1 INFO-VAX	Thu, 27 Jun 2002	Volume 2002 : Issue 351       Contents:! Re: $CRMPSC and memory management ! Re: $CRMPSC and memory management P %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020, PCP Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020P Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020P Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020P Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020" Re: 7.3-1 new backplane controller; Re: A possible shift in the status of VMS ar HP ???? ERRATA ; Re: A possible shift in the status of VMS ar HP ???? ERRATA ; Re: A possible shift in the status of VMS ar HP ???? ERRATA 8 Re: Amdrew wants numbers, here they are!  Blow out baby!8 Re: Amdrew wants numbers, here they are!  Blow out baby!8 Re: Amdrew wants numbers, here they are!  Blow out baby!8 Re: Amdrew wants numbers, here they are!  Blow out baby!' Re: Andrew wan'ts the numbers, here ... G Re: BACKUP/INCREMENTAL not parsing directories properly during restores < Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC@ Re: Deutsche Bank would like to outsource there IT to IBM or CSC Encompass - DECUS VMS survey1 Re: how detect all batch queues in a VMS system ? ' Re: How detect Batch queues in Pascal ?  HP HPTC in Oil & Gas Re: HP HPTC in Oil & Gas Re: HP HPTC in Oil & Gas Re: HP HPTC in Oil & Gas Re: IPSEC ?  Re: IPSEC ? , Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ..., Re: My conversation with Linus about VMS ... Re: Open Letter to HP  Re: Open Letter to HP  Re: Open Letter to HP  parsing >255 Re: Quorum discussion/questions * Re: remote node is not currently reachable* Re: remote node is not currently reachable3 Re: Reuters test - Itanium II blows away Sparky III 3 Re: Reuters test - Itanium II blows away Sparky III 3 Re: Reuters test - Itanium II blows away Sparky III 3 Re: Reuters test - Itanium II blows away Sparky III ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? ) Re: RMS : preserving a RAB across $OPEN ? " Re: Sun benchmarketeering campaign" Re: Sun benchmarketeering campaign" Re: Sun benchmarketeering campaign
 Re: TCO study 
 Re: TCO study  Re: Trouble with BACKUP/RECORD Re: Trouble with BACKUP/RECORD Re: Trouble with BACKUP/RECORD Re: Trouble with BACKUP/RECORD  F ----------------------------------------------------------------------  # Date: Thu, 27 Jun 2002 11:44:07 GMT  From: system@SendSpamHere.ORG * Subject: Re: $CRMPSC and memory management0 Message-ID: <00A10136.E150D5A9@SendSpamHere.ORG>  O In article <3D1A47CD.39CC6B81@gte.net>, Robert Davis <res057y3@gte.net> writes:  > C >Thanks for the reply. That makes sense that VIRTUALPAGECNT doesn't B >affect the address range for P0, only how many pages can be used. > G >But, now I'm confused as to why we're getting VASFULL errors trying to F >map sections at x50000000 on VAX (when it seems to work OK on Alpha).F >Are there other params that affect the P0 range? Any idea why I wouldG >get a VASFULL error when trying to map a section at any address higher ( >than x10000000 on a VAX system? Thanks. >  >Rob  E Perhaps it's because OpenVMS Alpha memory management allows a sparse  E population of the virtual address space whereas OpenVMS VAX will need E to fill in the region with NA or DZRO pages.  The PT needs to be pop-  ulated on VAX.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              5   "Well my son, life is like a beanstalk, isn't it?"     ------------------------------    Date: 27 Jun 2002 08:23:35 -0600 From: briggs@encompasserve.org* Subject: Re: $CRMPSC and memory management3 Message-ID: <L8vwxPsZZzXK@eisner.encompasserve.org>   F In article <slrnahhpqm.d7o.danco@pebble.org>, danco@pebble.org writes:= > In article <3D18D1F8.2F42998B@gte.net>, Robert Davis wrote: G >> Sorry I wasn't clear ... I do know the size of the second section at I >> the fixed address and it's relatively small. So once I know what's the I >> 'safe' top end, I can easily decide the start address. I'm mostly just E >> a little unclear about the relationship between VIRTUALPAGECNT and @ >> P0/P1, and where that makes P0/P1 begin/end, and exactly what >> addresses are available.  > C > So far as I know, VIRTUALPAGECNT has no bearing on where P0 ends, 7 > at least not in the normal OpenVMS user mode program.   A VIRTUALPAGECNT does have a bearing.  It is essentially sum of the % lengths of the P0 and P1 page tables.   ? If VIRTUALPAGECNT = 100000 then you can't have more than 100000 F pages in the P0 page table.  And since the P0 page table is contiguous3 (on VAX where VIRTUALPAGECNT matters) and starts at = page 0, that means that the high end of the usable portion of > P0 space would be somewhere prior to virtual address 030D4000.  ? If you burn up 50000 pages of P1 space on CLI, stack and system E overhead then the high end of the usable portion of P0 space would be > somewhere around 0186A000 (again, assuming a relatively meager VIRTUALPAGECNT of 100000).  D I agree with the general sentiment here.  Don't base images.  If youD think you have to base, try using relative pointers instead.  If you@ really have to base and you have to allow for additional address= space of unknown but potentially huge size, base low.  If you 9 can't base low then base high and document your resulting  VIRTUALPAGECNT requirements.   	John Briggs   ------------------------------  % Date: Thu, 27 Jun 2002 20:48:28 +0530 # From: "Vivek Soni" <visoni@bmc.com> Y Subject: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020, PC / Message-ID: <uhmb2m7ara0e57@corp.supernews.com>    Hi,   7 I have an excutatable made for VAX and Alpha Platforms.   F It takes two string parameters  The second string parameter is used to create a file by its name.  L If  i by mistake all "\" as the last letter to the Second String ( filename) then I have following Error.  ; %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual  address=20202020, PC =80000010, PSL=03C00004    For VAX:  2   Improperly handled condition, image exit forced.  4         Signal arguments              Stack contents  1         Number = 00000005                873EAE4D 1         Name   = 0000000C                00000002 1                  00000001                7FE9B5E4 1                  20202020                7FE9B5CC 1                  80000010                00000004 1                  03C00004                7FE9C168 1                                          00000002 1                                          040003DD 1                                          7FE9B730 1                                          05000001            Register dump B         R0 = 03C00000  R1 = 20202020  R2 = 000003DD  R3 = 7FE9C4F7B         R4 = 00000200  R5 = 00000001  R6 = FFFFFFFF  R7 = 00000001B         R8 = 0005242C  R9 = 7FFECC50  R10= 7FFED7D4  R11= 7FFE2BDCB         AP = 7FE9B580  FP = 7FE9B540  SP = 7FE9B5BC  PC = 80000010         PSL= 03C00004   
 For Alpha:  ; %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual  address=000000007AF & C003, PC=FFFFFFFF8083E688, PS=0000001B  2   Improperly handled condition, image exit forced.1     Signal arguments:   Number = 0000000000000005 1                         Name   = 000000000000000C 1                                  0000000000010000 1                                  000000007AF9C003 1                                  000000008083E688 1                                  000000000000001B        Register dump:J     R0  = 00000000000003DD  R1  = 000000007AF9B245  R2  = 000000007AF9AE68J     R3  = 000000007AF9BC2F  R4  = 0000000000000200  R5  = 0000000000000200J     R6  = FFFFFFFFFFFFFFFF  R7  = 0000000000000001  R8  = 0000000000126A2CJ     R9  = 000000007FFAC410  R10 = 000000007FFAD238  R11 = 000000007FFCE3E0J     R12 = 000000007AF9AD5C  R13 = 000000007AF9AD34  R14 = 000000007AF9AD34J     R15 = 000000007AF9AE68  R16 = 000000007AF9C004  R17 = 008101001B202020J     R18 = 000000007AF9B239  R19 = 8101001B20202020  R20 = 0000000000000004J     R21 = 00000000020EFFFF  R22 = 0000000000063C84  R23 = 000000007AF9A890J     R24 = 000000007AF9AE68  R25 = 0000000000000003  R26 = FFFFFFFF8080BABC    K I have read about a similar problem.  But could not lead to any conclusion.   + If you have any idea what the prolem is ...    Thanks in Advance  Vivek    ------------------------------  + Date: Thu, 27 Jun 2002 15:35:09 +0000 (UTC) - From: lewis@spyder.mitre.org (Keith A. Lewis) Y Subject: Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020 . Message-ID: <affbbd$rut$1@newslocal.mitre.org>  z "Vivek Soni" <visoni@bmc.com> writes in article <uhmb2m7ara0e57@corp.supernews.com> dated Thu, 27 Jun 2002 20:48:28 +0530:< >%SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual >address=20202020, PC  >=80000010, PSL=03C00004  I ACCVIO usually means you have a bad pointer.  In this case the pointer is I hex 20202020, which could be 4 spaces in a string somewhere that your app F is attempting to use as a pointer, or maybe your app overwrote its own pointer with a string...   >        Register dumpC >        R0 = 03C00000  R1 = 20202020  R2 = 000003DD  R3 = 7FE9C4F7    There's that 20202020 again.  J Register dumps don't contain as much useful info as tracebacks, so you mayG need to use the VMS debugger to figure out exactly where the program is  bombing.  + --Keith Lewis              klewis$mitre.org > The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Thu, 27 Jun 2002 10:39:22 -0500 C From: "Craig A. Berry" <craig.berry@nospam.SignalTreeSolutions.com> Y Subject: Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020 H Message-ID: <craig.berry-0BFD22.10392027062002@news.directvinternet.com>  / In article <uhmb2m7ara0e57@corp.supernews.com>, %  "Vivek Soni" <visoni@bmc.com> wrote:   9 > I have an excutatable made for VAX and Alpha Platforms.  > H > It takes two string parameters  The second string parameter is used to > create a file by its name. > N > If  i by mistake all "\" as the last letter to the Second String ( filename) > then I have following Error. > = > %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual  > address=20202020, PC > =80000010, PSL=03C00004   D A memory location that should contain the address of another memory E location instead contains spaces (0x20), or in common parlance, your  G pointer is hosed.  