INFO-VAX Fri, 13 Jul 2007 Volume 2007 : Issue 380 Contents: Re: %LICENSE-F-INTERINJ, internal LMF error was encountered 1994... Alphaserver ES45 Special Offer Re: Alphaserver ES45 Special Offer Re: Alphaserver ES45 Special Offer Re: Alphaserver ES45 Special Offer Re: Alphaserver ES45 Special Offer RE: Cobol Link Issue... Re: EDT Replacement Re: EDT Replacement Re: EDT Replacement Re: EDT Replacement Experience building Perl modules on VMS Itanium? Re: Experience building Perl modules on VMS Itanium? RE: FTSO/FASTCopy replacement Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? RE: HoffmanLabs: High-performance Interconnects, MPI? RE: HoffmanLabs: High-performance Interconnects, MPI? RE: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? Re: HoffmanLabs: High-performance Interconnects, MPI? RE: HoffmanLabs: High-performance Interconnects, MPI? Re: How many people here use Itanium w VMS Re: How many people here use Itanium w VMS Re: How many people here use Itanium w VMS Re: How many people here use Itanium w VMS Re: Install VMS on a Alphastation 200 ? Re: Install VMS on a Alphastation 200 ? Re: Install VMS on a Alphastation 200 ? Re: Install VMS on a Alphastation 200 ? IRT: Cobol Link Issue... Re: IRT: Cobol Link Issue... Re: IRT: Cobol Link Issue... Re: Is VMS losing the Financial Sector, also? Re: Is VMS losing the Financial Sector, also? Re: Is VMS losing the Financial Sector, also? Re: Is VMS losing the Financial Sector, also? RE: Linux (was Re: How many people here use Itanium w VMS) Now that's what I'm talking about! (was:Re: Is VMS losing the Financial Sector, RE: Now that's what I'm talking about! (was:Re: Is VMS losing the Financial Sec Re: Now that's what I'm talking about! (was:Re: Is VMS losing the Financial Sec Re: OpenVMS - When downtime is not an option Re: Solved: SAN switch 2/8 on MSA1000 - cannot connect to serial of rj-45 port Re: Solved: SAN switch 2/8 on MSA1000 - cannot connect to serial of rj-45 port Re: Solved: SAN switch 2/8 on MSA1000 - cannot connect to serial of rj-45 port Re: SYSMAN oddity (username on same node) UK report says data protection breaches "out of control" UK report says data protection breaches "out of control" Re: UK report says data protection breaches "out of control" Re: UK report says data protection breaches "out of control" Re: UK report says data protection breaches "out of control" Re: [OT] July the 4th Re: [OT] July the 4th ---------------------------------------------------------------------- Date: Fri, 13 Jul 2007 11:09:42 +0200 From: "Walter Kuhn" Subject: Re: %LICENSE-F-INTERINJ, internal LMF error was encountered Message-ID: <46974159$0$23941$91cee783@newsreader01.highway.telekom.at> Thank you, the patch solved the problem. Walter "Jilly" schrieb im Newsbeitrag news:46967290$0$5561$ec3e2dad@news.usenetmonster.com... > > "Walter Kuhn" wrote in message > news:46961385$0$18449$91cee783@newsreader02.highway.telekom.at... >> Hello Group, >> >> we are building a Cluster with 3 Itaniums under OpenVMS 8.3 with the >> OPENVMS-I64-MCOE license. >> When we reboot one machine, we get the error >> >> %LICENSE-F-INTERINJ, internal LMF error was encountered. Checkpoint 16419 >> -SYSTEM-F-IVLOCKID, invalid lock ID >> >> when issuing the $LIC LOAD command. To make the cluster work again, we >> have to reboot all machines! >> Does someone know that problem? >> >> We are just making a call for that problem... >> >> Regards >> Walter >> > > Have you installed the VMS83I_LMF-V0200 patch kit to make sure this issue > has not been corrected already? ------------------------------ Date: Fri, 13 Jul 2007 13:30:58 +0200 From: "Zibri" Subject: 1994... Message-ID: <4697626d$0$36442$4fafbaef@reader5.news.tin.it> http://www.youtube.com/watch?v=-1l6aBgX5UY ------------------------------ Date: Fri, 13 Jul 2007 09:28:01 -0400 From: "David Turner, Island Computers" Subject: Alphaserver ES45 Special Offer Message-ID: <139evemfde80vb5@news.supernews.com> C.O.V. Special for July Get 'em whilst they hot ! Alphaserver ES45 Model 2 (works with 120 and 240V 50/60Hz) 4 x 1Ghz CPU EV68CB with 8MB Cache 8GB Memory (8 x 1GB) 3 Power Supplies BA610-6D 6 Slot Disk Cage U160 Dual Channel RAID Controller with 128MB Cache ATI Radeon 7500 Graphics Card PCI DE602-AA Dual Channel 10/100 Ethernet 2Gbit Fiberchannel Controller PCI Pedestal or Rackmount Kit for H9A Cabinets Brand NEW HP Authorized OPenVMS base, SMP and EIP Licenses OpenVMS 7.3-2 preloaded if required. Only $18,995 Add $7500 for a NEW unlimited user license Other licenses available upon request This must be a better deal than a dual cpu Itanium !!!!!!!!!!!!!!!!!!!!!?????????????????????!!!!!!!!!! Call or email -- David B Turner Island Computers US Corp 2700 Gregory St, Suite 180 Savannah GA 31404 T: 877-6364332 x201 Intl: 001 912 447 6622 E: dturner@islandco.com F: 912 201 0402 W: http://www.islandco.com ------------------------------ Date: 13 Jul 2007 17:56:23 +0200 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOeGER) Subject: Re: Alphaserver ES45 Special Offer Message-ID: <4697bcc7$1@news.langstoeger.at> In article <139evemfde80vb5@news.supernews.com>, "David Turner, Island Computers" writes: >C.O.V. Special for July > >Only $18,995 > >This must be a better deal than a dual cpu Itanium >!!!!!!!!!!!!!!!!!!!!!?????????????????????!!!!!!!!!! Are you really sure? :-) I had an offer for an rx2600 (2x 1.3GHz, 4GB RAM, ...) Eur 1500.- (not 15k) (and missed it because of my vacation, so must hope for the next bargain ;-) Ok, there is Alpha S/W while the Itanic S/W is still rare, but a factor of 10 for only ~twice the CPU power (and the big minus electric power costs)? -EPLAN PS: there's a zx6000 on ebay item# 110148065528 for $1599.- (VMS support?) -- Peter "EPLAN" LANGSTOEGER Network and OpenVMS system specialist E-mail peter@langstoeger.at A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist ------------------------------ Date: Fri, 13 Jul 2007 12:07:01 -0400 From: "David Turner, Island Computers" Subject: Re: Alphaserver ES45 Special Offer Message-ID: <139f8orfe5iheb4@news.supernews.com> Right ... but is that realistic I got a quote from a distributor for over $20,000 without any thing but 2gb memory and 1 72gb disk No raid. Licenses were for a workstation I believe Etc etc "Peter 'EPLAN' LANGSTOeGER" wrote in message news:4697bcc7$1@news.langstoeger.at... > In article <139evemfde80vb5@news.supernews.com>, "David Turner, Island > Computers" writes: >>C.O.V. Special for July >> >>Only $18,995 >> >>This must be a better deal than a dual cpu Itanium >>!!!!!!!!!!!!!!!!!!!!!?????????????????????!!!!!!!!!! > > Are you really sure? :-) > > I had an offer for an rx2600 (2x 1.3GHz, 4GB RAM, ...) Eur 1500.- (not > 15k) > (and missed it because of my vacation, so must hope for the next bargain > ;-) > > Ok, there is Alpha S/W while the Itanic S/W is still rare, but a factor of > 10 for only ~twice the CPU power (and the big minus electric power costs)? > > -EPLAN > > PS: there's a zx6000 on ebay item# 110148065528 for $1599.- (VMS support?) > -- > Peter "EPLAN" LANGSTOEGER > Network and OpenVMS system specialist > E-mail peter@langstoeger.at > A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist ------------------------------ Date: Fri, 13 Jul 2007 13:06:57 -0400 From: Stephen Hoffman Subject: Re: Alphaserver ES45 Special Offer Message-ID: Peter 'EPLAN' LANGSTOeGER wrote: > PS: there's a zx6000 on ebay item# 110148065528 for $1599.- (VMS support?) Official support from HP for use with OpenVMS I64, no. But as discussed here regularly, it works -- do check what's in the PCI bus against what OpenVMS I64 recognizes, as usual. And you can buy a license for however many cores, and off you go, or you can use it as a hobbyist box. http://64.223.189.234/node/9 -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: 13 Jul 2007 19:29:10 +0200 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOeGER) Subject: Re: Alphaserver ES45 Special Offer Message-ID: <4697d286@news.langstoeger.at> In article <139f8orfe5iheb4@news.supernews.com>, "David Turner, Island Computers" writes: >Right ... but is that realistic It was (at least for me). Currently no Itanic offer here (not even on ebay). >I got a quote from a distributor for over $20,000 without any thing but 2gb >memory and 1 72gb disk Either they are more expensive in US than in Europe or I had a really good offer. Ok, it was a 2 year old used machine, and I assume, this your quote was for a new one... >No raid. "Mine" had had a RAID, too >Licenses were for a workstation I believe Ok. No licenses here. (but I don't need them, I'm a VMS hobbyist) And I bet this is a HUGE difference in price... Nevertheless, keep the heads up and have a nice weekend -- Peter "EPLAN" LANGSTOEGER Network and OpenVMS system specialist E-mail peter@langstoeger.at A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist ------------------------------ Date: Fri, 13 Jul 2007 15:45:38 +0000 From: "Main, Kerry" Subject: RE: Cobol Link Issue... Message-ID: > -----Original Message----- > From: Paul Raulerson [mailto:paul@raulersons.com] > Sent: July 13, 2007 8:20 AM > To: Info-VAX@Mvb.Saic.Com > Subject: IRT: Cobol Link Issue... > > I am not sure if the original note I sent posted or not; I don't > seem to > have a record of it so I may not have sent it. > > I posted a question yesterday about a link error with COBOL028 > under OpenVMS > Alpha 8.3. I got to playing last night and produced a listing with > a Map > and Machine_code in it, and found the issue. Or what I think the > issue is. > > It's a little embarrassing; when I installed the COBOL, it has > instructions > about which runtime to install for various versions of VMS- so I > selected > the one that appeared to be for the highest level VMS version. > There were > errors during the install complaining about not be able to replace > modules > that were newer than what was in the install kit. > > It really meant that. I tracked down the cussed module > giving me > trouble in a file (library) named sys$library:starlet.olb. Examing > this and > a few other libraries laying around led me, along with clue of > replacement > errors during install, to the idea that perhaps 8.x versions of VMS > *already > had the COBOL runtime installed*. > > I restored the system disk and went hunting, and by golly that > appears to be > true. I installed the compiler *but the not the runtime* and > everything runs > just perfect. > > Guess it was one of those things you just have to know about. Now > about the > Fortran runtime... :) > > Anyways, all is well that ends well! > > -Paul Good to hear the issue was resolved. Btw, not related to this issue, but for future reference, the following lin= ks are a good one to bookmark, as they are a listing of the latest OpenVMS = V8.3 public patches available for download off the web. Reference: ftp://ftp.itrc.hp.com/openvms_patches/alpha/V8.3/ALPHA_V83A_MASTER_ECO_LIST= txt (Alpha listing) ftp://ftp.itrc.hp.com/openvms_patches/alpha/V8.3/ (directory) ftp://ftp.itrc.hp.com/openvms_patches/i64/V8.3/I64_V83I_MASTER_ECO_LIST.txt= (Integrity) ftp://ftp.itrc.hp.com/openvms_patches/i64/V8.3/ (directory) Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Thu, 12 Jul 2007 23:24:48 GMT From: =?ISO-8859-1?Q?Jan-Erik_S=F6derholm?= Subject: Re: EDT Replacement Message-ID: <4Lyli.4007$ZA.1945@newsb.telia.net> Michael D. Ober wrote: > Our log files are too big for it. For small logfiles I use TYPE with a large scrollback buffer in my term-emulator. För large logfiles I use SEARCH with different combination of switches. I've selldom any need to load them into some text editor. Regards, Jan-Erik. ------------------------------ Date: 13 Jul 2007 07:48:10 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: EDT Replacement Message-ID: In article <1184283095.587900.37410@r34g2000hsd.googlegroups.com>, AEF writes: > Here is yet another advantage of EDT for me: I currently have 29 > MicroVAX systems on line, some in NYC, some in London, and 2 in Hong > Kong. Now, what's faster when I update my init file: copying the EDT > init file to 28 systems, two overseas, or copying an EVE section file > (1200 blocks IIRC) to the same systems?! > > (OK EVE experts -- if there's a better way, I'm all ears.) I have things in my TPU section that EDT can't do. I find copying both my TPU source and the section file once per new system not too painfull, but I generally copy it once and then use the system for years, or decades. After all, these are VMS systems, we don't replace them every couple minutes. ------------------------------ Date: Fri, 13 Jul 2007 19:33:16 +0200 From: Martin Krischik Subject: Re: EDT Replacement Message-ID: <5488295.oOCMlrJulC@linux1.krischik.com> Michael D. Ober wrote: > Is there a free or low cost replacement for EDT on VMS 8.3? Preferably > one that can be configured to use the same 10-key keypad (Gold) keys as > EDT. Well I use Vim [1] on VMS which works very well and is absolutely up to date. That's unlike most other VMS editors which have not seen an update let alone an upgrade in decades. As for your problem: 42Mb Log file - no problem at all - Vim will open anything (even binary file - got a hex dump converter as well). Unless you run out of virtual memory that is. And yes: The Key-Mapping can be completely replaced as fine tuned the way you need them. Martin [1] http://www.polarhome.com/vim/ -- mailto://krischik@users.sourceforge.net Ada programming at: http://ada.krischik.com ------------------------------ Date: Fri, 13 Jul 2007 19:43:16 +0200 From: Martin Krischik Subject: Re: EDT Replacement Message-ID: <2105112.ZxnBL7AjCN@linux1.krischik.com> Ron Johnson wrote: > And the lines are all less than 255 char. ROFL! How many years ago that this has been an issue on any editor I use. Ahh the last I remember is EPM on OS/2 but with a 2499 character limit. No the last editor with a 255 char limit must have been on my Atari 130 XE 8bit computer and that's is probably 15 years ago. If I ever needed poove that I did the right thing and never bothered to learn EDT that's it. Note that Vim [1] can even open binary files. One line with 30.000 chars - not problem at all. But it thru the build in hex-dumper and you can even edit the file (don't forget to un-dump before save). Martin [1] http://www.polarhome.com/vim/ -- mailto://krischik@users.sourceforge.net Ada programming at: http://ada.krischik.com ------------------------------ Date: Fri, 13 Jul 2007 13:02:03 -0400 From: Terrie Linker Subject: Experience building Perl modules on VMS Itanium? Message-ID: Anyone out there tried building Perl modules on OpenVMS Itanium? We need to build quite a few of them, and many have errors. Just wondered if this is something anyone has experience with. Terrie Linker ------------------------------ Date: Fri, 13 Jul 2007 13:42:47 -0400 From: Stephen Hoffman Subject: Re: Experience building Perl modules on VMS Itanium? Message-ID: Terrie Linker wrote: > Anyone out there tried building Perl modules on OpenVMS Itanium? > We need to build quite a few of them, and many have errors. > Just wondered if this is something anyone has experience with. "Many errors"? OpenVMS I64 version? Compiler versions? ECOs? Error message(s)? Module(s)? Command(s) used? My experience was that the Perl stuff typically "just built". You did need to have the Perl environment set up correctly, and be familiar with a few of the various Perl-isms. Check the VMS Perl mailing list and its archives? As for your question, might you post some details on your adventures into Perl on OpenVMS? -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: Fri, 13 Jul 2007 16:28:50 +0000 From: "Main, Kerry" Subject: RE: FTSO/FASTCopy replacement Message-ID: > -----Original Message----- > From: Pierre [mailto:pierre.bru@gmail.com] > Sent: July 12, 2007 9:55 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: FTSO/FASTCopy replacement > > On Jul 12, 3:51 pm, IanMiller wrote: > > On Jul 12, 2:42 pm, Pierre wrote: > > > > > hi, > > > > > I want restartable file copy from OpenVMS to OpenVMS, with > pre/post > > > actions on success/failure and queing system. restartable is > not the > > > most important feature. pre/post action and queuing are more > important > > > for me. > > > > > there once were FTSO and FASCopy. what product now can do that > ? > > > > > TIA, > > > Pierre. > > > > FTSV and FSTO are on Freeware > V7http://h71000.www7.hp.com/freeware/freeware70/ > > yes. we are currently using FTSO but we exprient some problems with > it > (for ex. sometimes copy jobq vanish) and would like some more > robust > and supported product. > > Pierre. Are you sure you are not running into some process quota issues ? Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: 13 Jul 2007 10:39:45 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <5fp33hF3clt1cU1@mid.individual.net> In article <4696CCFB.4162ADB9@spam.comcast.net>, David J Dachtera writes: > Bill Gunshannon wrote: >> >> In article <1184282066.484375.205310@k79g2000hse.googlegroups.com>, >> Doug Phillips writes: >> > >> > Since "super computing" can involve different needs depending upon the >> > application, VMS certainly is worth considering for any CS lab that >> > wishes to offer a broad spectrum of experience to the student. >> >> Sorry guy, that ship has sailed. > > I thought it was moth-balled in dry dock. Actually, it is most likley that it was burned with the other boats. bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: 13 Jul 2007 10:41:23 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <5fp36jF3clt1cU2@mid.individual.net> In article <1184288045.482997.134100@57g2000hsv.googlegroups.com>, Doug Phillips writes: > On Jul 12, 7:00 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: >> In article <1184282066.484375.205...@k79g2000hse.googlegroups.com>, >> Doug Phillips writes: >> >> >> >> > Since "super computing" can involve different needs depending upon the >> > application, VMS certainly is worth considering for any CS lab that >> > wishes to offer a broad spectrum of experience to the student. >> >> Sorry guy, that ship has sailed. >> > > That's *great* news! What with all the rumors of its sinking! Did you > get its itinerary by any chance? Something about The Bermuda Triangle or Sargaso Sea!! bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Fri, 13 Jul 2007 10:01:47 +0200 From: "P. Sture" Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: In article , Ron Johnson wrote: > On 07/12/07 17:10, Anton Shterenlikht wrote: > [snip] > > > > Some extra stability would be very welcome. I know of at least one case > > when a finite element code crashed after running for 4 months on > > a particular solid mechanics problem. This might, of course, have > > nothing to do with the OS, which I believe was some unix, but with > > the FE code, or with hardware, or whatever. > > I know this sounds like 20-20 hindsight, but why didn't the code > write out checkpoints so that it could be restarted without wasting > 4 months of effort? I'm not too sure that 20/20 hindsight is required here. I was putting checkpoints into batch jobs for restart purposes nearly 30 years ago. -- Paul Sture ------------------------------ Date: Fri, 13 Jul 2007 09:33:50 +0200 From: Joseph Huber Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: Anton Shterenlikht wrote: > Perhaps it's time I read the RMS manual, though I thought the use > of RMS is optional. In other words I was not going to use RMS in > my fortran programs. > But if doing Fortran I/O, You DO USE RMS of course. VMS Fortran I/O is not bypassing RMS by doing QIOs directly ! -- Joseph Huber - http://www.huber-joseph.de ------------------------------ Date: Thu, 12 Jul 2007 19:53:15 -0500 From: David J Dachtera Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <4696CCFB.4162ADB9@spam.comcast.net> Bill Gunshannon wrote: > > In article <1184282066.484375.205310@k79g2000hse.googlegroups.com>, > Doug Phillips writes: > > > > Since "super computing" can involve different needs depending upon the > > application, VMS certainly is worth considering for any CS lab that > > wishes to offer a broad spectrum of experience to the student. > > Sorry guy, that ship has sailed. I thought it was moth-balled in dry dock. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Thu, 12 Jul 2007 17:54:05 -0700 From: Doug Phillips Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <1184288045.482997.134100@57g2000hsv.googlegroups.com> On Jul 12, 7:00 pm, b...@cs.uofs.edu (Bill Gunshannon) wrote: > In article <1184282066.484375.205...@k79g2000hse.googlegroups.com>, > Doug Phillips writes: > > > > > Since "super computing" can involve different needs depending upon the > > application, VMS certainly is worth considering for any CS lab that > > wishes to offer a broad spectrum of experience to the student. > > Sorry guy, that ship has sailed. > That's *great* news! What with all the rumors of its sinking! Did you get its itinerary by any chance? ************ So as one door closes, another opens.;-) ------------------------------ Date: Fri, 13 Jul 2007 01:41:21 GMT From: "John E. Malmberg" Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <5LAli.9549$Xa3.6437@attbi_s22> JF Mezei wrote: > > Out of curiosity, if VMS writes to disks hosted by one of them fancy SAN > subsystems, does VMS really know when the write was absolutely written > to the disk platter ? No. The fancy raid subsystems have battery backed up RAM that may be holding the write for a few seconds. -John wb8tyw@qsl.network Personal Opinion Only ------------------------------ Date: Thu, 12 Jul 2007 20:53:47 -0500 From: Ron Johnson Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: On 07/12/07 17:10, Anton Shterenlikht wrote: [snip] > > Some extra stability would be very welcome. I know of at least one case > when a finite element code crashed after running for 4 months on > a particular solid mechanics problem. This might, of course, have > nothing to do with the OS, which I believe was some unix, but with > the FE code, or with hardware, or whatever. I know this sounds like 20-20 hindsight, but why didn't the code write out checkpoints so that it could be restarted without wasting 4 months of effort? -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Fri, 13 Jul 2007 02:10:21 GMT From: "John E. Malmberg" Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: David Mathog wrote: > Anton Shterenlikht wrote: > >> from AskHL: High-performance Interconnects, MPI? >> http://64.223.189.234/node/396 >> >> Posted July 3rd, 2007 by Hoff >> >> "How about the use of OpenVMS in High Performance computing? A recent >> question from a scholar on the possibility of using OpenVMS in this >> area forced me to give the rather embarrasing answer that you're >> pretty >> much on your own if you want to explore this on OpenVMS." >> >> Good! What could be better for a research project on parallel or >> distributed computing on VMS.. That means anything we can do in this area >> will be publishable. > > > Unless things have changed greatly in the last few releases of VMS, > which I doubt, then I addressed many of these issues years ago in a > series of posts about IO on VMS. In a nutshell, VMS is very much > concerned with data integrity, and so it goes to great lengths to see > that the data actually hits the disk. This causes an immense hit in > IO performance compared to the default situation on > Unix/Linux/Windows/(every other OS more or less) where the default is > to use a lot of memory caching and let the data hit the disk eventually, > whenever convenient, if then! Even if disk caching is enabled VMS > is still at a disadvantage because of RMS overhead, which all by itself > accounted for about a factor of 3 slow down of VMS compared to linux. > This was all carried out on DS10 boxes, google to find the posts. Things have changed. How much they would impact your tests, I do not know. Since you did your tests, a clusterwide filesystem cache was implemented, and as I understand it, the CRTL got some enhancements for doing I/O to stream files. I do not know the details of those enhancements. > So, the bottom line is, VMS is actually not very well suited for High > Performance computing for scientific and engineering applications, where > speed is everything, and data integrity is usually not that much of a > concern. Ie, if the system does happen to crash during a run the job > is just restarted, and nobody is out millions of dollars in lost > transactions. It would probably be possible to write some sort of > high performance subsystem for VMS, but it would need to bypass RMS, > and by the time you've done that, why bother using VMS? Now VMS > may have a place in "High Performance Transaction computing", > which would be some sort of distributed billing or other record keeping > system, but that seems to be pretty much where VMS clustering is right > now. RMS calls allow direct DMA access between the application and the controller to the storage device if the hardware supports it. VMS can drive high end disk controllers to the point where they will need to back off, and some of those controllers needed enhancements in order for them to handle the I/O load from OpenVMS. Using either asynchronous I/O via either QIO or RMS calls, you can queue up I/O for keeping the disk subsystem saturated. -John wb8tyw@qsl.network Personal Opinion Only ------------------------------ Date: 13 Jul 2007 07:51:43 -0500 From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: In article , Keith Parris writes: > Stephen Hoffman wrote: >> HPTC isn't a market where DEC, Compaq nor HP has classically aimed >> OpenVMS. > > VMS was a leader in HPTC early on -- IIRC, Parallel Virtual Machine > (PVM) was first developed on an LAVC. IIRC in the early days the VAXcluster was invented to enable a collection of systems to provide more compute power than any one VAX could muster, but appear in many ways similar to a single system. The terrific implementation led to VMScluster being so usefull in so many other ways, which have long since been the marketting target. ------------------------------ Date: Fri, 13 Jul 2007 13:20:55 +0000 From: "Main, Kerry" Subject: RE: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: > -----Original Message----- > From: David Mathog [mailto:mathog@caltech.edu] > Sent: July 12, 2007 11:52 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? > > Anton Shterenlikht wrote: > > from AskHL: High-performance Interconnects, MPI? > http://64.223.189.234/node/396 > > > > Posted July 3rd, 2007 by Hoff > > > > "How about the use of OpenVMS in High Performance computing? A > recent > > question from a scholar on the possibility of using OpenVMS in > this > > area forced me to give the rather embarrasing answer that > you're pretty > > much on your own if you want to explore this on OpenVMS." > > > > Good! What could be better for a research project on parallel or > > distributed computing on VMS.. That means anything we can do in > this area > > will be publishable. > > Unless things have changed greatly in the last few releases of VMS, > which I doubt, then I addressed many of these issues years ago in a > series of posts about IO on VMS. In a nutshell, VMS is very much > concerned with data integrity, and so it goes to great lengths to > see > that the data actually hits the disk. This causes an immense hit > in > IO performance compared to the default situation on > Unix/Linux/Windows/(every other OS more or less) where the default > is > to use a lot of memory caching and let the data hit the disk > eventually, > whenever convenient, if then! Even if disk caching is enabled VMS > is still at a disadvantage because of RMS overhead, which all by > itself > accounted for about a factor of 3 slow down of VMS compared to > linux. > This was all carried out on DS10 boxes, google to find the posts. > > So, the bottom line is, VMS is actually not very well suited for > High > Performance computing for scientific and engineering applications, > where > speed is everything, and data integrity is usually not that much of > a > concern. Ie, if the system does happen to crash during a run the > job > is just restarted, and nobody is out millions of dollars in lost > transactions. It would probably be possible to write some sort of > high performance subsystem for VMS, but it would need to bypass > RMS, > and by the time you've done that, why bother using VMS? Now VMS > may have a place in "High Performance Transaction computing", > which would be some sort of distributed billing or other record > keeping > system, but that seems to be pretty much where VMS clustering is > right > now. > > Regards, > > David Mathog > David, The write through OpenVMS default (data integrity) had an option added in t= he OpenVMS V7.3-1 or V7.3-2 timeframe that allowed you to set the default o= n OpenVMS to write back (speed not integrity) as is the default on most oth= er OS platforms. SYSGEN> help sys RMS_SEQFILE_WBH Its also a dynamic parameter, so testing the benefits should be easy to det= ermine. Just remember that data integrity is not your priority when you tur= n this feature on. Based on your comments, I suspect your testing is quite old predates this t= ime period. In addition, as you know, many additional (and significant) per= formance enhancements have been made since V7.3-2 days as well. Btw, everyone here should know that there are a number of different styles = of parallel computing, so one approach is not necessarily the best for all = parallel computing requirements. Note - one aspect that I am always amazed at when talking about high perfor= mance computing is all of the up front emphasis on cheap, cheap, cheap entr= y level platforms, but then when you ask those involved, they spend a huge,= huge amount of their staff time customizing their software to: 1. Try and keep all these parallel servers busy, 2. Manage all of the code (patching, updating etc) Yes, they all have some custom tools to assist with this, but in the end, d= epending on who you talk to, these are the 2 big areas that always seem to = emerge as big pain points. My point is that using common system disks (2 per cluster for availability)= with cluster generic queues that feed server queues automatically based on= priorities and a file system that uses logicals so that any job written ca= n run on any node in the cluster has to be of some interest to some areas o= f high performance computing. Note: while some have made mention of the fact that availability and securi= ty is not a typical concern, in some emerging areas of high performance com= puting it most definitely is. What is happening is a number of environments are consolidating and offerin= g their technical computing services to other departments and even other co= mpanies. There are big, big $'s in offering these services as lots of impor= tant and sensitive stuff is being done on them, so imho, there is an emergi= ng requirement for rock solid, ultra secure and easy to manage parallel tec= hnical computing. I thing some research in this area would definitely be interesting. Btw - overall licensing would likely be custom quote, so do not base pricin= g assumptions on list pricing. This is no different than existing parallel = technical environments - they are all heavily customized quotes. Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Fri, 13 Jul 2007 13:25:18 +0000 From: "Main, Kerry" Subject: RE: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: > -----Original Message----- > From: JF Mezei [mailto:jfmezei.spamnot@vaxination.ca] > Sent: July 12, 2007 9:18 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? > > Stephen Hoffman wrote: > >> File systems are a significant performance issue for HPTC on > OpenVMS as > > Mr Mathog has referenced; writing to disk is massively more slow > than > > the far faster in-memory lazy-write capabilities. > > > Out of curiosity, if VMS writes to disks hosted by one of them > fancy SAN > subsystems, does VMS really know when the write was absolutely > written > to the disk platter ? Or does the disk array system report a > succesfull > write once the request has been made and before the disk arrays has > had > time to actually physically write the data to whatever physicical > disk > drive onto which the disk is mapped ? > > IF VMS insists on getting the "real" write confirmation from the > disk > array, how difficult would it be to add some sort of flag in the > MOUNT > command to let VMS be happy with just the disk subsystem's > confirmation > of receipt of the request ? > > Wouldn't that solve the performance problem since the disk arrays > would > then report an almost instantaneous write confirmation to VMS ? > (with > the risk of the data not being written if there is a power failure > in > the meantime) From earlier thread, but repeated for emphasis: The write through OpenVMS default (data integrity) had an option added in t= he OpenVMS V7.3-1 or V7.3-2 timeframe that allowed you to set the default o= n OpenVMS to write back (speed not integrity) as is the default on most oth= er OS platforms. SYSGEN> help sys RMS_SEQFILE_WBH Its also a dynamic parameter, so testing the benefits should be easy to det= ermine. Just remember that data integrity is not your priority when you tur= n this feature on. Also, remember that in SAN's, there is disk controller write back/write thr= ough features as well. Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Fri, 13 Jul 2007 14:23:26 +0000 From: "Main, Kerry" Subject: RE: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: > -----Original Message----- > From: bill@triangle.cs.uofs.edu [mailto:bill@triangle.cs.uofs.edu] > On Behalf Of Bill Gunshannon > Sent: July 12, 2007 8:01 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? > > In article <1184282066.484375.205310@k79g2000hse.googlegroups.com>, > Doug Phillips writes: > > > > Since "super computing" can involve different needs depending > upon the > > application, VMS certainly is worth considering for any CS lab > that > > wishes to offer a broad spectrum of experience to the student. > > Sorry guy, that ship has sailed. > > bill > Well, based on your previous comments, OpenVMS handles all of the mission c= ritical business admin, scheduling and financial stuff in your University (= Banner), so other readers on this list need to understand that you are only= talking about one part of your specific environment. Other Universities are also using OpenVMS in various ways: http://wasd.vsm.com.au/other/uma_2tb.html http://h71000.www7.hp.com/news/ospp_turin.html And I regularly see University of.. or College of.. on participant .sig fil= es in this newsgroup and others like Multinet, Rdb etc. So, it should not be a big stretch to offer technical services on OpenVMS a= s long as it is positioned in the right way with modern wrappings e.g. UNIX= or English friendly style interface options, active-active clustering of s= ame workloads with common data and no Oracle RAC in the mix (cost), Common = file system whereby the same device names can be used cluster wide at the s= ame time, write back IO (for speed where data integrity not as big a priori= ty), cluster wide batch queues and overall ultra secure environments with c= luster wide auditing and security, Web based, Java or Windows based managem= ent GUI`s. All now available on OpenVMS. :-) And if one looks at the future of education, ultra high availability and se= curity is coming big time, because of all the remote courses being offered = all year around. As an example - My son is taking a computer animation course (Animationment= or.com). Totally online. It is a course being driven by instructors from Di= sney, Pixar etc as they could not get enough of the right types of skills t= hey needed. His classmates are from around North America, Europe and Asia P= acific. It is not uncommon where he has to be online for class discussions = and lectures at 9:00pm at night. It is a compressed 18 month course with 6 = x 3 month sessions and 1 week between sessions. Huge advantage though is th= at he can literally do this course anywhere there is a high speed Internet = connection. (He has moved several times and it does not impact his course). Imho, that is where Universities are also heading for many of their course = offerings (not all, of course) as they can not afford to have huge infrastr= uctures sitting idle for numerous months with no student activity generatin= g $`s. Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Fri, 13 Jul 2007 11:08:44 -0400 From: Stephen Hoffman Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: The speed seen with grids comes from not writing to the disk and from not having a fairly heavyweight open and close and such; the fastest I/O is one that isn't committed to disk. OpenVMS AFAIK still doesn't have anything similar in this regard -- open and close are heavyweight, and I/O is deliberately careful. Grids certainly do care about disk speed during some operations, though HPTC can be quite different than a traditional commercial environment in its implementation. There can be financial and commercial applications that fit onto grids; there isn't a clear application distinction within HPTC. SunOS and ULTRIX (VAX first, then MIPS) and other such were the early targets for HPTC, and there was a time when hardware accelerators ruled the industry; there was the vector unit on a few of the VAX boxes, and multiple vendors offered compute engines to make your VAX go faster. There was the Intel-based DECmpp (which itself had a console named "SRM"; go figure) box. Eons ago, ULTRIX MIPS became the go-fast processor strategy for HPTC at DIGITAL, given the VAX boxes were behind MIPS, SPARC, PA RISC and other designs then in the market. (At the time, SunOS SPARC was clobbering VAX -- VMS and ULTRIX -- in the market, and that's arguably where the VAXcluster scheme arose.) Around 25 years ago, there were also a number of vendors of hardware accelerators for VAX and other systems. The early SPARC boxes saw bolt-on acceleration, and there was some really weird stuff lurking out on SCSI buses. Now, the same sorts of hardware accelerator schemes are lurking within Cell processor. AMD Hypertransport compute accelerators are widely discussed; DRC Computer provides this sort of thing. But I digress... More recently, there was the AlphaServer SC series. And various Itanium grids have been installed and publicized, such as those based on the HP Cluster Platform 6000 boxes; on rack-n-stack rx-series boxes or blades. Kerry is correct here, grids can and do have multiple-use requirements; whether it's different groups using the grid, or using desktops as grid engines; Linux, Mac OS X and various commercial operating systems can implement SETI-like background processing. If the screensaver is up and showing on a desktop and the grid is configured, the box is probably chewing away on part of some problem on behalf of the grid. And vendors do have distributed authentication, security and tracking, as customers can and do want that. All discussions of I/O and authentication and such aside, grid capabilities are bolt-on or baked into various Linux distros, and are built into Mac OS X client and server. Oracle offers a product line targeting database applications. Microsoft offers a variant of Windows Server with its Compute Cluster Windows Server. HP-UX hptc/ClusterPack or HP Linux XC software and various of the Cluster Platform boxes (http://www.hp.com/go/hptc). With OpenVMS, you're porting stuff and/or rolling your own components to form a grid. I/O. Files. Communications. Management. Certainly none of which is infeasible, though in aggregate likely quite non-trivial. Or you do what most HPTC sites do, and load and run a grid built for whatever platform you're using: Linux, Windows, Mac OS X, Oracle, etc. -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: 13 Jul 2007 16:12:46 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: <5fpmjtF3e2hc1U1@mid.individual.net> In article , "Main, Kerry" writes: >> -----Original Message----- >> From: bill@cs.uofs.edu [mailto:bill@cs.uofs.edu] >> On Behalf Of Bill Gunshannon >> Sent: July 12, 2007 8:01 PM >> To: Info-VAX@Mvb.Saic.Com >> Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? >> >> In article <1184282066.484375.205310@k79g2000hse.googlegroups.com>, >> Doug Phillips writes: >> > >> > Since "super computing" can involve different needs depending >> upon the >> > application, VMS certainly is worth considering for any CS lab >> that >> > wishes to offer a broad spectrum of experience to the student. >> >> Sorry guy, that ship has sailed. >> >> bill >> > > > Well, based on your previous comments, OpenVMS handles all of the mission c= > ritical business admin, scheduling and financial stuff in your University (= > Banner), so other readers on this list need to understand that you are only= > talking about one part of your specific environment. And what was the original poster talking about? "VMS certainly is worth considering for any CS lab that wishes to offer a broad spectrum of experience to the student" Doesn't look like administrative use to me. And it was the academic that I was talking about when I I said "that ship has sailed". Name one school today where VMS is the primary (or even secondary) computer system for teaching computing in the CS department? > > Other Universities are also using OpenVMS in various ways: > http://wasd.vsm.com.au/other/uma_2tb.html Looks like they use VMS for their webserver. Doesn't say anything about students actually learning about VMS in the CS department. Or even having access to a VSM machine beyond the web interface which could be anything without the user knowing or caring. Now, Spanish is not one of my stronger languages but unless they hide it under Mathematics, Electronics or Engineering I could not see where they even have a CS Department. Oh wait, they have an English language web page (well, sort of.) A bit of computer engineering but no Computer Science or even Computer Information Systems. Sorry, Kerry, but looks like they have no academic VMS either. > http://h71000.www7.hp.com/news/ospp_turin.html That's HP talking about something that Politecnico di Torino doesn't think is important enough to even have on their web page. No real CS department, but they have a Computer Engineering departmetn which seems to stress hardware design rather than software or or pure CS. No mention of this project there either. I would guess they had about as much luck getting their project going as I did here. So, how long does something have to be dead before HP stops touting it as a success? > > And I regularly see University of.. or College of.. on participant .sig fil= > es in this newsgroup and others like Multinet, Rdb etc. Maybe like all the people who have responded to Dadid Turner we could do an unofficial poll of c.o.v users. How many people here with .edu in their email address are at schools that still use VMS as their primary (Secondary? Tertiary?) computing systems for teaching Computer Science or even Computer Inofrmatuion Systems? <.......