1 INFO-VAX	Thu, 02 Sep 2004	Volume 2004 : Issue 486       Contents:! Re: A whopping 50 percent...  ??? ! RE: A whopping 50 percent...  ??? ! Re: A whopping 50 percent...  ??? ! Re: A whopping 50 percent...  ??? ! Re: A whopping 50 percent...  ??? ! Re: A whopping 50 percent...  ??? ! Re: A whopping 50 percent...  ??? ! Re: A whopping 50 percent...  ??? + Alphastation 200 storage controller options / Re: Alphastation 200 storage controller options / Re: Alphastation 200 storage controller options ( Re: Big IA64 test coming in a few months Re: charon vax emulator??? Re: charon vax emulator??? Re: charon vax emulator???< Re: CSWS_JAVA v2.1 - Tomcat v4.1.24 - update/upgrade Tomcat?< Re: CSWS_JAVA v2.1 - Tomcat v4.1.24 - update/upgrade Tomcat?< Re: CSWS_JAVA v2.1 - Tomcat v4.1.24 - update/upgrade Tomcat? Re: Flat panel VT replacement? Re: Flat panel VT replacement? Re: Flat panel VT replacement? Re: Flat panel VT replacement? Re: foto> Re: Host-Based Minimerge for OpenVMS Alpha V7.3-2 is available Re: HPworld - I Survived" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" RE: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device" Re: Itanium: EFI... no fs0: device  Re: misc questions about OpenVMS  Re: misc questions about OpenVMS  Re: misc questions about OpenVMS  Re: misc questions about OpenVMS  Re: misc questions about OpenVMS  Re: misc questions about OpenVMS
 Re: need help 
 Re: need help   RE: OpenVMS running on HP 3000's( Re: Planned port of freeware to Itanium?& SMTP: stray messages left in directory* Re: SMTP: stray messages left in directory& Re: Somewhat OT: Is AMD the new Intel?& Re: Somewhat OT: Is AMD the new Intel?# Re: strange decnet-plus problems...   F ----------------------------------------------------------------------  % Date: Wed, 01 Sep 2004 14:00:17 -0400 - From: Jf Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: A whopping 50 percent...  ???, Message-ID: <41360E29.23023D31@teksavvy.com>   Bob Koehler wrote:H >    It takes a certina amount of money to build a computer.  If you putJ >    a slow chip in it you save a small amount.  RISC designs were running) >    orders of magnitude faster than VAX.   M I know that RISC chips had much higher clock rates. But did they actually run ( software "orders of magnitudes" faster ?  K And while you have a fair point about the minimum cost to build a computer, M the fact remains that Digital did not lower prices of its older hardware, and A because of that, prices its performing hardware much much higher.   L There were vax chips that performed well. But they were priced out of range,E so when you compared machines with resonably similar prices, you were 6 comparing the older VAX ones with the newer RISC ones.  F DEC was happy to claim that with each new VAX, they were improving theM price-performance compared to the old VAXes. The problem is that they weren't A matching the trend in the industry and the gap widened with time.   N Remember that at that time, Digital had made the mkstake of hiring quite a fewN ex-IBMers and had set its sights on attacking IBM's high margin sales, and hadL totally dismissed the lowe margin stuff and this is exactly where Apollo andN Sun ate Digital. And this didn't even learn their lesson because they then notE only allowed but fistered Wintel boxes eating the remains of Digital.    ------------------------------  $ Date: Wed, 1 Sep 2004 11:08:56 -0700# From: "Tom Linden" <tom@kednos.com> * Subject: RE: A whopping 50 percent...  ???9 Message-ID: <NDEMLKKEBOIFBMJLCECIKENFDLAA.tom@kednos.com>    < -----Original Message-----C < From: Michael Kraemer [mailto:kraemer@biors6a.gsi.de]On Behalf Of  < m.kraemer@gsi.de. < Sent: Wednesday, September 01, 2004 10:46 AM < To: Info-VAX@Mvb.Saic.Com + < Subject: Re: A whopping 50 percent... ???  <  < 5 < In article <TG256h$K08Ue@eisner.encompasserve.org>, ? < koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: 9 < > In article <4135F96D.9CCB6CCF@teksavvy.com>, Jf Mezei ( < <jfmezei.spamnot@teksavvy.com> writes: < > > C < > > It wasn't because RISC machines were faster that DEC lost. It  < was because DEC 5 < > > didn't price its slower machines competitively.  < > J < >    It takes a certina amount of money to build a computer.  If you put? < >    a slow chip in it you save a small amount.  RISC designs  < were running+ < >    orders of magnitude faster than VAX.  < > L < >    As a comparison:  I had VAXstations running around 1 MIP at about theJ < >    same time I got MIPS based DECstations running around 13 MIP.  MostB < >    of the hardware was the same.  The cost was about the same. < > @ < >    You really think DEC could make VAX system at an order of < magnitude lower / < >    cost than the same box with a MIPS chip?  < >  < J < If the hardware was mostly the same, why should the price be different ?K < BTW, the "order of magnitude" is a bit exaggerated if you compare systems K < of the same time frame. At the time MIPSen became popular the competition E < was VAX 3176 and later 4000. Although the MIPSen were significantly C < faster and cheaper, it wasn't an order of magnitude between them. A < It was, however, enough for me (and others) to leave VMS (apart  < from other reasons).   And those were ... <  < --- ( < Incoming mail is certified Virus Free.< < Checked by AVG anti-virus system (http://www.grisoft.com).A < Version: 6.0.735 / Virus Database: 489 - Release Date: 8/6/2004  <  --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).? Version: 6.0.735 / Virus Database: 489 - Release Date: 8/6/2004    ------------------------------  % Date: Wed, 01 Sep 2004 21:07:56 +0200 0 From: Keith Cayemberg <keith.cayemberg@arcor.de>* Subject: Re: A whopping 50 percent...  ???B Message-ID: <41361e0c$0$19553$9b4e6d93@newsread2.arcor-online.net>   Jf Mezei wrote:    >>>  1   VMS             837 >>>  2   IBM VM          651I >>>Please also notice that nearly all of the entries are from academic or  >>>research institutions.  >  > N > Which is quite telling at how VMS used to have very big presence in academia > and no longer has any.  , I wouldn't say it doesn't have any presence.  B Over the past several years, I have collected a non-validated but I relatively up-to-date (only sources within 5 years) list of academic and  F research institutes which have OpenVMS systems. My list currently has H 216 academic institutions and 126 (many well-known) research institutes ? listed. The list continues to grow at the rate of about 3 to 9   institutions per month.   E And no, I'm not at liberty to specify who they are. But, you all are  C welcome to perform your own research. I think some people would be  B suprised where OpenVMS can be found and what it is being used for.   Cheers!    Keith Cayemberg    ------------------------------  $ Date: Wed, 1 Sep 2004 15:30:13 -0400# From: "John Smith" <a@nonymous.com> * Subject: Re: A whopping 50 percent...  ???, Message-ID: <KMWdnXqv_NPcvqvcRVn-hQ@igs.net>   m.kraemer@gsi.de wrote: 5 > In article <TG256h$K08Ue@eisner.encompasserve.org>, ? > koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: 8 >> In article <4135F96D.9CCB6CCF@teksavvy.com>, Jf Mezei) >> <jfmezei.spamnot@teksavvy.com> writes:  >>> E >>> It wasn't because RISC machines were faster that DEC lost. It was ? >>> because DEC didn't price its slower machines competitively.  >>E >>    It takes a certina amount of money to build a computer.  If you G >>    put a slow chip in it you save a small amount.  RISC designs were 2 >>    running orders of magnitude faster than VAX. >>G >>    As a comparison:  I had VAXstations running around 1 MIP at about G >>    the same time I got MIPS based DECstations running around 13 MIP. F >>    Most of the hardware was the same.  The cost was about the same. >>? >>    You really think DEC could make VAX system at an order of > >>    magnitude lower cost than the same box with a MIPS chip? >> > > > If the hardware was mostly the same, why should the price be
 > different ? C > BTW, the "order of magnitude" is a bit exaggerated if you compare 	 > systems ? > of the same time frame. At the time MIPSen became popular the 
 > competition E > was VAX 3176 and later 4000. Although the MIPSen were significantly C > faster and cheaper, it wasn't an order of magnitude between them. F > It was, however, enough for me (and others) to leave VMS (apart from > other reasons).     G It wasn't the price that was different -- it was the price/performance.   L Sometimes you just have to put stuff  'on sale' in order to retain customersI until the next generation of stuff is ready for the market. DEC never did K that. Had they done so, unix would have not gained ground so quickly in any  marketspace.   ------------------------------   Date: 1 Sep 2004 20:27:34 GMT ( From: bill@cs.uofs.edu (Bill Gunshannon)* Subject: Re: A whopping 50 percent...  ???* Message-ID: <2pmplmFm9b76U1@uni-berlin.de>  , In article <KMWdnXqv_NPcvqvcRVn-hQ@igs.net>,& 	"John Smith" <a@nonymous.com> writes: > m.kraemer@gsi.de wrote: 6 >> In article <TG256h$K08Ue@eisner.encompasserve.org>,@ >> koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:9 >>> In article <4135F96D.9CCB6CCF@teksavvy.com>, Jf Mezei * >>> <jfmezei.spamnot@teksavvy.com> writes: >>>>F >>>> It wasn't because RISC machines were faster that DEC lost. It was@ >>>> because DEC didn't price its slower machines competitively. >>> F >>>    It takes a certina amount of money to build a computer.  If youH >>>    put a slow chip in it you save a small amount.  RISC designs were3 >>>    running orders of magnitude faster than VAX.  >>> H >>>    As a comparison:  I had VAXstations running around 1 MIP at aboutH >>>    the same time I got MIPS based DECstations running around 13 MIP.G >>>    Most of the hardware was the same.  The cost was about the same.  >>> @ >>>    You really think DEC could make VAX system at an order of? >>>    magnitude lower cost than the same box with a MIPS chip?  >>>  >>? >> If the hardware was mostly the same, why should the price be  >> different ?D >> BTW, the "order of magnitude" is a bit exaggerated if you compare
 >> systems@ >> of the same time frame. At the time MIPSen became popular the >> competitionF >> was VAX 3176 and later 4000. Although the MIPSen were significantlyD >> faster and cheaper, it wasn't an order of magnitude between them.G >> It was, however, enough for me (and others) to leave VMS (apart from  >> other reasons). >  > I > It wasn't the price that was different -- it was the price/performance.  > N > Sometimes you just have to put stuff  'on sale' in order to retain customersK > until the next generation of stuff is ready for the market. DEC never did M > that. Had they done so, unix would have not gained ground so quickly in any  > marketspace.   C Oh pullezzze.....  One of the thngs that led to the rapid spread of C Unix was free software.  Contrary to popular belief the idea wasn't A invented by Linus Torvalds or Richard Stallman.  Unix people were ? giving away the stuff they wrote pretty much from the beginning C and as many have tried to say here over and over, it's the software A that drives the sale.  A box without applications is just a space  heater.   E Anecdotal evidence:  I remember working on a bid at a certain college A (that shall remain nameless) and loosing to DEC who had bid a VAX A against our Prime although the Prime had thouroughly trounced the F VAX at the specified benchmark.  The driving force behind the decision@ to take the VAX over the Prime was the fact that the head of the@ University's procurement team wanted a machine on which to run a? certain astronomy package that was available from the Kitt Peak ? Observatory. (He was actually quoted as saying, "I don't really B care which one wins the bid as long as it says VAX on the front.")> We had the last laugh, though.  The system bid by DEC that wonF was a VAX running VMS and the Kitt Peak package was for BSD Unix.  :-)    bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 1 Sep 04 16:09:35 EDT ) From: cook@wvnvms.wvnet.edu (George Cook) * Subject: Re: A whopping 50 percent...  ???! Message-ID: <4MIgNGq5q7HY@wvnvms>   R In article <dY2dnXibcoKsJ6jcRVn-pA@igs.net>, "John Smith" <a@nonymous.com> writes:. > "Alan Winston - SSRL Admin Cmptg Mgr" wrote:< >> In article <08mdndOgk6KvuqjcRVn-uA@igs.net>, "John Smith" >> <a@nonymous.com> writes:  >>> Mikko Putkonen wrote: A >>>> Bob Koehler <koehler@eisner.nospam.encompasserve.org> wrote: F >>>>>   I wish somebody did have the numbers, I'd expect the total VMSI >>>>>   systems outnumbered the total UNIX systems for a few years in the  >>>>>   early to mid 80's. >>>>E >>>> I'm absolutely clueless and inexperienced, as always, but could, E >>>> say, an old list of systems once connected to ... umh ... Bitnet H >>>> provide anything of interest?  Does this reflect anything about the >>>> bigger picture? >>>> >>>> --- >>>>: >>>>                        CONNECTED NODES AS OF 11/11/874 >>>>                              TOTAL NODES = 2063 >>
