large numbers of CPU's / storage / etc?

Thanks for any info you can provide.

Re:And nobody cared....

[Anonymous Coward]

We use Solaris for huge database servers that are too big to run Linux (mostly Oracle DB.)

Yahoo have over 2PB, yes PetaBytes, in PostgreSQL on Linux. I doubt your extremely slow Oracle is anywhere near that.

Re:And nobody cared....

[otis wildflower]

Folks who do opensource NAS care, as ZFS on OpenSolaris is currently superior to anything else in the open sphere, and most if not all of the closed.

But as a Solaris admin, I would much prefer to see a more aggressive improvement of stock Solaris, particularly when it comes to package and patch management.. Nobody here ever did anything with OSol, but watched it to see what would be coming down the pike for Solaris 11..

That said, I'm sure Nexenta and Illumos will fully fork, and presumably if there are enough disgruntled-with-Sunacle devs who want to hack on it in their spare time, then they can make a go of it..

Re:And nobody cared....

[Anonymous Coward]

I don't know why you're getting modded down; a well tuned Linux or BSD install can do just as well if not better than a Solaris install.

Re:And nobody cared....

[Korin43]

That page looks like licensing info, not technical info (certain licenses allow unlimited CPUs). The Linux kernel has a limit on memory and processors. I think the memory limit is insanely high, the CPU one is less so. I've been trying to find info but all anyone seems to talk about is PAE.

Re:And nobody cared....

[codepunk]

Oracle DB to large to run on Linux? The days of going out and purchasing (big hardware) a 64 way box to run oracle or any application on for that matter is over and has been for some time.

Re:And nobody cared....

[StuartHankins]

Same experience here. I'm interested in finding out more about how it scales for very large loads which if I understand correctly require z series or POWER. I only have one RHEL Advanced server and it's x86_64 so I don't really have a comparison point.

Re:And nobody cared....

[Anonymous Coward]

No, its only over in your little corner of the world. Not all problems can be neatly distributed across clusters of computers. Not everyone has the luxury of being a Google, Facebook or Amazon.

Re:And nobody cared....

[Anonymous Coward]

I'm still working on my undergraduate, so I'm not exactly an expert, but I'd like to know as well what things besides just memory or cpu limits would prevent RHEL linux, or even linux in general from scaling this high? Would something as simple as switching out some kernel modules solve the problem?

Also, how does *BSD handle super crazy big size scaling?

Re:And nobody cared....

[Billly Gates]

PostgreSQL can't even handle a column with more than 53k rows until recently?? I doubt the whole Petabyte was all PostgreSQL based. It was probably used for caching to clients while the real work came from an Oracle system.

Re:And nobody cared....

[Billly Gates]

So you can buy a few wimpy Dell Poweredge servers with a cheap switch and expect it to perform trillions of operations a day with mysql or some dumb free database? Even if you had Oracle your Poweredges will have limited ram per node and add tools that use views and other SQL oriented features that enterprise software (not mysql) use and you will need the extra CPUs and ram quick.

Also CPU's like the UltraSparc have had hardware based threads for some time. This means that if an app like a database on a x86 overwhelm the system it slows down to an unusable state very fast. With Sun hardware the high number of threads are run on the cpu so even under an insane load you can launch an xterm and fix the issue. If you need a 64 way box you need lots of slots for fiber connections which is something intel based servers lack. The days are still far from over.

 

Re:And nobody cared....

[Korin43]

I was looking at kernel config options and here's what I found.

Max CPUs:

- x86: 512
- x86_64: 256 (not sure why this is smaller)
- sparc64: 1024
- ia64: 4096

It also notes that these are the max, hardware limits may make it lower. These numbers seem much smaller than I expected.

There's no citation but this Wikipedia page says Linux can support up to 64 TB of memory on x86_64.

Re:And nobody cared....

[codepunk]

I guess, if I where totally ignoring the existance of RAC as you are I would come to the same conclusion.