> >:-) > > And if one looks at the future of education, ultra high availability and se= > curity is coming big time, because of all the remote courses being offered = > all year around. Again, you are talking about administrative use and I (and Doug) were talking about academic use. of course, without the academic use less and less people will even know what VMS is and we can already see the results of that. > > As an example - My son is taking a computer animation course (Animationment= > or.com). Totally online. It is a course being driven by instructors from Di= > sney, Pixar etc as they could not get enough of the right types of skills t= > hey needed. His classmates are from around North America, Europe and Asia P= > acific. It is not uncommon where he has to be online for class discussions = > and lectures at 9:00pm at night. It is a compressed 18 month course with 6 = > x 3 month sessions and 1 week between sessions. Huge advantage though is th= > at he can literally do this course anywhere there is a high speed Internet = > connection. (He has moved several times and it does not impact his course). And how much has he learned about VMS so far? Does he even know what the underlying technology is (and I would bet it ain't VMS). What's more, does he care. > > Imho, that is where Universities are also heading for many of their course = > offerings (not all, of course) as they can not afford to have huge infrastr= > uctures sitting idle for numerous months with no student activity generatin= > g $`s. None of which requires VMS and being as none of the people involved have ever heard of VMS it is likely going to be a hard sell. How are you going to sell VMS to a Professor teaching animation when you don't even have graphics capability? Or are you announcing the return of VMS to the desktop workstation market? While everything you said may be true, it has nothing to do with what the original poster said or my response to it. HP has all but abandoned the academic world and that means the next generation of CIO's and IT managers will have never even heard of VMS. They will, however, know what Windows is and that it worked just fine when they were in college. Which one do you think they will bet the farm on? bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Fri, 13 Jul 2007 17:40:59 +0000 From: "Main, Kerry" Subject: RE: HoffmanLabs: High-performance Interconnects, MPI? Message-ID: > -----Original Message----- > From: bill@triangle.cs.uofs.edu [mailto:bill@triangle.cs.uofs.edu] > On Behalf Of Bill Gunshannon > Sent: July 13, 2007 12:13 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? > > In article > p.net>, > "Main, Kerry" writes: > >> -----Original Message----- > >> From: bill@cs.uofs.edu [mailto:bill@cs.uofs.edu] > >> On Behalf Of Bill Gunshannon > >> Sent: July 12, 2007 8:01 PM > >> To: Info-VAX@Mvb.Saic.Com > >> Subject: Re: HoffmanLabs: High-performance Interconnects, MPI? > >> > >> In article > <1184282066.484375.205310@k79g2000hse.googlegroups.com>, > >> Doug Phillips writes: > >> > > >> > Since "super computing" can involve different needs depending > >> upon the > >> > application, VMS certainly is worth considering for any CS lab > >> that > >> > wishes to offer a broad spectrum of experience to the student. > >> > >> Sorry guy, that ship has sailed. > >> > >> bill > >> > > > > > > Well, based on your previous comments, OpenVMS handles all of the > mission c=3D > > ritical business admin, scheduling and financial stuff in your > University (=3D > > Banner), so other readers on this list need to understand that > you are only=3D > > talking about one part of your specific environment. > > And what was the original poster talking about? > I just wanted to ensure that readers understand that OpenVMS plays a big pa= rt in your University. You only support the technical side, so the Admin si= de is not important to you as it is likely run by another group. And by the way, there is likely very few Universities that have a single OS= platform as their primary technical platform on the server side. It is ver= y much College specific as each College within the University tends to do w= hatever they want. In the US University (20K+ students) where I just completed a server optimi= zation and DC planning engagement, for their tech platform of choice they u= sed multiple versions of MAC, Linux, Solaris, Windows, Fedora, NetWare and = even a few instances of Irix. Yes, no OpenVMS. Bottom line is that the technical environment in most Universities is extre= mely fragmented based on the College and-or Dept. you are talking to. That = was one of the challenges I was asked to address i.e. how to bring all of t= hese various environments back under some type of centralized control. > "VMS certainly is worth considering for any CS lab that > wishes to offer a broad spectrum of experience to the student" > > Doesn't look like administrative use to me. And it was the > academic > that I was talking about when I I said "that ship has sailed". > Just wanted to emphasize that as my note stated. > Name one school today where VMS is the primary (or even secondary) > computer system for teaching computing in the CS department? > I never stated it was a primary platform. It may be, but I do not have acce= ss to Cust data. Course, see above about the hugely fractured technical environments at most= Universities. It usually depends on the College or Dept. > > > > Other Universities are also using OpenVMS in various ways: > > http://wasd.vsm.com.au/other/uma_2tb.html > > Looks like they use VMS for their webserver. Doesn't say anything > about students actually learning about VMS in the CS department. > Or even having access to a VSM machine beyond the web interface > which could be anything without the user knowing or caring. > > Now, Spanish is not one of my stronger languages but unless they > hide > it under Mathematics, Electronics or Engineering I could not see > where > they even have a CS Department. > > Oh wait, they have an English language web page (well, sort of.) > A bit of computer engineering but no Computer Science or even > Computer Information Systems. Sorry, Kerry, but looks like they > have no academic VMS either. > You may be right, but since you have zero knowledge of the University, that= is pure FUD based on your UNIX OS religion and preferences. > > http://h71000.www7.hp.com/news/ospp_turin.html > > That's HP talking about something that Politecnico di Torino > doesn't > think is important enough to even have on their web page. No real > CS department, but they have a Computer Engineering departmetn > which > seems to stress hardware design rather than software or or pure > CS. > No mention of this project there either. I would guess they had > about > as much luck getting their project going as I did here. > > So, how long does something have to be dead before HP stops touting > it > as a success? > Give it a rest. You have next to zero interest in hearing anything but that= UNIX is the answer for all University technical computing requirements. > > > > And I regularly see University of.. or College of.. on > participant .sig fil=3D > > es in this newsgroup and others like Multinet, Rdb etc. > > Maybe like all the people who have responded to Dadid Turner we > could do > an unofficial poll of c.o.v users. How many people here with .edu > in > their email address are at schools that still use VMS as their > primary > (Secondary? Tertiary?) computing systems for teaching Computer > Science > or even Computer Inofrmatuion Systems? > As I stated earlier, c.o.v. is only a small, micro subset of Custs who use = OpenVMS. Most of the Rdb and Multinet newsgroup users do not even participate on com= p.os.vms. > <.......> > > >:-) > > > > And if one looks at the future of education, ultra high > availability and se=3D > > curity is coming big time, because of all the remote courses > being offered =3D > > all year around. > > Again, you are talking about administrative use and I (and Doug) > were > talking about academic use. of course, without the academic use > less > and less people will even know what VMS is and we can already see > the > results of that. > > > > > As an example - My son is taking a computer animation course > (Animationment=3D > > or.com). Totally online. It is a course being driven by > instructors from Di=3D > > sney, Pixar etc as they could not get enough of the right types > of skills t=3D > > hey needed. His classmates are from around North America, Europe > and Asia P=3D > > acific. It is not uncommon where he has to be online for class > discussions =3D > > and lectures at 9:00pm at night. It is a compressed 18 month > course with 6 =3D > > x 3 month sessions and 1 week between sessions. Huge advantage > though is th=3D > > at he can literally do this course anywhere there is a high speed > Internet =3D > > connection. (He has moved several times and it does not impact > his course). > > And how much has he learned about VMS so far? Does he even know > what the > underlying technology is (and I would bet it ain't VMS). What's > more, does > he care. > Again, you are missing the point. > > > > Imho, that is where Universities are also heading for many of > their course =3D > > offerings (not all, of course) as they can not afford to have > huge infrastr=3D > > uctures sitting idle for numerous months with no student activity > generatin=3D > > g $`s. > > None of which requires VMS and being as none of the people involved > have > ever heard of VMS it is likely going to be a hard sell. How are > you going > to sell VMS to a Professor teaching animation when you don't even > have > graphics capability? Or are you announcing the return of VMS to > the > desktop workstation market? > > While everything you said may be true, it has nothing to do with > what > the original poster said or my response to it. HP has all but > abandoned > the academic world and that means the next generation of CIO's and > IT managers will have never even heard of VMS. They will, however, > know > what Windows is and that it worked just fine when they were in > college. > Which one do you think they will bet the farm on? > I have never stated better marketing is not an issue (the same could be sta= ted for all enterprise OS platforms, as opposed to departmental), but the t= rend in computing requirements from a senior BU perspective is delivering I= T as ultra high HA, ultra high security services on radically fewer server = OS instances and they do not care what the HW/OS platform is. Now you can continue to promote Windows/Linux servers with their 5-20 secur= ity patches per month and one-app, one server cultures, as good back end ce= ntralized computing platforms, but I really have to disagree. Course, feel free to argue that they are .. :-) Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Fri, 13 Jul 2007 10:27:31 +0200 From: "P. Sture" Subject: Re: How many people here use Itanium w VMS Message-ID: In article <4696CCCD.4854827A@spam.comcast.net>, David J Dachtera wrote: > "David Turner, Island Computers" wrote: > > > > sad > > they might learn something if they hung out in cov once in a while > > > > ;0) > > True, but how much of it would be VMS and how much would be debating skills? And how much time devoted to finding a newsreader with decent killfile capabilities? Sorry, could not resist. :-) -- Paul Sture ------------------------------ Date: Fri, 13 Jul 2007 00:07:53 GMT From: John Santos Subject: Re: How many people here use Itanium w VMS Message-ID: John Santos wrote: > Rich Jordan wrote: > >> On Jul 12, 7:40 am, "David Turner, Island Computers" > islandco.com> wrote: >> >>> Please .... if you have an Itanium running VMS or soon will have in a >>> commerical environment we would love to hear from you. >>> >>> We are not trying to accrue contact names etc, just trying to get an >>> idea of >>> how many users are transitioning or will transition to Itanium VMS >>> >>> -- >>> David B Turner >>> Island Computers US Corp >>> 2700 Gregory St, Suite 180 >>> Savannah GA 31404 >>> >>> T: 877-6364332 x201 >>> Intl: 001 912 447 6622 >>> >>> E: dtur...@islandco.com >>> F: 912 201 0402 >>> W:http://www.islandco.com >>> >>> -- >>> David B Turner >>> Island Computers US Corp >>> 2700 Gregory St, Suite 180 >>> Savannah GA 31404 >>> >>> T: 877-6364332 x201 >>> Intl: 001 912 447 6622 >>> >>> E: dtur...@islandco.com >>> F: 912 201 0402 >>> W:http://www.islandco.com >> >> >> >> >> One in use here, but just starting. We have two customers looking at >> moving on from Alpha so we needed it here to be ready for them. We >> also want to try some of the newer tools (like netbeans) which our >> Alpha is too strained to support well. >> >> We still need to decide if we're going to leave it standalone and keep >> the Alphaserver and the MicroVAX clustered, or if we de-cluster the >> VAX and cluster the itanium. >> > > Similar situation, except our intended customer postponed a decision > until the next fiscal year, but in the meanwhile we ported the app > over and are using it for other stuff... It's a development system, Ambiguous "it": "... we are using the Itanium (rx2620) for other stuff" > not a production system, and we just clustered it in with our Alphas > and single remaining VAX. It may not be officially supported in > a cluster with VAXes, but it works fine. (Some issues with MON CLUS > and the VAX can't see all the files on ODS-5 disks, but those are > really V8.x vs. V7.3 issues, rather than Itanium vs. VAX issues.) > -- John Santos Evans Griffiths & Hart, Inc. 781-861-0670 ext 539 ------------------------------ Date: Thu, 12 Jul 2007 19:52:29 -0500 From: David J Dachtera Subject: Re: How many people here use Itanium w VMS Message-ID: <4696CCCD.4854827A@spam.comcast.net> "David Turner, Island Computers" wrote: > > sad > they might learn something if they hung out in cov once in a while > > ;0) True, but how much of it would be VMS and how much would be debating skills? -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: Fri, 13 Jul 2007 08:49:53 -0700 From: Galen Subject: Re: How many people here use Itanium w VMS Message-ID: <1184341793.612813.302440@k79g2000hse.googlegroups.com> Dave, Although you specified "commercial environment" I thought I'd respond anyway, even though we don't fit that description: Our customer is a US Government agency. We are transitioning one of our customer's critical applications (developed and maintained by us at Booz Allen) from VMS Alpha on roughly 20 standalone DS10 workstations to VMS on roughly 20 clustered rx1620 pairs of Integrity servers. The code has moved over pretty easily. There were the usual problems with build procedures not recognizing the I64 (vs. Alpha or Vax) and a few minor problems with the application code itself. One of the harder ones to find and figure out was a dependency on undocumented, unspecified behavior when executing an integer shift by more than 32 bits in HP C. The Alpha version's unspecified behavior is not the same unspecified behavior we observed on the I64. :-) ------------------------------ Date: 12 Jul 2007 22:21:03 GMT From: healyzh@aracnet.com Subject: Re: Install VMS on a Alphastation 200 ? Message-ID: Niels S. Eliasen wrote: > I have just aquired an Alphastation 200/233 .. and would like to install > VMS on it... (Just for old time's sake....) What are the gotcha's in doing > this? Off the top of my head you need the SRM console rather than the ARC console, and you need to check and see if the video card is supported. I've run these in the past both as a workstation and headless with a VT420. In spite of what the doc's say, they'll run with 128MB SIMM's if you can get them, which makes them quite nice. The downsides are that the built-in SCSI is Narrow SCSI, and the built-in ethernet is only 10Mbit. I ran one with a DE500 for a couple years, and again briefly with a DE500 (100Mbit) and a KZPBA (UW-SCSI) when I was having issues with my PWS433au. Personally they're one of my favorite systems, and if I had room I'd consider having one clustered in with a 24-bit card and a 100Mbit NIC. They might not be that fast, but they're a nice size, and are rock solid. As a workstation the biggest problem I can see is that you can't run Mozilla or Netbeans on one. Zane ------------------------------ Date: 12 Jul 2007 22:21:47 GMT From: "Niels S. Eliasen" Subject: Re: Install VMS on a Alphastation 200 ? Message-ID: On 2007-07-12, Bob Gezelter wrote: > On Jul 12, 4:58 pm, "Niels S. Eliasen" konsult.com> wrote: >> Hi >> I have just aquired an Alphastation 200/233 .. and would like to install VMS on it... >> (Just for old time's sake....) >> What are the gotcha's in doing this? >> >> -- >> kind regards/mvh >> Niels > > Niels, > > Really, there are no obstacles. I would recommend verifying that you > are running the SRM console, and I would verify that the console > firmware is the last one issued. ok ....two questions... 