 >> [snippage]  >> >>>  >>> Thanks for the list. >>> F >>> Any bets as to just how many of these organizations still have any' >>> current VMS systems up and running?  >> >>D >> We were on that list (as SSRL750) and we've still got VMS systems >> running and  >> are doing active development. >>F >> And we learned to stop naming nodes after the hardware they run on;E >> it was embarrassing when SSRL750 became an 8700 and later an 8800.  >  > ' > That's one institution....any others?  >    : WVNMVS    West Virginia Network                     MVS/SP7 WVNVAXA   West Virginia Network                     VMS 7 WVNVAXB   West Virginia Network                     VMS 9 WVNVM     West Virginia Network                     VM/SP 7 WVNSVC    West Virginia Network                     VMS 7 WVNVMS    West Virginia Network                     VMS   = The IBM machines are long gone but we still have multiple VMS = machines (I'm sending this article from an upgraded WVNVAXA). = The following nodes added after 1987 still exist (having been > upgraded several times) at thirteen colleges and universities:  B WVNBSC, WVNCC, WVNFSC, WVNGSC, WVNNCAXP, WVNWVUPA, WVNPSC, WVNSC1,1 WVNSC3, WVNSCC, WVNWLSC, WVNWVSC, WVNWVIT, MUINFO   < Note that it has been a never ending battle to keep VMS withA DEC/Compag/HP at times working against us or totally ignoring us. > When/if VMS goes we very likely will never do business with HP5 again unless they buy (or are bought by) IBM or Dell.      George Cook  WVNET    ------------------------------  $ Date: Wed, 1 Sep 2004 22:20:35 +0200  From: "Dr. Dweeb" <dr@dweeb.com>* Subject: Re: A whopping 50 percent...  ???- Message-ID: <ch5auk$1g38$1@news.cybercity.dk>    Andrew Harrison wrote: > David Froble wrote:  >  >> Andrew Harrison wrote:  >> >>> David Froble wrote:  >>>  >>>>> Kenneth Farmer wrote:  >>>> >>>> >>> C >>>> How good or bad the itanic was, is, and becomes isn't the real  >>>> issue, at least for me. >>>>H >>>> Starting with DEC, and the painful path to HP, I've always had thisE >>>> issue.  With VAX and Alpha under the control of the same company D >>>> providing VMS there has never been an issue with VMS losing theD >>>> hardware on which it ran.  Once the hardware was abandoned, the4 >>>> fate of VMS depends upon whoever makes the CPU. >>>>C >>>> Alpha was the best CPU.  I doubt many will question this.  Not : >>>> speaking for him, (and who could since he never quitsF >>>> talking/posting), but even Andrew might admit that the CPU itselfD >>>> was great.  We won't get into packaging of the CPU. Regardless,9 >>>> Alpha is no more, as far as the future is concerned.  >>>> >>> F >>> You are right, Alpha was/is a good CPU let down by packaging which8 >>> except for a few honourable exceptions sucked rocks. >> >> >>F >> Note that the EV7 developers addressed this issue, by placing stuff@ >> on the chip and bypassing the system developers.  I've alwaysG >> wondered if this caused them to turn on Alpha.  Then again, it could @ >> have just been trying to keep their jobs.  The itanic in it'sF >> current form would seem to cause a need for good system developers.D >> I also think I read a while back that those system developers whoE >> turned against Alpha were given their appropriate reward for being  >> good company men, the AX. >> > E > Yes the did, the GS1280 is a very well balances system. The problem B > is that it is the first large Alpha system that can be describedC > as well balanced and it appeared as Alpha's demise was announced.  > H > Before that Digital and Compaq had 2 or 3 attempts at getting it right= > starting with the 8400 and ending miserably with the GS320.  > B > Ironically HP have delivered a system which is too much too lateC > in the sense that the GS1280 is actually a rather better balanced E > system than the SuperDome while also being a system that has little  > or no future.  >   I They also did lots of tweaks to VMS which resulted in massive performance 4 improvements at customer sites - even for the GS320.  B But sadly you are right - too much too late for HP and VMS I fear.  	 Dr. Dweeb   	 > Regards  > Andrew Harrison E >> I've not tried to keep up with the latest Alpha systems.  But from G >> some of the things you've said, it appears that the on-chip SMP glue D >> and on-chip memory controllers may have shown their worth.  Don'tD >> know what else would keep a 3-year old processor running with theB >> rest of the pack, and beating the itanic.  Well, the last isn't >> quite as astounding.  :-) >> >>E >>> Itanium seems to be an indifferent CPU unless you count being 4th G >>> fastest on Integer behind Opteron, Xeon, Power4 and 2and fastest on B >>> FP for single CPU workloads as something other than mid table. >>> D >>> Worse this mid table possition slides even further if you have aC >>> whiff of SMP involved and Itanium slides back to 5th on Integer  >>> and 4th on FP. >>> 9 >>> Put an app on there and the position worsens further.  >>> H >>> Or put another way the Itanium package like the Alpha package beforeD >>> it sucks because this is exactly what happened with Alpha boxes. >>> 9 >>> Frontside is not a bus that is designed to assist you 9 >>> building large SMP systems (SGI will vouch for that).  >>> B >>> Couple that with a large system infrastructure where your best? >>> server has a bisectional bandwidth which is 1/2 that of the ? >>> slowest competitors, a tiny application portfolio, confused D >>> marketing messages, a pissed off customer base for your existing2 >>> systems and you have Alphacide all over again. >>>  >>> Regards  >>> Andrew Harrison  >>>  >>>  >>> H >>>> If the itanic never becomes profitable for Intel, what have they toG >>>> lose by dropping it?  Not much.  The big losers will be you and me G >>>> and the rest of those who continue to believe that VMS is the best # >>>> commercial OS available today.  >>>>H >>>> That's my issues with the itanic.  In my opinion the EPIC based CPUG >>>> will never touch IBM's OOO Power architecture.  Thus, except for a H >>>> few things that it may do well, it will not succeed at the top end.B >>>> Opteron will kill it below the top end, and maybe even there. >>>>E >>>> It may be adequate for most VMS users.  I hope it is.  I hope it D >>>> stays around. But if it doesn't, the actions of some idiots and7 >>>> Intel will have hurt VMS very much, maybe fatally.  >>>>	 >>>> Dave    ------------------------------  % Date: Wed, 01 Sep 2004 17:16:35 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> * Subject: Re: A whopping 50 percent...  ???, Message-ID: <41363C1F.63F43577@teksavvy.com>   Bill Gunshannon wrote:E > Oh pullezzze.....  One of the thngs that led to the rapid spread of E > Unix was free software.  Contrary to popular belief the idea wasn't 2 > invented by Linus Torvalds or Richard Stallman.   M Humm, how quickly people forget that the DECUS VMS libraries were some of the L largest repositories of free software available in the 1980s. It was a greatL asset to Digital sales when the time came to show total cost of ownership of/ VMS was low due in part to the DECUS libraries.    ------------------------------   Date: 1 Sep 2004 11:18:33 -0700 & From: jordan@ccs4vms.com (Rich Jordan)4 Subject: Alphastation 200 storage controller options< Message-ID: <cc5619f2.0409011018.83def34@posting.google.com>  @ I want to hang a couple of big disks on an AS200 I use for localC webserving; they will be IDEs with Acard SCSI adapters (which I can E get from storage in narrow or wide, or ultra wide type).  I'd like to B get a little better performance from them so instead of connectingC them to the embedded FNSE SCSI bus, I'd like to put a more advanced  controller in it.   B The system will be running OpenVMS V7.3-2, and is running the most recent firmware.  > The only SOC I have that covers the AS200 only lists the KZPAAF controller (FNSE), but its a 1995 edition; I don't have any later SOCs@ until the AS200 was discontinued.  The online SOC archives don't= provide much info, so I don't know if any other card was ever 
 supported.  D I've got a KZPSA (FWD) and a few KZPDA (FWSE) units.  Installing one@ of either, in either PCI slot, with or without the graphics cardF (ZLXp-E2) installed causes the machine to hang on power up.  The KZPDA7 cards do work in an AS600-5/333, so I know they are OK.   C Has anyone had any successes using alternate SCSI controllers in an F AS200?  I could run another KZPAA (at least the big drives would be onC a second channel) but I'd prefer something faster, if its possible.    Thanks   Rich CCS    ------------------------------   Date: 1 Sep 2004 20:51:53 GMT  From: healyzh@aracnet.com 8 Subject: Re: Alphastation 200 storage controller options, Message-ID: <ch5cp902jo0@enews3.newsguy.com>  ' Rich Jordan <jordan@ccs4vms.com> wrote: E > Has anyone had any successes using alternate SCSI controllers in an H > AS200?  I could run another KZPAA (at least the big drives would be onE > a second channel) but I'd prefer something faster, if its possible.   K I ran a KZPBA in one of my AlphaStation 200 4/233's for a week or two until / I figured out what was wrong with my PWS 433au.    	Zane    ------------------------------  # Date: Wed, 01 Sep 2004 22:04:03 GMT 1 From: Keith Parris <keithparris_NOSPAM@yahoo.com> 8 Subject: Re: Alphastation 200 storage controller options2 Message-ID: <nHrZc.9395$RB4.1119@news.cpqcorp.net>  I The Alphastation 200 is similar in most respects to the Alphaserver 300,  F and I've had good luck so far in using the Supported Options List for  the latter. See F http://h18002.www1.hp.com/alphaserver/options/as300/as300_options.html   ------------------------------  % Date: Wed, 01 Sep 2004 18:53:29 +0100 < From: Andrew Harrison <andrew_remove__s._harrison@s_u_n.com>1 Subject: Re: Big IA64 test coming in a few months 0 Message-ID: <ch52ar$c8m$1@new-usenet.uk.sun.com>   Dirk Munk wrote: > Jf Mezei wrote:  >  >> Dirk Munk wrote:  >>G >>> It seems Instel is to late already. According to The Register, AMD   >>> demonstrated" >>> a dual core Opteron yesterday. >> >> >>J >> Demosntrated on an HP server to boot ! But making a demo of a beta chipK >> doesn't equate being able to ship in quantities. That demo was the moral H >> equivalent of the VMS engineers making first boot on that IA64 thing. >  > I > You're right, but it does mean development of the dual core Opteron is  J > well on its way, and it seems AMD is ahead of Intel. I can't wait for a G > dual AMD64 in my PC. Just imagine has fast I will be able to type my   > e-mails !!A According to the AMD roadmap the first dual core AMD64 processors @ will be Opterons not Athlon-64's though you could have a desktop using Opterons Sun does.  A The dual core Opterons will plug into the current 940 socket used  by the existing processor.  > The first dual core Intel processors look to be the Smithfield< processor which will have a new FrontSide Bus to support it.   Regards  Andrew Harrison    ------------------------------   Date: 1 Sep 2004 16:13:21 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) # Subject: Re: charon vax emulator??? 3 Message-ID: <vN7w2Xl7KS2u@eisner.encompasserve.org>   d In article <41338b73$0$13924$ba620dc5@nova.planet.nl>, Wilm Boerhout <w3.boerhout@planet.nl> writes: > Bob Koehler wrote: > L >>>Sneakernet may be disabled as well by Charon-VAX. Just assign the floppy G >>>and the CDROM drive to their VMS counterpart within Charon-VAX, and  L >>>$ALLOCATE the VMS drive for those devices. Then Windows won't be able to  >>>read from them. >>   >>  & >>    I hope you don't have a USB bus. > I > Now here's a real concern. Yes, we have a USB bus. We have to, because  3 > Charon-VAX is licenced through a Aladdin USB key.   / How does that work on Alphas that have no USB ?    ------------------------------   Date: 1 Sep 2004 16:10:26 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) # Subject: Re: charon vax emulator??? 3 Message-ID: <8aMERxhsoJgw@eisner.encompasserve.org>   U In article <41339A7C.8010903@MMaz.com>, "Barry Treahy, Jr." <Treahy@MMaz.com> writes:  > Bob Koehler wrote: > e >>In article <41338d45$0$2416$ba620dc5@nova.planet.nl>, Wilm Boerhout <w3.boerhout@planet.nl> writes:  >>   >>E >>>To -maybe- conclude this discussion: VMS is an insecure operating  J >>>system, because anyone can walk up to the system console, type CTRL/P, % >>>and cause major denial of service.  >>>      >>>  >>A >>   I think we already agreed on the need for physical security.  >> >>   >>G > Amen, common ground which does eliminate all prior concerns directly  H > related to physical hacking of the box (ie. floppy, CD, DVD, USB, FW, H > plugging in alternate boot drives, or, as mentioned above, CTRL/P and ' > booting with an alternate SYSTARTUP)!   A There is a further requirement for such a room full of physically @ secured computers -- LABEL the various consoles for those who do have access.   ------------------------------  % Date: Wed, 01 Sep 2004 17:54:07 -0400 2 From: "Stanley F. Quayle" <squayle@insight.rr.com># Subject: Re: charon vax emulator??? - Message-ID: <41360CBF.3012.A5F3005@localhost>   . On 1 Sep 2004 at 16:13, Larry Kilgallen wrote:= >> Now here's a real concern. Yes, we have a USB bus. We have @ >> to, because Charon-VAX is licenced through a Aladdin USB key. > 1 > How does that work on Alphas that have no USB ?   A The Alpha version is running VMS.  So, a PAK license is supplied  ! instead of a USB hardware dongle.     