Why this might be happening depends entirely on how  G the code in question parses the string, and since you haven't shown us  H the code or even mentioned what language it's in, the possibilities are B limitless.  I will say, though, that if the language is C and the C string is fed through one of the printf() family of functions, the  E backslash might well be getting interpreted as an "escape" assigning  + special meaning to the following character.    ------------------------------   Date: 27 Jun 2002 15:31:10 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann) Y Subject: Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020 0 Message-ID: <affb3u$5d4$1@n.ruf.uni-freiburg.de>  U In article <uhmb2m7ara0e57@corp.supernews.com>, "Vivek Soni" <visoni@bmc.com> writes: 8 >I have an excutatable made for VAX and Alpha Platforms. > G >It takes two string parameters  The second string parameter is used to  >create a file by its name.  > M >If  i by mistake all "\" as the last letter to the Second String ( filename)  >then I have following Error.  > < >%SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual >address=20202020, PC  >=80000010, PSL=03C00004 > 	 >For VAX:  > 3 >  Improperly handled condition, image exit forced.  > 5 >        Signal arguments              Stack contents  > 2 >        Number = 00000005                873EAE4D2 >        Name   = 0000000C                000000022 >                 00000001                7FE9B5E42 >                 20202020                7FE9B5CC2 >                 80000010                000000042 >                 03C00004                7FE9C1682 >                                         000000022 >                                         040003DD2 >                                         7FE9B7302 >                                         05000001 >  >        Register dumpC >        R0 = 03C00000  R1 = 20202020  R2 = 000003DD  R3 = 7FE9C4F7 C >        R4 = 00000200  R5 = 00000001  R6 = FFFFFFFF  R7 = 00000001 C >        R8 = 0005242C  R9 = 7FFECC50  R10= 7FFED7D4  R11= 7FFE2BDC C >        AP = 7FE9B580  FP = 7FE9B540  SP = 7FE9B5BC  PC = 80000010  >        PSL= 03C00004 >  >For Alpha:  > < >%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual >address=000000007AF' >C003, PC=FFFFFFFF8083E688, PS=0000001B  > 3 >  Improperly handled condition, image exit forced. 2 >    Signal arguments:   Number = 00000000000000052 >                        Name   = 000000000000000C2 >                                 00000000000100002 >                                 000000007AF9C0032 >                                 000000008083E6882 >                                 000000000000001B >  >    Register dump: K >    R0  = 00000000000003DD  R1  = 000000007AF9B245  R2  = 000000007AF9AE68 K >    R3  = 000000007AF9BC2F  R4  = 0000000000000200  R5  = 0000000000000200 K >    R6  = FFFFFFFFFFFFFFFF  R7  = 0000000000000001  R8  = 0000000000126A2C K >    R9  = 000000007FFAC410  R10 = 000000007FFAD238  R11 = 000000007FFCE3E0 K >    R12 = 000000007AF9AD5C  R13 = 000000007AF9AD34  R14 = 000000007AF9AD34 K >    R15 = 000000007AF9AE68  R16 = 000000007AF9C004  R17 = 008101001B202020 K >    R18 = 000000007AF9B239  R19 = 8101001B20202020  R20 = 0000000000000004 K >    R21 = 00000000020EFFFF  R22 = 0000000000063C84  R23 = 000000007AF9A890 K >    R24 = 000000007AF9AE68  R25 = 0000000000000003  R26 = FFFFFFFF8080BABC  >  > L >I have read about a similar problem.  But could not lead to any conclusion. > , >If you have any idea what the prolem is ...  N Compile & link your program with the /DEBUG and /TRACEBACK qualifier to see atL which point it is failing. In addition, please post the relevant part of the code.    Regards,    Christoph Gartmann   H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, Germany                                           |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  % Date: Thu, 27 Jun 2002 11:52:01 -0400 1 From: Michael Austin <maustin@firstdbasource.com> Y Subject: Re: %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=20202020 2 Message-ID: <3D1B34A1.7A238BC7@firstdbasource.com>   Vivek Soni wrote:  >  > Hi,  > 9 > I have an excutatable made for VAX and Alpha Platforms.  > H > It takes two string parameters  The second string parameter is used to > create a file by its name. > N > If  i by mistake all "\" as the last letter to the Second String ( filename) > then I have following Error. > = > %SYSTEM-F-ACCVIO, access violation, reason mask=01, virtualC > address=20202020, PC > =80000010, PSL=03C00004s > 
 > For VAX: > 4 >   Improperly handled condition, image exit forced. > 6 >         Signal arguments              Stack contents > 3 >         Number = 00000005                873EAE4D03 >         Name   = 0000000C                00000002s3 >                  00000001                7FE9B5E4n3 >                  20202020                7FE9B5CCM3 >                  80000010                00000004t3 >                  03C00004                7FE9C168 3 >                                          00000002s3 >                                          040003DDw3 >                                          7FE9B730r3 >                                          05000001  >  >         Register dumpeD >         R0 = 03C00000  R1 = 20202020  R2 = 000003DD  R3 = 7FE9C4F7D >         R4 = 00000200  R5 = 00000001  R6 = FFFFFFFF  R7 = 00000001D >         R8 = 0005242C  R9 = 7FFECC50  R10= 7FFED7D4  R11= 7FFE2BDCD >         AP = 7FE9B580  FP = 7FE9B540  SP = 7FE9B5BC  PC = 80000010 >         PSL= 03C00004u >  > For Alpha: > = > %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual  > address=000000007AFn( > C003, PC=FFFFFFFF8083E688, PS=0000001B > 4 >   Improperly handled condition, image exit forced.3 >     Signal arguments:   Number = 0000000000000005u3 >                         Name   = 000000000000000Ci3 >                                  0000000000010000B3 >                                  000000007AF9C003e3 >                                  000000008083E688B3 >                                  000000000000001Bh >  >     Register dump:L >     R0  = 00000000000003DD  R1  = 000000007AF9B245  R2  = 000000007AF9AE68L >     R3  = 000000007AF9BC2F  R4  = 0000000000000200  R5  = 0000000000000200L >     R6  = FFFFFFFFFFFFFFFF  R7  = 0000000000000001  R8  = 0000000000126A2CL >     R9  = 000000007FFAC410  R10 = 000000007FFAD238  R11 = 000000007FFCE3E0L >     R12 = 000000007AF9AD5C  R13 = 000000007AF9AD34  R14 = 000000007AF9AD34L >     R15 = 000000007AF9AE68  R16 = 000000007AF9C004  R17 = 008101001B202020L >     R18 = 000000007AF9B239  R19 = 8101001B20202020  R20 = 0000000000000004L >     R21 = 00000000020EFFFF  R22 = 0000000000063C84  R23 = 000000007AF9A890L >     R24 = 000000007AF9AE68  R25 = 0000000000000003  R26 = FFFFFFFF8080BABC > M > I have read about a similar problem.  But could not lead to any conclusion.w > - > If you have any idea what the prolem is ...l >  > Thanks in Advance  > Viveke    C a bad program that does not determine if the file being create is aw valid filename? > show us the code snippet that 1) takes in the parameter and 2) creates/opens the file   -- M Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163 7 Sr. Consultant            http://www.firstdbasource.comsE                           http://www.firstdbasource.com/donation.html / 704-947-1089 (Office)     704-236-4377 (Mobile)    ------------------------------  % Date: Thu, 27 Jun 2002 09:43:18 -0500 1 From: "Dave Gudewicz" <david.gudewicz@abbott.com> + Subject: Re: 7.3-1 new backplane controllerr1 Message-ID: <aff8da$s4m$1@fizban.pprd.abbott.com>E  J Will _all_ the SmartArray controllers be supported?  And whats the word on the MSA1000.   -- Dave...-  3 More than one cigar at a time is excessive smoking.m -----Mark Twain   A "John.Malmberg" <Malmberg@dskwld.zko.dec.compaq> wrote in message1. news:3D1A2BAE.2090401@dskwld.zko.dec.compaq... > Dave Gudewicz wrote:I > > VMS web page mentions support of a new backplane controller in 7.3-1.  > > K > > Anyone know what new backplane controller.  No model or part number was 
 > > given. >eI > It is from the "Smart Array" line of Backplane controllers according tog > the web page.n >tI > For those not familiar with the "Smart Array" line, here is the generalr > link.a >c >hL http://www.compaq.com/products/servers/proliantstorage/arraycontrollers/inde x.html >y > -Johnh# > malmberg@dskwld.zko.dec.compaq.hpe > Personal Opinion Onlye >p   ------------------------------  % Date: Thu, 27 Jun 2002 09:47:15 -0500n1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>iD Subject: Re: A possible shift in the status of VMS ar HP ???? ERRATA1 Message-ID: <aff8kn$s4n$1@fizban.pprd.abbott.com>P  K This reminds me of the recent editorial page cartoon I saw about the re-org  at the FBI.    Before:     Moe, Curly, Larryt   After:        Larry, Curly, Moe-   -- Dave...J  3 More than one cigar at a time is excessive smoking.  -----Mark TwainP  < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3D1A5AFD.D693088D@fsi.net...h > Dirk Munk wrote: > >>J > > Can someone please tell me when the Monthy Pyton team took over the US > > government ? >rC > When they inherited it from (the estate of) Curly, Moe and Larry.> >e > -- > David J. Dachteraa > dba DJE Systemsc > http://www.djesys.com/ >e* > Unofficial Affordable OpenVMS Home Page:! > http://www.djesys.com/vms/soho/    ------------------------------  % Date: Thu, 27 Jun 2002 09:53:56 -0500h1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>>D Subject: Re: A possible shift in the status of VMS ar HP ???? ERRATA1 Message-ID: <aff916$s7n$1@fizban.pprd.abbott.com>h  K Don't forget the free shipping for all the stereo equipment you accumulated L during your stay.  I still can't believe my speakers, amp and turntable made9 it home, still worked AND were not the least bit damaged.a  B I never got a steak dinner though.  My Mom took care of that.  ;-)   -- Dave...I  3 More than one cigar at a time is excessive smoking.s -----Mark Twaind  < "Terry C. Shannon" <terryshannon@attbi.com> wrote in message( news:dqsS8.317884$352.34840@sccrnsc02... >5: > "John E. Malmberg" <wb8tyw@qsl.network> wrote in message& > news:3D1A4476.2090407@qsl.network... > > John Smith wrote:oI > > > In the on-line newspaper world, its the latest news that's free andT the0K > > > archives that cost money. For Terry, it's the inverse. It costs moneyu to > fly I > > > to these places, and the man has to eat.....how else does Terry payb forn > > > that?  > > G > > Well Terry keeps refering being an ex-employee of some intelligencel	 > agency.  >aL > 509th RRG USASA 70-72. USASA being a military subsidiary of a better-known	 > agency.  >h > >gJ > > According to the movies and the tv shows in the U.S., ex-employee's ofI > > intelligence services have stashes of cash and "defensive" weapons in G > > safehouses all over the globe.  But of course they only use them inm" > > defense of the free world. :-) > L > Would that this was the case. Truth is, us lowly enlisted types didn't getB > much when we mustered out. I got a steak dinner, travel pay, and accumulatedmK > leave when I departed sunny Southeast Asia and stopped defending the freenL > world (or whatever the heck the reason was that we got ourselves into Viet > Nam). That was it. >rK > As for paying for travel and eating dinners these days, I have to rely ondJ > SKHPC and such consulting work as might rear its head from time to time. >o >n >  >    ------------------------------  # Date: Thu, 27 Jun 2002 15:04:50 GMT01 From: "Terry C. Shannon" <terryshannon@attbi.com>0D Subject: Re: A possible shift in the status of VMS ar HP ???? ERRATA. Message-ID: <mQFS8.336305$cQ3.20006@sccrnsc01>  I Point taken on "hold baggage." Great deals on stereo stuff over there. My J ex-mother in law still has my Pioneer receiver, and it still works fine 30 years later. ;-}  < "Dave Gudewicz" <david.gudewicz@abbott.com> wrote in message+ news:aff916$s7n$1@fizban.pprd.abbott.com... A > Don't forget the free shipping for all the stereo equipment youD accumulated4I > during your stay.  I still can't believe my speakers, amp and turntable  made; > it home, still worked AND were not the least bit damaged.  > D > I never got a steak dinner though.  My Mom took care of that.  ;-) >0 > --	 > Dave...l >:5 > More than one cigar at a time is excessive smoking.  > -----Mark Twain0 >0> > "Terry C. Shannon" <terryshannon@attbi.com> wrote in message* > news:dqsS8.317884$352.34840@sccrnsc02... > >a< > > "John E. Malmberg" <wb8tyw@qsl.network> wrote in message( > > news:3D1A4476.2090407@qsl.network... > > > John Smith wrote:0K > > > > In the on-line newspaper world, its the latest news that's free and  > the G > > > > archives that cost money. For Terry, it's the inverse. It costs0 money  > to > > flyrK > > > > to these places, and the man has to eat.....how else does Terry pay9 > for 
 > > > > that?0 > > >FI > > > Well Terry keeps refering being an ex-employee of some intelligenceF > > agency.0 > >0A > > 509th RRG USASA 70-72. USASA being a military subsidiary of a0 better-known > > agency.0 > >3 > > > L > > > According to the movies and the tv shows in the U.S., ex-employee's ofK > > > intelligence services have stashes of cash and "defensive" weapons in I > > > safehouses all over the globe.  But of course they only use them in $ > > > defense of the free world. :-) > > J > > Would that this was the case. Truth is, us lowly enlisted types didn't get D > > much when we mustered out. I got a steak dinner, travel pay, and
 > accumulated H > > leave when I departed sunny Southeast Asia and stopped defending the freeI > > world (or whatever the heck the reason was that we got ourselves into5 Viet > > Nam). That was it. > >rJ > > As for paying for travel and eating dinners these days, I have to rely onL > > SKHPC and such consulting work as might rear its head from time to time. > >. > >  > >v > >n >< >o   ------------------------------  # Date: Thu, 27 Jun 2002 08:12:18 GMTh. From: peter@langstoeger.at (Peter LANGSTOEGER)A Subject: Re: Amdrew wants numbers, here they are!  Blow out baby! 2 Message-ID: <CNzS8.61543$IR.770797@news.chello.at>   In article <rdeininger-2706020110500001@11cust138.tnt2.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes:a@ >All of the preview material I have seen names the entire familyA >GS-whatever.  I haven't seen any plans to continue the DS and ES K >designations in the EV7 line.  The systems will start at 2P, and the names2F >will start with GS.  Unless the marketing folks change their minds, I	 >suppose.m  L I've a nice little picture of an AlphaServer ES80. So it is GS1280/ES80, butL I don't know how the DS25 fits in the picture. Is the DS25 the standalone 2P5 box ? If yes, then EV7 has GS/ES/DS and not GS alone.-   -- - Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialistA E-mail  peter@langstoeger.atP A-1030 VIENNA  AUSTRIA              I'm looking for (a) Network _and_ VMS Job(s)   ------------------------------    Date: 27 Jun 2002 06:26:02 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)tA Subject: Re: Amdrew wants numbers, here they are!  Blow out baby!m3 Message-ID: <QzpwsOiTUstG@eisner.encompasserve.org>X  c In article <CNzS8.61543$IR.770797@news.chello.at>, peter@langstoeger.at (Peter LANGSTOEGER) writes:  > In article <rdeininger-2706020110500001@11cust138.tnt2.nashua.nh.da.uu.net>, rdeininger@mindspring.com (Robert Deininger) writes:nA >>All of the preview material I have seen names the entire family B >>GS-whatever.  I haven't seen any plans to continue the DS and ESL >>designations in the EV7 line.  The systems will start at 2P, and the namesG >>will start with GS.  Unless the marketing folks change their minds, Is
 >>suppose. > N > I've a nice little picture of an AlphaServer ES80. So it is GS1280/ES80, butN > I don't know how the DS25 fits in the picture. Is the DS25 the standalone 2P7 > box ? If yes, then EV7 has GS/ES/DS and not GS alone.h  B I would expect any system named DS25 to bear the same relationshipC to the current DS20 as the ES45 does to  the ES40.  In other words,t it would be based on EV6*.  @ The systems under discussion in this thread are the EV7 systems.   ------------------------------  % Date: Thu, 27 Jun 2002 07:43:06 -0400n2 From: rdeininger@mindspring.com (Robert Deininger)A Subject: Re: Amdrew wants numbers, here they are!  Blow out baby!sK Message-ID: <rdeininger-2706020743060001@11cust17.tnt2.nashua.nh.da.uu.net>h  N In article <CNzS8.61543$IR.770797@news.chello.at>, peter@langstoeger.at wrote:   >In articlemA <rdeininger-2706020110500001@11cust138.tnt2.nashua.nh.da.uu.net>,r4 rdeininger@mindspring.com (Robert Deininger) writes:A >>All of the preview material I have seen names the entire familyrB >>GS-whatever.  I haven't seen any plans to continue the DS and ESL >>designations in the EV7 line.  The systems will start at 2P, and the namesG >>will start with GS.  Unless the marketing folks change their minds, Ia
 >>suppose. >,M >I've a nice little picture of an AlphaServer ES80. So it is GS1280/ES80, but M >I don't know how the DS25 fits in the picture. Is the DS25 the standalone 2P06 >box ? If yes, then EV7 has GS/ES/DS and not GS alone.  J Hmm.  Maybe I'm mis-remembering the preview stuff I've seen.  Or maybe the renaming has already started.   A The DS25 will be an EV68 system.  It's the DS20 follow-on, and is8I architechturally very similar to half an ES45.  The box looks like a DS20 + box. I expect it will show up by this fall.   @ The EV7 2P box is not a DS25.  From 1 to 4 of these boxes can beI connected, yielding a 2, 4, 6 or 8 processor system.  I guess the maximum=J 8P configuration is called ES80.  I haven't seen names for the non-maximum configs.   ------------------------------  % Date: Thu, 27 Jun 2002 13:48:49 -04007- From: JF Mezei <jfmezei.spamnot@videotron.ca>=A Subject: Re: Amdrew wants numbers, here they are!  Blow out baby! , Message-ID: <3D1B5001.4074560A@videotron.ca>   Larry Kilgallen wrote:D > I would expect any system named DS25 to bear the same relationshipE > to the current DS20 as the ES45 does to  the ES40.  In other words,a > it would be based on EV6*.  A In the same logic, EV7 marvels shoudl not be called "GS" systems.   K If DS was low end, ES midrange and GS was those big wildfires, then I don't V see any attachement to EVx. They certaintly could continue the same naming convention.   ------------------------------  / Date: Thu, 27 Jun 2002 08:52:55 +0200 (MET DST)D& From: Rudolf Wingert <win@fom.fgan.de>0 Subject: Re: Andrew wan'ts the numbers, here ...6 Message-ID: <200206270652.IAA13212@sinet1.fom.fgan.de>   Hallo,  = Andrew did wrote a lot of number, but did not write, where we0; can find this number to check the conditions. Anybody could < write a lot os arguments. But I would like to see verifyable< statements. We do have an AlphaStation ES40 with four 600MHz: EV6 CPUs and 16GB memory and a SunFire 880 with the 900MHz; UltraSparc III. Now I can compare performance/price. Sorry,=; but we do not have an application running on both. The onlyA= one I could test is I/O. There was a small difference between== both. Alpha a little bit faster (under OpenVMS!!!!). Also themA boot sequence was much faster (I know this is meaningless because ? OpenVMS have to be booted only once). But the big difference isE7 the price. We did pay the double value for the SunFire.    Regards Rudolf Wingert   ------------------------------  % Date: Thu, 27 Jun 2002 07:34:02 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)P Subject: Re: BACKUP/INCREMENTAL not parsing directories properly during restoresK Message-ID: <rdeininger-2706020734020001@11cust17.tnt2.nashua.nh.da.uu.net>   1 In article <3d19c710.58847689@news.easynews.com>,t/ skulker@easynews.com.yourpants (skulker) wrote:b  1 >On Wed, 26 Jun 2002 13:38:17 GMT, "Mark E. Levy" " ><mlevy70-nospam@attbi.com> wrote: >d >o >>N >>When you're going through the incrementals, are you doing so in the order inL >>which they were created, or do you restore the newest first? Yes, you readN >>that right - when doing a volume restore from images/incrementals, you startM >>with the image, and then the incrementals in reverse chronological order. IoJ >>don't know if that would help your problem, but it's a common error with >>incremental restores.s >> >>Mark Levy  >>SMAo >s >n >Mark, >  >Thanks for the reply! > C >Yeah - I'm aware that you "should" perform incremental restores ine= >reverse chronological order - but I can't in this particular  >situation.C >OG >The "original" cluster is at a customer's site (we are the outsourcing8G >company). We have built an exact duplicate (bubble) of this cluster atiD >our site (2000 miles away). Our task is now to copy all the data onA >the original cluster (14Tbytes) to the bubble. We have a 24 hour4 >outage window.e >=