Re:And nobody cared....

[fishbowl]

I think you underestimate the capabilities of PostgreSQL, and you're also not making a fair apples-to-apples comparison.

We could run Oracle on our "wimpy" Poweredge. (Quad Xeon, 32GB RAM, local SSD's, SAS-attached SAN). We don't make that choice -- everyone who works here has had years and years of experience with Oracle and no desire to repeat it.

In the price range, I'd love for you to show me hardware that makes my servers look "wimpy" though.

Re:And nobody cared....

[Billly Gates]

PostgreSQL could not even support more than 32k rows per column in the last release. It is much better than MySQL and I plan to use it for my business until I have enough money for DB2. PostgreSQL does not support OLAP nor can it store more than a TB before having major issues with things like auto-vacuum. 9.x will have more of these features from what I read. I have not heard of any successful implementations of more than 1/2 a TB. Many slashdotters state that it is just not caught up to the big players yet. I could be wrong.

My knowledge of poweredge servers may be weak as I have been out of the loop for 4 years, but I was under the impression they had a limit of 8GB of ram each.

Re:And nobody cared....

[TyFoN]

[citation needed]

Re:And nobody cared....

[erice]

OpenSolaris was the playground for new features that would, hopefully, be folded into Solaris when they matured. If Oracle doesn't care about OpenSolaris then they probably don't care about Solaris either. This is sad but not really surprising. Solaris hasn't had a major release in five years. Even the patches don't seem to be coming out at a plausible rate.

Re:And nobody cared....

[chef_raekwon]

yes, too big for RHEL. Running RHEL on RISC = dumb. Running Solaris on RISC/SPARC (read M4000/M5000) = bankable.

Re:And nobody cared....

[mikechant]

Currently they are claiming production environments in excess of 4Tb, and maximum rows is now unlimited.
http://www.postgresql.org/about/

Re:And nobody cared....

[fishbowl]

I'm not really arguing. I've worked with Sun and Oracle quite a lot. I've done systems programming on Solaris for long enough that I never stopped calling it SunOS. I know you can't beat their IO for certain applications. The last time I used the Solaris+Oracle combination was to develop a billing system for a company with 5 million customers. When the place where I work now crosses the threshold from a 30 million dollar company to a 100 million dollar company, some of those same considerations might come up. But right now it would be grossly unprofessional of me to say that we need more horsepower than the Intel-based (Linux and OSX) servers we use can deliver.

You're the one who introduced the requirement of trillions of ops, etc.

Today I'm running a factory on these wimpy systems and MySQL, and I'll be pretty happy if we start to outgrow that...

Re:And nobody cared....

[fishbowl]

>My knowledge of poweredge servers may be weak as I have been out of the loop for 4 years, but I was under the impression they had a limit of 8GB of ram each.

I have a rack of them, they are pretty nice. I never meant to imply that a 2U Dell server running linux VMs under QEMU is the equivalent to a SPARC M9K.
I also don't care that Oracle has features that neither MySQL nor Postgres will ever match. Someone could offer to pay the licensing and consulting and I could get on board with it, equally competent as a developer or a DBA.

Would love to have the budget to get out of the Intel-based realm, thanks.

It's happened before

[KnightBlade]

Buy your competition and then kill their products. This seems familiar. Where have I heard this one before....

Re:It's happened before

[Korin43]

Oracle had its own operating system before this?

Re:It's happened before

[richlv]

they had. http://en.wikipedia.org/wiki/Oracle_Enterprise_Linux

well... not really "their own", more like they "borrowed" it :)

Another instance of BSD vs. GPL licensing...

[Anonymous Coward]

...and why the GPL is superior. With the GPL, it is prohibited to take work private that has been built by the community. The BSD license /* encourages */ it. I see on the website of an deeply involved OpenSolaris developer where he is complaining about Oracle not adhering to the spirit of the open source license. I suggest that there is only the /* letter */ of an agreement whenever "push comes to shove." Spirit goes out the window.