1. I akways dread dealing with firmware.... how do I make certain I do not fall into some bear-trap with firmware that needs to be upgraded and then won't .. ?? 2. Perhaps a lame question.... How do I check whether I use the SRM console ? > > Then boot from the CDROM and away you go. > > On the hardware end, I would recommend investigating upgrading the > memory capacity to the maximum supported, particularly if you are > planning on playing with a Java or other related tools. > > A second disk is also a good idea. > > In any event, the AlphaStation 200 series is a good workhorse system. > It does not make too much noise, burn too much power, or take up too > much space. It is also generally compatible with many of the KVM > switches made for PC systems, so it is possible to avoid having more > monitors, keyboards, and mice than necessary. > > - Bob Gezelter, http://www.rlgsc.com > -- kind regards/mvh Niels ------------------------------ Date: Thu, 12 Jul 2007 18:55:44 -0400 From: "Richard B. Gilbert" Subject: Re: Install VMS on a Alphastation 200 ? Message-ID: <4696B170.4050806@comcast.net> Niels S. Eliasen wrote: > Hi > I have just aquired an Alphastation 200/233 .. and would like to install VMS on it... > (Just for old time's sake....) > What are the gotcha's in doing this? > > > Gotcha's? What gotcha's? Assuming that you have the VMS CD and the Alphastation has a CDROM drive, you put the CD in the drive, boot from it, and away you go. There ARE instructions for doing this somewhere; something like an "Installation and Upgrade" manual for whatever release you are dealing with. It has been a few years since I've done this so forgive me for being a little foggy on the details. My Alphastation 200 4/233 is still ticking away. . . . I started at 6.2-1H3 and subsequently upgraded to V7.2-1 as practice for doing the same upgrade at work. Having your very own test platform is nice! If you need the manual, try googling for "VMS Installation" site:HP.COM. HP's search engine is hopeless!!! You will probably want the manual for the version you are trying to install, though just about any of them will give you a clue. You will probably want the "Layered Product" kit or whatever it's called; you'll need it to install TCP/IP Services and anything else you are licensed for. ------------------------------ Date: Fri, 13 Jul 2007 16:39:26 +0000 (UTC) From: moroney@world.std.spaamtrap.com (Michael Moroney) Subject: Re: Install VMS on a Alphastation 200 ? Message-ID: Chris Scheers writes: >Niels S. Eliasen wrote: >> Hi >> I have just aquired an Alphastation 200/233 .. and would like to install VMS on it... >> (Just for old time's sake....) >> What are the gotcha's in doing this? Beware that the power supplies on these frequently fail, and replacements are hard to find or expensive. Other than that, they're pretty good. >A good number of AlphaStation 200s were sold with "half-flash". These >only have enough memory to hold ARC or SRM, but not both. If you have a >half-flash machine with ARC, it can be changed to SRM, but it is a bit >tricky. >To switch a half-flash machine, you need to do a firmware upgrade, even >if you are at the latest version. Check the firmware upgrade directions >for the exact incantation necessary. >FWIW: All of the half-flash machines I know of came from Onsale a while >back. They were part (all?) of a special order of machines that never >got deployed. There are two empty PLCC sockets. They can be converted >to full flash machines by putting the correct chips in these sockets. I >don't remember the chip number anymore, but it is the same as the other >two nearby flash chips. The AS 200 was "value engineered" to reduce cost after it was first released. Part of the "value engineering" was to eliminate those two chips, since it was realized that 99.99% of those systems, once configured to use a certain console, would never ever be reconfigured to use the other. In other words, the earlier systems will be full-flash, and most if not all the later ones will be half-flash, and half-flash systems are much more common than full-flash ones. (I did qualification work for DEC for the AS200) >Likewise, my AlphaStation 255 is half-flash. Can that be converted to >full flash? I don't know about the 255, but do you really need it? Do you run ancient versions of WNT or early fieldtests of Windows 2000? ------------------------------ Date: Fri, 13 Jul 2007 07:20:09 -0500 From: "Paul Raulerson" Subject: IRT: Cobol Link Issue... Message-ID: <000f01c7c548$28117b60$78347220$@com> I am not sure if the original note I sent posted or not; I don't seem to have a record of it so I may not have sent it. I posted a question yesterday about a link error with COBOL028 under OpenVMS Alpha 8.3. I got to playing last night and produced a listing with a Map and Machine_code in it, and found the issue. Or what I think the issue is. It's a little embarrassing; when I installed the COBOL, it has instructions about which runtime to install for various versions of VMS- so I selected the one that appeared to be for the highest level VMS version. There were errors during the install complaining about not be able to replace modules that were newer than what was in the install kit. It really meant that. I tracked down the cussed module giving me trouble in a file (library) named sys$library:starlet.olb. Examing this and a few other libraries laying around led me, along with clue of replacement errors during install, to the idea that perhaps 8.x versions of VMS *already had the COBOL runtime installed*. I restored the system disk and went hunting, and by golly that appears to be true. I installed the compiler *but the not the runtime* and everything runs just perfect. Guess it was one of those things you just have to know about. Now about the Fortran runtime... :) Anyways, all is well that ends well! -Paul ------------------------------ Date: Fri, 13 Jul 2007 13:38:43 GMT From: Robert Deininger Subject: Re: IRT: Cobol Link Issue... Message-ID: In article <000f01c7c548$28117b60$78347220$@com>, "Paul Raulerson" wrote: > I am not sure if the original note I sent posted or not; I don't seem to > have a record of it so I may not have sent it. > > I posted a question yesterday about a link error with COBOL028 under OpenVMS > Alpha 8.3. I got to playing last night and produced a listing with a Map > and Machine_code in it, and found the issue. Or what I think the issue is. > > It's a little embarrassing; when I installed the COBOL, it has instructions > about which runtime to install for various versions of VMS- so I selected > the one that appeared to be for the highest level VMS version. There were > errors during the install complaining about not be able to replace modules > that were newer than what was in the install kit. > > It really meant that. I tracked down the cussed module giving me > trouble in a file (library) named sys$library:starlet.olb. Examing this and > a few other libraries laying around led me, along with clue of replacement > errors during install, to the idea that perhaps 8.x versions of VMS *already > had the COBOL runtime installed*. > > I restored the system disk and went hunting, and by golly that appears to be > true. I installed the compiler *but the not the runtime* and everything runs > just perfect. > > Guess it was one of those things you just have to know about. Now about the > Fortran runtime... :) VMS generally includes the run-time libraries for the traditional compiled languages. Sometimes there's a language run-time library kit to help you use NEW languages features on OLD VMS versions. Are you reading the release notes and installation instructions for these products before you jump in? -- Robert ------------------------------ Date: Fri, 13 Jul 2007 10:20:25 -0400 From: John Reagan Subject: Re: IRT: Cobol Link Issue... Message-ID: Robert Deininger wrote: > In article <000f01c7c548$28117b60$78347220$@com>, > "Paul Raulerson" wrote: > >>It's a little embarrassing; when I installed the COBOL, it has instructions >>about which runtime to install for various versions of VMS- so I selected >>the one that appeared to be for the highest level VMS version. There were >>errors during the install complaining about not be able to replace modules >>that were newer than what was in the install kit. >> > > Are you reading the release notes and installation instructions for > these products before you jump in? > > -- Robert Well, to give Paul a break, the COBOL V2.8 release notes were written before V8.3 existed. Little did it know that the COBOL RTL included with V8.3 was just fine and NO RTL from the kit needed to be installed. Even I could get confused with the release notes. COBOL V2.9 (coming on the next quarterly CDROMs) has better release notes. -- John Reagan OpenVMS Pascal/Macro-32/COBOL Project Leader Hewlett-Packard Company ------------------------------ Date: Fri, 13 Jul 2007 10:30:27 +0200 From: "P. Sture" Subject: Re: Is VMS losing the Financial Sector, also? Message-ID: In article <3phzoYozE$rD@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) wrote: > In article , Rob > Brown writes: > > On Wed, 11 Jul 2007, Larry Kilgallen wrote: > > > >> In article , Rob > >> Brown writes: > >>> On Wed, 11 Jul 2007, Larry Kilgallen wrote: > >>> > >>>> In article , Rob > >>>> Brown writes: > >> > >>>>> I don't see any protection against that if the potential victim > >>>>> does not avoid the trap. > >>>> > >>>> That is what SET TERMINAL/PERMANENT/SECURE_SERVER is about. > >>> > >>> That merely give the potential victim a means to avoid the trap, but > >>> does not enforce it. > >> > >> So long as you don't have /AUTOBAUD, BREAK is the only way for the > >> user to get a Username prompt. > > > > No it's not. My password grabber presents a Username prompt. > > Ok, but they will not have any experience getting a _valid_ username > prompt without hitting break, so your scenario will not present the > normal experience. That would have caught me out many years ago, as I was using some third party terminals which didn't have a break key. -- Paul Sture ------------------------------ Date: 12 Jul 2007 21:56:38 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: Is VMS losing the Financial Sector, also? Message-ID: <3phzoYozE$rD@eisner.encompasserve.org> In article , Rob Brown writes: > On Wed, 11 Jul 2007, Larry Kilgallen wrote: > >> In article , Rob Brown writes: >>> On Wed, 11 Jul 2007, Larry Kilgallen wrote: >>> >>>> In article , Rob Brown writes: >> >>>>> I don't see any protection against that if the potential victim >>>>> does not avoid the trap. >>>> >>>> That is what SET TERMINAL/PERMANENT/SECURE_SERVER is about. >>> >>> That merely give the potential victim a means to avoid the trap, but >>> does not enforce it. >> >> So long as you don't have /AUTOBAUD, BREAK is the only way for the >> user to get a Username prompt. > > No it's not. My password grabber presents a Username prompt. Ok, but they will not have any experience getting a _valid_ username prompt without hitting break, so your scenario will not present the normal experience. ------------------------------ Date: 13 Jul 2007 08:55:10 -0500 From: Kilgallen@SpamCop.net (Larry Kilgallen) Subject: Re: Is VMS losing the Financial Sector, also? Message-ID: In article , "P. Sture" writes: > In article <3phzoYozE$rD@eisner.encompasserve.org>, > Kilgallen@SpamCop.net (Larry Kilgallen) wrote: >> Ok, but they will not have any experience getting a _valid_ username >> prompt without hitting break, so your scenario will not present the >> normal experience. > > That would have caught me out many years ago, as I was using some third > party terminals which didn't have a break key. Any shop planning to make their VMS system more secure must consider things like that. In some environments there is a combination like SHIFT/SPACE that create a Break signal. Note that the specifications for Break are pretty wide, which is why you lose autobaud capability. ------------------------------ Date: Fri, 13 Jul 2007 10:49:30 -0600 From: Mark Berryman Subject: Re: Is VMS losing the Financial Sector, also? Message-ID: <46974aac@mvb.saic.com> Larry Kilgallen wrote: > In article , "P. Sture" writes: >> In article <3phzoYozE$rD@eisner.encompasserve.org>, >> Kilgallen@SpamCop.net (Larry Kilgallen) wrote: > >>> Ok, but they will not have any experience getting a _valid_ username >>> prompt without hitting break, so your scenario will not present the >>> normal experience. >> That would have caught me out many years ago, as I was using some third >> party terminals which didn't have a break key. > > Any shop planning to make their VMS system more secure must consider > things like that. In some environments there is a combination like > SHIFT/SPACE that create a Break signal. Note that the specifications > for Break are pretty wide, which is why you lose autobaud capability. On current versions of VMS, if you mix AUTOBAUD and SECURE_SERVER, pressing the break key is simply guaranteed to terminate any existing process, it does not start the login procedure. One can then use the normal autobaud procedure to set the terminal speed and get a login prompt. I do not know how long this has been the case. In my case I never had to worry about this. The only place we would have used autobaud was on modem lines and we always used modems that could present a fixed speed to the