Forgot your password?
typodupeerror
Sun Microsystems IBM Operating Systems Software

Mainframe OpenSolaris Now Available 135

Posted by ScuttleMonkey
from the better-late-than-never dept.
BBCWatcher writes "When Sun released Solaris to the open source community in the form of OpenSolaris, would anyone have guessed that it would soon wind up running on IBM System z mainframes? Amazingly, that milestone has now been achieved. Sine Nomine Associates is making its first release of OpenSolaris for System z available for free and public download. Source code is also available. OpenSolaris for System z requires a System z9 or z10 mainframe and z/VM, the hypervisor that's nearly universal to mainframe Linux installations. (The free, limited term z/VM Evaluation Edition is available for z10 machines.) Like Linux, OpenSolaris will run on reduced price IFL processors."
This discussion has been archived. No new comments can be posted.

Mainframe OpenSolaris Now Available

Comments Filter:
  • by tjstork (137384) <todd,bandrowsky&gmail,com> on Friday October 17, 2008 @12:27PM (#25414027) Homepage Journal

    I have a big old z in my basement that I've been itching to upgrade!

  • Okay... (Score:4, Funny)

    by LWATCDR (28044) on Friday October 17, 2008 @12:33PM (#25414115) Homepage Journal

    And IBM mainframe running Solaris...
    Now I have seen everything. Next AIX on the Sparc.

  • by TheLink (130905) on Friday October 17, 2008 @12:38PM (#25414181) Journal
    The way I see it, IBM is in the business of providing so much choice to customers that they need expensive IBM consultants to help them decide.

    Microsoft will sell you the Microsoft Way of doing things.

    Whereas IBM will say "You want a Active Directory server, a Z mainframe with RedHat, OpenSolaris and Oracle, Cisco switches, and there must be full J2EE buzzword compliance? No problem, just sign here".

    Careful to make sure they will actually do the job though, and not outsource it to a bunch of fresh PHP coders in India ;).
  • IBM has a strangle-hold on the high-margin mainframe world. This is causing issues in the Big Blue God Pod right now, be certain.

    • by Amarok.Org (514102) on Friday October 17, 2008 @12:51PM (#25414375)

      How so? If customers have a need for Solaris, would IBM rather see them go buy some Sparc gear from Sun, or a few extra processors for their System z complex?

    • by LWATCDR (28044) on Friday October 17, 2008 @01:00PM (#25414519) Homepage Journal

      Actually they are jumping for joy.
      Now if a Solaris shop needs some big Iron IBM can walk right in and sell a Z to them.
      If an IBM shop wants Solaris then IBM can say hey no need to by Sun hardware just put in on your Z.
      This is a happy day in Armonk.

      • by fm6 (162816)

        All spelling and grammar errors are intentional. Grammar Nazis' need entertainment.

        That should be "grammatical" and "Nazis" (no apostrophe).

        Thank you for the entertainment.

  • Seriously, I've run Suse 10 on an IFL engine. It's so slow, I don't know how anyone could run anything serious on it. I have an old laptop that matches or exceeds the performance in about every measurable way. Mainframe Linux and now Mainframe Solaris is a joke.
    • Re: (Score:3, Insightful)

      Seriously, I've run Suse 10 on an IFL engine. It's so slow, I don't know how anyone could run anything serious on it.

      That's why their called 'z's zzzzz zzzzz zzzzz....

      No, seriously, mainframes aren't about performance. They're about stability. Think about 16-core server with 40 GB of RAM running Solaris, AIX or Linux as a Ferrari Testerosa, while the Z10 is more like Abrams M1A1. Not as fast the Testerosa, but pretty quick for something that weights over 60 metric tons....

      • Stability. That's what everyone always comes back with, but it seems like the Mainframe side of my company has as many unplanned outages as the distributed side. Not to mention we run circles around them in terms of data processing.
        • Re: (Score:2, Insightful)

          Got any data to back this up? Usually I find people who say such things have a distorted view of reality. Not saying that you do, but I hear people say that and almost none of them have real evidence with which to backup their statement.

          • by Znork (31774)

            From what I've seen, weekly IPLs seems to be close to standard practice within the mainframe world. With a reboot frequency like that you're not going to even see most stability issues.

            Has it even become possible to switch to and from daylight savings yet without rebooting the mainframe?

            almost none of them have real evidence with which to backup their statement.

            Yes, well, rather like mainframe benchmarks, eh?

