INFO-VAX Sun, 19 Aug 2007 Volume 2007 : Issue 454 Contents: Re: Alpha/Integrity Dead Pool RE: Alpha/Integrity Dead Pool Festivals and Events Re: Gzip 1.3.12? Re: Linux Happiness Re: Linux Happiness (was:Re: Free to good home. Microvaxes, Vaxstations... Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion Re: Wonderful things happen to an OS when it has an internal champion RE: Wonderful things happen to an OS when it has an internal champion ---------------------------------------------------------------------- Date: Sun, 19 Aug 2007 03:34:04 -0500 From: Ron Johnson Subject: Re: Alpha/Integrity Dead Pool Message-ID: <1gTxi.152349$dA7.132621@newsfe16.lga> On 08/18/07 19:44, Paul Raulerson wrote: > >> -----Original Message----- >> From: Main, Kerry [mailto:Kerry.Main@hp.com] >> Sent: Saturday, August 18, 2007 1:00 PM >> To: Info-VAX@Mvb.Saic.Com >> Subject: RE: Alpha/Integrity Dead Pool [snip] >> >> Btw, not sure if you knew this but z/OS runs on Integrity. >> > > Yep- we had lunch with those folks last Tuesday, and got a good long look > at the brand new 64 core model they put out. > > It is definitely sweet, since the PSI guys are all Amdahl folks, and really > know what they are talking about. > > The box is limited to 2000 Mainfame MIPS though (with 64 cores running!), > less if you are running concurrent Windows on it. That's kind of iffy for > a lot of installations. And of course, IBM will NOT license anything to run > on the beastie, so legally, all you can run is zLinux. > > But it is SWEET. You can move processor cores from Windows to > zArch and back seamlessly. Applications never even know it happened, except > that > they suddenly find they have more or less processors. > > So you *can* run z/OS on Integrity, just requires the PSI firmware > "enhancements." How can POWER binaries run on ia64 without there being an emulator or FX!32-like incremental translator? > They are pretty adamant in claiming it is not an emulator, but I've yet to > see the > Itanium chip that can execute: > > MVC 0(40,R1),0(R2) > AHI R2,40 > ... > ... > and so on. :) Easy: have your BAL code generator generate ia64 opcodes. That's what DEC/Cpq/HP does with MACRO-32 that is to be run on Alpha or ia64. > >> Reference: >> http://www.platform-solutions.com/news/system64_announcement.pdf >> -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Sun, 19 Aug 2007 07:31:13 -0500 From: "Paul Raulerson" Subject: RE: Alpha/Integrity Dead Pool Message-ID: <001c01c7e25c$d4d6bd20$7e843760$@com> > -----Original Message----- > From: Ron Johnson [mailto:ron.l.johnson@cox.net] > Sent: Sunday, August 19, 2007 3:34 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Alpha/Integrity Dead Pool > > On 08/18/07 19:44, Paul Raulerson wrote: > > > >> -----Original Message----- > >> From: Main, Kerry [mailto:Kerry.Main@hp.com] > >> Sent: Saturday, August 18, 2007 1:00 PM > >> To: Info-VAX@Mvb.Saic.Com > >> Subject: RE: Alpha/Integrity Dead Pool > [snip] > >> > >> Btw, not sure if you knew this but z/OS runs on Integrity. > >> > > > > Yep- we had lunch with those folks last Tuesday, and got a good long > look > > at the brand new 64 core model they put out. > > > > It is definitely sweet, since the PSI guys are all Amdahl folks, and > really > > know what they are talking about. > > > > The box is limited to 2000 Mainfame MIPS though (with 64 cores > running!), > > less if you are running concurrent Windows on it. That's kind of iffy > for > > a lot of installations. And of course, IBM will NOT license anything > to run > > on the beastie, so legally, all you can run is zLinux. > > > > But it is SWEET. You can move processor cores from Windows to > > zArch and back seamlessly. Applications never even know it happened, > except > > that > > they suddenly find they have more or less processors. > > > > So you *can* run z/OS on Integrity, just requires the PSI firmware > > "enhancements." > > How can POWER binaries run on ia64 without there being an emulator > or FX!32-like incremental translator? > Mainframe (i.e. z Architecture ) != POWER. POWER