Sun rolls out OpenSolaris 2009.06 release As odd as it may seem, the OpenSolaris development release of Sun Microsystems' Unix operating system has only been available officially on x64 PCs, workstations, and servers. The OpenSolaris distribution has not been packaged up for Sparc workstations or servers. Starting today, with the OpenSolaris 2009.06 release, both x64 …
It is nice to see more features getting bundled into OpenSolaris!
OpenSolaris offered robust kernel and file system integrated CIFS for some time (something that no other operating system has done as well, besides Windows) - a beautiful thing for integrating Solaris, Linux, and Windows environments onto a single underlying file system.
Since OpenSolaris is the core infrastructure which the Sun Storage platforms are based upon, adding faster networking and processor enhancements (both CPU throughput as well as power efficiency) provides performance boosts for Sun integrated storage systems.
In the area of integrated storage systems, being able to release OpenSolaris under UltraSPARC T1/T2/T2+ means being able to leverage octal crypto engines for both encrypted network transfers of storage data as well as encrypted disk reads/writes for storage data. Additional performance on encryption from client to disk would give a great boost in performance to the storage line. If Sun decided that this would be of interest to the U.S. DoD. - an UltraSPARC T OpenStorage product would be sensible.
Seeing the inclusion of SPARC RocK code seems to indicate that the next generation silicon is moving forward, otherwise programmers would not have wasted their time including code for a processor that would not be released (to undergo another set of silicon revisions.)
Also, seeing OpenSolaris boot under SPARC is a good indication that Solaris 11 is right around the corner, since OpenSolaris is basically the Solaris 11 release. The GUI install integration of OpenSolaris for SPARC is tantalizing - this would possibly make Solaris 11 one release-away.
The only other thing the market would want, on a future wish-list, is full clustering integrated into ZFS (with Sun's acquisition of HPC clustered file system, it is just a matter of time.) One would hope the market will not have to wait until Solaris 12 to run a zpool command sequence to configure a clustered ZFS file system! :-(
The author writes, "Fujitsu's Sparc64 platforms were not mentioned as being supported."
Fujitsu SPARC64 appears to be supported in OpenSolaris 2009.06 in this video presentation from CommunityOne.
http://link.brightcove.com/services/player/bcpid1640183659?bctid=24573586001
As per an earlier El Reg story (http://www.theregister.co.uk/2009/04/30/sun_solaris_11/), Solaris 11 is not due until mid-2010.
Also, OpenSolaris is *NOT* 'basically the Solaris 11 release'. That would be Solaris Express & Solaris Express: Community Edition. Those are OpenSolaris with all of the Sun proprietary bits tacked back on. There are LOTS of pieces of code that Sun can't give out because they don't own the full copyright to, which is why most of the OpenSolaris distros use GNU code to fill in the gaps.
Which makes me wonder. if SX:CE is OpenSolaris+Sun goodies, what is the point in using OpenSolaris?
Caveat: I have no knowledge of Sun hardware or Solaris, but ....
--"Before, a fast 10 Gigabit Ethernet or 20 Gigabit InfiniBand adapter card would be tied to a CPU or VM, which it could easily flood"--
I fail to see how virtualization in software would change this. Surely this depends on the hardware architecture. e.g. If a NIC is sitting on a bus that all CPUs have equal access to (SMP) then a modern multi-threaded kernel driver will allow all CPUs equal access to that NIC. If the NIC however is attached to an IO bus which is attached to a system/cpu "box" which is part of a larger NUMA type system, then IO for that NIC is tied to the cpu(s) in the system/cpu "box" and no amount of virtualization can change that. So this statement doesn't make any sense to me, unless Solaris has been sporting funneled NIC device drivers all this time and has only just got round to making them thread safe: and I can't believe that (it would be astonishing).
Anon posts, "As per an earlier El Reg story... Solaris 11 is not due until mid-2010."
Yep, so the rumor mill has it!
Anon posts, "Also, OpenSolaris is *NOT* 'basically the Solaris 11 release'."
"OpenSolaris"... "is the evolution beyond Solaris 10"... "the next generation of Solaris that is being built in to the OpenSolaris community"
http://www.sun.com/jsp_utils/vid.jsp?bctid=24573586001&size=360x250&autostart=true
Well, it seems the Group Manager over the OpenSolaris Marketing Organization from Sun is publicly disagrees with this anonymous poster. I think I will believe the proverbial "horses mouth" over what the next generation of Solaris is over an anonymous poster.