Re:Another instance of BSD vs. GPL licensing...

[VGPowerlord]

...and why the GPL is superior. With the GPL, it is prohibited to take work private that has been built by the community. The BSD license /* encourages */ it. I see on the website of an deeply involved OpenSolaris develor where he is complaining about Oracle not adhering to the spirit of the open source license. I suggest that there is only the /* letter */ of an agreement whenever "push comes to shove." Spirit goes out the window.

This is slightly off topic, but there's a certain form of irony here.

Licensing issues is the reason that Linux has no ZFS support. But the GPL is the ultimate "can't re-license" license. No, seriously, you are forbidden by the license to re-license GPL code under non-GPL licenses, even if they are stricter... see sections 1, 2, 6, and 10 of the GPLv2.

One could argue that the GPL doesn't adhere to the spirit of the open source license.

The ZFS issues? The term for this is "hoist by your own petard."

Note: You can still bypass this by getting permission from the original author(s), but that has nothing to do with the GPL; it is a general statement about copyrights.

Re:Another instance of BSD vs. GPL licensing...

[vadim_t]

This is slightly off topic, but there's a certain form of irony here.

Licensing issues is the reason that Linux has no ZFS support.

True. But what's the reason for that?

In the words of Danese Cooper, who is no longer with Sun, one of the reasons for basing the CDDL on the Mozilla license was that the Mozilla license is GPL-incompatible. Cooper stated, at the 6th annual Debian conference, that the engineers who had written the Solaris kernel requested that the license of OpenSolaris be GPL-incompatible. "Mozilla was selected partially because it is GPL incompatible. That was part of the design when they released OpenSolaris. [...] the engineers who wrote Solaris [...] had some biases about how it should be released, and you have to respect that".

Quoted from wikipedia, original source for the quote.

IMO it's most likely that there was an explicit intention of being Linux incompatible as well (meaning, not just because it happens to be GPL licensed). After all, why would Sun give such a gift to its greatest competitor? I think that it doesn't really matter what Linux was licensed under, the license for ZFS would be guaranteed to be incompatible with it anyway.

No, seriously, you are forbidden by the license to re-license GPL code under non-GPL licenses, even if they are stricter... see sections 1, 2, 6, and 10 of the GPLv2.

That's kind of the point of it, yes. I'm not sure what you mean by "stricter" here though. Additional restrictions, like for instance non-commercial usage only would bring it closer to a proprietary license, which is entirely against the intent of it. If you mean things like the AGPL, it would be very difficult to figure out which additional restrictions would be allowable due to favouring what the GPL tries to accomplish, and which wouldn't due to going counter to it, and write some sort of rule that would allow the former but not the later.

The ZFS issues? The term for this is "hoist by your own petard."

Nope. The term for this is "hoist by Sun's very intentional decision to make it be that way".

After all, if Sun were all about complete freedom they would have went with a BSD license. It would have been very easy and they wouldn't have needed to spend time on making yet another license. There must be a reason why that wasn't suitable.

Re:Another instance of BSD vs. GPL licensing...

[VGPowerlord]

The ZFS issues? The term for this is "hoist by your own petard."

Nope. The term for this is "hoist by Sun's very intentional decision to make it be that way".

and yet... it's the same kind of thing the GPL does. As you said yourself

That's kind of the point of it, yes.

How many people here realize that while there are lots of GPL-compatible licenses, the GPL is compatible with no other licenses? Not even newer versions of itself unless you specifically grant it, which GNU strongly encourages. The LGPL is slightly better... it's compatible with the same version GPL; LGPLv2 to GPLv2 and LGPLv3 to GPLv3.

After all, if Sun were all about complete freedom they would have went with a BSD license. It would have been very easy and they wouldn't have needed to spend time on making yet another license. There must be a reason why that wasn't suitable.