 --Stan Quayle  Quayle Consulting Inc.  
 ----------- Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363 3 8572 North Spring Ct., Pickerington, OH  43147  USA 0 stan-at-stanq-dot-com       http://www.stanq.com   ------------------------------  % Date: Wed, 01 Sep 2004 20:32:16 -0400 * From: Chuck Chopp <ChuckChopp@rtfmcsi.com>E Subject: Re: CSWS_JAVA v2.1 - Tomcat v4.1.24 - update/upgrade Tomcat? : Message-ID: <tRtZc.51240$N11.33791@bignews5.bellsouth.net>   Now I'm really puzzled.   L I took my struts-based webapp and modified it.  I moved all of my JSP files G from the "WEB-INF/form" subfolder and placed them in the webapp's root  I folder.  After rebuilding the webapp's WAR file and deploying it, it now  L executes properly on the Tomcat implementation on OpenVMS.  It appears that K Tomcat on OpenVMS is having problems locating auto-generated ".java" files  K that are created for JSP files in a webapp when those JSP files are not in  H the root folder of the webapp.  Other Tomcat implementations on WinXP & J NetWare v6.5 don't exhibit this problem, and all of the Struts examples & J tutorials that I've seen or read about demonstrate placing JSP files in a L subfolder under the "WEB-INF" subfolder as a means of implementing security M in a webapp.  Placing the JSP files under "WEB-INF" makes it so that the JSP  J files cannot be directly accessed by a web browser; instead, they must be / sent to the browser by an action in the webapp.      --   Chuck Chopp   8 ChuckChopp (at) rtfmcsi (dot) com http://www.rtfmcsi.com  @ RTFM Consulting Services Inc.     864 801 2795 voice & voicemail2 103 Autumn Hill Road              864 801 2774 fax Greer, SC  29651  , Do not send me unsolicited commercial email.   ------------------------------  % Date: Wed, 01 Sep 2004 19:57:26 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>E Subject: Re: CSWS_JAVA v2.1 - Tomcat v4.1.24 - update/upgrade Tomcat? + Message-ID: <41366FF6.39827398@comcast.net>    Chuck Chopp wrote: >  > Now I'm really puzzled.  > M > I took my struts-based webapp and modified it.  I moved all of my JSP files H > from the "WEB-INF/form" subfolder and placed them in the webapp's rootJ > folder.  After rebuilding the webapp's WAR file and deploying it, it nowM > executes properly on the Tomcat implementation on OpenVMS.  It appears that L > Tomcat on OpenVMS is having problems locating auto-generated ".java" filesL > that are created for JSP files in a webapp when those JSP files are not inI > the root folder of the webapp.  Other Tomcat implementations on WinXP & K > NetWare v6.5 don't exhibit this problem, and all of the Struts examples & K > tutorials that I've seen or read about demonstrate placing JSP files in a M > subfolder under the "WEB-INF" subfolder as a means of implementing security N > in a webapp.  Placing the JSP files under "WEB-INF" makes it so that the JSPK > files cannot be directly accessed by a web browser; instead, they must be 1 > sent to the browser by an action in the webapp.   F Is there perhaps a logical name or environment variable that is needed to be defined?   D.J.D.   ------------------------------  % Date: Wed, 01 Sep 2004 21:24:09 -0400 * From: Chuck Chopp <ChuckChopp@rtfmcsi.com>E Subject: Re: CSWS_JAVA v2.1 - Tomcat v4.1.24 - update/upgrade Tomcat? : Message-ID: <6CuZc.51251$N11.41155@bignews5.bellsouth.net>   David J Dachtera wrote:   H > Is there perhaps a logical name or environment variable that is needed > to be defined?  G I'll be damned if I know... but I doubt it.  I'm going to do some more  M testing with moving JSP files around into different subfolders in the webapp  K to see which configurations cause the webapps to fail on OpenVMS but still  % function properly on WinXP & NetWare.      --   Chuck Chopp   8 ChuckChopp (at) rtfmcsi (dot) com http://www.rtfmcsi.com  @ RTFM Consulting Services Inc.     864 801 2795 voice & voicemail2 103 Autumn Hill Road              864 801 2774 fax Greer, SC  29651  , Do not send me unsolicited commercial email.   ------------------------------   Date: 1 Sep 2004 14:02:04 -0700 & From: jordan@ccs4vms.com (Rich Jordan)' Subject: Re: Flat panel VT replacement? = Message-ID: <cc5619f2.0409011302.45044a9c@posting.google.com>   y Rob Brown <mylastname@gmcl.com> wrote in message news:<Pine.LNX.4.44.0409010851240.26432-100000@localhost.localdomain>...  > On 1 Sep 2004, Galen wrote:  > J > > Does anyone know of a flat panel device that could be used in place of > > a  VT2xx or later terminal?  > " > Is the VT520(?) still available? > F > The one I am thinking of had a DEC Keyboard and a box, to which you + > attached the SVGA monitor of your choice.   ? VT525.  The 'box' is the same size as the Multia/UDB, and has a C standard HD15 VGA connector, so perhaps a low end LCD monitor would B work with it.  I think you can still get all VT5xx terminal models# from Boundless (www.boundless.com).   F There used to be a plasma-screen VT320 variant (ELS320???) by Planar.  I have no information on these.   C DEC also used to sell a plasma screen (VRE01-xx) that could be used B with VT1200 and monochrome VAXstations.  A VT1200 is a small pizzaB box, self booting early X-terminal with the ability to connect via> serial, LAT, or TELNET (if I recall correctly).  It used LK201@ keyboards and old-style hockey puck mice.  Display quality of myD VT1200 was adequate on a VR150 monochrome monitor; I've never seen a@ VRE01 live so no idea how they look, but it (the VT1200, not theB VRE01!) was a cheap and easy way to have multiple DECterms open to various systems.   Rich Jordan  CCS    ------------------------------   Date: 1 Sep 2004 20:59:12 GMT  From: healyzh@aracnet.com ' Subject: Re: Flat panel VT replacement? , Message-ID: <ch5d7012jo0@enews3.newsguy.com>  1 Keith Cayemberg <keith.cayemberg@arcor.de> wrote:  > Try the following link...   " > MIDCOM - Planar ELT320 Terminals/ > http://www.midcomdata.com/terminal/planar.htm   L ISTR, that these are not only no longer made, but quite expensive.  What I'dK like to do is get a VT525, an LK451 (I think that's the right keyboard) and  a cheap 15" LCD.  I Compaq made some "Thin Clients" that are supposed to be a VT replacement, L you could hook a LCD up to these as well, but I don't know how good their VTA emulation is.  I think they're called the Compaq EVO T10 and T20.    		Zane   ------------------------------  # Date: Wed, 01 Sep 2004 23:17:45 GMT " From:   VAXman-  @SendSpamHere.ORG' Subject: Re: Flat panel VT replacement? 0 Message-ID: <00A373E0.D80DB058@SendSpamHere.ORG>  f In article <cc5619f2.0409011302.45044a9c@posting.google.com>, jordan@ccs4vms.com (Rich Jordan) writes:z >Rob Brown <mylastname@gmcl.com> wrote in message news:<Pine.LNX.4.44.0409010851240.26432-100000@localhost.localdomain>... >> On 1 Sep 2004, Galen wrote: >>  K >> > Does anyone know of a flat panel device that could be used in place of   >> > a  VT2xx or later terminal? >>  # >> Is the VT520(?) still available?  >>  G >> The one I am thinking of had a DEC Keyboard and a box, to which you  , >> attached the SVGA monitor of your choice. > @ >VT525.  The 'box' is the same size as the Multia/UDB, and has aD >standard HD15 VGA connector, so perhaps a low end LCD monitor wouldC >work with it.  I think you can still get all VT5xx terminal models $ >from Boundless (www.boundless.com). > G >There used to be a plasma-screen VT320 variant (ELS320???) by Planar.    >I have no information on these. > D >DEC also used to sell a plasma screen (VRE01-xx) that could be usedC >with VT1200 and monochrome VAXstations.  A VT1200 is a small pizza C >box, self booting early X-terminal with the ability to connect via ? >serial, LAT, or TELNET (if I recall correctly).  It used LK201 A >keyboards and old-style hockey puck mice.  Display quality of my E >VT1200 was adequate on a VR150 monochrome monitor; I've never seen a A >VRE01 live so no idea how they look, but it (the VT1200, not the C >VRE01!) was a cheap and easy way to have multiple DECterms open to  >various systems.  >  >Rich Jordan >CCS    F New iMac G5!  Run eXodus on it and you have the ultimate Xterminal! ;)   --  < http://www.ProvN.com  for the *best* OpenVMS system security=                       solutions that others only claim to be.  --  , Cyber-Terrorism (si'-ber tayr'-or-iz-em) n.:M   The release of, the sale of, or the use of any Micro$oft software product!   --  K VAXman- A Bored Certified VMS Kernel Mode Hacker   VAXman(at)TMESIS(dot)COM    ------------------------------   Date: 1 Sep 2004 23:40:32 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) ' Subject: Re: Flat panel VT replacement? 3 Message-ID: <HD5obc4+QqOM@eisner.encompasserve.org>   U In article <00A373E0.D80DB058@SendSpamHere.ORG>,   VAXman-  @SendSpamHere.ORG writes: h > In article <cc5619f2.0409011302.45044a9c@posting.google.com>, jordan@ccs4vms.com (Rich Jordan) writes:{ >>Rob Brown <mylastname@gmcl.com> wrote in message news:<Pine.LNX.4.44.0409010851240.26432-100000@localhost.localdomain>...  >>> On 1 Sep 2004, Galen wrote:  >>> L >>> > Does anyone know of a flat panel device that could be used in place of! >>> > a  VT2xx or later terminal?   H > New iMac G5!  Run eXodus on it and you have the ultimate Xterminal! ;)  4 But what would you add for booting the workstation ?   ------------------------------  % Date: Wed, 01 Sep 2004 21:08:10 -0400 - From: "John E. Malmberg" <wb8tyw@qsl.network>  Subject: Re: foto 1 Message-ID: <DcWdna_TSrri76vcRVn-rw@adelphia.com>    Mark Berryman wrote: > 4 > I can go one better.  The message originated from:1 > 24.107.134.203 (24.107.134.203.charter-stl.com)  > 2 > which appears to be an address in St. Louis, MO. > K > I'll try to add a new anti-spam rule that blocks any message coming into  > > me with a message-id that claims to have originated with me.  > http://www.moensted.dk/spam/?addr=24.107.134.203&Submit=Submit   Dynamic address:  $      http://njabl.org/dynablock.html  ;      http://www.dnsbl.sorbs.net/lookup.shtml?24.107.134.203 . 	24.107.128.0/19 (24.107.128.0-24.107.159.255)  <      http://www.mail-abuse.com/cgi-bin/lookup?24.107.134.203  9      Not currently listed in the http://www.pan-am.ca/pdl    -John  wb8tyw@qsl.network Personal Opinion Only    ------------------------------  # Date: Wed, 01 Sep 2004 20:23:06 GMT / From: brooks@cuebid.zko.dec.nospam (Rob Brooks) G Subject: Re: Host-Based Minimerge for OpenVMS Alpha V7.3-2 is available - Message-ID: <k36EJblnOHPN@cuebid.zko.dec.com>   5 Michael Unger <spam.to.unger@spamgourmet.com> writes:  > Rob Brooks wrote:   D >> Host-Based Minimerge for OpenVMS Alpha V7.3-2 is now available byE >> the standard patch distribution mechanism (http://www.itrc.hp.com)   & >> The kit is named VMS732_HBMM-V0100.   > | 6.1.1.2  Documentation  ; > | VS-HBMM-VMS732-V0100.PDF, VS-HBMM-VMS732-V0100.TXT, and 9 > | HBMM_PRO*.HTML are copied to SYS$HELP when the kit is  > | installed. > H > Can these manuals be made available on the "OpenVMS Documentation" web: > pages? (Which would avoid a download of about 12 MB ...)  I The documentation for HBMM will be put on the VMS documentation web page.   M I'm not sure when it'll happen (it may be as soon as Thursday, 9/2/2004), but  it is going to happen.  M Please note that it'll be rather important to actually read the documentation > before using the host-based minimerge feature of the HBMM kit.  I The DCL syntax for the placement of bitmaps can look a bit odd; we needed I a flexible mechanism, and to get that flexibility, the command syntax can C appear to be imposing.  The online HELP for SET and SHOW SHADOW has 0 a fair amount of examples that should, uh, help.  = HBMM uses the bitmap technology that appeared with mini copy.   ' There will not be any HBMM for the VAX.    --    M Rob Brooks    VMS Engineering -- I/O Exec Group     brooks!cuebid.zko.dec.com    ------------------------------   Date: 1 Sep 2004 16:26:36 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) ! Subject: Re: HPworld - I Survived 3 Message-ID: <M6hK$rWDtZ3Z@eisner.encompasserve.org>   U In article <2pgvg8Fklj2aU3@uni-berlin.de>, bill@cs.uofs.edu (Bill Gunshannon) writes:   C > Most intereting.  The only train I have ever ridden in the US was G > between Boston and NYC.  It was diesel.  Had seats like an old school F > bus (not like all those adds for Amtrak I see on TV). And smelled ofH > exhaust like an Greyhound bus all the way.  Not likely to make me wantE > travel that way again. Granted, that was quite a while ago, but the / > trains I see in my travels all look the same.   F What _I_ see in the media are trains in The Harvey Girls and The Music= Man, where the cowcatcher on the front was really functional.   F But I have also ridden Acela and it is quite different from those.  MyG major complaint was a woman who thought it was ok to use her cell phone  in the quiet car.    ------------------------------  * Date: Wed, 1 Sep 2004 17:57:58 +0000 (UTC)7 From: moroney@world.std.spaamtrap.com (Michael Moroney) + Subject: Re: Itanium: EFI... no fs0: device ( Message-ID: <ch52j6$sat$1@pcls4.std.com>  / Jf Mezei <jfmezei.spamnot@teksavvy.com> writes:   M >And if the drive will truly be partitioned into different logical drives, it N >is still "kosher" for VMS to have some file whose blocks physically reside in >another partition ?  