            • Re: (Score:3, Interesting)

              by BBCWatcher (900486)

              I doubt there's a reason for the IPLs (reboots). If your mainframe operators are doing it, they're probably just doing it because (seriously) somebody had a memory leak 30 years ago and that was how they "fixed" it. And nobody bothered to update the procedures manual. Nor did anybody ask them, "Hey, can we improve the SLA (Service Level Agreement) here?" "Sure boss, I'll just stop IPLing. Let's try skipping the next one." That's usually how that conversation goes, seriously.

              In fact, if you've got a Couplin

              • by Znork (31774)

                "Hey, can we improve the SLA (Service Level Agreement) here?"

                Usually, the 'scheduled' IPL's don't count into the SLA uptime measurements, so it doesn't seem like there's much pressure to avoid them.

                • Sure they do, at least in most mainframe operational organizations. (Non-mainframe operations are a different story. You ask "what about avoiding planned outages?" and they look at you funny.) In fact, it's even possible the reboots aren't actually happening, but the operators have reserved the right (in the SLA) to do them, so they declare that early Sunday mornings (for example) are planned outages. If you want a different SLA, ask for it. This certainly isn't a technical problem.
            • by mikechant (729173)

              From what I've seen, weekly IPLs seems to be close to standard practice within the mainframe world. With a reboot frequency like that you're not going to even see most stability issues.

              We run IBM z Series mainframes on behalf of a large number of customers and the typical IPL frequency is now about every 2-3 months. It could probably be less frequent on some systems. If you *have* to IPL every week, you're not running your system properly.

              Has it even become possible to switch to and from daylight savings ye

              • by Znork (31774)

                I would wonder how well any software on a non-mainframe platform (e.g. Oracle on Linux) would cope with time going backwards - surely this would screw up the logs etc?

                Time on Unix platforms basically doesn't go backwards. It's defined as seconds since 1970, and the number of seconds since then goes only one way.

                How that information is _presented_ to human consumers is a different matter, and it's at that layer the DST stuff is done. So presentations for human consumption like time stamps in logs can be ambi

        • by dedazo (737510)

          With all due respect, your mainframe people must be idiots. Nothing comes close to big iron in terms of processing capabilities and uptime.

          • That's a distinct possibility. It's something I have believed for quite a while...
            • by dedazo (737510)

              Heh. Think about this: At a former client of mine (your typical $LARGE_CORP) they had three physical boxes, I forget how many sysplexes and a number of LPAR instances. Two LPARs (one each ona separate physical box) where responsible for production (one hot, one spare). Between them they had something like seven years of actual service uptime.

              I'll never cease to be amazed at that culture and how different it is from ours. Changing a network card (or whatever) on those things required a gaggle of IBM consulta

        • by Tony Hoyle (11698)

          If a mainframe doesn't have damned near 100% uptime then someone needs to be fired. Those things just don't break. Hell, they phone the service engineer for you *before* they break and you suddenly get the IBM guy turn up on the doorstep with a replacement CPU (which is hotswapped.. no downtime).

          • It's not the hardware fails, it's the third party software that fails.
            • by LWATCDR (28044)

              Ahh. Well then that is your problem. That shouldn't happen at all. Also when you talk about through put you should look at strengths of the two systems.
              If you are doing anything with a lot of floating point. The Zmachine will lose to Intel every time. If you want to do that then get one of the Big Power boxes, a bunch of X86, or one of the big Itantium boxes with a lot of cores.
              If you are doing millions of database transactions and you want to make sure that it NEVER goes down. Get a Zmachine with an applic

        • If this is true then your mainframe guys need to be fired ASAP.

          It used to be a standing joke that nobody got fired for buying IBM but they did if it EVER went down unplanned.

          Also the throughput on a mainframe is truly astounding. I hate to think just how bad the software must be for a mainframe to be considered slow!

      • by $RANDOMLUSER (804576) on Friday October 17, 2008 @01:42PM (#25415101)
        Stability and I/O (particularly disc) bandwidth. Very important.
        • Right. I forgot the part about I/O bandwidth. It's more about the utilization rates than about the speed of any one task/transaction/etc.

      • by steelfood (895457)

        That having been said, the next logical step would be to compare the amount of compensation that mainframe operators are getting versus that of other server operators.

      • by samsonov (581161)
        Mainframes are typically for number crunching apps. I had to run some comparisons and for the RHEL5/WebSphere Process Server (ESB) combination we ran - it was keeping up quite well with transactions. I had to pull them back when the business was suggesting to run their web server there. The cost was not justified.
    • Re: (Score:2, Insightful)

      by bws111 (1216812)
      Your laptop can meet or exceed the IO performance? How about the memory access performance? Your laptop has MTBF measured in decades? Or by 'every measurable way' do you mean simple CP performance? These machines are not about CP performance.
      • Yes, even the IO performance which was surprising to me. It's running on an older Mainframe, i.e. not a new Z10, but still. Granted, I'm totally at the mercy of the Mainframe admins that control DASD access. It doesn't matter though, even if I had faster IO, you still have to have the CPU to process the data once it's been retrieved. My tests showed that the CPU spent less time in IO wait, but they were so slow to do anything else, it didn't matter.
      • by Znork (31774)

        From what I've seen in a SAN in a mixed environment, it certainly isn't the mainframe that's exposing the FC bottlenecks. It's a long time ago that the mainframe had any special hardware.