You Sunshiners have to look forward seeing as you are so far behind Linux, let alone WIndows!
".....OpenSolaris offered robust kernel and file system integrated CIFS for some time (something that no other operating system has done as well, besides Windows)...." Where do you get this male bovine manure? CIFS is basically the follow up to the original M$ SMB product, which was the competitior to Sun's NFS - you are crowing about having to integrate your competitor's product because it has become the de facto standard yours failed to become! And as for Slowaris being the only OS to integrate CIFS, have you forgotten that thing called Samba? It's been available both as a free download and as an integrated offering for many OSs for years. For example, hp-ux has had a CIFS product for so long I can't remember when it wasn't there. Please put the marketing piece back in the sales pack and go do some research.
"....Seeing the inclusion of SPARC RocK code seems to indicate that the next generation silicon is moving forward, otherwise programmers would not have wasted their time including code for a processor that would not be released (to undergo another set of silicon revisions.)...." LOL! Is it new code? Or were they just worried if they stripped it out everyone would know Rock has been cancelled? More likely it was just the usual sloppy Sun coding, with not enough staff left after the cuts to go back and weed out all the orphan code.
Honestly, Novatose, that's the worst bit of marketeering posing as comment you've posted for all of... well, a week at least! Please try something original rather than just cut 'n' pasting from the Sun press releases.
/shaking my head in disbelief, SP&L.
david ~ OpenSolaris offered robust kernel and file system integrated CIFS
matt ~ have you forgotten that thing called Samba?
samba is not integrated at the kernel level, matt
matt ~For example, hp-ux has had a CIFS product
hp cifs server a.02.04 is based on samba 3.0.30 - sun bundled free samba for some time
no comparison to a true kernel based system - you clearly don't know what a kernel is
by the way, it is spelled S - O - L - A - R - I - S
one day, you'll finish grammar school ;-)
Macka asks, "If a NIC is sitting on a bus that all CPUs have equal access to (SMP) then a modern multi-threaded kernel driver will allow all CPUs equal access to that NIC. If the NIC however is attached to an IO bus which is attached to a system/cpu "box" which is part of a larger NUMA type system, then IO for that NIC is tied to the cpu(s) in the system/cpu "box" and no amount of virtualization can change that."
This video may help clarify with discussion about OpenSolaris 2009.06
http://link.brightcove.com/services/player/bcpid1640183659?bctid=24579687001
Infrastructure has been re-written at the NIC, Driver, and Socket levels - all the way up the stack.
Macka posts, "So this statement doesn't make any sense to me, unless Solaris has been sporting funneled NIC device drivers all this time and has only just got round to making them thread safe: and I can't believe that (it would be astonishing)"
Network Virtualization has to do with dedicated resources and isolation of network resources. They are talking about multiple: Hardware Ring Buffers in a NIC, TCP/IP Stacks in a Kernel, Kernel Ring Buffers in a Stack.
http://www.opensolaris.com/use/ProjectCrossbow.pdf
"Crossbow is designed as a fully parallelized network stack structure. If you think of a physical network link as a road, then Crossbow allows dividing that road into multiple lanes. Each lane represents a flow of packets, and the flows are architected to be independent of each other — no common queues, no common threads, no common locks, no common counters."
Some of the more interesting results of this integration: create networks with no physical NIC cards; create switches in software; assign bandwidth to a virtual NIC card (vNIC); assign CPU resources to a vNIC; assign quality of service (QoS) attributes to a vNIC; throttling protocols on a vNIC; virtualize dumb NIC's via the kernel to look like smart NIC's; switch automatically between interrupt and polled modes.
The implications are staggering:
- Modeling of applications and their performance can be done completely on a laptop, all application tiers, including H-A, without ever leaving the laptop - allowing architects to test the system performance implications by making live configuration settings
- If there is a DoS attack against a virtual server's vNIC card, the other virtual servers do not necessarily have to be impacted on the main system due to isolation and resource management.