I I just hope that what happened when you ran both NT and VMS (or Unix) on  J older Alphas doesn't happen on Windows/VMS Itaniums.  On those Alphas, NT K would see the VMS disk and would ask to write a "harmless signature" to it. G Unfortunately, this "signature" overwrote the bootblock, making the VMS  drive unbootable.  --   -Mike    ------------------------------  % Date: Wed, 01 Sep 2004 14:04:14 -0400 - From: Jf Mezei <jfmezei.spamnot@teksavvy.com> + Subject: Re: Itanium: EFI... no fs0: device , Message-ID: <41360F16.DC885F2C@teksavvy.com>   John Reagan wrote:H > If you have this level of access, why bother mucking around inside theG > FAT partition?  Just provide any new execlet that reformats all disks  > when loaded.   Well, someone was asking...   L One reason for this is that the FAT partition contains code and formats thatI are totally foreign to a VMS system manager, so it would be far easier to K insert malicious stuff inside that without the system manager being able to I trace back what actually happened and who changed what (since the changes ) would occur outside the VMS file system).    ------------------------------  % Date: Wed, 01 Sep 2004 20:15:05 +0200 3 From: Michael Unger <spam.to.unger@spamgourmet.com> + Subject: Re: Itanium: EFI... no fs0: device * Message-ID: <2pmi7hFmhr77U1@uni-berlin.de>  - On 2004-09-01 19:57, "Michael Moroney" wrote:    > [...]  > K > I just hope that what happened when you ran both NT and VMS (or Unix) on  L > older Alphas doesn't happen on Windows/VMS Itaniums.  On those Alphas, NT M > would see the VMS disk and would ask to write a "harmless signature" to it. E                                                    ^^^^^^^^^^^^^^^^^^ I > Unfortunately, this "signature" overwrote the bootblock, making the VMS  > drive unbootable.   F Isn't that a "feature" of Winwoes totally independent of the CPU it is running on?    Michael    --  ; Real names enhance the probability of getting real answers. 5 My e-mail account at DECUS Munich is no longer valid.    ------------------------------  # Date: Wed, 01 Sep 2004 18:53:29 GMT 4 From: "Fred Kleinsorge" <fred.nospam@nospam.dec.com>+ Subject: Re: Itanium: EFI... no fs0: device 2 Message-ID: <JUoZc.9334$cd4.7430@news.cpqcorp.net>  : "Jf Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message& news:413603C0.D4BBBF5B@teksavvy.com... > Fred Kleinsorge wrote:I > > might need it.  The real information on where the console file system  lives J > > on the disk is in the GUID partition table - which occupies the blocks! > > following the traditional MBR  >  > J > If/when Windows becomes available for IA64, and you have a system with 2L > physical drives, one loaded with Windows, will Windows see the "VMS" drive asE > 2 distinct drives (one with the EFI FAT file system, and one with a  strange, > unknown file system on it ?) > K > Or is the GUID a truly specialized data structure in some formerly unused  diskE > location that will be seen only be the IA64 primitive rom program ?  >   I Windows *is* available on IA64 (if in Beta form) - it uses EFI to boot as  well. L Windows (as will Linux and HPUX) will see the partition record - it will see the J GUID table when present - and will see the EFI partition (the GUID for the typeJ says its EFI) and it will see the VMS partitions as unknown types.  I have no idea C if Windows would allow mapping the EFI partition as a drive or not.   J > Currently, when a VAX or Alpha initialise a drive, the drive is 100% VMS withJ > all its own proprietary structures, and VMS has full ownership of all of
 the drive. > K > But with IA64, will an Alpha or VAX be able to use its INITIALIZE command  toE > create a VMS drive destined for IA64, or will there be some special 
 utility toF > first partition a drive with the microsoft formats, and then use the normalH > INIT program to populate the big partition with the ODS-x structures ? >   E INITIALIZE and SETBOOT handles all of this.  See HELP INITIALIZE/GPT. K The default on IPF is /GPT, and /NOGPT on Alpha.  Backup will recognize the L presence of the system files (GPT.SYS) and will also do the right thing when( backing up and restoring system volumes.    K > And if the drive will truly be partitioned into different logical drives,  itL > is still "kosher" for VMS to have some file whose blocks physically reside in > another partition ?   K We do not support paritioned disks *in VMS* we've creatively found a way to K co-exist.  We do *not* support multiple OS partitions on the same disk.  So  you L won't see WINDOWS.SYS or LINUX.SYS as container files for other file systems on the disk.   ------------------------------  $ Date: Wed, 1 Sep 2004 12:05:09 -0700# From: "Tom Linden" <tom@kednos.com> + Subject: RE: Itanium: EFI... no fs0: device 9 Message-ID: <NDEMLKKEBOIFBMJLCECIAENIDLAA.tom@kednos.com>    < -----Original Message-----; < From: Fred Kleinsorge [mailto:fred.nospam@nospam.dec.com] . < Sent: Wednesday, September 01, 2004 11:53 AM < To: Info-VAX@Mvb.Saic.Com - < Subject: Re: Itanium: EFI... no fs0: device  <  <  < < < "Jf Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message( < news:413603C0.D4BBBF5B@teksavvy.com... < > Fred Kleinsorge wrote:K < > > might need it.  The real information on where the console file system  < lives L < > > on the disk is in the GUID partition table - which occupies the blocks# < > > following the traditional MBR  < >  < > L < > If/when Windows becomes available for IA64, and you have a system with 2B < > physical drives, one loaded with Windows, will Windows see the
 < "VMS" drive  < asG < > 2 distinct drives (one with the EFI FAT file system, and one with a 
 < strange,  < > unknown file system on it ?) < > = < > Or is the GUID a truly specialized data structure in some  < formerly unused  < diskG < > location that will be seen only be the IA64 primitive rom program ?  < >  < K < Windows *is* available on IA64 (if in Beta form) - it uses EFI to boot as  < well. B < Windows (as will Linux and HPUX) will see the partition record -
 < it will see  < the L < GUID table when present - and will see the EFI partition (the GUID for the < typeL < says its EFI) and it will see the VMS partitions as unknown types.  I have	 < no idea E < if Windows would allow mapping the EFI partition as a drive or not.   / FWIW, I think AdvFS under Tru64 would permit it  < L < > Currently, when a VAX or Alpha initialise a drive, the drive is 100% VMS < withL < > all its own proprietary structures, and VMS has full ownership of all of < the drive. < > : < > But with IA64, will an Alpha or VAX be able to use its < INITIALIZE command < toG < > create a VMS drive destined for IA64, or will there be some special  < utility toH < > first partition a drive with the microsoft formats, and then use the < normalJ < > INIT program to populate the big partition with the ODS-x structures ? < >  < G < INITIALIZE and SETBOOT handles all of this.  See HELP INITIALIZE/GPT. ? < The default on IPF is /GPT, and /NOGPT on Alpha.  Backup will  < recognize the C < presence of the system files (GPT.SYS) and will also do the right  < thing when* < backing up and restoring system volumes. <  < = < > And if the drive will truly be partitioned into different  < logical drives,  < it< < > is still "kosher" for VMS to have some file whose blocks < physically reside  < in < > another partition ?  < > < We do not support paritioned disks *in VMS* we've creatively < found a way toC < co-exist.  We do *not* support multiple OS partitions on the same  < disk.  So  < you A < won't see WINDOWS.SYS or LINUX.SYS as container files for other  < file systems < on the disk. <  <  <  < --- ( < Incoming mail is certified Virus Free.< < Checked by AVG anti-virus system (http://www.grisoft.com).A < Version: 6.0.735 / Virus Database: 489 - Release Date: 8/6/2004  <  --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).? Version: 6.0.735 / Virus Database: 489 - Release Date: 8/6/2004    ------------------------------  % Date: Wed, 01 Sep 2004 15:41:00 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> + Subject: Re: Itanium: EFI... no fs0: device + Message-ID: <413625BE.306D697@teksavvy.com>    Fred Kleinsorge wrote:N > Windows (as will Linux and HPUX) will see the partition record - it will see > the L > GUID table when present - and will see the EFI partition (the GUID for the > typeE > says its EFI) and it will see the VMS partitions as unknown types.    K So, since Windows is able to access FAT drives, then theoretically, Windows N will be able to play around with the contents of the EFI "drive" that would be4 physically stored on a disk that also contains VMS.   M And since the GUID is a microsoft structure, wouldn't Windows then be able to M resize the partitions and give the EFI a lot more room which would then quash  the VMS partition ?   N I have no problems with the EFI stored as a container file inside a VMS drive.M But I do have a problem with a VMS drive having windows structures (GUID) ate 9 the very base/core of the disk drive under the VMS stuff.   M I take it data drives will not require to have that microsoft partitioning of 	 a drive ?    > I have	 > no idea E > if Windows would allow mapping the EFI partition as a drive or not.   L Worse case scenario MUST be expected from anything coming out of microsoft.   G > INITIALIZE and SETBOOT handles all of this.  See HELP INITIALIZE/GPT.   G I tried HELP INITIALIZE/GPT on my all mighty Microvax II, but it didn't , recognize that qualifier :-) :-) :-) :-) :-)  M > The default on IPF is /GPT, and /NOGPT on Alpha.  Backup will recognize the N > presence of the system files (GPT.SYS) and will also do the right thing when* > backing up and restoring system volumes.  M > We do not support paritioned disks *in VMS* we've creatively found a way to J > co-exist.  We do *not* support multiple OS partitions on the same disk.   K Ok, this is where I am perplexed. If the drive is partitioned, wouldn't VMS L just see its own partition on that logical drive ? Wouldn't it see the otherN partitions as different drives on the system ? If so, how come you wouldn't beC able to have multiple logical  drives on the same physical drive ?    J Or have you guys found a way to cheat and and just mount the real physicalP drive, bypassing the partitioning stuff the IA64 hardware is presenting to you ?  D If you bypass the partitioning stuff, doesn't that conflict with theF positioning of ODS structures with the positioning of the partitioning* structures at the beginning of the drive ?  M Or is VMS smart enough to sequentially read blocks from block 0 onwards until M it finds the ODS structure ? (thus allowing it to transparently skip over the  microsoft partitioning garbage)    ------------------------------  # Date: Wed, 01 Sep 2004 20:15:12 GMT 4 From: "Fred Kleinsorge" <fred.nospam@nospam.dec.com>+ Subject: Re: Itanium: EFI... no fs0: device 2 Message-ID: <k5qZc.9358$fi4.4340@news.cpqcorp.net>  . "Tom Linden" <tom@kednos.com> wrote in message3 news:NDEMLKKEBOIFBMJLCECIAENIDLAA.tom@kednos.com...  >   I have  > < no idea G > < if Windows would allow mapping the EFI partition as a drive or not.  > 1 > FWIW, I think AdvFS under Tru64 would permit it    I don't think so.   L AdvFS is a file system.  Tru64 or HP-UX, or Linux, or whatever would need toI map it as a FAT file system not AdvFS.  But I'm sure that many UNIX's and 9 Linuxes have FAT file system code (legally or illegally).   + But I'm not a UNIX file system kind of guy.    ------------------------------  * Date: Wed, 1 Sep 2004 20:35:14 +0000 (UTC)7 From: moroney@world.std.spaamtrap.com (Michael Moroney) + Subject: Re: Itanium: EFI... no fs0: device ( Message-ID: <ch5bq2$97v$1@pcls4.std.com>  5 Michael Unger <spam.to.unger@spamgourmet.com> writes:   . >On 2004-09-01 19:57, "Michael Moroney" wrote:  L >> I just hope that what happened when you ran both NT and VMS (or Unix) on M >> older Alphas doesn't happen on Windows/VMS Itaniums.  On those Alphas, NT  N >> would see the VMS disk and would ask to write a "harmless signature" to it.F >                                                   ^^^^^^^^^^^^^^^^^^J >> Unfortunately, this "signature" overwrote the bootblock, making the VMS >> drive unbootable.  G >Isn't that a "feature" of Winwoes totally independent of the CPU it is  >running on?  G It is, but the only time it would be a problem is when there is a drive D with an OS on it that is not partitioned the way Windows expects it.E Not a problem with PCs (no VMS disks there!) except maybe for certain I flavors of Linux/Unix.  Linuxes I've seen all live on a partitioned drive # that Windows understands/tolerates.  --   -Mike    ------------------------------  # Date: Wed, 01 Sep 2004 20:11:54 GMT 4 From: "Fred Kleinsorge" <fred.nospam@nospam.dec.com>+ Subject: Re: Itanium: EFI... no fs0: device 1 Message-ID: <e2qZc.9357$fi4.842@news.cpqcorp.net>   : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message% news:413625BE.306D697@teksavvy.com...  > Fred Kleinsorge wrote:L > > Windows (as will Linux and HPUX) will see the partition record - it will see  > > the J > > GUID table when present - and will see the EFI partition (the GUID for the  > > typeF > > says its EFI) and it will see the VMS partitions as unknown types. > E > So, since Windows is able to access FAT drives, then theoretically,  Windows G > will be able to play around with the contents of the EFI "drive" that  would be5 > physically stored on a disk that also contains VMS.  >   H OK imagine it's a FC disk so a second system running Windows can see it. "Could" I Windows hack the stuff around?  Maybe.  But it would mean that someone on  the E Windows system was deliberately repartitioning a disk with *existing* 