        Frankly, I've heard so many sales pitches for so long that there's only one thing that matters. Publish the benchmarks or it's just hot air.

        In the case of mainframes I haven't seen any serious benchmarks for more than a decade. I expect performance to be entirely predictable from that point only.

  • Ive never seen a System Z, and then only Sun boxes I've ever seen in real life are the ones I've owned. But from my experience with Solaris, and Sparc, Its the best damn arch/OS ever made. Im a big Linux fan (my /. nickname is a joke ;) ) But if it were up to me, id run everything Sun Way to go boys! now, There real Question is: Solaris and System z: How well does it run crysis ;)
    • by thogard (43403)

      How long can you run solaris 10 without having to add a security update that requires a reboot? How long can you do the same with a real mainframe os?

  • I'd still like to see Solaris and/or BSD come to the IBM Power Systems line. I think it'd be pretty cool to run Solaris or *BSD in an LPAR next to i5/OS and/or AIX.

  • Or maybe Java with plain text paragraphs?
  • When the PowerPC CPU was first introduced, everyone was going to play on the new platform. IBM AIX was trivial, of course, because the PowerPC is based on the POWER CPU. But there was Windows NT 3.51, Mac OS of course, and this thing from Sun called Solaris.

    Sun decided to stay with their own chips and then branched out to Intel x86 and AMD64, Microsoft eventually went back to an all-Intel code base (dropping Alpha support as well). The real killer for those boxes? IBM's port of OS/2 failed. Failed h

    • by dboyes99 (1388249)

      Except it wasn't Sun or IBM's idea.

      Sun and IBM were bystanders. SNA did it alone, with no development assistance from either.

  • Customer: Hello, IBM, I want to run *ix on my mainframe.

    IBM: Sure. Are you wanting Novell SUSE Enterprise Linux Server or Red Hat Enterprise Linux?

    Customer: Sorry, I meant Unix...

    IBM: Sure. So you are wanting Unix System Services?

    Customer: No I want to run that new openSolaris on it.

    IBM: Let me get this straight, you are wanting to run an unsupported hobbiest Unix variant on your multi-million dollar mainframe, correct?

    Customer: Uhh... no, I want openSolaris... oh... wait a minute, I see your point. .

    • IBM: "Open source? Let me transfer you to our Global Technology Services organization who would be happy to write you a very special support contract at a fair price. They'll support just about anything you can imagine, including whole business processes. Would you like them to take over welfare payments, digital TV transition consumer rebate processing, and backoffice call center support functions too?"
  • by BBCWatcher (900486) on Friday October 17, 2008 @03:08PM (#25416189)

    Every time a mainframe story comes up on Slashdot we seem to get the skeptics who point out that an X86 processor core can add or multiply two numbers (stored in registers anyway) about as fast as a single System z10 core, at least as long as they're integers. (z10s have hardware decimal floating point.) Based on this brilliant SPECint-y observation, combined with the fact that a System z10 EC Linux processor has an advertised one-time charge of $125K, these "experts" thus conclude that no one could possibly buy a mainframe because it's just so darn expensive. (Note that's one-time charge, folks: if you do a hardware model upgrade typical IBM practice is to charge you something for the frame swap but not to charge you again for turning on the processors.) Of course, in the same discussion people don't bother to explain why the same argument also holds for SPARC CPUs. Heck, why not run business applications on Sony Playstation 3s or ARMs? They're even "cheaper."

    May I humbly point out that IBM just posted (yesterday) another record quarter for mainframe sales. Revenues were up 25 percent, with double digit growth in every region of the world. Because prices are higher? No, the opposite: shipped capacity was up 49 percent; specialty capacity (including Linux processors) was up 120 percent. And IBM has been posting quarters like this for years now. This mainframe stuff is wildly successful and gaining marketshare.