- Heavy consumption of network resources by an application does not necessarily have to step-on other mission critical applications running in another virtual server
- Multiple physical ports can be aggregated into a single virtual port and then re-subdivided into multiple virtual NIC's so many applications or many virtual servers can experience load sharing and redundancy in a simplified way
- Priorities for latency sensitive protocols (ex. VoIP) can be specified for traffic based upon various packet policies, like Source IP, Destination IP, MAC address, Port, or Protocol
- Security is enhanced since Solaris 10 containers no longer have to share IP stacks for the same physical NIC, but physical NIC's can now have multiple IP stacks for each container
- Older systems can be retained for D-R or H-A since their dumb NIC's would be virtualized in the kernel and the newer NIC's with newer equipment can be added into the application cluster for enhanced performance
- Heavily used protocols will switch a stack into "polled mode" to remove the overhead of interrupts to the overall operating system, providing better overall system performance, as well as providing faster network throughput to competing operating systems
- Enhanced performance at a lower system resource expense is achieved by tuning the vNIC's to more closely match the clients mean flow control can happen at the hardware or NIC card level (instead of forcing the flow control higher in the TCP stack)
Usually, adding & leveraging features like QoS and Virtualization will decrease performance to an operating system, but with OpenSolaris, adding these feature with a substantial re-write of subsystems, enables a substantial increase in read & write throughput over Solaris as well as substantial increase in read throughput (with close to on-par write throughput) in comparison to Linux on the same hardware.
http://www.opensolaris.com/learn/features/networking/networkperformance/
What is really appealing is that the OpenSolaris performance and feature enhancements can be leveraged today through a live migration of an existing Solaris 10 LDOM onto an OpenSolaris system. (Yes, enhanced support can be purchased.)
http://webcast-west.sun.com/interactive/09D02126/index.html
This OpenSolaris technology is truly ground-breaking for the industry.
As usual, the Sunshiners have no grasp on reality.
"....samba is not integrated at the kernel level, matt...." Why does it have to be? SMB's a protocol and works just fine outside the kernel, and has done so for years for Linux, hp-ux, AIX, Mac OS et al. Oh, sorry, did I just p*ss all over your latest Slowaris feature sell? "Hey, no-one else has CIFS integrated into the kernel, so buy Slowaris 'cos having it in the kernel is just VITAL, dude!" Puh-lease! So, I'm supposed to give up my tried-and-tested and trusted Linux Samba for what is an untried and no-doubt bug-ridden piece of Sun code done on a shoestring ('cos Ponytail doesn't have any budget to spend on Slowaris) by developers that failed keep Slowaris on SPARC competitive with hp-ux and AIX ('cos all the good developers saw the writing on the wall and left Sun long ago)? Yeah, I can really see that one flying with the board - not!
"....hp cifs server a.02.04 is based on samba 3.0.30 - sun bundled free samba for some time...." And your point is? Oh, your point is that it isn't new, and bundling it into the kernel has no real world value. To prove otherwise, please post some cohearent argument and performance data to back up your drivel. Otherwise you can go perform fellatio on a the nearest power socket.
"....you clearly don't know what a kernel is...." The problem for you is not only do I know, but anyone with a modicum of tech know-how will too, which covers a large swathe of the customers that are going to laugh at you when you try your new "CIFS-in-the-kernel-for-the-win" feature sell.
".....by the way, it is spelled S - O - L - A - R - I - S..." Not to the customers, they christened it Slowaris, not me, and surely now you'll have to start spelling it "O r a c l e E n t e r p r i s e L i n u x"? By the way, since that's based on RedHat EL, it also already has a tried-and-tested Samba/CIFS content and won't need the Slowaris CIFS-in-the-kernel, which will no doubt come loaded with bugs and unwanted "features" aka limitations common to the whole OpenSlowaris nonproduct.
"....one day, you'll finish grammar school...." Long since, Sunshiner, and it was a real Grammar, not a jumped up secondary. And got my degree many years back too. Oh, and I have a job with a future because I work with OSs that businesses actually want, unlike you - enjoy your pinkslip when Slowaris dies!
/SP&L
Matt Bryant asks, "Why does it [CIFS] have to be [at the kernel level of an operating system]?"
Kernel based CIFS offers benefits with cross-vendor compatibility and performance unseen with a userland SMB protocol. Solaris is now a first-class citizen in the CIFS world, leaving SaMBa reliant operating systems suffering with unresolved POSIX/Windows compatibility issues.