 >Our plan: > 9 >While the "original" cluster is doing Business As Usual: G >- Run IMAGE BACKUPs on "original" cluster per schedule (begins Friday) " >- Ship IMAGE tapes to bubble site! >- Restore IMAGEs to bubble disks05 >- Run DIFFERENTIAL BACKUPs daily on original cluster 4 >- Ship DIFFERENTIAL tape to bubble site on Thursday. >- Retore DIFFERENTIAL BACKUPS on bubble disks >  > D >Shutdown applications/user/etc of original cluster Friday (begin 24
 >hour outage) 1 >- Perform INCREMENTAL BACKUP on original cluster # >- Ship INCREMENTALs to bubble site4 >- Restore INCREMENTALs0 >RA >OK - We "think" that will work and we can do it with 1 - 24 hour B >outage. However, I cannot perform the INCREMENTAL restores in the >traditional way.P >7< >All that taken into consideration - is there a problem with9 >BACKUP/INCREMENTAL doing directories or it is my method?=    I I can't speak to the particular problem you are having here.  It would be I useful to have a bit more detail -- in particular the exact commands usedmI for the backup and restore operations.  Maybe I missed that in an earlier0 post.0  G You are on the right track here.  You need to completely test the whole ( operation before you actually deploy it.  F I find RAM disks, as provided by the DECram product or otherwise, veryD valuable for this kind of testing.  You can set up a number of smallH virtual disks containing all the features you care about, and then run aH complete series of backup and restore operations in a few seconds.  Your scripts go something like this:0  ' 1. Create and INIT the virtual devices. I 2. Populate the devices with directories, files, aliases, etc. that mimic H your real disks.  These files don't need to be large.  A block or two is  enough for most test situations.D 3. Run your BACKUP operations, perhaps interspersed with changes and& concurrent access on the source disks.M 4. Make a new set of virtual disks and run your restore operations onto them.0I 5. Run whatever situation-specific tests are needed to prove the restoredF disks meet your needs.  G You can very quickly learn the effect of various combinations of BACKUP D qualifiers.  You can understand what happens to files that are open,B marked /NOBACKUP, deleted during the operation, etc. etc.  You canE automate the whole test sequence, and save the scripts for regressiontJ testing whenever you deploy a new VMS version, a BACKUP ECO, or whatever. I If you do bump into a bug in BACKUP, these scripts are also a nice way to.G submit a bug report.  About the only things you CAN'T test this way arefA the physical data paths and the speed of a real BACKUP operation.u  H For anyone who takes their BACKUPs seriously and values their time, thisF is probably sufficient justification to buy a DECram license.  In yourF case, I'd at least request one of the free 30-day evaluation licences.  I In short, if you aren't comfortable with your own understanding of BACKUPEH and its features, limitations, and perhaps bugs, you haven't done enoughF testing.  You aren't ready to deploy a mission-critical BACKUP scheme.   ------------------------------  / Date: Thu, 27 Jun 2002 09:21:51 +0200 (MET DST)a& From: Rudolf Wingert <win@fom.fgan.de>E Subject: Deutsche Bank would like to outsource there IT to IBM or CSCm6 Message-ID: <200206270721.JAA13261@sinet1.fom.fgan.de>   Hello,  C if this is right, what mean that for OpenVMS? Do Deutsche Bank haveaF OpenVMS? If yes, could IBM or CSC give them fully support for this OS?   TIA and regards Rudolf Wingert   ------------------------------  # Date: Thu, 27 Jun 2002 08:11:12 GMT ? From: Jim.Johnson@software-exploration.nospam.com (Jim Johnson)rI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCg/ Message-ID: <3d1ac85a.5513898@news.demon.co.uk>l  C I don't know if they do or don't, but I'd be amazed if either would7F not support a VMS installation under the correct incentives (much cash) now, or control now and more cash later).a   Jim.  < On Thu, 27 Jun 2002 09:21:51 +0200 (MET DST), Rudolf Wingert <win@fom.fgan.de> wrote:   >Hello,a >aD >if this is right, what mean that for OpenVMS? Do Deutsche Bank haveG >OpenVMS? If yes, could IBM or CSC give them fully support for this OS?t >  >TIA and regards Rudolf Wingerts >i   Jim Johnson  Software Exploration, Ltd.) (remove '.nospam' from the reply address)    ------------------------------  % Date: Thu, 27 Jun 2002 05:50:34 -0400n- From: JF Mezei <jfmezei.spamnot@videotron.ca>kI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCt, Message-ID: <3D1ADFE9.7EBFD2F0@videotron.ca>   Rudolf Wingert wrote: E > if this is right, what mean that for OpenVMS? Do Deutsche Bank have H > OpenVMS? If yes, could IBM or CSC give them fully support for this OS?  E DB owns what used to be Bankers Trust. BT was instrumental in gettingdL significant VMS-MVS data connectivity software developped. However, in laterM years, execs from BT were also instrumental in refusing VMS solutions because M "VMS will die soon". So I don't know how much VMS Bankers Trust still has. WeyK don't hear about their disaster recovery plans with VMS not having missed a,N tick on sept 11 (they NYC building was across the street from WTC and suffered serious damage).  6 BT in new York was a big Compaq shop for wintel crap.   M Yes, IBM is fully capable of supporting non-IBM platforms, and as a matter ofrK fact, thet probably will want to make that as public as possible since they.N want their outsourcing arm to get as many contracts as possible. Also, if theyK get a contract to support non-IBM gear, in the longer term, they stand morep% chances of converting it to IBM gear.e  K This would be most interesting in the current environment of Tru64, VMS andoS HP-UX whose future is uncertain exept for a certain need to port to a new hardware.e   ------------------------------    Date: 27 Jun 2002 06:23:31 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)gI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSC 3 Message-ID: <sNv+1I61fHA3@eisner.encompasserve.org>t  q In article <3d1ac85a.5513898@news.demon.co.uk>, Jim.Johnson@software-exploration.nospam.com (Jim Johnson) writes:vE > I don't know if they do or don't, but I'd be amazed if either wouldoH > not support a VMS installation under the correct incentives (much cash+ > now, or control now and more cash later).y >  > Jim. > > > On Thu, 27 Jun 2002 09:21:51 +0200 (MET DST), Rudolf Wingert > <win@fom.fgan.de> wrote: >  >>Hello, >>E >>if this is right, what mean that for OpenVMS? Do Deutsche Bank havetH >>OpenVMS? If yes, could IBM or CSC give them fully support for this OS? >>  >>TIA and regards Rudolf Wingert  G Both IBM and CSC have quite competent VMS people among their employees. ; One of the CSC folk is a major contributor to comp.os.vms !    ------------------------------    Date: 27 Jun 2002 07:17:40 -0600- From: koehler@encompasserve.org (Bob Koehler)tI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCA3 Message-ID: <pr7E5Nc4$8AV@eisner.encompasserve.org>   _ In article <200206270721.JAA13261@sinet1.fom.fgan.de>, Rudolf Wingert <win@fom.fgan.de> writes:h > Hello, > E > if this is right, what mean that for OpenVMS? Do Deutsche Bank havezH > OpenVMS? If yes, could IBM or CSC give them fully support for this OS?  B    Assuming the initials CSC stands for us, we certainly can fully1    support VMS.  IIRC, that's about all Nic does.    ------------------------------  % Date: Thu, 27 Jun 2002 14:02:51 +0100iU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>.I Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSC 0 Message-ID: <aff2c4$7vm$1@new-usenet.uk.sun.com>   Rudolf Wingert wrote:    > Hello, > E > if this is right, what mean that for OpenVMS? Do Deutsche Bank havesH > OpenVMS? If yes, could IBM or CSC give them fully support for this OS? >   > TIA and regards Rudolf Wingert >     > Deutsche Bank bought Bankers Trust which was a Tru64/HP-UX and OpenVMS house.  D Deutsche Bank have two UNIX suppliers Sun and IBM, their investementE in OpenVMS via BT has declined and I currently only know of one majore7 app in the bank on OpenVMS, though there may be others.a   Regards  Andrew Harrison    ------------------------------  % Date: Thu, 27 Jun 2002 15:41:55 +0200e' From: Brass Christof <welcome@spam.not>SI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCy( Message-ID: <3D1B1623.7E6FD709@spam.not>   Rudolf Wingert wrote:  >  > Hello, > E > if this is right, what mean that for OpenVMS? Do Deutsche Bank havevH > OpenVMS? If yes, could IBM or CSC give them fully support for this OS? >   > TIA and regards Rudolf Wingert  C The Deutsche Bank is an almost exclusively IBM shop wrt HW and SW. l< They still use a lot of desktops with OS/2 as far as I know.  ? OTOH the normal process of outsourcing in Germany in this case  @ would require IBM to employ the former workers of Deutsche Bank @ and probably buying the equipment. Sometimes the working places @ are even kept, i.e. the building or facitlities of the Deutsche  Bank would be used by IBM.   -- s? According to the Quality Assurance Institute C/C++/ObjC, PERL,  @ UNIX (incl. Linux) and Windows/XY are regarded as harmful. Java 0 is slow and the class library is badly designed.7 moc dot slupofni at ssarb - please reverse the sequence(   ------------------------------  % Date: Thu, 27 Jun 2002 09:37:57 -0500 1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>.I Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSC.1 Message-ID: <aff83a$s1m$1@fizban.pprd.abbott.com>i  J From what little I understand about this, (heard at LUG meetings) they areH building several large VMS systems.  A couple in Chicago and in Germany.   -- Dave...n  3 More than one cigar at a time is excessive smoking.a -----Mark Twain   4 "Brass Christof" <welcome@spam.not> wrote in message" news:3D1B1623.7E6FD709@spam.not... > Rudolf Wingert wrote:r > >.