    Why? Because, with all due respect, you're an idiot if you stop your careful business case analysis at the first sentence above. Unless you're running SETI@Home, rendering the next Pixar movie, or simulating nuclear explosions, business applications across many users just don't run that way. Companies (particularly CFOs) and big data center managers are not (generally) idiots. They buy this stuff because it works wonderfully and because it's cost-effective, taking all costs into consideration. Think $125K (once) is a lot of money? What's your salary, dude? Who are the richest single human beings in the software industry, and did they get that way because software is free? And how much did it cost the London Stock Exchange when they couldn't trade? Are you the guy who wants to explain why you have to build another $20M data center because you can't power or cool yet another X86 chip? In the real world, there are single companies running hundreds of these mainframe CPUs. And they run at 80%+ busy 24 hours a day, by the way.

    Honestly, there are way too many Slashdotters who are much more the stubborn non-thinkers that they probably accused mainframe-skilled people of being a few years ago. It's a different world: grow up. The boring but wonderful truth is that -- surprise! -- different servers are good at different things! Intel/AMD X86 servers are useful in certain ways, and so are System z servers. Even in the same data center. Wow, what a concept!

    • Re: (Score:3, Insightful)

      by TheRaven64 (641858)
      Even discounting all of the good points you've made, System/Z has one other advantage. It's a direct descendant of System/360 and still runs software written in 1960 without a recompile. The same mainframe can now also run virtual Linux and Solaris instances. Sure, you could run something like Hercules on your x86 machine, but who would support it, and would you trust it on that bit of critical software that has been supporting your business for almost fifty years?
      • Re: (Score:3, Informative)

        by Richard Steiner (1585)

        I wish Unisys had the foresight of IBM when it comes to running POSIX software on their mainframes. Or maybe they do and I just don't know about it. They are seemingly a stealth company, after all. But they still manage to sell some very good mainframe server hardware (Clearpath Dorado and Libra servers running OS2200 and MCP respectively), and both of those OSes run fairly old software as well (the 2200 stuff requires a recompile as of some point in the 1970's when the ABS format changed, but I'm not as

    • Part of the reason for the sales line is also that people are finding it impossible to move their code off of mainframes. You need to continue to support the environment, and that means upgrades. Plus, as long as it's sitting there, a lot of CIOs feel they should try to leverage it more or add some pizazz. We had a Linux partition on our box for 2 years that never even was installed with the OS. Our mainframe team wouldn't let anyone touch it and they didn't know anything about Linux.

      Our shop still has

    • by segedunum (883035)

      Of course, in the same discussion people don't bother to explain why the same argument also holds for SPARC CPUs.

      Decent analysis, apart from this one tiny flaw. SPARCs have traditionally competed in areas where there is a lot of crossover. Indeed, a lot of SPARC's business post-2000 has been eaten for breakfast by x86, and x86 running on Linux. This is why Sun has suffered more than most at the hands of x86 and Linux. SPARC simply cannot keep up with x86 for performance, and although PowerPCs and other pro

    • by LWATCDR (28044)

      Trust me they will not get it.
      Mainframes are dull. They just work. Most people on slashdot have only used PCs. They can not imagine a computer that you never have to shut down. That you can upgrade CPUs and RAM with out any downtime.
      All they know is that it will not transcode any faster than their X86 and costs as much as a house.

      • by IvyKing (732111)

        All they know is that it will not transcode any faster than their X86 and costs as much as a house.

        Hmmmm, I'd love to be able to buy a house like the one I'm in for what the low-end Z-series runs... Then again, I remember when moderate size mini-computers (e.g. CDC-1700) cost as much as a really nice house.

    • Unless you're running SETI@Home, rendering the next Pixar movie, or simulating nuclear explosions, business applications across many users just don't run that way.

      And if you are running them, you still might want IBM hardware:

      "Hello IBM? I need a Cell based cluster." or if you need it really cheap and don't need a HUGE amount of processing power.

      "Hello Terrasoft? I need a 8-Node PS3 cluster."

    • by Znork (31774)

      In the real world, there are single companies running hundreds of these mainframe CPUs. And they run at 80%+ busy 24 hours a day, by the way.

      See, here's a perfect example of why people tend to be skeptical. You're saying hundreds like it's a lot. There are single companies running on thousands or tens of thousands of x86 CPUs, consolidated with VMware, Xen and other similar technologies and running at 80% load. It's not exceptional any more.

      Then we get marketing from mainframe people about running a hundred

      • IBM publishes lots of mainframe benchmarks and has for years. They're called LSPR tables [ibm.com], and there's a ton of data available. There are several different types of workloads measured. But there's this mythology out there, including among many on Slashdot, that benchmarks give you a single number and then you just pick the higher number and you're done. Oh, that'd be so simple and wonderful, but it just doesn't work that way. First of all, your workloads (and time of day patterns) won't match the ones in LSP

Nobody said computers were going to be polite.

Working...