 partitions -- L NOT unassigned space.  In which case you will most likely have a corrupt VMS disk.   G But hey.  Give physical access to a drive, and unlimited privs - and it  doesn't matterJ what the OS is if someone wants to be malicious.  But I don't believe that the Windows H tools will "gratuitously" allow you to repartition a drive - or do it on	 it's own.   L > And since the GUID is a microsoft structure, wouldn't Windows then be able toI > resize the partitions and give the EFI a lot more room which would then  quash  > the VMS partition ?  >   K The GUID is *not* a Microsoft structure.  It is a specification produced by  Intel.  I > I have no problems with the EFI stored as a container file inside a VMS  drive.K > But I do have a problem with a VMS drive having windows structures (GUID)  ate ; > the very base/core of the disk drive under the VMS stuff.  >    It isn't a Windows structure.   L > I take it data drives will not require to have that microsoft partitioning of > a drive ?  >   K Correct.  But sometimes they aren't out to get you - it really is paranoia.    > I > Ok, this is where I am perplexed. If the drive is partitioned, wouldn't  VMS H > just see its own partition on that logical drive ? Wouldn't it see the other D > partitions as different drives on the system ? If so, how come you wouldn't be D > able to have multiple logical  drives on the same physical drive ? >   F Is it me or a comprehension problem?  VMS DOES NOT SUPPORT PARTITIONS.D PERIOD.  VMS has done some "magic" to map several "special files" asH contiguous container files.  The VMS IO subsystem and filesystems do NOTH create logical drives out of partition information.  If someone wants to	 replicate K what we do for "other" partitions than EFI and the Diagnostic partition --   have at : it - but it still won't amount to parition support in VMS.  L > Or have you guys found a way to cheat and and just mount the real physicalL > drive, bypassing the partitioning stuff the IA64 hardware is presenting to you ?  >   7 THE HARDWARE DOES NOT PRESENT PARTITIONING INFORMATION. E The physical disk contains information.  We are CHOOSING to read that  information J and interpret it in specific ways.  From an OS and file system perspective there IS NO I partitioning.  The ONLY things in VMS that know about it at all are INIT,  SETBOOT 2 and BACKUP (which knows it needs to call SETBOOT).  L EFI chooses to read that information and implement partitioning.  It creates logical I drives for the partitions that are identified as EFI.  Windows chooses to  read the partitioning J and for all I know it can map any partition ir wants as a known or unknown file system  with a logical device letter.   F > If you bypass the partitioning stuff, doesn't that conflict with theH > positioning of ODS structures with the positioning of the partitioning, > structures at the beginning of the drive ? >    Please read slowly.   L VMS creates the partition data for a VMS disk.  Nothing else does.  Nor will any L other partitioning tools work for VMS.  The VMS tools create structures that are L interpreted by EFI as partition records, and identify some parts of the disk as an J "unknown" partition type, and one for EFI and one for HP Diagnostics.  The toolJ then creates a home block that does not overlap the partition records, and creates K container files that are contiguous at specific locations where the EFI and  DiagJ partitions reside.  VMS itself (XQP,F11ACP,RMS,etc) only sees the files as files.I We've given them attributes to keep you from moving them or deleting them  without J jumping through hoops.  We've also made sure that backing up the disk will cause ? all the right pointers to get fixed up on the destination disk.     I > Or is VMS smart enough to sequentially read blocks from block 0 onwards  until K > it finds the ODS structure ? (thus allowing it to transparently skip over  the ! > microsoft partitioning garbage)   " Do you know anything about ODS2/5?  F The "root" of the fie system structure is usually at block 1 (the home block).  VMSI searches for a signature (suprise) in this block.  If it is not found, it  looks for it in H other locations (in the old days a physical block might go bad and *not*
 have fancy disk HW to redirect it).  L On IPF, the home block is not in block 1.  The GUID partition table (GPT) is there.   ------------------------------  % Date: Wed, 01 Sep 2004 17:13:42 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> + Subject: Re: Itanium: EFI... no fs0: device , Message-ID: <41363B73.F23F0C42@teksavvy.com>   Fred Kleinsorge wrote:M > The GUID is *not* a Microsoft structure.  It is a specification produced by  > Intel.  M But isn't this the standard method used on wintel machines to partition lareg % drives into multiple smaller drives ?     H > Is it me or a comprehension problem?  VMS DOES NOT SUPPORT PARTITIONS.F > PERIOD.  VMS has done some "magic" to map several "special files" as > contiguous container files.   I I have no problem with that. And I understand that perfectly from the VMS  point of view.  A However, from the harware point of view, things are no so simple.   M Consider VM from IBM. It was able to present a vartual hardware machine to an K OS and the OS didn't know about anything else running on the same mainframe  (unless it really wanted to).   M So while VMS may be totally oblivious to the various structures in and around J its system drive, the fact remains that those structures are there and areT accessivble from other operating systems who do have the ability to understand them.  . >  The VMS IO subsystem and filesystems do NOT6 > create logical drives out of partition information.   M And I never though that VMS would *create* them. But if VMS asks the hardware K for a list of disk drives, and the hardware provides a list of disk drives, $ then VMS will use that list. Right ?  K If hardware tells a windows OS that it has 2 disk drives, how come the same I hardware would tell VMS it has only one disk drive ? (eg: if the hardware K understand the partitions, how come windows is told about the multiple disk  drives, but VMS would not ?)    9 > THE HARDWARE DOES NOT PRESENT PARTITIONING INFORMATION.   ! Ok, let me rehprase the question:   I Lets say you wanted to install windows on an IA64 box, and split the hard @ drive into drives C and D. That is called partitioning , right ?  J Would you in effect have to create 3 partitions on that drive, one EFI for, booting, one for disk C and one for disk D ?  M Or would you create 2 GUID partitions (one for EFI, and one for the data) and 5 then partition the data partion into 2 disk drives ?)   N Again, these questions revolved around HP employee's use of the term PARTITIONI combined with microsoft/intel standards. So we assume that it is the same 3 partitioning which is so common on wintel machines.     N > On IPF, the home block is not in block 1.  The GUID partition table (GPT) is > there.  K Ok, so you have chosen to place the ODS home block in some of the alternate  positions, correct ?  I So if I were to physically plug a IA64-VMS system disk onlt my all mighty K microvax II with VMS 7.2, would I be able to mount the drive (eg: would old K VMS ODS-2 enventually find the home block ?, or is it now located somewhere 4 where old VMS systems wouldn't thik of looking at ?)    M Again, I understand how the container file issue is done from within VMS/RMS. H And I understand how VMS knows nothing about partitioning. But since theH hardaware loads its own software well before VMS is booted, then it does< become an issue if tha partitioning is usable from other OS.   ------------------------------  % Date: Wed, 01 Sep 2004 20:07:16 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>+ Subject: Re: Itanium: EFI... no fs0: device + Message-ID: <41367244.4E421945@comcast.net>    (DJ dons dental smock...)    Hoff Hoffman wrote:  > [snip]E >   The fs0 device is the first FAT partition found, and it typically E >   contains the EFI loader for OpenVMS.  (OpenVMS I64 can have three E >   to five disk partitions and one to two FAT partitions on a system E >   disk, and the order of the partitions is not necessarily static.)   F Sounds like a topic worthy of a session at the HP Tech Forum, Nawlins.  F >   If you are not seeing an fs0 or an fs0 and fs1 device,  then there; >   are no bootstrap structures written to the target disk.   > How does one fix that (step by step, and type slowly so we can
 follow...)  A >   The tool that writes bootstrap structures is sys$setboot.exe. C >   (There is an IA64_SETBOOT.COM wrapper around in the cross-tools B >   kit, and that can be used -- it sets up logicals that fool theC >   image activator into thinking the images have been moved to the B >   system directories.  I tend to simply move the images into theB >   appropriate system directories, and use the tool directly from3 >   there, though this technique is not supported.)   / ...and he wrapper.COM proc. is found ... where?   K >   I tend to use the boot options menu and not the fs devices, but that is I >   largely preference -- I tend to add boot aliases for the devices I am F >   interested in, and do not tend to use nor to reference the fsX FAT >   partition devices directly.   ! Lost me at the bakery. Try again.   E >   What materials do you have on-hand?  The OpenVMS I64 distribution F >   disk on DVD media?  E8.2, specifically?  An rx2600 or similar box,E >   and the associated disk brick(s)?  In other words, what materials 3 >   and what hardware do we have to work with here?   F Rather depends on what's available, what it costs, and what it take to1 order/get it. Details (or URLs at least), please?   H >   If you have the DVD media, I will assume you have booted it and have- >   installed OpenVMS I64 E8.2 onto the disk.    ?   G >   If you are working from a network download kit, I have instructions G >   for burning the distro to optical media, and some related pointers. E >   If you don't have a burner and an optical drive available, you'll 9 >   have to find, configure, and use common SCSI storage.   F Ah. Now we are getting to meat of it: Sounds like bootable OpenVMS DVD7 media will be rather more of a challenge than CDs were.   E Amazing. Budgets tend to fill up to the dollar amounts available, and 5 software tends to expand to fill the available media.   F >   If you are creating a CD or DVD installation kit, then you need toG >   use sys$setboot against the master partition -- before you burn it.   C ...and one finds SYS$SETBOOT(.EXE) and its documentation ... where?    D.J.D.   ------------------------------  # Date: Thu, 02 Sep 2004 04:00:02 GMT   From: John Santos <john@egh.com>+ Subject: Re: Itanium: EFI... no fs0: device ( Message-ID: <6VwZc.432$H26.190@trnddc07>   JF Mezei wrote:  > Fred Kleinsorge wrote: > N >>Windows (as will Linux and HPUX) will see the partition record - it will see >>the L >>GUID table when present - and will see the EFI partition (the GUID for the >>typeE >>says its EFI) and it will see the VMS partitions as unknown types.   >  > M > So, since Windows is able to access FAT drives, then theoretically, Windows P > will be able to play around with the contents of the EFI "drive" that would be6 > physically stored on a disk that also contains VMS.  > O > And since the GUID is a microsoft structure, wouldn't Windows then be able to O > resize the partitions and give the EFI a lot more room which would then quash  > the VMS partition ?  >   H Not if you don't run Windows on it.  Windows can (theoretically) ignore F or change the partition table, just like anything else.  If you don't G want it to do this, 1) don't run Windows or 2) if you insist on running H Windows, don't repartition, run an antivirus, watch out when downloading< anything, and backup your VMS disks before you boot Windows.  P > I have no problems with the EFI stored as a container file inside a VMS drive.O > But I do have a problem with a VMS drive having windows structures (GUID) ate ; > the very base/core of the disk drive under the VMS stuff.  >   F Data structures are DATA, not code.  