technology is currently used in the CP, but the actual instructions that a zArch machine sees and uses are very very much CISC. Actual mainframes implement some of the 3500 or so instructions in hardware, and others in millicode (firmware). This is a fun little issue; what the PSI firmware appears to do is use the EPIC chips to load additional firmware for the IA64 cores in "zArch" mode. PSI claims that isn't emulation, and I, like you, claim that if it quacks like a duck and it acts like a duck... Also, 2000 mainframe MIPS in a 64 core install means each core is driving only a little over 31 MIPS each. Perhaps more like 35MIPS each, since some cores have to be dedicated as SAPs. Those are mainframe MIPS of course, but that seems a little low per IA64 core. > > They are pretty adamant in claiming it is not an emulator, but I've > yet to > > see the > > Itanium chip that can execute: > > > > MVC 0(40,R1),0(R2) > > AHI R2,40 > > ... > > ... > > and so on. :) > > Easy: have your BAL code generator generate ia64 opcodes. > > That's what DEC/Cpq/HP does with MACRO-32 that is to be run on Alpha > or ia64. > > > > >> Reference: > >> http://www.platform-solutions.com/news/system64_announcement.pdf > >> > > -- > Ron Johnson, Jr. > Jefferson LA USA > > Give a man a fish, and he eats for a day. > Hit him with a fish, and he goes away for good! ------------------------------ Date: Sun, 19 Aug 2007 08:47:14 -0700 From: rvfulltime Subject: Festivals and Events Message-ID: <5rogc35vuhjpu99fenb0lholbemef6389s@4ax.com> As we wind down full timing and start part timing, we plan on visiting various festivals and events around the country in the 5th wheel. So we started to put together a list of top events, but came up a bit short. Additions to the list would be appreciated. (*) means "been there done that". (?) we think this is the place. Blank means we don't know where the best is. Any help would be appreciated. Auto Racing: Indianapolis 500 and Daytona 500 Horse Racing: Kentucky Derby Country Music: CMA Fest (*) Mardi Gras: Lafayette (?) and New Orleans Hot Air Balloons: Albuquerque Hydro Boat Racing: Seattle Shakespeare Festival: Ashland (*) Jazz Festival: Golf: US Open College Football: Rose Bowl Chili Cookoff: Beer Tasting: Wine Tasting: Air Races: Reno Rock and Gem Show: Quartzsite RV Show: Portland (?)(*) Dog Sled Races: Iditorod Anchorage Western Art Show: Oklahoma City (?)(*) Olympic Trails: Eugene 2008 Scottish Games: -- Posted via a free Usenet account from http://www.teranews.com ------------------------------ Date: Sun, 19 Aug 2007 08:28:04 -0700 From: "Tom Linden" Subject: Re: Gzip 1.3.12? Message-ID: On Sat, 18 Aug 2007 00:05:45 -0700, Steven M. Schweda = = wrote: > I don't know if it has any actual advantages over (or disadvantages= > under) previous versions on VMS, but thrill-seekers may wish to > investigate a VMS port of the current version (1.3.12) of gzip: > > http://antinode.org/dec/sw/gzip.html > > It does offer large-file support on newer non-VAX systems, but that= 's > new only if you're comparing it with version 1.2.x. > > As always, complaints are welcome. Getting it to build with VAX C > was more of an ordeal than before, and I didn't try it on any VMS V6.x= > systems, and I hardly tested it, so I won't offer more than my usual > guarantee (that is, none), but it's mostly portable code, so what coul= d > go wrong? > > If enough acclaim accumulates, I could try to get the VMS-related > changes pushed back into the main source stream. It doesn't look as i= f > anyone official has touched the VMS code since June 1993, but they did= > recently fix a spelling error in the VMS help file, so there just migh= t > be some interest in GNUville. (In which case, it'd be less embarrassi= ng > if all the VMS-specific bugs were found and fixed before I started > pleading for acceptance.) > > ----------------------------------------------------------------------= -- > > Steven M. Schweda sms@antinode-org > 382 South Warwick Street (+1) 651-699-9818 > Saint Paul MN 55105-2547 Just tried it on a large