A summary of the basic requirements for CIFS under Solaris underscores the issues that are apparent when kernel level work is not done - those holes, not addressed while solely running SaMBa in userland, can result in odd (inadvertent) problems when Linux, UNIX, and Solaris users manipulate certain files which were created/modified/copied/compressed/archived over CIFS (and visa-versa):
http://us1.samba.org/samba/ftp/slides/cifs-workshop-2007/cifs_workshop_2007_09_27.pdf
Some of the best direct answers to your question comes from Solaris CIFS developers, many of whom are also SaMBa developers. A few quotes to touch upon the answers with full articles to review at your convenience:
http://blogs.sun.com/afshinsa/entry/cifs_for_solaris_a_new
"ZFS now understands Windows style identities (SIDs), in addition to other enhancements including Windows style ACLs, DOS attributes, and mixed and case-insensitive operations."
http://blogs.sun.com/afshinsa/entry/solaris_cifs_server_and_zfs
"The CIFS server is a consumer of the Virtual File System (VFS), which provides a common interface to all file systems. The VFS provides a common abstraction layer that allows consumers to treat all file systems as generic resources, so the CIFS server provides interoperability with any file system that can be shared"
http://blogs.sun.com/amw/entry/cifs_in_solaris
"There is a common misconception that Windows interoperability is just a case of implementing file transfer using the CIFS protocol. Unfortunately, that doesn't get you very far. Windows interoperability also requires that a server support various Windows services, typically MSRPC services, and it is very sensitive to the way that those services behave: Windows inter- operability requires that a CIFS server convince a Windows client or server that it "is Windows". This is really only possible if the operating system supports those services at a fundamental level. "
"....Kernel based CIFS offers benefits with cross-vendor compatibility and performance unseen with a userland SMB protocol. Solaris is now a first-class citizen in the CIFS world, leaving SaMBa reliant operating systems suffering with unresolved POSIX/Windows compatibility issues....." Again, what issues? If you hadn't noticed (and you probably didn't because your head is so far up Ponytail's rectum), the rest fo the industry is doing just fine with Samba when needed, and real fine with standard Windows.
"....A summary of the basic requirements for CIFS under Solaris...." Expecting an independent discussion, especially given the Samba.org link? Nope, just another Sun marketeering slideset. But, it does give us the first hint of the real problem - it's also pushing ZFS. Suddenly the penny drops - this is just trying to paper over another of the holes in ZFS - poor Windows interoperability.
"....Some of the best direct answers to your question comes from Solaris CIFS developers...." And again, more Sunshiner material, nothing from the rest of the community saying "Hey, wouldn't it be nice to have CIFS in the kernel". In short, just more Sun hogwash, but it does admit the core problem for Slowaris - Windows is the king in the new area Slowaris has to play in. Seeing as Sun is being driven out fo the datacenter, for Slowaris to survive it has to be able to integrate better with Windows if it is to survive in the appliance world, and that means it has to acknowledge Windows is the prime OS that their applienaces will be tacked onto. Congratulations to Ponytail and McNeedy - they not only turned a $200bn company into a has-been, they also turned one of the leading UNIX OSs into a cheap appliance OS. Not happy with copying NetApp's WAFL to make ZFS, Sun is now turning Slowaris into a poor Data ONTAP clone.
The real merriment is that Sun and the rabid Sunshiners like Novatose are highlighting the appliance tech of the new OpenSlowaris rather than any advances in datacenter technologies like FCOE (not surprising seeing as the OpenSlowaris implementation is just a copy of the Linux OpenFCOE initiator). Just another example of how the Sunshiners are abandoning the high ground of the enterprise for the edge server market. The real laugh is this just makes them more vulnerable to Linux, which already has a massive advantage there.
/SP&L
"Again, what issues?"
Read the posted URL's. If you read, you may not be so ignorant.
"Expecting an independent discussion, especially given the Samba.org link? Nope, just another Sun marketeering slideset"
Nope - just a developer who works on open source projects from SaMBa and OpenSolaris. You should read & understand.
"But, it does give us the first hint of the real problem"
The issues surround interoperability between POSIX and Windows systems manipulating the same files, access lists, etc. Reading will help you understand.