 > > Hello, > >EG > > if this is right, what mean that for OpenVMS? Do Deutsche Bank havesJ > > OpenVMS? If yes, could IBM or CSC give them fully support for this OS? > >e" > > TIA and regards Rudolf Wingert >nD > The Deutsche Bank is an almost exclusively IBM shop wrt HW and SW.> > They still use a lot of desktops with OS/2 as far as I know. > @ > OTOH the normal process of outsourcing in Germany in this caseA > would require IBM to employ the former workers of Deutsche BankeA > and probably buying the equipment. Sometimes the working placessA > are even kept, i.e. the building or facitlities of the Deutsche0 > Bank would be used by IBM. >o > --@ > According to the Quality Assurance Institute C/C++/ObjC, PERL,A > UNIX (incl. Linux) and Windows/XY are regarded as harmful. Javac2 > is slow and the class library is badly designed.9 > moc dot slupofni at ssarb - please reverse the sequencee   ------------------------------  + Date: Thu, 27 Jun 2002 14:48:14 +0000 (UTC) % From: John Forkosh <john@invalid.com>dI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCm, Message-ID: <aff8jd$4in$1@reader2.panix.com>  . JF Mezei <jfmezei.spamnot@videotron.ca> wrote: : Rudolf Wingert wrote:hG : > if this is right, what mean that for OpenVMS? Do Deutsche Bank havefJ : > OpenVMS? If yes, could IBM or CSC give them fully support for this OS?  G : DB owns what used to be Bankers Trust. BT was instrumental in gettingoN : significant VMS-MVS data connectivity software developped. However, in laterO : years, execs from BT were also instrumental in refusing VMS solutions because-O : "VMS will die soon". So I don't know how much VMS Bankers Trust still has. WebM : don't hear about their disaster recovery plans with VMS not having missed ayP : tick on sept 11 (they NYC building was across the street from WTC and suffered : serious damage).  N I had a contract at BT several years ago (www.forkosh.com/resume.html#bt0697),J and can assure you their money transfer infrastructure is (was) so heavilyE tied to VMS architecture that porting it would be next-to-impossible. C Until they're willing to subsidize a total rewrite (and significantaA redesign to decouple the code from VMS features), they'll have to  keep a noticeable VMS presence.dF      Moreover, the code is not only spaghetti, but angel hair at that.? Outsourcing would only be possible if the outsourcer takes on alB significant fraction of the programmers who already know the code.? Of course, that's pretty much par for the course, so maybe it's D not a problem.  But it's certainly a risk (also par for the course).C      As customary with outsourcing, a lot of management and supportrA could be absorbed by the outsourcer.  But, as I recall, BT wasn'toF near perfect, but neither was it near as bad as some places I've been.4 So it's not clear to me what savings they'd realize. --  > John Forkosh  ( mailto:  j@f.com  where j=john and f=forkosh )   ------------------------------  # Date: Thu, 27 Jun 2002 15:03:50 GMT 2 From: "Sue Skonetski" <susan.skonetski@compaq.com>I Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSC-2 Message-ID: <qPFS8.21$%P6.357432@news.cpqcorp.net>  E Both HP and IBM have very large services departments with a number ofb OpenVMS specialists.  K JF there are a number of major banks/stock exchanges/financial institutions L that use OpenVMS.  There are several reasons that you may not see the actual names.  C 1.  Many customers view their computer environment as a competitiveuB advantage and if they disclosed what they used they would lose the
 advantage.  J 2.  Many customers feel that the more information you publicly provide the easier you make it for hackers.   G 3.  For any company to use a specific customer name vs something like a-K "well recognized chip manufacturer" requires that a document be written and J reviewed a number of times and then approved of by the legal department inD both companies.  Then after that is done either company can pull the testimonial at any time.  L If you do want to see what testimonials are publicly available please visit.K http://www.openvms.compaq.com/openvms/brochures/ scroll down, there are 36.    suei      3 "Rudolf Wingert" <win@fom.fgan.de> wrote in messageu0 news:200206270721.JAA13261@sinet1.fom.fgan.de... > Hello, >0E > if this is right, what mean that for OpenVMS? Do Deutsche Bank have H > OpenVMS? If yes, could IBM or CSC give them fully support for this OS? >   > TIA and regards Rudolf Wingert >w   ------------------------------  % Date: Thu, 27 Jun 2002 16:05:25 +0100sU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> I Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCl0 Message-ID: <aff9ht$a8r$2@new-usenet.uk.sun.com>   Brass Christof wrote:-   > Rudolf Wingert wrote:0 >  >>Hello, >>E >>if this is right, what mean that for OpenVMS? Do Deutsche Bank haveUH >>OpenVMS? If yes, could IBM or CSC give them fully support for this OS? >>  >>TIA and regards Rudolf Wingert >> > E > The Deutsche Bank is an almost exclusively IBM shop wrt HW and SW. >> > They still use a lot of desktops with OS/2 as far as I know. > A > OTOH the normal process of outsourcing in Germany in this case  B > would require IBM to employ the former workers of Deutsche Bank B > and probably buying the equipment. Sometimes the working places B > are even kept, i.e. the building or facitlities of the Deutsche  > Bank would be used by IBM. >  >     @ You are refering to the retail bank, the Financial services part@ of the bank which is mostly where BT for example went is Sun and? IBM with I think Compaq desktops though one peice of Wintel tin   looks the same as another to me.   Regardsa Andrew Harrison!   ------------------------------  % Date: Thu, 27 Jun 2002 13:57:15 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>hI Subject: Re: Deutsche Bank would like to outsource there IT to IBM or CSCs, Message-ID: <3D1B51FA.9B396F7D@videotron.ca>   Sue Skonetski wrote:M > JF there are a number of major banks/stock exchanges/financial institutionswN > that use OpenVMS.  There are several reasons that you may not see the actual > names. > E > 1.  Many customers view their computer environment as a competitivedD > advantage and if they disclosed what they used they would lose the > advantage.  I Then Sue, perhaps you should find a way for senior execs outside of IT inlM those banks to be aware of the importance VMS plays in their own bank insteadRN of allowing them to make departmental decisions with the idea that VMS is dead and should be gotten rid of.   ------------------------------  % Date: Thu, 27 Jun 2002 12:40:32 -0500P1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>S% Subject: Encompass - DECUS VMS surveys0 Message-ID: <affiqa$hk$1@fizban.pprd.abbott.com>  * There is a VMS / Itanium survey posted at:   http://www.encompassus.org/t  ' Part of the online advocacy initiative.l   Dave...    ------------------------------    Date: 27 Jun 2002 06:18:21 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)d: Subject: Re: how detect all batch queues in a VMS system ?3 Message-ID: <9acnXJg$Tbkr@eisner.encompasserve.org>a  m In article <ddf392ea.0206261605.5289bbad@posting.google.com>, contracer11@uol.com.br (Shiva MahaDeva) writes:aJ > Im looking for a Pascal program to detect and show me all Batch queues  > in a VMSsystem.   = I am looking for a Pascal program to detect a duplicate postsi in a VMS newsgroup...c   ------------------------------   Date: 27 Jun 2002 07:27:36 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann)t0 Subject: Re: How detect Batch queues in Pascal ?0 Message-ID: <afeep8$njp$1@n.ruf.uni-freiburg.de>  l In article <ddf392ea.0206261558.7e55f4a@posting.google.com>, contracer11@uol.com.br (Shiva MahaDeva) writes:J >Im looking for a Pascal program to detect all Batch queues in my system. >How can I make this ?   Seel4    SYS$COMMON:[SYSHLP.EXAMPLES.PASCAL]SYS$GETQUI.PAS   Regards,    Christoph Gartmann,  H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, Germany                                           |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  % Date: Thu, 27 Jun 2002 09:32:29 -0500s$ From: "Art Beane" <beane@petris.com> Subject: HP HPTC in Oil & Gas,7 Message-ID: <007501c21de7$7720e2d0$352810ac@petris.com>n  , This is a multi-part message in MIME format.  + ------=_NextPart_000_0076_01C21DBD.8E4ADAD0  Content-Type: text/plain;h 	charset="us-ascii"  Content-Transfer-Encoding: 7bite  E You know I only go to these "seminars" for the free lunch, shirt, and F binder: HP held a marketing event yesterday to promote HPTC in the O&GH sector yesterday at the "Solutions Center of the Americas" (what we used to call the "Compaq Campus").   A There was a huge focus on linux and high performance (1000+ cpus)rD cluster capabilities, including a number of live demos including theF soon-to-be-fully-supported Landmark application suite running on LinuxE (32-bit). While mentioned, they did not push HP-UX very hard. I foundTC that interesting because there is a growing interest in large linuxuF clusters to reduce the cost of seismic processing. It's one of the fewH times I've seen a major vendor ready to "flow with the crowd" instead of9 trying to convince them that the vendor's way is better. d  C There was also a lot of talk about why HP McKinley products runningoG linux will outperform Sun. There seem to be two primary reasons: the HP:C proprietary "ZX" chipset (boosts system performance over pure IntelhD chipsets) and highly optimizing compilers. And of course they showedF many benchmarks to prove the point. Most comparisons were to Itanium 1E (700 MHz)  (I2 will be about 1.8 to 2.0 times faster for 1GHz chips).pH The also showed Sun US III 1GHz at about a wash for fp and 30-40% faster7 than the Itanium 1. Sun vs I2 numbers were "estimates".n  	 Art Beane1 Petris Technology, Inc.W 1900 St. James Place, Suite 700  Houston, TX 77056  Phone: 713-403-8423e Fax: 713-956-2185? http://www.petris.com     + ------=_NextPart_000_0076_01C21DBD.8E4ADAD0y Content-Type: text/html; 	charset="us-ascii"d+ Content-Transfer-Encoding: quoted-printablep  1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">e <HTML> <HEAD>9 <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =0 charset=3Dus-ascii">@ <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version = 6.0.4630.0">' <TITLE>HP HPTC in Oil &amp; Gas</TITLE>3 </HEAD>E <BODY>' <!-- Converted from text/rtf format -->   > <P><FONT SIZE=3D2 FACE=3D"Arial">You know I only go to these =G &quot;seminars&quot; for the free lunch, shirt, and binder: HP held a =OA marketing event yesterday to promote HPTC in the O&amp;G sector = H yesterday at the &quot;Solutions Center of the Americas&quot; (what we =8 used to call the &quot;Compaq Campus&quot;). </FONT></P>  F <P><FONT SIZE=3D2 FACE=3D"Arial">There was a huge focus on linux and =H high performance (1000+ cpus) cluster capabilities, including a number =A of live demos including the soon-to-be-fully-supported Landmark = H application suite running on Linux (32-bit). While mentioned, they did =G not push HP-UX very hard. I found that interesting because there is a = H growing interest in large linux clusters to reduce the cost of seismic =I processing. It's one of the few times I've seen a major vendor ready to =tI &quot;flow with the crowd&quot; instead of trying to convince them that = ' the vendor's way is better. </FONT></P>n  I <P><FONT SIZE=3D2 FACE=3D"Arial">There was also a lot of talk about why = J HP McKinley products running linux will outperform Sun. There seem to be =H two primary reasons: the HP proprietary &quot;ZX&quot; chipset (boosts =D system performance over pure Intel chipsets) and highly optimizing =J compilers. And of course they showed many benchmarks to prove the point. =J Most comparisons were to Itanium 1 (700 MHz)&nbsp; (I2 will be about 1.8 =I to 2.0 times faster for 1GHz chips). The also showed Sun US III 1GHz at =.E about a wash for fp and 30-40% faster than the Itanium 1. Sun vs I2 = . numbers were &quot;estimates&quot;.</FONT></P>  F <P><B><I><FONT SIZE=3D2 FACE=3D"Arial">Art Beane</FONT></I></B><I></I>  ? <BR><I><FONT COLOR=3D"#008000" SIZE=3D2 FACE=3D"Arial">Petris =s Technology, Inc.</FONT></I>   H <BR><FONT SIZE=3D2 FACE=3D"Arial">1900 St. James Place, Suite 700</FONT>  : <BR><FONT SIZE=3D2 FACE=3D"Arial">Houston, TX 77056</FONT>  < <BR><FONT SIZE=3D2 FACE=3D"Arial">Phone: 713-403-8423</FONT>  : <BR><FONT SIZE=3D2 FACE=3D"Arial">Fax: 713-956-2185</FONT>  B <BR><A HREF=3D"http://www.petris.com"><U><FONT COLOR=3D"#0000FF" =< SIZE=3D2 FACE=3D"Arial">http://www.petris.com</FONT></U></A> </P>   </BODY>  </HTML>m- ------=_NextPart_000_0076_01C21DBD.8E4ADAD0--    ------------------------------    Date: 27 Jun 2002 09:53:52 -0600+ From: young_r@encompasserve.org (Rob Young)o! Subject: Re: HP HPTC in Oil & Gas 3 Message-ID: <hcJFOXAk8ngM@eisner.encompasserve.org>   ^ In article <007501c21de7$7720e2d0$352810ac@petris.com>, "Art Beane" <beane@petris.com> writes:   > G > You know I only go to these "seminars" for the free lunch, shirt, andkH > binder: HP held a marketing event yesterday to promote HPTC in the O&GJ > sector yesterday at the "Solutions Center of the Americas" (what we used  > to call the "Compaq Campus").  > C > There was a huge focus on linux and high performance (1000+ cpus) F > cluster capabilities, including a number of live demos including theH > soon-to-be-fully-supported Landmark application suite running on LinuxG > (32-bit). While mentioned, they did not push HP-UX very hard. I foundnE > that interesting because there is a growing interest in large linuxrH > clusters to reduce the cost of seismic processing. It's one of the fewJ > times I've seen a major vendor ready to "flow with the crowd" instead of; > trying to convince them that the vendor's way is better. t >     = 	This is what is called "smart business practices."  The kindU> 	of Business Intelligence is as simple as surveying the target< 	audience about their needs and intended directions.  If 90%? 	put "cost containment" as a top priority in a survey, it makesrC 	little sense to steer them elsewhere.  Especially given the highly0? 	(embarassingly?) parallelizable nature of their app.  That andjG 	performance as a top criteria.... Given those conditions, pretty much eF 	dictates P4 boxes (no better bang for the buck), so then which OS to H 	run, certainly not HP/UX.  I'd say they are forced to push Linux today.  