And that applies equally to every other open source project that isn't under the BSD/MIT license. Including the Linux kernel, GNU userland, Apache web server, XFree86, X.org, KDE, GNOME, Firefox web browser... you get the point.

Re:Another instance of BSD vs. GPL licensing...

[vadim_t]

How many people here realize that while there are lots of GPL-compatible licenses, the GPL is compatible with no other licenses? Not even newer versions of itself unless you specifically grant it, which GNU strongly encourages. The LGPL is slightly better... it's compatible with the same version GPL; LGPLv2 to GPLv2 and LGPLv3 to GPLv3.

Me for instance?

Look, I choose the GPL for my software very intentionally. It does things in a certain way to achieve a certain outcome, and there are tradeoffs with that, which I'm willing to live with.

And that applies equally to every other open source project that isn't under the BSD/MIT license. Including the Linux kernel, GNU userland, Apache web server, XFree86, X.org, KDE, GNOME, Firefox web browser... you get the point.

Right. Because being a rewording of the WTFPL isn't really what the GPL tries to do.

Anyway. You're trying to say "It's the GPL's fault that Linux doesn't have ZFS right now". But that's not true. It doesn't matter what Linux is licensed under, because I don't think Sun would be stupid enough to just give that to a competitor for free. If Linux's license was different, Sun would simply have choosen something else that would be unsuitable for integration.

I think if everything was the same, except Linux was BSD licensed, there are two likely possibilities: Either ZFS would have been GPL licensed, which in theory is BSD compatible, but which in practice a BSD project wouldn't accept, or there would be no source release at all.

The BSD is ideal in regards to integrating things licensed under it into other software. But it doesn't work that way in the opposite direction, because the very reason you license something under the BSD is that you think people should do whatever the hell they want to with it. And that mindset conflicts with accepting a change that would mean adding a restriction on the usage of the entire project.

BSD works a bit like type O blood. Anybody can accept it. But it will only accept transfusions from its own type.

They are toying with powerful forces here.

[zill]

http://xkcd.com/419/

If anyone asks for me ...

[mseeger]

If anyone asks for me, i am in my room with some candles, a Larry Ellison doll and tons of needles.....

CDDL killed it

[Anonymous Coward]

The CDDL killed it as much as anything else. That license sucked. If Ellison really wanted to make some blood vessels burst in Redmond, he'd put all CDDL material out as GPLv3.

64-way DB Servers

[Anonymous Coward]

I'm retired now, but at my last job we had 20,000+ UNIX servers. My projects - I was a technical architect - had about 300 of those servers. Trying to compare the throughput for most x86 servers to a P590 or HP Superdome or Sun E25K just shows your ignorance of the larger UNIX system capabilities. From a compute standpoint, Intel CPUs are hard to beat, but when you need 10 fibre connections to storage and 20 10GigE connections and have 10,000 concurrent DB users, RH stuff ain't gonna cut it. Sorry, but those are the facts. These servers aren't for a website.

Then you have the issue of getting a vendor that only certifies their program on HP systems to bother with RedHat. The $3M that the DB server HW costs is nothing compared to the software costs for another platform to be supported by some vendors. The software ran on HP-UX - that's all. The software cost $25M for 2 prod instances and 5 non-prod instances (DR, Test1, Test2, pre-prod ... ) This is software you run your business on AND not very well designed. Bigger hardware is always the answer over system design changes. It is cheaper. Our prod DB servers were 64-way with 108GB of RAM. We had 4 of them - 2 production locations with 2 DB srvs each. An active/failover cluster model. We had 4 DR servers that were almost as large located in another data center that got data updates nightly. There were about 20 app servers inside data centers, about 40 app/GIS servers located in the same building as the users who were spread all over the USA for this project. Another 20 servers were used for the dev, test, pre-prod, test2, test3 environments. It was not possible to run all the software on the say system, at least 3 systems were required for each environment. Crap, I know. Back when I worked on it, VPARS were specifically not supported by the vendor, so we didn't use them - anywhere.