Random writes to the GUID are no I more or less likely to cause corruption than random writes to INDEXF.SYS.   O > I take it data drives will not require to have that microsoft partitioning of  > a drive ?  >  >  >>I have	 >>no idea E >>if Windows would allow mapping the EFI partition as a drive or not.  >  > N > Worse case scenario MUST be expected from anything coming out of microsoft.   G What comes from Microsoft other than a data structure?  FAT is a way of I laying out data on a disk, just like ODS-2 or RSTS/E's file structure or  F one of the Unix file structures.  It may be inefficient, and have lessF built-in redundancy than ODS-5, but for a small number of files, it isG a) simple and b) widely implemented, so it works with relatively little F effort for many OSes.  Intel could have chosen any file structure theyH wanted or invented a new one.  They chose FAT.  This does not mean it is" running any Microsoft code at all.   > G >>INITIALIZE and SETBOOT handles all of this.  See HELP INITIALIZE/GPT.  >  > I > I tried HELP INITIALIZE/GPT on my all mighty Microvax II, but it didn't . > recognize that qualifier :-) :-) :-) :-) :-) >    V8.2?  > M >>The default on IPF is /GPT, and /NOGPT on Alpha.  Backup will recognize the N >>presence of the system files (GPT.SYS) and will also do the right thing when* >>backing up and restoring system volumes. >  > M >>We do not support paritioned disks *in VMS* we've creatively found a way to J >>co-exist.  We do *not* support multiple OS partitions on the same disk.  >  > M > Ok, this is where I am perplexed. If the drive is partitioned, wouldn't VMS N > just see its own partition on that logical drive ? Wouldn't it see the otherP > partitions as different drives on the system ? If so, how come you wouldn't beE > able to have multiple logical  drives on the same physical drive ?   > L > Or have you guys found a way to cheat and and just mount the real physicalR > drive, bypassing the partitioning stuff the IA64 hardware is presenting to you ? >   G Huh?  Partitioning is a software construct.  The IA64 hardware doesn't  G present any partitioning.  The IA64 doesn't even present physical disk  H drives.  That job is done by the disk controllers (or by the MSCP serverF if the IA64 is looking at a disk on another cluster member.)  The SCSIE card doesn't know anything about partition tables (unless it's a RAID H card but that still presents the same virtual disks to the firmware that it presents to the O/S.)    F > If you bypass the partitioning stuff, doesn't that conflict with theH > positioning of ODS structures with the positioning of the partitioning, > structures at the beginning of the drive ? >   C Which is why the partition table and FAT partition are contained in D container files placed at particular LBN's on the disk.  (Presumably6 marked "Nomove" so defraggers won't try to move them!)  O > Or is VMS smart enough to sequentially read blocks from block 0 onwards until O > it finds the ODS structure ? (thus allowing it to transparently skip over the ! > microsoft partitioning garbage)   F No, VMS treats the whole entire disk as having ODS structure.  INIT isH smart enough to reserve space for the partition table and FAT structuresE just as does the boot block (in INDEXF.SYS) and as it did the factory  bad-block file on RP06's...   E If you mount the disk on a VAX or Alpha, it will just show a new file " in [0,0] containing the FAT stuff.  ' It's a dessert topping and a floor wax!   G (There are viruses, real viruses made out of DNA and protein, not those D fake computer things, that encode more than one protein in the same G stretch of DNA.  You know how 3 DNA bases encode one amino acid?  Well, F if you start part way into the sequence of one protein, at a base thatC is not a multiple of 3 from the start, you'll see a "protein start  C sequence" followed by the encoding for a totally different protein. E Being efficient at protein encoding is enough of a survival advantage E for these viruses that once a random mutation allowed this to happen, G it has been conserved ever since.  Same way with IA64 system disks.  If H you start at the VMS home block (block 1, IIRC), you'll see a consistentF VMS ODS-2 or ODS-5 disk.  If you start at the partition table (pointedC to by a longword at a particular offset in the boot block, block 0, C IIRC, you'll see two partitions, one containing the FAT stuff and a E reserved partition for the rest of the disk.  The partition table and A the FAT partition are both in the same physical blocks as the new 6 reserved file in [0,0] in the ODS-2/5 view of things.)   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Wed, 01 Sep 2004 19:05:30 +0100 + From: Rodrigo Ventura <yoda@isr.ist.utl.pt> ) Subject: Re: misc questions about OpenVMS ) Message-ID: <m3vfexamqd.fsf@pixie.isrnet>   E Thank you Michael for your elucidating explanations. I'd just like to  followup on a bit of issues:  L >>>>> "Michael" == Michael Moroney <moroney@world.std.spaamtrap.com> writes:  ?     Michael> Something that old will probably be discussing the A     Michael> console processors of old VAXes and not the built-in A     Michael> consoles of later systems, and certainly not Alphas. E     Michael> Many of the details are similar but so much is different      Michael> now.   D Yeah, quite right. I recall it discussed really old VAX system, like
 11/something.     B     Michael> MSCP is a disk server protocol that operates over theF     Michael> cluster communications protocol (SCS).  It has nothing toD     Michael> do with DECnet.  MSCP is a disk block serving protocol,@     Michael> unlike NFS.  NFS is probably more equivalent to the(     Michael> DECnet file access methods.  F What do you mean by "disk block serving protocol"? Is it like "give meE block offset so-and-so"? Does this imply that only one node can mount @ a volume? Multiple mounts with block-level access would kill the filesystem structure...   B In my previous cluster instalation (which is being overriden as weF speak) I had a USERS volume in one node, and each node mounted it withA a MOUNT/SYSTEM command in the SYSTARTUP_VMS.COM. Is it ok to have D multiple mounts in a cluster? Should it be a MOUNT/CLUSTER once? I'm? confused with the simultaneous idea of MSCP being a block level E protocol, and multiple mounts in a cluster system... (multiple mounts D is how unix would do, although there is no such concept in unix as a cluster...)     D     Michael> CLUSTER_CONFIG*.COM configure a cluster by adding a newF     Michael> node, removing a node etc. They are almost the same, withE     Michael> the difference having to do with whether DECnet or LANCP F     Michael> is the agent controlling MOP (for satellite nodes bootingE     Michael> over a LAN).  Use LANCP (CLUSTER_CONFIG_LAN.COM) for new ?     Michael> configurations.  They do configure DECnet on a new B     Michael> cluster node, but when using LANCP DECnet is entirely     Michael> optional.  F By the way, the reason I suspect I damaged my previous instalation was; that I ran NET_CONFIGURE.COM (or something like that) after @ configuring the cluster and having it up and running. I did thatF because I wasn't able to get the MONITOR program working: when I tryedE to do a MONITOR/CLUSTER it said it was unable to communicate with the > other nodes. Do you know how do MONITOR contacts other cluster members?    ?     Michael> It is pretty much equivalent to a hard link to the #     Michael> [VMSCOMMON] directory.   D Is it an alias? I read somewhere about the distinction among primary links, hard links, and aliases.     B     Michael> If you do a SYSBOOT> SET WRITESYSPARAMS 0 just beforeD     Michael> exiting SYSBOOT the changes you made will apply only to1     Michael> that boot, otherwise they are saved.   ! Where are they saved, by the way?    Cheers,    Rodrigo    --    : *** Rodrigo Martins de Matos Ventura <yoda@isr.ist.utl.pt>. ***  Web page: http://www.isr.ist.utl.pt/~yoda0 ***   Teaching Assistant and PhD Student at ISR:7 ***    Instituto de Sistemas e Robotica, Polo de Lisboa 4 ***     Instituto Superior Tecnico, Lisboa, PORTUGALH *** PGP fingerprint = 0119 AD13 9EEE 264A 3F10  31D3 89B3 C6C4 60C6 4585   ------------------------------  % Date: Wed, 01 Sep 2004 14:38:40 -0400  From: John Doe <jdoe@doe.org> ) Subject: Re: misc questions about OpenVMS ' Message-ID: <41361726.85EB0AF4@doe.org>    Rodrigo Ventura wrote:H > What do you mean by "disk block serving protocol"? Is it like "give meG > block offset so-and-so"? Does this imply that only one node can mount B > a volume? Multiple mounts with block-level access would kill the > filesystem structure...   M Yes, it is a disk block serving protocol. And no, it does not imply that only K one node can mount the drive. As a matter of fact, there are configuratiosn I where the disk drive has multiple direct paths to each node, so each node L sends the SCSI commands directly to the drive and there is no "single" ownerK of the drive through which all MSCP requests pass. But with the distributed + lock manager, this all works out perfectly.   N VMS has a distributed lock manager. So when one node mounts a drive, any locksK it takes on the drive, files or records on that drive are seen by all other N nodes. This is what makes VMS clustering orders of magnitudes more elegant and- functional than on any other cluster-wannabe.   J Consider a cluster that has a POP server on one node and an SMTO server onL another node. Because of the distributed lock manager, if the SMTP server isJ writing a new record to a mail index, the POP server will automatically beC told that the record is locked and "please retry in a few seconds".   K Without a distributed lock manager, such a setup requires the POP server to K create a file telling anytone else not to touch the mail file while the pop L server is accessing it, and if the pop server dies, that file remains there,F blocking delivery to that user until some deamon comes to clean it up.  D > In my previous cluster instalation (which is being overriden as weH > speak) I had a USERS volume in one node, and each node mounted it withC > a MOUNT/SYSTEM command in the SYSTARTUP_VMS.COM. Is it ok to have  > multiple mounts in a cluster?   N You can have MOUNT/SYSTEM done by each node, or a MOUNT/CLUSTER/SYSTEM done by. one node (usually the one that owns the node).  N If the drive is acessed physically only by one node, it is a good idea to haveJ that node perform a MOUNT/CLUSTER on the drive when it boots. Consider theJ case where you brought the node down and it is being brought back up whileG other nodes are purring right along. The other nodes won't be executing K systartup_vms.com since they are already booted, but when the one node that J boots does the mount/cluster, then all other nodes will remount the drive.  N But if Node A stays up while NOde B goes down, then when Node B comes back up,J Node B will need to mount Node A's drives manually because Node A won't be" running any procedures to do that.  K This is where with experience and thinkering, you will concuct some elegant I ways of managing this type of activity. SYSMAN exists to allow you to ask I other nodes to execute tasks for you, so when node B boots, you could use 5 SYSMAN on node B to ask Node A to perform some tasks.   K Similarly, it coudl also ask all nodes to perform some task, at which point < SYSMAN sends commands to all nodes at roughly the same time.  A > confused with the simultaneous idea of MSCP being a block level G > protocol, and multiple mounts in a cluster system... (multiple mounts F > is how unix would do, although there is no such concept in unix as a
 > cluster...)   I Once booted, if you do a SHOW DEV D, you will see that VMS considers disk N drives that belong to other nodes to be just your regular disk drive, and fromH your perspective, they are the same as if they had been mounted/attached< locally even though that drive could be hundreds of km away.    H > because I wasn't able to get the MONITOR program working: when I tryedG > to do a MONITOR/CLUSTER it said it was unable to communicate with the @ > other nodes. Do you know how do MONITOR contacts other cluster