 	Yes, forced.M    E > There was also a lot of talk about why HP McKinley products running I > linux will outperform Sun. There seem to be two primary reasons: the HPME > proprietary "ZX" chipset (boosts system performance over pure IntelaF > chipsets) and highly optimizing compilers. And of course they showed& > many benchmarks to prove the point.   F 	Ah you went and spoiled all the McKinley FUD fun.  Are you sure?  AndA 	what were they comparing it to?  Old Sun kit.... blah blah blah.i  A 	When does it become necessary to go 64-bit?  Until that time, P4n 	makes more sense, right?I  @ 	So there is a bit of technical prowess in the ZX.  Can you shed% 	more light as to whys and wherefors?    				Robi   ------------------------------  % Date: Thu, 27 Jun 2002 11:54:02 -0500q$ From: "Art Beane" <beane@petris.com>! Subject: Re: HP HPTC in Oil & Gas57 Message-ID: <009e01c21dfb$3d7156f0$352810ac@petris.com>m   -----Original Message-----3 From: Rob Young [mailto:young_r@encompasserve.org] s& Sent: Thursday, June 27, 2002 10:54 AM To: Info-VAX@Mvb.Saic.ComI! Subject: Re: HP HPTC in Oil & GasU   <snip>A 	When does it become necessary to go 64-bit?  Until that time, P4o 	makes more sense, right?e </snip>oE 64-bit for seismic has been around for quite a while. Current seismic G surveys are 1.5 to 2 TB in size. Processing is embarrassingly parallel,7C and generally requires significant display capabilities (two-headed E workstations). In the "old days" everything ran on Cray. SGI/MIPS and E SUN have been very popular in more recent years. It's only with large?F scale linux clusters that the P4 has even been considered. My guess isF that large P4 linux clusters are useful as a learning experience until 64-bit systems are available.k  C My personal opinion is that 2-P 64-bit blade servers (256+ CPUs peraG rack) would make killer seismic systems, but the server guys don't want B to build them because of power and heat dissipation problems while running at full speed.   <snip>@ 	So there is a bit of technical prowess in the ZX.  Can you shed% 	more light as to whys and wherefors?n </snip>uH Nope. They didn't expand on anything more than existence of the chipset.C "Real" system details may be available under NDA, but without one I + guess we have to wait another couple weeks.e   ------------------------------  # Date: Thu, 27 Jun 2002 17:34:55 GMT:* From: "Bill Todd" <billtodd@metrocast.net>! Subject: Re: HP HPTC in Oil & GashB Message-ID: <31IS8.152928$_j6.8292962@bin3.nnrp.aus1.giganews.com>  / "Art Beane" <beane@petris.com> wrote in messagee1 news:009e01c21dfb$3d7156f0$352810ac@petris.com...o   ...   E > My personal opinion is that 2-P 64-bit blade servers (256+ CPUs per I > rack) would make killer seismic systems, but the server guys don't want D > to build them because of power and heat dissipation problems while > running at full speed.  F 64 bits?  Low cost?  Dual-processor?  Smoking Bricks of Death need not apply?  - Sounds like the ideal Clawhammer application.e   - bill   ------------------------------  + Date: Thu, 27 Jun 2002 10:31:55 +0000 (UTC)  From: david20@alpha1.mdx.ac.uk Subject: Re: IPSEC ?+ Message-ID: <afepir$kcj$1@aquila.mdx.ac.uk>e  W In article <afd2gl$t5u$2@web1.cup.hp.com>, "Leo Demers" <leo_dot_demers@HP.COM> writes:o
 >Hi Larry,I >   In the meantime we are making a port of OpenSSL 0.9.6b available withuH >7.3-1 we also have a port of Stunnel done that ships on the Open SourceL >Tools CD.  So folks will be able to secure telnet, ftp(control channel onlyM >no data encryption), RCP, POP etc. without the applications being rewritten.rH > OpenVMS engineering is feeding back the changes to the respective open% >source teams for these two projects.t >b  M I'm already using OpenSSL 0.9.6d on my systems. I presume that replacing yoursK shipping OpenSSL version in 7.3-1 with a later release won't break anythingfL else in VMS which you've built to rely on OpenSSL ie The shipping OpenSSL is3 a standard release without any special alterations..  M Great to hear about the Stunnel port. Will it install on previous versions ofaN VMS ? The timing of the VMS 7.3-1 release may not allow me to put it on beforeN the students come back - which would mean I then I'd be stuck at VMS 7.3 untilK next summer. I also have some systems stuck at VMS 6.2 - I don't suppose ita would install on those.-    
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Thu, 27 Jun 2002 08:55:43 -0400u* From: "Leo Demers" <Leo_dot_Demers@HP.COM> Subject: Re: IPSEC ?* Message-ID: <aff2fp$kgk$1@web1.cup.hp.com>   > J > I'm already using OpenSSL 0.9.6d on my systems. I presume that replacing yourD > shipping OpenSSL version in 7.3-1 with a later release won't break anythingK > else in VMS which you've built to rely on OpenSSL ie The shipping OpenSSLe is5 > a standard release without any special alterations.  >oD The OpenSSL ships with OpenVMS 7.3-1 as an optional layered product.L We will not intergrate it into the base O/S until OpenSSL 1.0 this is due to thecA fact that the OpenSSL API's will not be backward compatible untilpI OpenSSL 1.0.  So products using SSL today either look for the SSL code onc
 the systemG or ship with thier own version of OpenSSL bundled into the product likes CSWS. (aka Apache)J So support for your configuration could be an issue as the API's in 0.9.6dI may not be the same and 0.9.6.b what we've tested with. Now I don't thinkhK this is the case but we don't have the ability/time to test all versions toa- determine if all of our products work or not.   L > Great to hear about the Stunnel port. Will it install on previous versions ofI > VMS ? The timing of the VMS 7.3-1 release may not allow me to put it onh beforeJ > the students come back - which would mean I then I'd be stuck at VMS 7.3 untiltJ > next summer. I also have some systems stuck at VMS 6.2 - I don't suppose it > would install on those.  >aE Stunnel ships as a full Open Source project with the source and buildl instructions that we usedoL to make it work on 7.3-1.  An executable image will be shipping with the kit as well that worksG for 7.3-1. The rest is up to you as an Open Source consumer.  Note: Theo version that we areeH providing with 7.3-1 calls the OpenSSL 0.9.6.b shareable images from the OpenSSL kit.   - Leov   ------------------------------  # Date: Thu, 27 Jun 2002 07:25:53 GMTO1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie) 5 Subject: Re: My conversation with Linus about VMS ... : Message-ID: <56zS8.11530$4a2.606706@typhoon.austin.rr.com>  . JF Mezei (jfmezei.spamnot@videotron.ca) wrote: : Lyndon Bartels wrote:nK : > Of course, VMS doesn't have a mail command named after the programmer'sn : > dog either.y : ? : How do you know that no VMS engineer named his dog "mail" or , : "all-in-1" ???????  7 I worked with a programmer who named her dog "Fortran".   H --Jerry Leslie   leslie@clio.rice.edu  (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for emailS   ------------------------------  % Date: Thu, 27 Jun 2002 18:57:32 +0010o% From: paddy.o'brien@zzz.tg.nsw.gov.aud5 Subject: Re: My conversation with Linus about VMS ... 5 Message-ID: <01KJFP806K0I0000AG@tgmail.tg.nsw.gov.au>i   Christof Brass wrote:o   >Lyndon Bartels wrote: >> aK >> It's been my experience, it seems that many people are biggotted towards9D >> whatever operating system they "grew up on". Same seems to go forG >> vehicles (Ford, versus Chevy, versus Dodge, versus whatever.) People K >> build an emotional bond with it, and there's nothing that can be said or. >> done to disuade them. >oA >This is not my personal experience. VMS was my fourth OS and it o >*convinced* me.  L Nor me either.  I worked on IBM JCL, etc. from 1960's to 1982.  I worked on 2 CDC NOS/BE from 1983 to 1985, thereonafter on VMS.  M After my earlier bigotry towards IBM, I initialially thought VMS was "Mickey  N Mouse" and something I suffered reluctantly.  I am now an OpenVMS bigot.  Not H the system (I'll forget cars) that I first started on, but probably the I inverse (could/would be the same with cars -- I'm happier with my 2000's c- model than my 1960's model of whatever make).$   Regards, Paddy   ------------------------------  % Date: Thu, 27 Jun 2002 08:15:14 -0400m2 From: Atlant Schmidt <atlantnospam@mindspring.com>5 Subject: Re: My conversation with Linus about VMS ...a. Message-ID: <3D1B01D2.CA9BA4AE@mindspring.com>   Lyndon Bartels wrote:l  I > I said that no matter what, as long as it's alpha, I can take my system I > disk and move it from one box to another, and it'll work. I may have to.= > retune, but I can get a basic OS running with no changes...a >a9 > The person's response.... "Wow,  that's *REALLY* nice."d  & It'd be even better if it were *TRUE*. 'Shame it's not, ehh?r   Atlant   ------------------------------  + Date: Thu, 27 Jun 2002 13:23:43 +0000 (UTC)- From: david20@alpha1.mdx.ac.uk5 Subject: Re: My conversation with Linus about VMS ...p+ Message-ID: <aff3kv$nhk$1@aquila.mdx.ac.uk>u  c In article <3D1B01D2.CA9BA4AE@mindspring.com>, Atlant Schmidt <atlantnospam@mindspring.com> writes:e >Lyndon Bartels wrote: >cJ >> I said that no matter what, as long as it's alpha, I can take my systemJ >> disk and move it from one box to another, and it'll work. I may have to> >> retune, but I can get a basic OS running with no changes... >>: >> The person's response.... "Wow,  that's *REALLY* nice." >-' >It'd be even better if it were *TRUE*.a >'Shame it's not, ehh? >  >Atlant  >-  > Well it is TRUE for VMS subject to just four qualifications :-  ; 1) The disk is supported on the system you are moving to eg L    If your system disk is a wide scsi disk and you are moving it to a system:    with only narrow scsi shelves you might have a problem.  J 2) VMS is actually supported on the destination Alpha box ie it wasn't oneK    of the infamous "whitebox" systems. (Though for a number of systems thise#    limitation can now be overcome).   K 3) The destination system supports that version of VMS which is being moveduI    across ie moving an old version of VMS say VMS 6.2 to a much newer boxDH    which is only qualified to work with say VMS 7.2 or above won't work.I    Moving a new version of VMS to an older Alpha system should currently ,    always work - see below.   K 4) The firmware of the destination system has been updated to be compatible ;    with the version of VMS on the disk being moved across.  3    (If it isn't then this is fairly quickly fixed).     K There will shortly be a further qualification since the VMS 7.3 SPD for theiM first time listed some Alpha systems which are supported on this version and uN previous versions of VMS but which will not be supported on future versions ofK VMS. Whether this "future" version is VMS 7.3-1 or a still future VMS 7.4 , 0 VMS 8.0 awaits publication of the VMS 7.3-1 SPD.  'N Once you have got the basic VMS system booting you will then obviously need to1 do some customization for your new environment eg'M Tell it about what other disks are available, retune for the amount of memoryt% available, setup new IP addresses etcT    sG VMS is a dream compared to many other systems when it comes to disastersJ recovery (Though with disaster recovery it is usually restoring the systemL disk of a destroyed system to a new system via tapes rather than by having aN spare copy of the original system disk. However the principals are exactly theN same - just one extra step ie boot from CD to enable you to restore the system disk from tape).  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  # Date: Thu, 27 Jun 2002 17:45:46 GMT " From: Alfred Falk <falk@arc.ab.ca>5 Subject: Re: My conversation with Linus about VMS ...09 Message-ID: <Xns923A77A851307falkarcabca@205.233.108.180>t  * Brass Christof <welcome@spam.not> wrote in! news:3D1AA5EF.AC8B80D6@spam.not: I   > Lyndon Bartels wrote:" >> HC >> It's been my experience, it seems that many people are biggotted H >> towards whatever operating system they "grew up on". Same seems to goD >> for vehicles (Ford, versus Chevy, versus Dodge, versus whatever.)G >> People build an emotional bond with it, and there's nothing that canT# >> be said or done to disuade them.n   Hmmm.... Lemme see:;     	Monitor I (IBM 1620) 