file, don't know if this is a bug or not. = Anything you want me to look at? HAFNER> gunzip13 MULTICS$TAR.GZ HAFNER::TOM 08:14:18 GZIP1_3_1 CPU=3D00:03:58.31 PF=3D3023 IO=3D85029 ME= M=3D196 gzip1_3_12: DISK$COMMON:[MULTICS]MULTICS$TAR: non-translatable vms error= = code: 0 x1828A %rms-e-flk, file currently locked by another user %NONAME-E-NOMSG, Message number 00000002 HAFNER> ll Directory DISK$COMMON:[MULTICS] MULTICS$TAR.;1 8772920 19-AUG-2007 08:08:02.43 [SYSTEM] -- = PL/I for OpenVMS www.kednos.com ------------------------------ Date: Sun, 19 Aug 2007 18:08:50 +0930 From: Mark Daniel Subject: Re: Linux Happiness Message-ID: <13cg0dp7hfjah03@corp.supernews.com> Phillip Helbig---remove CLOTHES to reply wrote: > In article , > bradhamilton@comcast.net (Brad Hamilton) writes: > > >>>Shame on you for not moving to Linux. You'd be happier. >> >>In general, I'm happier with Linux than with Windows on my laptop (sorry, Bill >>G., but wireless network "support" for *BSD is chancy - otherwise, I'd be there >>with my laptop :-)) > > > On the other hand, another Bill G. is happier. We don't know that. Of course, as my maternal Grand-Mother was want to say, "Money may not buy you happiness but it's better to be rich and miserable than poor and miserable." ------------------------------ Date: Sun, 19 Aug 2007 08:22:21 +0000 (UTC) From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) Subject: Re: Linux Happiness (was:Re: Free to good home. Microvaxes, Vaxstations... Message-ID: In article , bradhamilton@comcast.net (Brad Hamilton) writes: > >Shame on you for not moving to Linux. You'd be happier. > > In general, I'm happier with Linux than with Windows on my laptop (sorry, Bill > G., but wireless network "support" for *BSD is chancy - otherwise, I'd be there > with my laptop :-)) On the other hand, another Bill G. is happier. ------------------------------ Date: Sun, 19 Aug 2007 09:11:48 +0200 From: "P. Sture" Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: In article <46C7A15C.7050801@comcast.net>, "Richard B. Gilbert" wrote: > Ron Johnson wrote: > > On 08/18/07 11:55, Main, Kerry wrote: > > [snip] > > > >>After you do OS virtualization using solutions like VMware, Zen or any other > >>solution, the next question out of the CIO's mouth will be "Great. Now how > >>are you going to reduce the number of OS's, so I can cut my FTE numbers?" > >> > >>And that is where App stacking, Workload Mgmt comes in. > > > > > > What exactly *is* App Stacking, other than "running multiple apps on > > the same machine"? > > > > > The whole problem, as I understand it, is that Windows has traditionally > been not very good at protecting applications from each other! > > Windows has gotten a great deal better in the last seven years or so but > it would still take a brave man to run two applications simultaneously > on one server. > > Running multiple virtual servers on one physical server seems to solve > this problem; I guess VMWare provides the protection that Windows cannot! The other problem is that Windows application developers seem to think that they can take control of a complete machine. The sad side of that is that people let them. -- Paul Sture Sue's OpenVMS bookmarks: http://eisner.encompasserve.org/~sture/ovms-bookmarks.html ------------------------------ Date: Sun, 19 Aug 2007 03:28:15 -0500 From: Ron Johnson Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: On 08/18/07 20:48, Richard B. Gilbert wrote: > Ron Johnson wrote: >> On 08/18/07 11:55, Main, Kerry wrote: >> [snip] >> >>> After you do OS virtualization using solutions like VMware, Zen or >>> any other >>> solution, the next question out of the CIO's mouth will be "Great. >>> Now how >>> are you going to reduce the number of OS's, so I can cut my FTE >>> numbers?" >>> >>> And that is where App stacking, Workload Mgmt comes in. >> >> >> What exactly *is* App Stacking, other than "running multiple apps on >> the same machine"? >> > > > The whole problem, as I understand it, is that Windows has traditionally > been not very good at protecting applications