Just because RH claims to run on 20+ way systems, doesn't mean any of the software will. BTW, Oracle RAC was not supported by the sw vendor, so lots of small Oracle Nodes wasn't gonna work.

Anyway, you wanted some background on why anyone uses non-RH machines. Oh - we were seeing about 4k TPS on each production system during business hours. Transactions came from client tools, app servers, reporting tools, and ad hoc queries from blackberries and other portable devices.

We had an outage 1 day for about 6 hours around 2004 due to DB corruption - over 10,000 people couldn't do their jobs (half the users). It wasn't good. I'm glad only 1 production site was impacted.

Re:64-way DB Servers

[hibiki_r]

Against all odds, AC knows about the realities of life. When your production environment includes a dozen Superdomes, RH might not be the right solution for you. And on something like a bank or a big telco, you'll find said big systems all the time, not a very wide array if little machines.

Re:64-way DB Servers

[StuartHankins]

RHEL Advanced server runs on mainframes (including IBM z systems) and POWER platforms, not just x86 platforms. I just assumed you'd have to use one of these to get massive scalability and we weren't going to try to compare x86 systems.

Re:64-way DB Servers

[TheRaven64]

Saying RHEL runs on mainframes is a bit misleading. RHEL runs in smallish partitions of a mainframe. You can run a few thousand RHEL instances on an IBM Z-series (or whatever they're called this week) mainframe, but you can't just run a single instance of RHEL as you can with z/VM. The RHEL support is there for people who have a mainframe and want to use some of the spare capacity to replace some x86 servers. It's not something you would buy a mainframe for.

Re:64-way DB Servers

[Anonymous Coward]

AC here ... "vendor only certifies on HP-UX", so trying anything else was worthless. The server programs were written in C/C++, so we weren't gonna get binaries for any other platform anyway.

Also, Z-Systems are great for IO, not so much for raw MIPS. Yes, I've watched the 1,000 linux servers running apache on a single mainframe reports. Interesting, but not really useful. BTW, I was a mainframe dev for 5+ years - MVS, JCL, clists, TSO. None of that rocks. Ruby - rocks. Perl - rocks.

I LOVE IBM-POWER-systems, but I would never consider dropping RH on them, except to play with. IBM's LPAR technology is the best there is in the virtualization world, IMHO, and I do a lot of virtualization on UNIX and x64 servers.

When your company is running critical systems and spends millions of dollars on HW and millions of dollars on software, you want as few throats to choke as possible. IBM is one. Oracle is two. The Software vendor is three. During an outage, getting 2 vendors to talk is nearly impossible. I don't want any of them to claim that RH is the issue and have to add a forth party to the issue to blame. Honestly, I'd prefer to remove Oracle and use DB2 on IBM hardware if just for the 1 less throat to look for during an outage.

I've been there with a "very large storage vendor" claiming all sorts of reasons that the outage couldn't be caused by their equipment. We were able to prove the issue was with background mirroring for DR purposes ("BCV copy") that impacted foreground read/write needs for the DB in question after building an accurate timeline for the issue. Basically, we proved it began at the instant the copy began and ended when the copy finished (about 15 min later). Teaching a vendor SE something that he should have known. Our environments tended to break vendor hardware and software. Yes, it was a very large telco.

When you need lots of IO, x64 is not the place to look for 8+ redundant backplanes.

Re:64-way DB Servers

[StuartHankins]

Thank you, this helps me understand. It makes all the difference when it has to be partitioned.

Re:64-way DB Servers

[Anonymous Coward]

Against all odds, AC knows about the realities of life. When your production environment includes a dozen Superdomes, RH might not be the right solution for you. And on something like a bank or a big telco, you'll find said big systems all the time, not a very wide array if little machines.

If you say so.

I worked at one of the largest banks in the state - and I don't mean South Dakota. We were turning off Solaris systems right and left when I departed. Replacing them with RH.