"ZFS - poor Windows interoperability"
ZFS has better interoperability than any other non-windows filesystem - LOL! How about a reference to your funny FUD?
By the way, you should learn how to spell Solaris. There was no one named Novatose in the thread, not sure why you talk to yourself.
Can't read, can't spell, talk to people who do not exist - not the sharpest knife in the drawer.
".....Read the posted URL's. If you read, you may not be so ignorant....." The link you posted with ref to the claimed "POSIX issues" is to the Sun marketerring slideset and has no details on any POSIX issues. Try proof-reading your own marketing before posting. With over a decade of using Samba in real production I have never seen any POSIX/Windoze compatibility issues that have been such show-stoppers as to make me want something more built into the kernel. A quick straw poll with other users I know (some whom have been using Samba since it was called nbserver back in '93) drew similar blanks for any such "issues". Same for a scan through the Samba release notes and a look at Samba.org. Not that I expected there to be anything, but I thought I'd best look just to confirm my suspicion that this is all just Sunshiner make bovine manure and feature sell. Even stretching the definition of "issue" to cover the known problem where a Windows users can cause timestamp differences with owners files, this can be handled in the smb.conf file with the DOS filetimes setting. Please go read the Samba man pages to realise how ignorant you sound.
".....Nope - just a developer who works on open source projects from SaMBa and OpenSolaris. You should read & understand....." Here's your problem - I already have good sources for advice on opensource projects I turst, and Sun and Sun-sponsored developers are not any of them. You'll find that theme quite common amongst what you no doubt refer to as "freetards", namely that we treat Sun with the same derision and suspicion as M$. Still, I went and read through the material at all the links you provided and it hasn't changed anything - there is still no information on the supposed "POSIX issues" you claim, and no case for pushing CIFS into the kernel. For a start, it means that any changes to the CIFS package have to be made as kernel changes, unlike with a Samba server package where it is outside and independent. And by sticking CIFS into the kernel on Slowaris, Sun are tieing it to ZFS - ah, now I see the point! This is just Sun trying to hijack opensource CIFS development and muddy the waters.
"....ZFS has better interoperability than any other non-windows filesystem - LOL! How about a reference to your funny FUD?...." Really? Ignoring the many opensource options such as BTRFS, want to compare to WAFL? Should be a fair comparison seeing as ZFS is just a WAFL rip off. Maybe if you did a bit more reading on a wider scope (i.e., not just swallowing the Sunshine) you might realise Slowaris brings very little new to the table, a lot that has been done before and better by others, and nothing of real value compared to established competitors such as RHEL or SLES. Every time you open your mouth an squeal "the OpenSolaris option is the ONLY option because I say so" you sound just like the old Sun mantra ""Solaris on SPARC and nothing else", and look where that ended!
"....By the way, you should learn how to spell Solaris. There was no one named Novatose in the thread, not sure why you talk to yourself...." Looks like your memory is as short as your objectivity. We've already been over the fact that Sun's own customers came up with the Slowaris moniker, not me. But I admit I christened you Novatose after a particularly nauseating bit of Sunshiner blather you posted on The Reg not too long ago, where you stupidly cherrypicked SAP benchmark figure in an attempt to FUD Itanium. That was fun, it left you wide open to a counter showing how poor Sun's own kit did on the same benchmark. For those wanting to relive the comedy, take a gander at http://www.theregister.co.uk/2009/02/27/hp_sun_oem_comment/comments/ and enjoy a real Novatose moment (and please note Jake's correction on the whole petard thing). I chose Novatose due to your complete and unquestioning faith in Sun's Nova project, and how it mimic the objectivity of the comatose. I seem to remember around that tie you were posting a whole load of male bovine manure about how Sun was doing just fine, no need for a sale, Nova was going to save/change the World, and Slowaris on SPARC was going to rise again and crush the competition..... Yeah, you got that one so right - not!
"...Can't read, can't spell, talk to people who do not exist - not the sharpest knife in the drawer." Oh dear, the usual Sunshiner insults in lieu of counters. Seeing as you obviously have both memory issues and problems with reality, I'd advise anyone to keep you and knives far apart, if only for your own safety.
/the Novatose Workout - I'm getting bulging biceps from all this pointing and laughing!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