 > members?    K By default, Monitor uses DECNET. And by default, SYSMAN does use DECNET for L certain types of tasks. There are ways to tweak Monitor to use TCPIP insteadK of DECNET. But monitor does require a high level protocol (decnet or tcpip) ) and not a low level protocol such as SCS.   F > Is it an alias? I read somewhere about the distinction among primary! > links, hard links, and aliases.   * On VMS, there is just one thing: an alias.  L create chocolate.recipe   will create a real file with its own file id (fileM number) and its own entry in indexf.sys and its own allocated blocks, as well  as its own directory entry  . set file chocolate.recipe/enter=chocolate.cake  K All this does is create a new entry in the directory file. "chocolate.cake" E will appear in the directory listings. But all the info will point to L "chocolate.recipe" because .cake will have the same file id (file number) asK .recipe and thus point to the same entry in indexf.sys and this to the same  disk blocks etc.  Z help SET FILE/ENTER has pretty good description. In the context of the system directories:  . $ dir/fileid $disk4:[sys0]syscommon.dir/nohead< $2$DKA200:[SYS0]SYSCOMMON.DIR;1                     (11,1,0)  1 $ dir/fileid $disk4:[000000]vms$common.dir/nohead < $2$DKA200:[000000]VMS$COMMON.DIR;1                  (11,1,0)   ------------------------------  * Date: Wed, 1 Sep 2004 19:52:50 +0000 (UTC)7 From: moroney@world.std.spaamtrap.com (Michael Moroney) ) Subject: Re: misc questions about OpenVMS ( Message-ID: <ch59ai$qp7$1@pcls4.std.com>  - Rodrigo Ventura <yoda@isr.ist.utl.pt> writes:     F >Thank you Michael for your elucidating explanations. I'd just like to >followup on a bit of issues:    No problem.   C >    Michael> MSCP is a disk server protocol that operates over the G >    Michael> cluster communications protocol (SCS).  It has nothing to E >    Michael> do with DECnet.  MSCP is a disk block serving protocol, A >    Michael> unlike NFS.  NFS is probably more equivalent to the ) >    Michael> DECnet file access methods.   G >What do you mean by "disk block serving protocol"? Is it like "give me F >block offset so-and-so"? Does this imply that only one node can mountA >a volume? Multiple mounts with block-level access would kill the  >filesystem structure...  G MSCP operates at a very low level.  It really does operate at the level F "give me block offset so-and-so".  The remote system just acts as a goJ between between the actual disk and the client system.  MSCP knows nothingF of file structure or anything like that.  As to "multiple mounts with D block-level access would kill the filesystem structure", this is theG magic of the VMS distributed lock manager.  It operates on the file (or D record) level, and MSCP knows nothing about it.  (MSCP will happily ? corrupt a disk if there was no locking coordinating everything)   C >In my previous cluster instalation (which is being overriden as we G >speak) I had a USERS volume in one node, and each node mounted it with B >a MOUNT/SYSTEM command in the SYSTARTUP_VMS.COM. Is it ok to have >multiple mounts in a cluster?  ; Yes, see above, multiple mounts within a cluster work fine.   $ > Should it be a MOUNT/CLUSTER once?  I MOUNT/CLUSTER is just a shorthand for doing a MOUNT/SYSTEM on every node   in the cluster at the time.    > I'm @ >confused with the simultaneous idea of MSCP being a block levelF >protocol, and multiple mounts in a cluster system... (multiple mountsE >is how unix would do, although there is no such concept in unix as a  >cluster...)  D Again, this is the lock manager.  It coordinates access in a clusterC so the nodes don't corrupt everything.  One node directly connected K to a drive and also serving it via MSCP to other nodes logically works the  H same as several nodes on a fibrechannel fabric accessing a single drive.G The only difference is the disk "controller" is also another full node   in the cluster.   A >configuring the cluster and having it up and running. I did that G >because I wasn't able to get the MONITOR program working: when I tryed F >to do a MONITOR/CLUSTER it said it was unable to communicate with the? >other nodes. Do you know how do MONITOR contacts other cluster 	 >members?     B I believe it operates over DECnet.  Which means you have to set up? network "objects" to handle remote MONITOR requests.  NETCONFIG : should do that but you have to answer the right questions.  @ >    Michael> It is pretty much equivalent to a hard link to the$ >    Michael> [VMSCOMMON] directory.  E >Is it an alias? I read somewhere about the distinction among primary   >links, hard links, and aliases.  D I know of links in the Unix world.  A hard link is when a file is inC more than one directory.  Both directory entries point to the exact H same file ("inode" in Unix terms) and are equal.  A soft link is when a F special file exists that really contains the path to another file and H nothing else.  Accessing this file tells the file system to look at its H contents which tells it where the real file is.  Windoze .LNK files are K really soft links. VMS links are hard links except for the fact that there  J is only one "backlink" (points to the directory a file is in) so all hard  links are not totally equal.  C >    Michael> If you do a SYSBOOT> SET WRITESYSPARAMS 0 just before E >    Michael> exiting SYSBOOT the changes you made will apply only to 2 >    Michael> that boot, otherwise they are saved.  " >Where are they saved, by the way?  G SYS$SYSTEM:ALPHAVMSSYS.PAR on Alpha, SYS$SYSTEM:VAXVMSSYS.PAR on VAXes.  Dunno Itanium. --   -Mike    ------------------------------  % Date: Wed, 01 Sep 2004 20:36:25 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>) Subject: Re: misc questions about OpenVMS + Message-ID: <41367919.E8EBA5E2@comcast.net>    Michael Moroney wrote: > / > Rodrigo Ventura <yoda@isr.ist.utl.pt> writes:  > [snip]& > > Should it be a MOUNT/CLUSTER once? > J > MOUNT/CLUSTER is just a shorthand for doing a MOUNT/SYSTEM on every node > in the cluster at the time.   2 Sort of. Ideally, that's how its supposed to work.  F There are a number of "known" problems with (DIS)MOUNT/CLUSTER. One ofF the more troublesome is easiest to explain (to the seasoned VMS folks)	 this way:   F When a volume was DISMOUNTed/CLUSTER, the HSJ pair insisted that there, were still channels open to it, even though:   $ MC SYSMAN  SET ENV/CLU  DO SHOW DEV ddcu:   F ...clearly showed that the volume was not MOUNTed on any cluster node.   The fix?   $ MC SYSMAN  SET ENV/CLU . DO MOUNT/SYSTEM ddcu volume_label logical_name DO DISMOUNT ddcu  B ...and "magically", the HSJ pair now allowed deletion of the unit,  mirror-set and the stripeset(s).   > > I'm B > >confused with the simultaneous idea of MSCP being a block levelH > >protocol, and multiple mounts in a cluster system... (multiple mountsG > >is how unix would do, although there is no such concept in unix as a  > >cluster...)  6 Well, yes and no. Don't get Andrew Harrison started...  # > Again, this is the lock manager.    F Eh, well, yes and no. Yes, the lock manager does co-ordinate activity.C However, this alone does not account for the ability to MOUNT disks E cluster-wide. If it did, there would be no need for the VMSclustering E SIP that was once installed separately but is now integrated into the 
 distribution.   $ > It coordinates access in a clusterE > so the nodes don't corrupt everything.  One node directly connected L > to a drive and also serving it via MSCP to other nodes logically works theJ > same as several nodes on a fibrechannel fabric accessing a single drive.H > The only difference is the disk "controller" is also another full node > in the cluster.   C Note also that the VMS system itself can be the "controller", or it ' could be an HSJ, HSZ or HSC (very old).   C > >configuring the cluster and having it up and running. I did that I > >because I wasn't able to get the MONITOR program working: when I tryed H > >to do a MONITOR/CLUSTER it said it was unable to communicate with theA > >other nodes. Do you know how do MONITOR contacts other cluster  > >members?  > D > I believe it operates over DECnet.  Which means you have to set upA > network "objects" to handle remote MONITOR requests.  NETCONFIG < > should do that but you have to answer the right questions.  , Quite so - NETCONFIG should do this for you.  B > >    Michael> It is pretty much equivalent to a hard link to the& > >    Michael> [VMSCOMMON] directory. > G > >Is it an alias? I read somewhere about the distinction among primary " > >links, hard links, and aliases.  H That's UN*X lingo. recent revisions to ODS for ODS-2 and ODS-5 introduce. some similar concepts, but not quite the same.  F > I know of links in the Unix world.  A hard link is when a file is inE > more than one directory.  Both directory entries point to the exact I > same file ("inode" in Unix terms) and are equal.  A soft link is when a G > special file exists that really contains the path to another file and I > nothing else.  Accessing this file tells the file system to look at its I > contents which tells it where the real file is.  Windoze .LNK files are L > really soft links. VMS links are hard links except for the fact that thereK > is only one "backlink" (points to the directory a file is in) so all hard  > links are not totally equal.  F Careful there. The "directory backlink" ("DID") is in the file header,H not the directory entry. The directory entry points to the file header(sH - multiple versions, remember?). A file can have more than one directoryF entry; however, there is still no known way to track all the directoryE entries for any given file, AFAIK. A counter is kept of the number of G known "links" (read: directory entries) for each file, similar to UN*X. F Thus, if a file's link counter is non-zero, deleting the file could beD the wrong thing to do. However, and again AFAIK, there is no list of9 directory entries that can be traversed to find them all.   E > >    Michael> If you do a SYSBOOT> SET WRITESYSPARAMS 0 just before G > >    Michael> exiting SYSBOOT the changes you made will apply only to 4 > >    Michael> that boot, otherwise they are saved. > $ > >Where are they saved, by the way? > I > SYS$SYSTEM:ALPHAVMSSYS.PAR on Alpha, SYS$SYSTEM:VAXVMSSYS.PAR on VAXes.   F The .PAR file he mentioned is the system parameters that are loaded atH boot time. The default value for WRITESYSPARAMS is 1 (true or yes: writeH the booted parameters back to the parameter file). Thus, it retains thisE value after AUTOGEN, unless specifically overridden in MODPARAMS.DAT.    D.J.D.   ------------------------------  % Date: Wed, 01 Sep 2004 20:46:28 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>) Subject: Re: misc questions about OpenVMS + Message-ID: <41367B74.8749AB5E@comcast.net>    "Keith A. Lewis" wrote:  > J > My Answers are from practical experience and may not be exactly correct. > ~ > Rodrigo Ventura <yoda@isr.ist.utl.pt> writes in article <m38ybuarb8.fsf@pixie.isrnet> dated Wed, 01 Sep 2004 17:26:35 +0100:I > >2. what's the relationship between MOP, MSCP and DECnet? I believe MOP G > >is really low level, in terms of only supporting booting and console G > >commands. MSCP is a kind of NFS (unix) for VMS. Does MSCP run on top  > >of DECnet? Or not?  > K > Yes, MOP is a boot loading protocol.  It can be used to boot a VMS system M > from a disk in a cluster on a LAN, or it can be used to boot lesser devices  > such as terminal servers.   E Well, yes and no. M.O.P. stands for "Maintenance Operation Protocol". H The most commonly used sub-protocols were downline load, upline dump andG virtual console. MOP is non-routable - must be bridged between ethernet * segments transparently at decent priority.  L > MSCP is a disk sharing protocol used within a VMS cluster.  It runs on topM > of SCS, not Decnet.  It differs from NFS in that it can only be used within ? > a cluster; whereas NFS can be used to share disks between two ! > otherwise-independent machines.  > K > DECnet can also be used to share files, but I would only recommend it for  > small files.  C It should be noted that DECnet was used to "share files" before the   emergence of VAX/VMS clustering.  D > >3. does this last question has anything to do with the differenceI > >between the CLUSTER_CONFIG.COM and the CLUSTER_CONFIG_LAN.COM scripts? 9 > >Does any one of these automatically configures DECnet?  > 7 > CLUSTER_CONFIG calls CLUSTER_CONFIG_LAN in one place.  > G > >4. regarding the boot flags on the alpha platform, in the form "boot G > >-fl x,y device", the x number is only used when booting from a local H > >disk, and when booting from LAN (MOP?) the x is ignored (in favour of@ > >the [SYSn] directory indicated by MOP; the y number is always? > >respected. Is this right? In what document is this throughly 
 > >explained?  > H > My personal experience supports that, I don't know if it's documented.  H Try the clustering manuals - if its anywhere, that probably a good place to start looking.   G > >5. inside the [SYSn] directory there is a SYSCOMMON subdirectory; is ; > >it a copy of SYS$COMMON, or some sort of soft/hard-link?  > ? > [SYSn]SYSCOMMON.DIR is an "alias" for [000000]VMS$COMMON.DIR. ' > "help set file /enter" for more info.  > C > >6. when booting with "-fl 0,1" the SYSBOOT> (or SYSGEN>, I can't ? > >recall) appears, which looks like s stripped-down version of H > >SYSGEN. The modifications made there (e.g., USE DEFAULT, SET, etc) doF > >seem permanent, in the sense that in the next boot the settings areG > >the same; how can I revert to a previous setting. Or in other words, 9 > >after a minimal boot, how can I make it boot normally?  > > > Those parameters are stored in SYS$SYSTEM:ALPHAVMSSYS.PAR orM > SYS$SYSTEM:VAXVMSSYS.PAR.  I don't know offhand if it creates a new version 0 > of the file or modifies the existing version.    Modifies the existing version.   > Either way, you can undo theB > changes by restoring that file from a backup or by using SYSMAN. > 
