Talk:VM (operating system)
Is it just me, or does this article seem very strongly biased?
- It's not just you. Admittedly I work for a competitor, but this reads like marketing literature. --184.108.40.206 23:04, 2 May 2006 (UTC)
- Agreed. I'm mildly surprised this one hasn't gotten an NPOV complaint. In any event, the merge seems like a Good Idea. Jay Maynard 16:11, 9 May 2006 (UTC)
Does "VM/390" exist? I thought VM/ESA was the initial S/390 VM offering. TimP 03:38, 13 January 2006 (UTC)
- No. The term VM/390 is one of a dozen terms devised by someone who didn't know the nomenclature and took a wild stab in the dark. What existed were
Not released to customers.
- Virtual Machine Facility/370
- VM/SP HPO
- VM/XA MA
- VM/XA SF
- VM/XA SP
- SEPP and BSEPP installed on top of the free base; somewhere along the line the free base disappeared and the program product versions were self contained. Shmuel (Seymour J.) Metz Username:Chatul (talk) 21:48, 21 September 2010 (UTC)
The VM nomenclenture has never been very tidy - VM/XA, VM/HPO and VM/SP have also been official names.
Both this article and the VM/CMS one are pretty sad. Henry Troup 19:38, 20 September 2006 (UTC)
- 1 Some thoughts from a z/VM sysprog
- 2 Restructuring of VM and CP/CMS material
- 3 AIX/Linux under VM
- 4 Dates missing
- 5 Restructure of articles about IBM mainframe operating systems
- 6 Image copyright problem with Image:ZVMCMS.png
- 7 He's dead, Jim - AIX
- 8 Provenance of term Hypervisor
- 9 Examples of (deeply) nested virtualization?
- 10 VM - the missing bits!
Some thoughts from a z/VM sysprog
Well, after all the advertising, some thoughts from a z/VM sysprog...
After some time of getting used to it z/VM is a great system to work on. But, of course, it also has its disadvantages.
On the downside, z/VM can be a real pain for the sysprog staff. Many things have to be done manually, especially when upgrading the z/VM environment to a new release which includes tedious editing sessions merging the old system files with the new ones. IBM doesn't offer a simple procedure for this task, while system maintenance can usually be done with a simple command, SERVICE.
Luckily z/VM features REXX, a great scripting language which is nowadays available for almost any system platform, helping to automate the upgrade process. Still, since z/VM had'nt been too popular before Linux for z/Series came to existance, there's a serious lack in support and education personnel - any sysprog dealing with z/VM knows most of the IBM people working on z/VM personally. While these guys do a great job supporting customers, IBM would be well advised to recrute new experts in the near future.
What's more: if a customer doesn't stick to the IBM way of doing things (like running only a few z/VM LPARS) it shows an obvious lack of systems management functionality, at least 'out of the box'. There are many inofficial (aka unsupported) tools available from IBM's z/VM website, but their evaluation and installation is completely up to the customer, and you still find yourself coding many small helper utilities, because the things you need are simply non-existant.
Of course, all of this is absolutely transparent to the system images you run on z/VM, but it can occupy the z/VM sysprog quite a bit. Admittedly, it's fun to code on z/VM, but it can be hard to explain this to someone's boss, who has been told (probably by some IBM consultant?) that z/VM can be run absolutely stable (which is true, at least most of the time) and without much effort (which, in my experience, really depends on the given situation).
Compared to PC-based solutions, z/VM is much more powerful, but at the same time, much more complex to maintain. —Preceding unsigned comment added by 220.127.116.11 (talk • contribs) 19:52, 3 November 2006
Moved from the article to here by RossPatterson 01:07, 4 November 2006 (UTC).