Re:64-way DB Servers

[Billly Gates]

As the A/C mentioned it is not only hardware. You think MySQL or even PostgreSQL can handle anywhere near the requirements of that job? lol

Middleware demanding these high end systems is the other reason. I am thinking of starting a business which will have an enormous demand for database transactions which need to be ACID based and can grow very fast if it takes off. I am starting out with PostgreSQL until I have a few thousand members but plan to switch fast to Oracle or DB2 as soon as I get some funding after I show I can generate revenue. 90% of most business is small to medium and do not require that insane computing power or storage so most systems do not support it. RHES and a few Dell Poweredge servers do not fit the other 10%.

Re:64-way DB Servers

[Anonymous Coward]

We had an outage 1 day for about 6 hours around 2004 due to DB corruption

You should have been using MySQL... 8^)

Re:64-way DB Servers

[Cato]

Very large RISC Unix servers may still be the way to go for DB servers, due to the I/O required, but for app servers you can scale out at very low cost with JEE clustering. x86 servers are now down to $20K or so for 48 cores (4 sockets x 12 core AMD CPU) - e.g. http://www.dell.com/us/en/enterprise/servers/poweredge-r815/pd.aspx?refid=poweredge-r815&cs=555&s=biz - enabling very high volumes.

Enterprise application vendors are seeing a lot of demand for Linux/x86 support from their customers, for this reason.

Re:64-way DB Servers

[Anonymous Coward]

Apparently, Linux runs pretty well on systems with 512 CPUs and 1TB RAM. The whole thing is a cluster of 20 such systems. I can't really imagine that there's any big difference between RHEL and SuSE regarding the number of CPUs. If your software only runs on HPUX, that's an entirely different problem.

NOPE

[WindBourne]

THe first time that SUN went OSS, I fully supported them. When they burned everybody, I KNEW that it was easy for Sun or any buyers to screw over OSS world. And I have spoken out against Sun's opening of the code, even though LOADS of Sun fanbois were pushing them. And in this case, Sun did not disappoint.

"...chance of becoming a true fork..."

[meerling]

when you're down to your last piece of tableware, you just have to make do with what's left, unless it's a butter knife and the meal is soup...

(if it was a real knife, you could at least carve a spoon out of something)

CDDL != BSD

[Fished]

The CDDL is not a BSD license. It's more like an MPL or NPL style license, in that it privileges a particular company (in this case, Sun) to a special status with regard to derivative works.

Issued an ultimatum to Oracle.

[nurb432]

Thanks, i needed a good laugh after a day like today.

Re:Issued an ultimatum to Oracle.

[lotho brandybuck]

hahaha... yeah, me too, perhaps they didn't realize, in Corporate America, Oracle Issues Ultimatum to YOU!

But then again...

[gwolf]

iBCS was not maintained because there was no real need for it anymore. 15 years ago, you often needed to run stuff that just would not exist on Linux, and today that's just not going to happen, except for very specific, very propietary stuff. And if you want propietary stuff, you will most likely pay for it. And the OS it runs on, becomes basically an afterthought.

Dangerous.

[unity100]

It takes aeons to build up a reputation, but it takes a few days to totally destroy it. oracle should watch its standing with open source community.

Re:Dangerous.

[Billly Gates]

Oracle doesn't care and looks at us as a threat.

Yes we have some influence in buying decisions with some bosses, but these bosses have sales reps from Oracle directly. Oracle does not like anything that undermines the value of its products. If Mysql and PostgreSQL did not exist it would mean 3x the amount of revenue for Oracle. In the 1990s businesses bought Oracle for small department, small business servers, and even tiny websites. lol. It was the only tool out there besides the new SQL-Server. We switched to Mysql and that pissed them off. Does Oracle-lite even exist anymore? I remember the error messages from Oracle Databases back then from websites because they were so hard for non DBAs to setup to host simple pages.