computer regardless of what speed was negotiated across the telephone line (either that or modems configured to only negotiate one specific speed - I remember the days of using one phone number for 1200 baud and a different one for 2400 baud). Mark Berryman ------------------------------ Date: Fri, 13 Jul 2007 16:05:47 +0000 From: "Main, Kerry" Subject: RE: Linux (was Re: How many people here use Itanium w VMS) Message-ID: > -----Original Message----- > From: Ron Johnson [mailto:ron.l.johnson@cox.net] > Sent: July 12, 2007 2:28 PM > To: Info-VAX@Mvb.Saic.Com > Subject: Linux (was Re: How many people here use Itanium w VMS) > > On 07/12/07 08:29, Jeffrey H. Coffield wrote: > > David Turner, Island Computers wrote: > >> Please .... if you have an Itanium running VMS or soon will have > in a > >> commerical environment we would love to hear from you. > >> > >> We are not trying to accrue contact names etc, just trying to > get an > >> idea of how many users are transitioning or will transition to > Itanium > >> VMS > >> > >> > > So far we have 6 running OpenVMS and 3 running SUSE Linux. I > expect the > > rest of the production systems (3) to be gone by year end. > > Why run Linux on Itanium? > > Any particular affirmative reason, or just because you have the > boxes and didn't want to waste assets? > > -- > Ron Johnson, Jr. > Jefferson LA USA > > Give a man a fish, and he eats for a day. > Hit him with a fish, and he goes away for good! Well, from a developer or ISV perspective, assuming low cost SANs are in pl= ace, you can develop and support your 64bit application on Windows, Linux, = OpenVMS, HP-UX - all using a single server. Simply redirect the console to = boot off a different SAN disk. That is a pretty good story from an ISV perspective that supports multiple = target platforms. And remember that many Customers buy solutions or services and do not care = about the OS and platform religion issues often discussed here. I would als= o suggest that the numbers of users (and companies for that matter) who do = not care what the back end is as long as it is always available and ultra s= ecure is going to exponentially increase in the future. For those interested in future directions, I highly recommend looking at Sa= aS (Software as a Solution) happenings. Simply google SaaS .. Reference: http://en.wikipedia.org/wiki/Software_as_a_Service Regards Kerry Main Senior Consultant HP Services Canada Voice: 613-592-4660 Fax: 613-591-4477 kerryDOTmainAThpDOTcom (remove the DOT's and AT) OpenVMS - the secure, multi-site OS that just works. ------------------------------ Date: Fri, 13 Jul 2007 13:53:57 +0800 From: "Richard Maher" Subject: Now that's what I'm talking about! (was:Re: Is VMS losing the Financial Sector, Message-ID: Hi Ken, > > If only there was some middleware out there that let you do this. . . > > For someone who does serious programming like yourself, it > seems like less than an afternoon's worth of design and coding. :-) Either my marketing attempts to date have been about as successful as Digital's, or my legendary subtlety has somehow managed to obscure the inherent sarcasm and facetiousness in my lament :-) Yes, there definitely is "middleware out there that [will] let you do this" and that middleware is called Tier3. I was rather hoping that you'd heard of our Client/Server middleware offering, and that my post would once more re-focus attention on exactly why Tier3 and its *connection-oriented*, *context-rich*, execution environment is so good; but no matter. > And it would be very easy to write a program to do that on > VMS Depends what you mean by "very easy" I suppose. Here's a couple of questions you may wish to consider before setting off: - 1) Would you let just anyone connect to your server and query the last login-times for User-X? 2) Would your error messages inform a dictionary attacker when it found a "valid" username? So now your duplicating the VMS login process to find out who the request is coming from and only letting them see their information? 3) Do you incorporate Intrusion Detection? 4) Do you update the SYSUAF entries with login failures? 5) When access is successful, do you update SYSUAF with last [non]interactive login time? You suggested something like INETd 6) Are you happy for a new process to be created every time a query is made? 7) Do you believe the ratio of (1 VMS Process per user) to be optimal? 8) Happy cleaning up all those log files? So you decide to go for an "easy" multi-threaded server and roll your own TCP/IP and/or DECnet programming 9) Threads or ASTs? So now you've completed the simple task of returning last-login-times to the user; why not use the same mechanism as a full-blown Application Server 10) The units-of-work have grown, and stuff like Rdb doesn't work in threads or ASTs anyway, so you need to farm off work to single-threaded processes 11) Would it be nice to have a TP Monitor capability will a pool of pre-initialised worker process and then to be able to grow/shrink that pool as client demand dictates? Also imagine what it'd be like to control server-affinity during message exchange and obtain the true benefits of parallelism between client and server? 12) How would the worker processes assume the persona of the client without endowing the client's request with elevated privileges? And when you've done all that make sure it's generic and re-usable enough that the next Application can just slot-in seemlessly with a single configuration record. Like rollin' off a log really :-) But you mentioned VMS running a web server? Ok, you go ahead and use that connectionless, session-hijacking prone, protocol with its cookies and session IDs and dodgy expiration times (God knows the engineer's comments in the VMS source code itself has us scared stiff that VMS security has already been dumbed-down for Java) and let me know how you got on. Then we can talk about "code-path", seemingly endless redundant layers of software crud, and finally, performance. But what about your non-browser-based clients are you really gonna make them also talk SOAP, IDL, STDL, WSDL, XML, HTTP, PHP, Axis, Tomcat, JSP, CGI and to re-perform the network-connection and authorization handshake everytime you want a subroutine to add two numbers together? And if you still like that idea, here's something else that may appeal to you, ODBC/JDBC and using a Database Stored Procedure as a poor-man's RPC! I mean how "easy" is that? Cheers Richard Maher PS. I don't see your name on our Hobbyists License list (unless you're using a hotmail account); how about we rectify that issue now and you can see first-hand how easy fronting your rich herritage of VMS Applications and Data with a full-function GUI was meant to be? PPS. WRT Kerry's post, here is the code for the "Welcome" dialogue box that appears once a user has successfully logged on, via the Web, to the Tier3 DEMO example program. (As I hope you will soon discover, the DEMO example provides a simple VMS Queue lookup and management facility.) You can find the rest of the Java, JavaScript, HTML and COBOL source files in your t3$examples directory once you've installed Tier3 V3.1 on your system. (Please let me stress once more that this code is just a suggestion! If you'd rather use Swing instead of AWT, or bypass the browser and use a native application instead then go crazy.) NB: *ALL* of this code is re-usable for any number of client/server applications! Use the same Java Applet as many times as you like; just change the parameters. import java.awt.*; import java.awt.event.*; import java.lang.*; import java.util.*; import java.text.*; public class Tier3Welcome extends Dialog implements ActionListener { Button okay = new Button("OK"); Label label1 = new Label(); Label label2 = new Label(); Label label3 = new Label(); Label label4 = new Label(); Font labelSmall = new Font("Helvetica", Font.BOLD, 12); Font labelBig = new Font("Helvetica", Font.BOLD, 14); String outLine; int logFails; public class DateVMS { int year; int month; int day; int hour; int minute; int second; int hsecs; DateVMS (String inDate) { this.year = Integer.parseInt(inDate.substring( 0, 4)); this.month = Integer.parseInt(inDate.substring( 4, 6)); this.month--; this.day = Integer.parseInt(inDate.substring( 6, 8)); this.hour = Integer.parseInt(inDate.substring( 8, 10)); this.minute = Integer.parseInt(inDate.substring(10, 12)); this.second = Integer.parseInt(inDate.substring(12, 14)); this.hsecs = Integer.parseInt(inDate.substring(14, 16)); } public int getYear() { return year; } public int getMonth() { return month; } public int getDay() { return day; } public int getHour() { return hour; } public int getMinute() { return minute; } public int getsecond() { return second; } public int getHsecs() { return hsecs; } } public Tier3Welcome(Frame wframe, String AppName, byte[] t3IdBuf) { super(wframe,true); setUndecorated(true); DateFormat fullDate = DateFormat.getDateTimeInstance(DateFormat.FULL, DateFormat.LONG); Calendar lastLogin = Calendar.getInstance(); outLine = "Welcome to the " + AppName + " application via TIER3 V" + Byte.toString(t3IdBuf[3]) + "." + Byte.toString(t3IdBuf[4]) + " on node " + new String(t3IdBuf, 5, 6); label1.setAlignment(Label.LEFT); label1.setText(outLine); label1.setFont(labelBig); DateVMS lastInt = new DateVMS(new String(t3IdBuf, 16, 16)); lastLogin.clear(); lastLogin.set(lastInt.year, lastInt.month, lastInt.day, lastInt.hour, lastInt.minute, lastInt.second); outLine = " Last interactive login on " + fullDate.format(lastLogin.getTime()); label2.setAlignment(Label.LEFT); label2.setText(outLine); label2.setFont(labelSmall); DateVMS lastNonInt = new DateVMS(new String(t3IdBuf, 32, 16)); lastLogin.clear(); lastLogin.set(lastNonInt.year, lastNonInt.month, lastNonInt.day, lastNonInt.hour, lastNonInt.minute, lastNonInt.second); outLine = " Last non-interactive login on " + fullDate.format(lastLogin.getTime()); label3.setAlignment(Label.LEFT); label3.setText(outLine); label3.setFont(labelSmall); logFails = Integer.parseInt(new String(t3IdBuf, 11, 5)); if (logFails == 0) { outLine = ""; } else { if (logFails == 1) { outLine = " 1 failure since last successful login"; } else { outLine = " " + Integer.toString(logFails) + " failures since last successful login"; } } label4.setAlignment(Label.LEFT); label4.setForeground(Color.RED); label4.setText(outLine); label4.setFont(labelSmall); Panel top = new Panel(); top.setLayout(new GridLayout(4, 1)); top.add(label1); top.add(label2); top.add(label3); top.add(label4); add("North", top); Panel bot = new Panel(); okay.addActionListener(this); bot.add(okay); add("South", bot); setBounds(280, 320, 300, 300); pack(); setResizable(false); setVisible(true); } public void actionPerformed(ActionEvent ae) { if (ae.getSource() == okay) setVisible(false); } } "Ken Fairfield" wrote in message news:5fnoibF3dsgb1U1@mid.individual.net... > Richard Maher wrote: > > Hi Kerry, > > > >> One thing I have always liked about OpenVMS is that on logon, > >> it states when you last logged on - both interactively and via > > [non-inter]. > >> At least that is one way which can be used to flag the user that > >> someone else may have access their account. > > > > Now imagine that after clicking on a web page and enetering your VMS > > credentials you could be presented with that very same information > > (including login failures since last successful login) - Now that'd be > > bloody useful! *Well wouldn't it?* > > And it would be very easy to write a program to do that on > VMS (I'm not offering! ) to query the SYSUAF and return that > information. You could implement it via whatever hook your > VMS middleware requires, like a service under TCPIP/IP (or a > DECnet object) if you want, or you if the VMS system is running > a web server, it could run under that. > > > If only there was some middleware out there that let you do this. . . > > For someone who does serious programming like yourself, it > seems like less than an afternoon's worth of design and coding. :-) > > -Ken > -- > Ken & Ann Fairfield > What: Ken dot And dot Ann > Where: Gmail dot Com ------------------------------ Date: Fri, 13 Jul 2007 06:57:49 -0500 From: "Paul Raulerson" Subject: RE: Now that's what I'm talking about! (was:Re: Is VMS losing the Financial Sec Message-ID: <000a01c7c545$09412cb0$1bc38610$@com> An awful lot of good points indeed - I tended to agree with most of what you said, till I got to the Java code. :) I have two problems to deal with when I have to deal with Java; those are I have users at the end of low speed connections (some of whom run odd machines, thin clients, or what amount to terminals...) and second, that's an awful lot of code just to do a login. Sure, VA-Java or some other another Java environment will let you design and build it with little fuss, but hey... a big application adds lots of code indeed. I am also a little unclear about how well that implements MVC separation. Perhaps if you were to post the website, which I assume also includes costs for a non-Hobby licesne, I could go explore a but more? Googling Tier3 comes up with the telecom company for me, unless that *is* you? Take the sarcasm as a joke, that's the way I meant it, but it would be great if you would not mind posting the URL. -Paul > -----Original Message----- > From: Richard Maher [mailto:maher_rj@hotspamnotmail.com] > Sent: Friday, July 13, 2007 12:54 AM > To: Info-VAX@Mvb.Saic.Com > Subject: Now that's what I'm talking about! (was:Re: Is VMS losing the > Financial Sector, also?) > > Hi Ken, > > > > If only there was some middleware out there that let you do this. . > . > > > > For someone who does serious programming like yourself, it > > seems like less than an afternoon's worth of design and coding. :-) > > Either my marketing attempts to date have been about as successful as > Digital's, or my legendary subtlety has somehow managed to obscure the > inherent sarcasm and facetiousness in my lament :-) > > Yes, there definitely is "middleware out there that [will] let you do > this" > and that middleware is called Tier3. I was rather hoping that you'd > heard of > our Client/Server middleware offering, and that my post would once more > re-focus attention on exactly why Tier3 and its *connection-oriented*, > *context-rich*, execution environment is so good; but no matter. > > > And it would be very easy to write a program to do that on > > VMS > > Depends what you mean by "very easy" I suppose. Here's a couple of > questions > you may wish to consider before setting off: - > > 1) Would you let just anyone connect to your server and query the last > login-times for User-X? > 2) Would your error messages inform a dictionary attacker