Restructuring of VM and CP/CMS material
I have just finished a reshuffling/rewriting of material from VM and CP/CMS articles. I have tried to incorporate the various comments applicable to this page, though of course there were many issues involved. If you don't see something that you feel was important, naturally please go ahead and make changes. Trevor Hanson 19:39, 3 December 2006 (UTC)
Also: If anybody can upload a public-domain VM teddy bear for inclusion here, that would be nice. I see plenty of bears on the web, but obviously they do not belong to me. I think this would be a nice addition. (In contrast, I did not feel that the huge VM/370 welcome screen from earlier did much for the page.) Trevor Hanson 19:39, 3 December 2006 (UTC)
AIX/Linux under VM
I had stated that VM was required for AIX and Linux, but this was just challenged. After some further investigation, I see that I was misinformed. I had been relying on some strong and credible statements to that effect (which of course I can't put my finger on at the moment), but I now see from other sources that this must have been bogus. IBM's installation docs clearly describe native configurations, e.g. http://www.redbooks.ibm.com/abstracts/sg246264.html. There probably have been specific environments or situations where VM was the only way to go, which led to the claims in the sources I had trusted. I agree that the blanket statements here needed to disappear. I have thus removed the remaining claims about VM being necessary. It would be helpful if some good references about native AIX/Linux configurations could be located, particularly from an historical perspective about AIX. Trevor Hanson 18:55, 17 December 2006 (UTC)
I am not competent to contribute here, but I wish to point out there are too little dates (years), in this article, so that the timeline of the whole VM story is blurry. Even the "History" section only has one year: 1972. You can't figure out what happened later and when. If the experts could fix this, I'd be very grateful. —The preceding unsigned comment was added by 18.104.22.168 (talk) 08:24, August 22, 2007 (UTC)
- Good point. I had been hoping that a current z/VM expert would step in on this page and add some more detail about recent history. I have added the z/VM original release date (October 2000) and a link to IBM's z/VM release history. Plenty of early dates are found in the CP/CMS articles. Details on the middle years between 1972 and 2000 could be beefed up by somebody with the expertise and interest. Most of the release dates should be close to the associated hardware announcements, e.g. the XA versions were linked with the release of XA hardware. On the other hand, there probably aren't too many people who care about specifics on those events. The IBM archives remain the best bet for a timeline of versions. Trevor Hanson 22:34, 22 August 2007 (UTC)
Restructure of articles about IBM mainframe operating systems
After a big edit of MVS I concluded that the whole set of articles about IBM mainframe operating systems from System/360 onwards needed to be re-structured to minimise overlap and to make clearer the evolutionary relationships between these operating systems (notably in memory management, which is historically a major distinguishing feature). There is already some support for this proposal. Please add comments at Talk: MVS. Philcha 23:56, 20 October 2007 (UTC)
Image copyright problem with Image:ZVMCMS.png
The image Image:ZVMCMS.png is used in this article under a claim of fair use, but it does not have an adequate explanation for why it meets the requirements for such images when used here. In particular, for each page the image is used on, it must have an explanation linking to that page which explains why it needs to be used on that page. Please check
- That there is a non-free use rationale on the image's description page for the use in this article.
- That this article is linked to from the image description page.
He's dead, Jim - AIX
The S/370 versions of AIX have been dead for well over a decade; essentially, since the availability of a Unix environment in MVS/ESA SP Version 4. Linux on z, of course, is alive and well.Shmuel (Seymour J.) Metz Username:Chatul (talk) 21:48, 21 September 2010 (UTC)
Provenance of term Hypervisor
The term Hypervisor did not come in with VM; it was used in the 1960's. I don't recall whether it was used in the context of CP-67 or in a different context. Shmuel (Seymour J.) Metz Username:Chatul (talk) 18:51, 16 February 2011 (UTC)
- And what, pray tell, is your reference? With respect to VM/370 and its successors, the term Hypervisor was coined by me in late 1971 or early 1972, while we were reorganizing and rewriting the Control Program functions. At the time, source module names and external symbols were limited to 8 characters. IBM internal standards required that the first three letters were the assigned Component Code; the next three characters were a module identifier, and the last two were reserved for entry point identification. The component code for VM/370-CP was 'DMK' and the code for VM/370-CMS was 'DMS'. John Seymour was responsible for the CP program interrupts handler, DMKPRG; page faults came in there, and he was also the author of DMKPAG. I was responsible for the privileged operation exception handler, DMKPRV. One of the privileged operations handled by DMKPRV was the Diagnose instruction, which CP-67/CMS and VM/370 used for "paravirtualized" requests from the guest system to the virtual machine monitor - i.e. to the Hypervisor. I named and wrote DMKHVC, the Hypervisor Call module, derived from the S/360 Supervisor Call instruction (SVC). [I cannot avoid violating the 'No Original Research' constraints of Wikipedia, for reasons that should be obvious.] If you look around online, you should be able to find "The Origin of the VM/370 Time Sharing System" by R. J. (Bob) Creasy, IBM Journal of Research & Development, Vol. 25 No. 5, September 1981, and/or one or another version of Melinda Varian's "VM and the VM Community, Past, Present, and Future" from 1989 through the 1990's. Dave Tuttle (talk) 23:19, 24 February 2011 (UTC)
- And what, pray tell, is wrong with the reference in Hypervisor?