 > $ MC SYSMAN  > SYSMAN> PARAM USE CURRENT ! > SYSMAN> PARAM SET STARTUP_P1 ""  > SYSMAN> PARAM WRITE CURRENT  > SYSMAN> EXIT  G ...or just rerun AUTOGEN from GETDATA through REBOOT, or from SAVPARAMS G through REBOOT if your feedback info would be valid (system has been up  at least 24 hours).   F > >7. does AUTOGEN uses the current system values or the saved ones --H > >for instance, after a minimal boot (i.e., USE DEFAULT; SET STARTUP_P1- > >"MIN"), does AUTOGEN "fubar"'s the system?  > M > AUTOGEN gets its data from feedback info (which may include current params) H > and MODPARAMS.DAT.  Some parameters aren't even controlled by AUTOGEN.7 > Always review the report before using the new params.   = AUTOGEN starts from the default parameters, and uses feedback H (conditionally) as well as the information in MODPARAMS.DAT to calculate a new parameter set.  C Some parameters are not calculated - string parameters for example. H Others are choices that the system administrator makes like the value ofG the clustering parameter, the value of SCSSYSTEMID, MSCP_LOAD, the SHDW  parameters, ...    D.J.D.   ------------------------------  % Date: Wed, 01 Sep 2004 20:53:11 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>) Subject: Re: misc questions about OpenVMS + Message-ID: <41367D06.CC89095A@comcast.net>    Jf Mezei wrote:  > [snip]K > At the protocol level, they are quite separate. However, MOP historically L > relied on the DECNET database to authorize a node to get served and defineM > while file would be served to that node, but this is not not 100% accurante 7 > since you can hve the LANCP database do that job now.   @ In practice, the node requesting the downline load can ask for aH specific file. Any node listening on that line/circuit with line serviceD enabled on that circuit and which ash / can find the requested file,# could and likely would serve it up.   N > MOP simply feeds a file to a remote device when that device send an ethernet3 > broadcast asking if anyone is willing to feed it.   G Actually, anyone who has it and has line service enabled appropriately.   D Where you NEED a DECnet or LANCP database entry is for those devicesE that simply ask for an MOP downline load and don't request a specific D file, or need to be fed multiple files (primary, secondary, tertiaryF bootstraps) or which need authorization or configuration assistance at& boot time (SNA Gateway comes to mind).   D.J.D.   ------------------------------  # Date: Wed, 01 Sep 2004 20:44:23 GMT + From: LESLIE@JRLVAX.HOUSTON.RR.COM (leslie)  Subject: Re: need help2 Message-ID: <HwqZc.3598$Dl4.2086@fe2.texas.rr.com>    supe_ulhas@rediffmail.com wrote:! : how do i access SETCIM records, . : what is the exact structure of setcim record :  : thanks :   / From an Aspentech (www.aspentech.com) employee:   D   "Point them to the Setcim API manual.  There are also some sample J    programs somewhere on the distribution (or at least there used to be)."     --Jerry Leslie9   Note: leslie@jrlvax.houston.rr.com is invalid for email    ------------------------------   Date: 1 Sep 2004 16:36:22 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen)  Subject: Re: need help3 Message-ID: <s4vDgXfUyI+o@eisner.encompasserve.org>   _ In article <c4fc6627.0409010614.31e9f5e9@posting.google.com>, supe_ulhas@rediffmail.com writes: ! > how do i access SETCIM records, . > what is the exact structure of setcim record   What is a SETCIM record ?    What is a SETCIM ?   ------------------------------   Date: 1 Sep 2004 16:30:47 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) ) Subject: RE: OpenVMS running on HP 3000's 3 Message-ID: <rgl+wCFiPAE7@eisner.encompasserve.org>   | In article <FD827B33AB0D9C4E92EACEEFEE2BA2FB3E0025@tayexc19.americas.cpqcorp.net>, "Main, Kerry" <kerry.main@hp.com> writes:   > Re: "legacy" systems ..  >  > Rant mode on>  > D > I always get a kick out of those media types that like to refer toF > "legacy" systems as necessary evils that does not warrant additional > expenditures.   I My wife had a boss (of the IBM mainframe persuasion) who defined "legacy"  as "stuff that works".   ------------------------------   Date: 1 Sep 2004 16:22:57 -0500 - From: Kilgallen@SpamCop.net (Larry Kilgallen) 1 Subject: Re: Planned port of freeware to Itanium? 3 Message-ID: <HEKnk3iFw1dA@eisner.encompasserve.org>   q In article <I$56foiIXvCc@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes: l > In article <79de16e3.0408270452.652f2557@posting.google.com>, Anders.Wallin@om.com (Anders Wallin) writes:	 >> Hello,  >>  G >> Are there any plans from HP for a port of any of the products on the  >> freeware CD to Itanium? >>  I >> The freeware products are very useful and it vould be valuable to have " >> access to them on Itanium also. > D >    I thinks that's up to the folks who proivided them in the first% >    place, or anyone who volunteers.   E I recall reading that some University got an Itanium machine for free - in return for agreeing to port some freeware.   D Personally I don't want much in the way of freeware, but I _do_ wantC students to get experience maintaining someone else's code.  I feel B certain that one of them will someday program an airplane on which I will fly.    ------------------------------  % Date: Wed, 01 Sep 2004 21:55:02 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> / Subject: SMTP: stray messages left in directory , Message-ID: <41367D52.99BD41C1@teksavvy.com>  , TCPIP Services, SMTO  5.3 ECO 2, VAX VMS 7.2  D I have found that for some reason, some messages, usually those withJ attachements, just fail to be delivered, usually outbound from my vax, andL just stay as stray files in the TCPIP$SMTP_COMMON: directory with no warningP or error messages being sent, so the sender assumes the messages have been sent.    J ANALYZE MAIL does tell me about some stray files with no queue entries but# doesn't do anything to those files.   9 Has anyone else had this problem ? Any simple solution ?    M It is really frustrating to consider that a message has been sent, but a week P or two later when you look at the directory, you realise that it was never sent.  L Right now, I have messages whose contents range from 15 to 305 blocks. Si it) isn't just very large messages that fail.    ------------------------------  # Date: Thu, 02 Sep 2004 05:01:31 GMT   From: John Santos <john@egh.com>3 Subject: Re: SMTP: stray messages left in directory ) Message-ID: <LOxZc.1695$Ch1.224@trnddc05>    JF Mezei wrote: . > TCPIP Services, SMTO  5.3 ECO 2, VAX VMS 7.2 > F > I have found that for some reason, some messages, usually those withL > attachements, just fail to be delivered, usually outbound from my vax, andN > just stay as stray files in the TCPIP$SMTP_COMMON: directory with no warningR > or error messages being sent, so the sender assumes the messages have been sent. >  > L > ANALYZE MAIL does tell me about some stray files with no queue entries but% > doesn't do anything to those files.  > ; > Has anyone else had this problem ? Any simple solution ?   > O > It is really frustrating to consider that a message has been sent, but a week R > or two later when you look at the directory, you realise that it was never sent. > N > Right now, I have messages whose contents range from 15 to 305 blocks. Si it+ > isn't just very large messages that fail.   D What's your uplink to the Internet?  If you are going through a NAT D router and PPPOE on a DSL circuit, you might be having MTU problems.G The MTU for Ethernet is 1500, but for PPPOE is 1492 (because of 8 bytes D of overhead.)  MTU negotiation is supposed to be automatic, but I'veE had lots of problems with various firmware revs on my home router not F propagating it correctly.  I've heard other reports (people with otherG brands of routers; I have a Linksys) of similar problems.  Anyway, what E happens is if the sender tries to send a packet larger than 1492, the D recipient never gets it (or only gets the 1st 1492 bytes of it), andF eventually times out waiting for the rest of the packet and closes theH connection.  I don't know how HP's SMTP server reacts when this happens.     --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  $ Date: Wed, 1 Sep 2004 15:24:40 -0400# From: "John Smith" <a@nonymous.com> / Subject: Re: Somewhat OT: Is AMD the new Intel? , Message-ID: <Pbadnf7tIY9ovKvcRVn-iw@igs.net>   Jf Mezei wrote:  > Tom Linden wrote: G >> Gee, that doesn't bode well for Itanium.  It is hard to imagine that F >> Itanium is going to get Intel's full attention, which is imperative" >> to making the chip competitive. > G > IA64 has an announced 3 year lifespan/roadmap for now. Engineers were D > given tasks. They'll deliver what they can when they can. The only7 > decisions management can make pertain to compromises.  > F > For instance, in order for AMD to deliver its dual core ships early,A > they have compromised on the cache, choosing to have 2 separate F > caches, instead of a single cache serving both cores (as is the case > for Power architecture). > F > Intel has also made that compromise for at least the dual core 8086, > not sure about IA64. > F >> I would hope that VMS amangement has the perspicacity to be workingF >> on porting VMS to Opteron and Power 5, so as not to be caught on an >> iceberg in a warm sea.  > C > I think that the engineers have stated a few times that this time F > around, they are restructuring the code to make it easier to port to( > other platforms should the need arise. > F > At this point in time, nobody within HP would ever reveal any covertE > plans to consider chips other than IA64 for ist enterprise systems, E > even if Hoff has been sequestered in some locked basement at ZKO to 3 > secretely work on porting VMS to the 64 bit 8086.     J I thought he was locked in the basement working on the VMS to EV8 port ;-)    K Seems to me that if HP EOL'd all the IA-64 stuff today and simply sunk some  bucks into EV8, they'd be:  K a) saving lots of money on IA-64 retirement costs simply because they don't = have very many machines in the field they'd have to swap out. L b) giving customers a proven CPU family (Alpha) at a price/performance point that was competitive. ) c) able to buy time to do VMS on Opteron. 2 d) able to blame the whole stinking mess on curly.   ------------------------------  % Date: Wed, 01 Sep 2004 15:50:02 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> / Subject: Re: Somewhat OT: Is AMD the new Intel? , Message-ID: <413627DC.966839AF@teksavvy.com>   John Smith wrote: M > Seems to me that if HP EOL'd all the IA-64 stuff today and simply sunk some  > bucks into EV8, they'd be: > M > a) saving lots of money on IA-64 retirement costs simply because they don't ? > have very many machines in the field they'd have to swap out.   C Correct. This is why the fewer IA64 are sold commercially, the less  liabilities HP will have.   N > b) giving customers a proven CPU family (Alpha) at a price/performance point > that was competitive.   H HP would have a problem with this. Remember that there is also HP-UX and Tandem Non-Stop.  K The only way restarting Alpha would make sense is if HP were to decide that N Alpha was to be its strategic chip to battle against IBM's power and make thisG a long term thing.  HP would have to make some difficult decisions with M regards to Tru64 versus HP-UX since it would be more cost efficient to simply % migrate some HP-UX features to Tru64.   + > c) able to buy time to do VMS on Opteron.   K Right now, I'd say that Alpha still has enough umph to last until a port to N Opteron is working. Remember that the next port shouldn't be so hard since theJ engineers have cleaned up the code to allow it to be easily ported. And itH would be far simpler to just do speed bumps and process shrinks of Alpha compared to EV8.  4 > d) able to blame the whole stinking mess on curly.  M No, blame it on carly and throw her out. Get Curly to hire Carly at Worldcom.    ------------------------------  % Date: Wed, 01 Sep 2004 22:02:27 +0200  From: Dirk Munk <munk@home.nl>, Subject: Re: strange decnet-plus problems...2 Message-ID: <ch59t5$5sg$1@news5.tilbu1.nb.home.nl>  @ Did you add your new node to the name space of the other nodes ?   (MC DECNET_REGISTER)   Carmine wrote:F > I can only set host to my new system via decnet ,if I first set hostA > from this box to others , has anyone seen anything like this ??  >  > Node1  > $ set hos Node2 ? > %SYSTEM-F-UNREACHABLE, remote node is not currently reachable  >  > Node1  > $ telnet node2 > (works fine...)  > Node2  > $ set hos node1  > (works fine...)  > logout from both nodes >  > Node1  > $ set hos node2  > (works for a while...) >  > Later in day   > Node1  > $ set hos node2 ? > %SYSTEM-F-UNREACHABLE, remote node is not currently reachable  >  > Node1 , > $ set hos node1(to itself , always works ) > = > This is a brand new system is there a patch I am missing ??  > VMS 7.3-2  > decnet-osi for VMS   ------------------------------   End of INFO-VAX 2004.486 ************************