when it found > a > "valid" username? > > So now your duplicating the VMS login process to find out who the > request is > coming from and only letting them see their information? > > 3) Do you incorporate Intrusion Detection? > 4) Do you update the SYSUAF entries with login failures? > 5) When access is successful, do you update SYSUAF with last > [non]interactive login time? > > You suggested something like INETd > > 6) Are you happy for a new process to be created every time a query is > made? > 7) Do you believe the ratio of (1 VMS Process per user) to be optimal? > 8) Happy cleaning up all those log files? > > So you decide to go for an "easy" multi-threaded server and roll your > own > TCP/IP and/or DECnet programming > > 9) Threads or ASTs? > > So now you've completed the simple task of returning last-login-times > to the > user; why not use the same mechanism as a full-blown Application Server > > 10) The units-of-work have grown, and stuff like Rdb doesn't work in > threads > or ASTs anyway, so you need to farm off work to single-threaded > processes > 11) Would it be nice to have a TP Monitor capability will a pool of > pre-initialised worker process and then to be able to grow/shrink that > pool > as client demand dictates? Also imagine what it'd be like to control > server-affinity during message exchange and obtain the true benefits of > parallelism between client and server? > 12) How would the worker processes assume the persona of the client > without > endowing the client's request with elevated privileges? > > And when you've done all that make sure it's generic and re-usable > enough > that the next Application can just slot-in seemlessly with a single > configuration record. > > Like rollin' off a log really :-) > > But you mentioned VMS running a web server? Ok, you go ahead and use > that > connectionless, session-hijacking prone, protocol with its cookies and > session IDs and dodgy expiration times (God knows the engineer's > comments in > the VMS source code itself has us scared stiff that VMS security has > already > been dumbed-down for Java) and let me know how you got on. Then we can > talk > about "code-path", seemingly endless redundant layers of software crud, > and > finally, performance. But what about your non-browser-based clients are > you > really gonna make them also talk SOAP, IDL, STDL, WSDL, XML, HTTP, PHP, > Axis, Tomcat, JSP, CGI and to re-perform the network-connection and > authorization handshake everytime you want a subroutine to add two > numbers > together? > > And if you still like that idea, here's something else that may appeal > to > you, ODBC/JDBC and using a Database Stored Procedure as a poor-man's > RPC! I > mean how "easy" is that? > > Cheers Richard Maher > > PS. I don't see your name on our Hobbyists License list (unless you're > using > a hotmail account); how about we rectify that issue now and you can see > first-hand how easy fronting your rich herritage of VMS Applications > and > Data with a full-function GUI was meant to be? > > PPS. WRT Kerry's post, here is the code for the "Welcome" dialogue box > that > appears once a user has successfully logged on, via the Web, to the > Tier3 > DEMO example program. (As I hope you will soon discover, the DEMO > example > provides a simple VMS Queue lookup and management facility.) You can > find > the rest of the Java, JavaScript, HTML and COBOL source files in your > t3$examples directory once you've installed Tier3 V3.1 on your system. > (Please let me stress once more that this code is just a suggestion! If > you'd rather use Swing instead of AWT, or bypass the browser and use a > native application instead then go crazy.) > > NB: *ALL* of this code is re-usable for any number of client/server > applications! Use the same Java Applet as many times as you like; just > change the parameters. > > import java.awt.*; > import java.awt.event.*; > import java.lang.*; > import java.util.*; > import java.text.*; > > public class Tier3Welcome extends Dialog > implements ActionListener > { > Button okay = new Button("OK"); > Label label1 = new Label(); > Label label2 = new Label(); > Label label3 = new Label(); > Label label4 = new Label(); > Font labelSmall = new Font("Helvetica", Font.BOLD, 12); > Font labelBig = new Font("Helvetica", Font.BOLD, 14); > > String outLine; > int logFails; > > public class DateVMS > { > int year; > int month; > int day; > int hour; > int minute; > int second; > int hsecs; > > DateVMS (String inDate) > { > this.year = Integer.parseInt(inDate.substring( 0, 4)); > this.month = Integer.parseInt(inDate.substring( 4, 6)); > this.month--; > this.day = Integer.parseInt(inDate.substring( 6, 8)); > this.hour = Integer.parseInt(inDate.substring( 8, 10)); > this.minute = Integer.parseInt(inDate.substring(10, 12)); > this.second = Integer.parseInt(inDate.substring(12, 14)); > this.hsecs = Integer.parseInt(inDate.substring(14, 16)); > } > public int getYear() > { > return year; > } > public int getMonth() > { > return month; > } > public int getDay() > { > return day; > } > public int getHour() > { > return hour; > } > public int getMinute() > { > return minute; > } > public int getsecond() > { > return second; > } > public int getHsecs() > { > return hsecs; > } > } > > public Tier3Welcome(Frame wframe, String AppName, byte[] t3IdBuf) > { > super(wframe,true); > setUndecorated(true); > > DateFormat fullDate = > DateFormat.getDateTimeInstance(DateFormat.FULL, DateFormat.LONG); > Calendar lastLogin = Calendar.getInstance(); > > outLine = "Welcome to the " + AppName + " application via TIER3 > V" + > Byte.toString(t3IdBuf[3]) + "." + > Byte.toString(t3IdBuf[4]) + > " on node " + new String(t3IdBuf, 5, 6); > > label1.setAlignment(Label.LEFT); > label1.setText(outLine); > label1.setFont(labelBig); > > DateVMS lastInt = new DateVMS(new String(t3IdBuf, 16, 16)); > lastLogin.clear(); > lastLogin.set(lastInt.year, lastInt.month, lastInt.day, > lastInt.hour, lastInt.minute, lastInt.second); > > outLine = " Last interactive login on " + > fullDate.format(lastLogin.getTime()); > label2.setAlignment(Label.LEFT); > label2.setText(outLine); > label2.setFont(labelSmall); > > DateVMS lastNonInt = new DateVMS(new String(t3IdBuf, 32, 16)); > lastLogin.clear(); > lastLogin.set(lastNonInt.year, lastNonInt.month, > lastNonInt.day, > lastNonInt.hour, lastNonInt.minute, > lastNonInt.second); > > outLine = " Last non-interactive login on " + > fullDate.format(lastLogin.getTime()); > label3.setAlignment(Label.LEFT); > label3.setText(outLine); > label3.setFont(labelSmall); > > logFails = Integer.parseInt(new String(t3IdBuf, 11, 5)); > if (logFails == 0) > { > outLine = ""; > } > else > { > if (logFails == 1) > { > outLine = " 1 failure since last successful login"; > } > else > { > outLine = " " + Integer.toString(logFails) + " > failures > since last successful login"; > } > } > > label4.setAlignment(Label.LEFT); > label4.setForeground(Color.RED); > label4.setText(outLine); > label4.setFont(labelSmall); > > Panel top = new Panel(); > top.setLayout(new GridLayout(4, 1)); > top.add(label1); > top.add(label2); > top.add(label3); > top.add(label4); > add("North", top); > > Panel bot = new Panel(); > okay.addActionListener(this); > bot.add(okay); > add("South", bot); > > setBounds(280, 320, 300, 300); > pack(); > setResizable(false); > setVisible(true); > } > > public void actionPerformed(ActionEvent ae) > { > if (ae.getSource() == okay) > setVisible(false); > } > } > > "Ken Fairfield" wrote in message > news:5fnoibF3dsgb1U1@mid.individual.net... > > Richard Maher wrote: > > > Hi Kerry, > > > > > >> One thing I have always liked about OpenVMS is that on logon, > > >> it states when you last logged on - both interactively and via > > > [non-inter]. > > >> At least that is one way which can be used to flag the user that > > >> someone else may have access their account. > > > > > > Now imagine that after clicking on a web page and enetering your > VMS > > > credentials you could be presented with that very same information > > > (including login failures since last successful login) - Now that'd > be > > > bloody useful! *Well wouldn't it?* > > > > And it would be very easy to write a program to do that on > > VMS (I'm not offering! ) to query the SYSUAF and return that > > information. You could implement it via whatever hook your > > VMS middleware requires, like a service under TCPIP/IP (or a > > DECnet object) if you want, or you if the VMS system is running > > a web server, it could run under that. > > > > > If only there was some middleware out there that let you do this. . > . > > > > For someone who does serious programming like yourself, it > > seems like less than an afternoon's worth of design and coding. :-) > > > > -Ken > > -- > > Ken & Ann Fairfield > > What: Ken dot And dot Ann > > Where: Gmail dot Com > ------------------------------ Date: Fri, 13 Jul 2007 12:54:46 GMT From: Robert Deininger Subject: Re: Now that's what I'm talking about! (was:Re: Is VMS losing the Financial Sec Message-ID: In article , "Richard Maher" wrote: > Hi Ken, > > > > If only there was some middleware out there that let you do this. . . > > > > For someone who does serious programming like yourself, it > > seems like less than an afternoon's worth of design and coding. :-) > > Either my marketing attempts to date have been about as successful as > Digital's, or my legendary subtlety has somehow managed to obscure the > inherent sarcasm and facetiousness in my lament :-) Maybe change the "or" to "and"? > Yes, there definitely is "middleware out there that [will] let you do this" > and that middleware is called Tier3. I was rather hoping that you'd heard of > our Client/Server middleware offering, and that my post would once more > re-focus attention on exactly why Tier3 and its *connection-oriented*, > *context-rich*, execution environment is so good; but no matter. Seriously, Richard, if you want to increase the uptake on your product, I think you need to improve your communication. I read your whole very long post, and understood less than 1/3 of it. You might have just the thing I need, but I have no way of knowing. I got lost in the jargon and acronyms; maybe only folks who already know all that stuff need to use your product. And the look-at-me-see-how-clever-I-am attitude in your post doesn't help me learn anything. At least HP has a (horrid, slow, confusing, intrusive, and annoying) web page. You didn't even mention a URL where we can find more information about your product. Do you have a web page for Tier3? Google gave me a hint, on the second page of results (how many people ever look at the second page?) It showed me your OpenVMS.org posting from June 24. Agony! Another long, jargon-laden description of Tier3 that leaves me as confused as ever. (Though I have to admit I didn't read the whole thing. I've spent too much time already, and have to go.) At least that post isn't filled with "attitude", and there is a well-hidden e-mail address that I could use if I was interested enough to try. Suggestions: 1. Put together some web pages for your product. I think the rule of thumb today is begin Company; if (no web page) then no product; end if; end Company; 2. Mention the URL often. (Have you noticed all the URLs in David Dachtera's posts? Or David Turner's?) 3. Write a description of the product that non-wizards can understand. Include some pictures and examples. 4. Include some live demos on your web site that use your product. Your lovely code for the Welcome dialog would be a lot more effective if I had just seen the live dialog in action. 5. I gather you have a demo kit and a hobbyist license available. Asking folks to e-mail you to obtain these is so --- DIGITAL! That's annoying, and inefficient. You're halfway around the world from me; I'd likely have to wait most of a day for an e-mail response. Web and ftp servers are actually pretty useful for distributing software. If you insist on gathering contact information before you show your product, use Tier3 to implement the interface. And brag about it. HP has a huge bureaucracy that keeps individual product groups from being customer-friendly. You don't have that handicap. Take advantage of your situation. -- Robert ------------------------------ Date: Fri, 13 Jul 2007 08:40:14 +0200 From: "P. Sture" Subject: Re: OpenVMS - When downtime is not an option Message-ID: In article , Keith Parris wrote: > John Smith wrote: > > HP should have also blown up current Solaris and AIX > > configurations -- configured as fairly as any of the others - as part of the > > demo. Maybe even a zOS config too. Keith.... does HP have any extra C4 to > > do this? > > NTS (National Technical Systems, http://ntscorp.com/) supplied the C-4. > If someone paid them to blow up Sun and IBM systems, they'd happily > provide the explosives. I'm thinking of a career change. Does anyone know if they have any vacancies? Oh, while I'm here, a bit of feedback. I found the displays of the NTS logo during the video somewhat distracting, as I was trying to work out what they meant, (Non-Stop Tandem? - nope, try again...) A quick introductory subtitle such as "Test conducted with the assistance of NTS - National Technical Systems" would have solved that. -- Paul Sture ------------------------------ Date: 13 Jul 2007 12:17:39 GMT From: bill@cs.uofs.edu (Bill Gunshannon) Subject: Re: Solved: SAN switch 2/8 on MSA1000 - cannot connect to serial of rj-45 port Message-ID: <5fp8r3F3dnmuaU1@mid.individual.net> In article <4696C87F.97B790F@spam.comcast.net>, David J Dachtera writes: > Bill Gunshannon wrote: >> >> In article <50474$46960ab2$cef8887a$2069@teksavvy.com>, >> JF Mezei writes: >> > Anton Shterenlikht wrote: >> >> I'm an idiot, I missed "This cable must be a straight through cable" in >> >> the manual. >> > >> > Note that ethernet also has the same concept of straight through versus >> > cross over. >> > >> > Worse yet, USB doesn't have that concept because it is client-server and >> > you cannot connect 2 clients together or 2 servers together no matter >> > how much you tsist the cable. >> >> Maybe not with a cable you make yourself, but you can connect two >> computers together with USB. Do it all the time to clone disks >> with GHOST. Just had to buy the cable. > > I'm sure you remember "InterLink" from the MS-DOS V6 days, eh? > > For the youngsters in the group that was parallel port to parallel port > connection. > > A third-party product called "LapLink" (presumably for laptops, prior to SMB and > such) also came about. Similar strategy. I still have and frequently use my Laplink serial and parallel cables!! bill -- Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves bill@cs.scranton.edu | and a sheep voting on what's for dinner. University of Scranton | Scranton, Pennsylvania | #include ------------------------------ Date: Fri, 13 Jul 2007 10:16:16 +0200 From: "P. Sture" Subject: Re: Solved: SAN switch 2/8 on MSA1000 - cannot connect to serial of rj-45 port Message-ID: In article <4696C87F.97B790F@spam.comcast.net>, David J Dachtera wrote: > Bill Gunshannon wrote: > > > > In article <50474$46960ab2$cef8887a$2069@teksavvy.com>, > > JF Mezei writes: > > > Anton Shterenlikht wrote: > > >> I'm an idiot, I missed "This cable must be a straight through cable" in > > >> the manual. > > > > > > Note that ethernet also has the same concept of straight through versus > > > cross over. > > > > > > Worse yet, USB doesn't have that concept because it is client-server and > > > you cannot connect 2 clients together or 2 servers together no matter > > > how much you tsist the cable. > > > > Maybe not with a cable you make yourself, but you can connect two > > computers together with USB. Do it all the time to clone disks > > with GHOST. Just had to buy the cable. > > I'm sure you remember "InterLink" from the MS-DOS V6 days, eh? > > For the youngsters in the group that was parallel port to parallel port > connection. > > A third-party product called "LapLink" (presumably for laptops, prior to SMB > and such) also came about. Similar strategy. Laplink and Interlink introduced a nice little business opportunity back in those days. The layman* who didn't know their cables thought they needed a "special" Laplink cable, when in fact a normal bi-directional parallel cable was all you needed. Only beer money, but it helped :-) * The word "layman" can include a surprising number of computer salesmen :-) -- Paul Sture ------------------------------ Date: Thu, 12 Jul 2007 19:34:07 -0500 From: David J Dachtera Subject: Re: Solved: SAN switch 2/8 on MSA1000 - cannot connect to serial of rj-45 port Message-ID: <4696C87F.97B790F@spam.comcast.net> Bill Gunshannon wrote: > > In article <50474$46960ab2$cef8887a$2069@teksavvy.com>, > JF Mezei writes: > > Anton Shterenlikht wrote: > >> I'm an idiot, I missed "This cable must be a straight through cable" in > >> the manual. > > > > Note that ethernet also has the same concept of straight through versus > > cross over. > > > > Worse yet, USB doesn't have that concept because it is client-server and > > you cannot connect 2 clients together or 2 servers together no matter > > how much you tsist the cable. > > Maybe not with a cable you make yourself, but you can connect two > computers together with USB. Do it all the time to clone disks > with GHOST. Just had to buy the cable. I'm sure you remember "InterLink" from the MS-DOS V6 days, eh? For the youngsters in the group that was parallel port to parallel port connection. A third-party product called "LapLink" (presumably for laptops, prior to SMB and such) also came about. Similar strategy. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ Date: 13 Jul 2007 15:17:01 +0200 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOeGER) Subject: Re: SYSMAN oddity (username on same node) Message-ID: <4697976d$1@news.langstoeger.at> In article <4543111D.D7C57886@teksavvy.com>, JF Mezei writes: >[SYSMAN> SET ENV/USER=xxx/NODE=xxx] >Why is it that the specified username is NOT used for the local node ? >Is there some magical logical that can be set to change this behaviour >so that the specified username is used on all specified nodes for DO >commands ? Is there perhaps any news on that (I do think V8.3 behaves as V7.3-2 did)? I still see it as a bug, which should be fixed (and not only documented)... Isn't this a problem for you, too? Or don't you use SYSMAN (maybe because you're a REAL sysmangler)? Did anyone -with a support contract - ask HPQ (and what did HPQ state)? Many TIA -- Peter "EPLAN" LANGSTOEGER Network and OpenVMS system specialist E-mail peter@langstoeger.at A-1030 VIENNA AUSTRIA I'm not a pessimist, I'm a realist ------------------------------ Date: Fri, 13 Jul 2007 08:43:52 -0700 From: ultradwc@gmail.com Subject: UK report says data protection breaches "out of control" Message-ID: <1184341432.812109.87800@m3g2000hsh.googlegroups.com> what are they on Andrew? slowaris, linux, windoze? certainly not OpenVMS ... http://news.bbc.co.uk/1/hi/business/4311073.stm http://www.sophos.com/security/blog/2007/07/356.html ------------------------------ Date: Fri, 13 Jul 2007 08:47:49 -0700 From: ultradwc@gmail.com Subject: UK report says data protection breaches "out of control" Message-ID: <1184341669.325315.141690@r34g2000hsd.googlegroups.com> well Andrew, looks like slowaris 10 and windoze and linux are not doing the job ... OpenVMS would not have these problems because it does not get viruses ... http://news.bbc.co.uk/1/hi/business/4311073.stm http://www.sophos.com/security/blog/2007/07/356.html ------------------------------ Date: Fri, 13 Jul 2007 12:53:34 -0400 From: Stephen Hoffman Subject: Re: UK report says data protection breaches "out of control" Message-ID: ultradwc@gmail.com wrote: > what are they on Andrew? slowaris, linux, windoze? > certainly not OpenVMS ... Certainly Solaris. And Linux. And Windows. These can and do arise on all platforms. Including on OpenVMS. I'm aware of and was directly involved with various breaches involving OpenVMS systems, and with the cleaning up from some. (These particular breaches completely and entirely outside my tenure at DIGITAL, Compaq and HP -- nor will I not provide specifics here.) DIGITAL VMS Engineering itself had a data exposure and security breach that was discussed widely and quite publicly, involving Mr. Mitnick. (Conversely, Mr Mitnick also testified that the only operating system that he was unable to directly breach was VMS. Though he did gain access.) Recovery from these breaches can be interesting and can be potentially involved, particularly when there is a savvy attacker involved. This involves finding and plugging the hole -- either in the organizational process, or in the software, too. And verification of data and system integrity. Been there, done that. Multiple times. Some of these general classes of vulnerabilities -- and depending on the value of the information involved -- are simply not pluggable within typical business (financial or political) economics; it can be more appropriate to use a strategy of detection and recovery. And there's the whole social-engineering wildcard. For some data, yes, you do need to manage data access and exposures due to regulatory or the cost of downtime and/or the cost of recovery. And you have the budget to purchase tokens and massive security. For other data, not so much. The change here is that various of these attacks are increasingly sophisticated, and increasingly are specifically targeted. And there are some clever and non-so-clever ways to target these attacks. There was an attack reported recently that targeted the access and information available (only) to specific personnel within multiple corporations, for instance. -- www.HoffmanLabs.com Services for OpenVMS ------------------------------ Date: Fri, 13 Jul 2007 10:15:20 -0700 From: ultradwc@gmail.com Subject: Re: UK report says data protection breaches "out of control" Message-ID: <1184346920.124742.143100@k79g2000hse.googlegroups.com> On Jul 13, 12:53 pm, Stephen Hoffman wrote: > ultra...@gmail.com wrote: > > what are they on Andrew? slowaris, linux, windoze? > > certainly not OpenVMS ... > > Certainly Solaris. And Linux. And Windows. > > These can and do arise on all platforms. Including on OpenVMS. > > I'm aware of and was directly involved with various breaches involving > OpenVMS systems, and with the cleaning up from some. (These particular > breaches completely and entirely outside my tenure at DIGITAL, Compaq > and HP -- nor will I not provide specifics here.) > > DIGITAL VMS Engineering itself had a data exposure and security breach > that was discussed widely and quite publicly, involving Mr. Mitnick. > (Conversely, Mr Mitnick also testified that the only operating system > that he was unable to directly breach was VMS. Though he did gain access.) > > Recovery from these breaches can be interesting and can be potentially > involved, particularly when there is a savvy attacker involved. This > involves finding and plugging the hole -- either in the organizational > process, or in the software, too. And verification of data and system > integrity. Been there, done that. Multiple times. > > Some of these general classes of vulnerabilities -- and depending on the > value of the information involved -- are simply not pluggable within > typical business (financial or political) economics; it can be more > appropriate to use a strategy of detection and recovery. And there's > the whole social-engineering wildcard. > > For some data, yes, you do need to manage data access and exposures due > to regulatory or the cost of downtime and/or the cost of recovery. And > you have the budget to purchase tokens and massive security. For other > data, not so much. > > The change here is that various of these attacks are increasingly > sophisticated, and increasingly are specifically targeted. And there > are some clever and non-so-clever ways to target these attacks. > There was an attack reported recently that targeted the access and > information available (only) to specific personnel within multiple > corporations, for instance. > > --www.HoffmanLabs.com > Services for OpenVMS VMS does not get viruses or trojans or worms, so what you are talking about are systems not properly secured or inside jobs of the hacker getting hold of someone elses password ... two different things ... this article is exposing the security weaknesses in unix/linux/windoze and validating the CERT site counts ... OS design is what is in question here, not poor setup or password stealing ... ------------------------------ Date: Fri, 13 Jul 2007 12:44:46 -0500 From: Ron Johnson Subject: Re: UK report says data protection breaches "out of control" Message-ID: On 07/13/07 12:15, ultradwc@gmail.com wrote: [snip] > > VMS does not get viruses or trojans or worms, so what > you are talking about are systems not properly secured > or inside jobs of the hacker getting hold of someone elses > password ... two different things ... this article is exposing > the security weaknesses in unix/linux/windoze and validating > the CERT site counts ... OS design is what is in question > here, not poor setup or password stealing ... How can you say that VMS doesn't get trojans or worms? (Note that I left out viruses.) -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Fri, 13 Jul 2007 03:01:58 -0500 From: Ron Johnson Subject: Re: [OT] July the 4th Message-ID: On 07/12/07 21:49, David J Dachtera wrote: > Ron Johnson wrote: >> On 07/12/07 18:43, David J Dachtera wrote: >>> Ron Johnson wrote: >>>> On 07/10/07 20:17, David J Dachtera wrote: >>>> [snip] >>>>> ...which brings us back to the discussion of what can be done vs. >>>>> what should be done. Given the MIB's tendency to go "ape shit" >>>>> these days, I'm not sure I'd want to be "seen" in a public forum >>>>> fomenting insurrection and/or siding with the sworn enemies of >>>>> the U.S. >>>>> >>>>> Hope you don't plan any air travel anytime during the remainder >>>>> of your lifetime... >> >> >>>> Bill can rant all he wants about the evils of our government, and I >>>> *guarantee* you that no MIBs are going to visit him and he'll be >>>> able to fly anywhere in this country he wants. >>> Go back and re-read the earlier post and you should notice that was not the >>> point: >>> >>> David J Dachtera wrote: >>>> ... I'm not sure I'd >>>> want to be "seen" in a public forum fomenting insurrection and/or >>>> siding with the sworn enemies of the U.S. >>> Big difference. >> Then what did "Hope you don't plan any air travel..." refer to? > > You CAN'T be serious???!!! You're not that dumb, Ron! Read it again, and read > further on down here. You'll get it. Keep your brain turned on and believe in > yourself. You'll get it. > >>>> How do I know? >>>> >>>> If the MIBs were that out of control, 3/4 of the people who >>>> regularly participate on the Daily Kos and all the tens (hundreds?) >>>> of thousands of people who horribly vilify W on a daily basis would >>>> have been detained by now. >>>> >>>> But they haven't. >>> Vilify "dubya" all you want - he and his lackeys deserve it and then some. Just >>> don't make threats or side with the enemy. > > There's your clue, right there! Maybe I've seen so many paranoid anti-W rants that I just automatically assume that all anti-W complaints are full of paranoid delusions of impending persecution. -- Ron Johnson, Jr. Jefferson LA USA Give a man a fish, and he eats for a day. Hit him with a fish, and he goes away for good! ------------------------------ Date: Thu, 12 Jul 2007 18:43:10 -0500 From: David J Dachtera Subject: Re: [OT] July the 4th Message-ID: <4696BC8E.BC62CF39@spam.comcast.net> Ron Johnson wrote: > > On 07/10/07 20:17, David J Dachtera wrote: > [snip] > > > > ...which brings us back to the discussion of what can be done vs. > > what should be done. Given the MIB's tendency to go "ape shit" > > these days, I'm not sure I'd want to be "seen" in a public forum > > fomenting insurrection and/or siding with the sworn enemies of > > the U.S. > > > > Hope you don't plan any air travel anytime during the remainder > > of your lifetime... > > Bill can rant all he wants about the evils of our government, and I > *guarantee* you that no MIBs are going to visit him and he'll be > able to fly anywhere in this country he wants. Go back and re-read the earlier post and you should notice that was not the point: David J Dachtera wrote: > > ... I'm not sure I'd > want to be "seen" in a public forum fomenting insurrection and/or > siding with the sworn enemies of the U.S. Big difference. > How do I know? > > If the MIBs were that out of control, 3/4 of the people who > regularly participate on the Daily Kos and all the tens (hundreds?) > of thousands of people who horribly vilify W on a daily basis would > have been detained by now. > > But they haven't. Vilify "dubya" all you want - he and his lackeys deserve it and then some. Just don't make threats or side with the enemy. -- David J Dachtera dba DJE Systems http://www.djesys.com/ Unofficial OpenVMS Marketing Home Page http://www.djesys.com/vms/market/ Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/ Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/ Unofficial OpenVMS Hobbyist Support Page: http://www.djesys.com/vms/support/ ------------------------------ End of INFO-VAX 2007.380 ************************