- I have been wrong before; no doubt it will happen again. I personally worked at MIT with a S/360-65 that had the IBM 7094 emulator microcode option, in late 1967 or early 1968. The "shared memory" in that case was actually the processor microcode, not main memory. The MIT Computation Center installed a S/360-40 in the same room as the (upgraded) S/360-65 and interconnected them via a Channel-to-Channel Adapter (CTCA). I worked during the summer of 1967 to modify a copy of IBM's Attached Support Processor (ASP) software, install it on the S/360-40, and set it up to run mixed-mode batch on the S/360-65. The ASP software on the Model 40 would recognize whether an incoming job (card deck) was an OS/360 job or an IBSYS (7094) job, then switch the main processor between OS/360 and IBSYS as necessary. I do not recall the use of Hypervisor in that context, but the circumstance does make sense. Dave Tuttle (talk) 16:48, 1 March 2011 (UTC)
- The ASP support for doing an IPL of the Main processor was not what I and Don Skiba were referinng to, and I've never heard it referred to by the term Hypervisor. RPQ E880801 allowed the emulator and OS/360 to be in memory concurrently, and had nothing to do with ASP. Shmuel (Seymour J.) Metz Username:Chatul (talk) 22:56, 7 March 2011 (UTC)
- I'm being to question the accuracy of this "History". Without a hard link back to RPQ E880801 the 2 citations provided here can't be considered accurate
Examples of (deeply) nested virtualization?
The paragraph about running VM inside VM mentions "The litterature" providing practical examples of nested virtualization "five levels deep". I'm curious to know which litterature that is so that I can find out what the examples were. 22.214.171.124 (talk) 16:13, 19 October 2011 (UTC)
If you look at page 55 of the document "VM and the VM Community" that is already referenced you'll find the following paragraph at the bottom:
<<The System/370 relocation architecture was different from the 360/67 architecture; it allowed both 2K and 4K pages and both 64K and 1M segments. So, Auroux modified his modified CP-67 to support 64K segments and the new System/370 instructions. He ran that system second-level, so he could run a virtual S/370 third-level. He developed a prototype “CP-370” in that third-level virtual machine. Then, to test this CP-370’s virtualization of System/370 virtual memory, he had to run it both third- and fourth-level, with a couple of CMS machines running fifth-level.>> — Preceding unsigned comment added by Deasington (talk • contribs) 11:58, 12 July 2012 (UTC)
VM - the missing bits!
In the late 1980's we did some interesting work running CICS directly on VM and also under VM/CMS and VM/PC. VM/PC doesn't seem to get a mention anywhere although it was a full VM implementation running on 370 order code on the XT or AT/370 hardware. There were also communications facilities between VM/PC and a host running either VM/CMS or CICS using a DIAG '78' (from memory...) interface. Does anyone remember more, or have documentation? — Preceding unsigned comment added by Deasington (talk • contribs) 15:26, 10 July 2012 (UTC)