from each other! > > Windows has gotten a great deal better in the last seven years or so but > it would still take a brave man to run two applications simultaneously > on one server. > > Running multiple virtual servers on one physical server seems to solve > this problem; I guess VMWare provides the protection that Windows cannot! It *constantly* amazes me that so many IT people such sheep to be suck into the MSFT mindset like this. There are two reasons people in the Linux world run one app per box: 1. They are Windows refugees who don't know any better, or 2. "don't put all your eggs in one basket". Just like VMS (and BSD and Solaris, etc, etc), it's fully capable of running multiple apps on a single box. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Sun, 19 Aug 2007 11:52:44 -0500 From: David J Dachtera Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <46C8755C.1527DBC0@spam.comcast.net> Ron Johnson wrote: > > On 08/18/07 20:48, Richard B. Gilbert wrote: > > Ron Johnson wrote: > >> On 08/18/07 11:55, Main, Kerry wrote: > >> [snip] > >> > >>> After you do OS virtualization using solutions like VMware, Zen or > >>> any other > >>> solution, the next question out of the CIO's mouth will be "Great. > >>> Now how > >>> are you going to reduce the number of OS's, so I can cut my FTE > >>> numbers?" > >>> > >>> And that is where App stacking, Workload Mgmt comes in. > >> > >> > >> What exactly *is* App Stacking, other than "running multiple apps on > >> the same machine"? > >> > > > > > > The whole problem, as I understand it, is that Windows has traditionally > > been not very good at protecting applications from each other! > > > > Windows has gotten a great deal better in the last seven years or so but > > it would still take a brave man to run two applications simultaneously > > on one server. > > > > Running multiple virtual servers on one physical server seems to solve > > this problem; I guess VMWare provides the protection that Windows cannot! > > It *constantly* amazes me that so many IT people such sheep to be > suck into the MSFT mindset like this. > > There are two reasons people in the Linux world run one app per box: > 1. They are Windows refugees who don't know any better, or > 2. "don't put all your eggs in one basket". > > Just like VMS (and BSD and Solaris, etc, etc), it's fully capable of > running multiple apps on a single box. I should think one major problem is the lack of a central co-ordinating agency for things like Registry key names. It's entirely possibly - and quite likely - that one app.'s key path/name could duplicate that of another, with potentially disasterous results. Another major limitation is the lack of VMS-like logical names. One way that multiple app.'s - even multiple instances of the same app. - are separated in VMS-land is by the use of group-level and privately-named logical name tables. That way my "C" drive could be anything from DKB100: to DAVIDS-INSTANCE: (DKB100:[david.]/trans=conc), while yours could be anything from DKB200: to YOUR-INSTANCE: (DKB100:[yours.]/trans=conc) as just one example. There are probably some ways around this using Windows environment symbols. However, this would be so far outside of the typical Windows programmer's paradigm that trying to introduce it now would be tantamount to introducing a whole new operating environment. Indeed, these critters have grown so accustommed to "I'll solve this problem MY way, regardless of whether it clashes with anything else" that the probability of changing it short of WMware or equiv. is indeed quite low. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Sun, 19 Aug 2007 12:00:34 -0500 From: David J Dachtera Subject: Re: Wonderful things happen to an OS when it has an internal champion Message-ID: <46C87732.C94EC917@spam.comcast.net> JF Mezei wrote: > > Main, Kerry wrote: > > Did you actually read the brochure? > > I even attended the Decus presentations. VMS doesn't need > virtualisation. If you want availability, you want your cluster on > separate hardware to begin with. Now, expand that paradigm a bit so that you have multiple instances on one box and multiple instances on another box at another location. That gives you multpile "Galaxies" within a cluster. Within one box, picture this: Suppose the console could present the physical "EWA0" port virtually to multiple Galaxy instances as, for example, VWA0. I could see some challenges allowing multiple MAC addresses that way (since DECnet-IV likes to do that). Otherwise, that allows sharing one ethernet port between multiple instances. Likewise, FGA0 could be shared as, for example, VGA0; PKA0 as VKA0, and so on. That's how virtualization can work on IBM System-P machines, if you want to set it up that way. Individual ethernet cards and HBAs may not be costly, but PCI drawers are. Why not maximize the utilization of your investment? -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Sun, 19 Aug 2007 14:00:00 +0000 From: "Main, Kerry" Subject: RE: Wonderful things happen to an OS when it has an internal champion Message-ID: > -----Original Message----- > From: Richard B. Gilbert [mailto:rgilbert88@comcast.net] > Sent: August 18, 2007 9:48 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: Wonderful things happen to an OS when it has an internal > champion > > Ron Johnson wrote: > > On 08/18/07 11:55, Main, Kerry wrote: > > [snip] > > > >>After you do OS virtualization using solutions like VMware, Zen or > any other > >>solution, the next question out of the CIO's mouth will be "Great. > Now how > >>are you going to reduce the number of OS's, so I can cut my FTE > numbers?" > >> > >>And that is where App stacking, Workload Mgmt comes in. > > > > > > What exactly *is* App Stacking, other than "running multiple apps on > > the same machine"? > > > > > The whole problem, as I understand it, is that Windows has > traditionally > been not very good at protecting applications from each other! > > Windows has gotten a great deal better in the last seven years or so > but > it would still take a brave man to run two applications simultaneously > on one server. > > Running multiple virtual servers on one physical server seems to solve > this problem; I guess VMWare provides the protection that Windows > cannot! Yes, VMware, ZEN and any other OS stacking virtualization scheme (including z/OS and Linux) does have some savings related to reduced hardware, and DC space/cooling. OS stacking is also politically easier to implement as one App is "isolated= " from the other so if APPA on VM-1 does something dumb, then APPB on VM-2 is not impacted (at least not directly). The downside is that OS stacking is a temporary measure as it does not addr= ess the 800lb gorilla of IT costs .. FTE numbers and associated staff costs whi= ch are typically 60-70% of most IT budgets today. Remember that even if an OS is virtualized, it still needs to be licensed, upgraded, monitored, secured, backed up, data archived, disk space for sys disk and page/swap files, user authorizations managed, print queues, batch jobs managed and monitored etc. Hence, migrating from 100 physical servers to 10 VM machines does not mean you can reduce the number of support FTE people, because the real workload of managing each OS instance remains the same (actually some increase as yo= u also need to manage, support and license the base OS's (Linux) associated with each VM server physical system. Yes, tools are available to assist in this, but the real workloads do not change. There is also an increase in the level of complexity which adds to the overall troubleshooting process as base host issues may reflect on all the VM's it is hosting. The "just reboot the server" mentality means 10 OS's will be rebooted so that is a whole different environment. Once companies are completed their current phase of reducing the up front HW / DC costs with OS virtualization, the next wave will be in reducing the number of OS's and that is where the fun and gritting of teeth will begin. :-) Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ End of INFO-VAX 2007.454 ************************