The point is business does not care about opensource. Just the results and Oracle does not care about offending us because if open software is less available they can charge more money for their existing products. Their customers do not care as we are not their customers. I pray that they do not be asses with Java. This will make .NET and Windows much more attractive if they do.

Re:Dangerous.

[mysidia]

Yes we have some influence in buying decisions with some bosses, but these bosses have sales reps from Oracle directly.

We have little bit more than "just some influence". The "sales reps" now find themselves confronted with dead air, because we don't really much care what an Oracle sales rep has to say anymore, at least not until they do a lot of explaining in regards to their company's recent underhanded actions as of late.

You think these types of things don't effect customers?

I pray that they do not be asses with Java.

They have already proven that they are.

Re:Dangerous.

[tehcyder]

It takes aeons to build up a reputation, but it takes a few days to totally destroy it. oracle should watch its standing with open source community.

What standing?

mod parent up

[unity100]

i posted in this thread. cant use points.

had a zfs scare, today (freebsd 8.0 to 8.1 hiccup)

[TheGratefulNet]

this was kind of scary; and I think it was *my* last nail in the coffin for zfs. for me, that is (ymmv).

I had a freebsd 8.0 system up for quite a long time (over 100 days). I often keep the /usr/src area current via cvsup but had no reason to reboot until the system hung under heavy load. when it rebooted, I found I was 'half 8.1' and half 8.0, still. this threw zfs out of sync and the pool would not be mounted! what a scare, let me tell you (I had all my valuable stuff on that 6TB pool).

had to reboot to kernel.old, do a zfs scrub (took overnight), then rebuild World, reboot to have World be seen, then rebuild kernel to 8.1. then reboot again. finally zfs and the kernel saw eye to eye and my 6TB pool was mountable again.

I know that 'better sysadmin' might have avoided this, but I don't think I've ever had anything on linux be this way. I've used reiser, jfs, xfs - all the fs's. don't think I've ever had to do things this drastic just to mount my 'raid' set like this. it was scary.

now, given the fact that opensolaris was the nearest upstream to freebsd for zfs and that opensolaris pretty much is no more, I have very little actual confidence in zfs and support, going forward. its scary enough with all the new concepts, as it is; but I need to feel good about it; now and years from now. this version incompatibility worries me! its fine to have new features be ignored if you run older systems but this 'we wont mount you AT ALL' stuff is just for the birds! why on earth would a fs be designed like that? its not real-world and I'm not sure I like this idea, to be honest.

I did think about going with pure opensolaris instead of freebsd8.* but I'm just not sure that there's a good compelling reason to stay with bsd anymore and zfs seems like a non-starter for me, given my experience with it.

(plus, on the same exact hardware, linux md-raid and freebsd zfs compared, linux keeps up (no pauses) and is overall faster on nfs and samba than bsd/zfs is. and zfs really 'wants' at least 4gb to run. that's asking too much, too, I think).

Re:had a zfs scare, today (freebsd 8.0 to 8.1 hicc

[mysidia]

As someone who has tested out ZFS on both BSD and Solaris, and is familiar with the subject, I can tell you quite plainly: The Solaris implementation is good and generally, the BSD / FreeBSD implementation sucks, is buggy, missing features (or key features are buggy).

The OpenSolaris implementation is mature, well maintained, performs well, and is up to date. The FreeBSD implementation is immature, lags way behind OSol development, and seems to be unstable.

If you want a mission critical system running ZFS, FreeBSD is not an option, period.

ZFS is Stable on Solaris, and to a great extent on OpenSolaris and NexentaCore. There may be some gotchas, in OSol, particularly if you use new advanced features such as dedup.

There may also be some performance gotchas with certain hardware in all implementations (e.g. 32-bit proc, less than 2GB of RAM), and if you use gzip compression.

Just because work has been done to port the filesystem to BSD, does not mean it has the same stability or performance characteristics. It might have similar ones in the future, right now it does not. I would call it "Alpha" quality, even if it is in stable versions of BSD, at this point.

Somebody forgot to tell Oracle that.

[Fished]

Why do you think Oracle bought Sun in the first place? It was for their presence in the "enterprise server market." Yes, some people are using Oracle RAC and the like, but there are still plenty of applications for that 64-way box running Oracle.

Re:had a zfs scare, today (freebsd 8.0 to 8.1 hicc

[anilg]

.. I have very little actual confidence in zfs and support ..

ZFS as it is today, even without any further feature additions, is miles ahead of any other FOSS filesystem in features and stability. As far as support goes, there is enough community and kernel expertize outside Oracle to atleast maintain the code (my employer Nexenta has made code contributions back into ZFS : the in-kernel iscsi code). So there is no reason for the doom and gloom.. that's the beauty of FOSS, you cant take back what you've provided.

~Anil

-_-

[Anonymous Coward]

OpenSolaris is dying!

Re:had a zfs scare, today (freebsd 8.0 to 8.1 hicc

[PipsqueakOnAP133]

(plus, on the same exact hardware, linux md-raid and freebsd zfs compared, linux keeps up (no pauses) and is overall faster on nfs and samba than bsd/zfs is. and zfs really 'wants' at least 4gb to run. that's asking too much, too, I think).

That's because linux mdraid doesn't provide you anything like checksumming amongst whatever other features you might be using.
So yeah, I'm sure ext3 and fat32 are plenty fast compared to ZFS, but if you're running ZFS, fast wasn't the point, was it?

(I'm using ZFS on 2GB of ram over gigE (EON Solaris NAS distro), works fine for me)

Using ZFS in production

[Cato]

ZFS is used in production on FreeBSD by some people and I generally like the ZFS features, but I don't view ZFS as really production ready on any OS.

This is because ZFS on any OS does sometimes lose all access to the zpool (i.e. you lose the entire set of RAID volumes and filesystems on them, all at once) due to ZFS bugs, and there is no fsck. I can't think of another filesystem where you can (a) lose all access to your files and (b) there is no fsck. Same goes for RAID - even if you use RAID-1 with ZFS you can still lose your entire zpool due to a ZFS bug.

Given that the "no data loss by design" (can't find the reference for this, perhaps Sun/Oracle has stopped saying this) hasn't really worked out for ZFS on Solaris or FreeBSD, there is still a need to have a complete backup of any ZFS zpool (as with any other RAID / filesystem). The checksumming, COW, snapshots, and self-healing data (for RAID) are very nice, but losing your whole pool due to a ZFS bug means it isn't really a high availability solution. On the plus side, it does make it very easy to snapshot in order to take a backup, and makes incremental backups easy with zfs send.

Here are a few samples of complete zpool loss:

http://opensolaris.org/jive/thread.jspa?threadID=132089&tstart=120

http://superuser.com/posts/130822/revisions - FreeBSD

http://www.mail-archive.com/zfs-discuss@opensolaris.org/msg39578.html

My point is not that ZFS is a bad idea, but it really needs an fsck (see http://www.osnews.com/story/22423/Should_ZFS_Have_a_fsck_Tool_ ) and anyone using ZFS must have full backups onto another server of the whole zpool, perhaps into a non-ZFS filesystem to avoid software bugs that hit both the main and backup zpool. The need for backups isn't unique to ZFS but I haven't seen other filesystems/RAID implementations promising "no data loss by design"

http://breden.org.uk/2009/05/01/home-fileserver-a-year-in-zfs/ has some good info on using ZFS for a home NAS, with a separate backup server also using ZFS.

btrfs isn't mature yet, but its designer has said he will always make fsck a priority over new features: http://lkml.indiana.edu/hypermail/linux/kernel/0706.2/1284.html

I'm not sure how this is good for Oracle

[Anonymous Coward]

My entire enterprise is now moving away from Oracle products.