     	KM-9 +     	OS/360 (nothing like proficient, ever)e     	TOPS-10*     	KRONOS (is that right?  CDC Cyber 70)     	DOS-15t.     	MTS (Michigan Terminal System) - UGGGGHH!     	Unix (V3, I think)t     	RDOS (DG Nova and Eclipse)>     	RSX-11M     	RSTS-11     	AOS (DG Eclipse)g     	VMS
     	RT-11     	AOS/VS (DG 32-bit Eclipse)l	     	CP/Mi     	MS-DOSd     	Windows 3.1 and its heirs  D No, I don't think I long to return to Monitor I or KM-9, not in the D least.  The only OSes I might get nostalgic for are TOPS-10 and AOS,? and they're certainly not where I began, nor where I'll end up.    -- t@ ----------------------------------------------------------------A   A L B E R T A         Alfred Falk               falk@arc.ab.ca  @ R E S E A R C H         Information Systems Dept   (780)450-5185+   C O U N C I L         250 Karl Clark Road<1                         Edmonton, Alberta, Canadaa http://www.arc.ab.ca/   T6N 1E4P  http://www.arc.ab.ca/staff/falk/   ------------------------------   Date: 27 Jun 2002 07:31:00 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren) Subject: Re: Open Letter to HP0 Message-ID: <afeevk$b9l$1@pegasus.csx.cam.ac.uk>  B In article <7fxS8.143551$_j6.7975180@bin3.nnrp.aus1.giganews.com>,) Bill Todd <billtodd@metrocast.net> wrote:n >;6 >"Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in message+ >news:afd88j$auj$1@pegasus.csx.cam.ac.uk... / >> In article <3D19F66F.7B72731D@videotron.ca>,a2 >> JF Mezei  <jfmezei.spamnot@videotron.ca> wrote: >> >Nick Maclaren wrote:D >> >> I would expect that 95% of the IA-64 port work could be reusedE >> >> if the target were changed to any other design, whether x86-64,o/ >> >> SPARC, PA-RISC, POWER4, MIPS or even ARM.D >> >G >> >I am not so sure about that. Consider Tandem that needs lockstep. Ii >wonder howTI >> >enthousiastic Tandem ISVs are with regards to the port to IA64. UntilN	 >they getFF >> >functional Tandem IA64 machines available to ISVs, not much native	 >softwareeK >> >could be written for it. So they are lucky in the sense that ISVs won'tt >have<K >> >wasted much money porting their tandem software IF HP decides to do theD >right5 >> >thing and abandon IA64 a few years down the road.2 >> >> All right - drop ARM :-)@ >>G >> All of the others are designed for SMP at least as much as IA-64 is.D >hH >That's true for the other platforms, but not for Tandem (which requiresL >special lock-step operation between two coupled processors that both ItanicJ >and Alpha had to be enhanced to provide) - which seemed to be JF's point.  " Massively unclear, again, Nick :-(  D Yes, that is so.  My point was that the lock-step operation needs as? a prerequisite the architectures to be cleaned up for multi-CPU C operation, and that there is no reason to say that the IA-64 is any ; better in this respect than any of the others (except ARM).c  B If the operation could be added to IA-64, it could be added to the1 others - unless I have missed something critical.i     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679e   ------------------------------  % Date: Thu, 27 Jun 2002 16:47:57 +0100hU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>i Subject: Re: Open Letter to HP0 Message-ID: <affc1l$b1h$1@new-usenet.uk.sun.com>   John Smith wrote:   B > "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message. > news:m1kS8.30$xi6.427156@news.cpqcorp.net... >  >>J >>System platform vendors will have to differentiate on more than just the >> > CPUr > K >>chip.  Price is obviously one aspect.  But in the server space especiallyiL >>there are many, many ways to add value.  Nor is performance purely the CPUA >>chip.  The core logic chipset also will determine memory and IOo >> > performance. >  > D > Sounds to me like we are coming full circle, only this time with aN > low-volume, pricey IA-64 vs. a low-volume, pricey Alpha. At least with Alpha0 > we had confidence that the chip would perform. >  >  >     ? You also had a portfolio of applications that run on Alpha, andd/ a set of apps that ran perfectly well emulated.e  ; You now have a platform with a much smaller native softwareu; portfolio and an overwhelming performance reason to require  ISV's to do a native version.   8 But then thats just one of the joys of Industry Standard
 Computing.   Regardss Andrew Harrison    ------------------------------  # Date: Thu, 27 Jun 2002 16:01:46 GMT 1 From: "Terry C. Shannon" <terryshannon@attbi.com>- Subject: Re: Open Letter to HP. Message-ID: <KFGS8.326486$352.35599@sccrnsc02>  # "Andrew Harrison SUNUK Consultancy"P> <andrew_nospam.harrison_remove_this@sun#.com> wrote in message* news:affc1l$b1h$1@new-usenet.uk.sun.com...   <snip> > : > But then thats just one of the joys of Industry Standard > Computing.  F There is only one real joy in "Industry Standard" Computing: Being big8 enough and wielding enough clout to set the standards...   ------------------------------  % Date: Thu, 27 Jun 2002 10:31:55 -0500b+ From: Chuck Aaron <caaron@ceris.purdue.edu>  Subject: parsing >255T0 Message-ID: <3D1B2FEB.6930742B@ceris.purdue.edu>   Group,  < It appears DCL command procedures cannot parse more than 255D characters at one time into a single form_fld buffer. Is there a way5 around this or is this still an internal restriction?n   Thanks,g   Chuck    ------------------------------  % Date: Thu, 27 Jun 2002 09:40:41 -0500e1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>n( Subject: Re: Quorum discussion/questions1 Message-ID: <aff88d$s1s$1@fizban.pprd.abbott.com>e  I I believe you can't host base shadow a Qdisk, but controller based is OK.u   -- Dave...   3 More than one cigar at a time is excessive smoking.b -----Mark Twainn  2 "Main, Kerry" <Kerry.Main@hp.com> wrote in messageL news:BE56C50EA024184DAF48F0B9A47F5CF4026607CC@kaoexc01.americas.cpqcorp.net. .. Lyndon,o  D I recommend option 1 - when ES45 is added, plan to get rid of quorum: disk (timing depends on when you can do a cluster reboot).  E Quorum disks add some complications (e.g. can not shadow quorum disk,-> adds additional IO's) that are not needed in a 3 node cluster.  
 Kerry Main Senior Consultant> Hewlett-Packard Canada! Consulting & Integration Services2 Voice: 613-592-4660n Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----5 From: Lyndon Bartels [mailto:lbartels@pressenter.com]  Sent: June 26, 2002 9:17 PMa To: Info-VAX@Mvb.Saic.Com($ Subject: Quorum discussion/questions    D I have some questions about possibly changing my quorum config on my cluster.   Here's the current config:  ' Two ES40s (currently v7.2-1, soon v7.3)t- The system disk and quroum disk are CI based.o  D Both ES40s are voting members. So with ES40s and Qdisk, I have three expected votes.-  H I'm probably going to be adding an ES45 (after v7.3) that will also be a voting member.  - The way I see it I have two possible options:D  G 1. I can remove the quorom disk from the config, and leave the expectedc votes at three (3).o  E 2. I can retain the quorum disk, add the ES45 as a voting member, andn6 increase the expected votes from three (3) to four (4)      E With these two possibilities... Could somebody care to comment on thefG positives, negatives, gotchas, or even "You've gotta be nuts!" of thesec approaches?,    H What about in short term...? Would it be good/bad/indifferent to add theG ES45 as a voting member (and have 4 votes) when the other two ES40s are H still configured for three (3)? Then, after a time, change the config onA the existing ES40s, to bring the votes and/or qdisk config to thea correct config.n     Thanks in advance,   Lyndon   --G My opinions are mine and mine alone. They seldom align with those of mya3 employer. And even if they did, it wouldn't matter.     H The only good thing about putting the cart before the horse is you don't! have to look at the horse's butt.o   ------------------------------    Date: 26 Jun 2002 23:23:43 -0700' From: piet@timmers-it.nl (Piet Timmers) 3 Subject: Re: remote node is not currently reachable2= Message-ID: <be44b12d.0206262223.21acd56d@posting.google.com>b  h koehler@encompasserve.org (Bob Koehler) wrote in message news:<M$iCL$0jbARd@eisner.encompasserve.org>...i > In article <be44b12d.0206260118.5ad4e1b1@posting.google.com>, piet@timmers-it.nl (Piet Timmers) writes:fG > > We have a serious problem reaching a decnet node wich according thesJ > > message is not reachable, but from all other nodes in our network this > > node is reachable. > >  > > $ DIR NODE2::e9 > > %DIRECT-E-OPENIN, error opening NODE2::*.*;* as inputn3 > > -RMS-E-FND, ACP file or directory lookup failediA > > -SYSTEM-F-UNREACHABLE, remote node is not currently reachablet > >  > E >    Been there, done that.  Had a typo when I registered that node's H >    address on that host.  Try the address, for a node who's address is >    area.node:p >  >       $ dir area.node::n  B We have tried this to, same result, node not reachable, on all our other nodes no problem.n   ------------------------------  % Date: Thu, 27 Jun 2002 09:01:17 +0200 * From: "Sven Tieste" <sven.tieste@sw-hb.de>3 Subject: Re: remote node is not currently reachablee2 Message-ID: <afecc7$av0$1@f40-3.zfn.uni-bremen.de>  % Maybe you using another user-account. 2 Then the problem could be solved by adding a proxy to the sys$system:authorize. (At our machines it helps :-))  < Piet Timmers <piet@timmers-it.nl> schrieb in im Newsbeitrag:2 be44b12d.0206260118.5ad4e1b1@posting.google.com...E > We have a serious problem reaching a decnet node wich according theVH > message is not reachable, but from all other nodes in our network this > node is reachable. >o > $ DIR NODE2::o7 > %DIRECT-E-OPENIN, error opening NODE2::*.*;* as inputh1 > -RMS-E-FND, ACP file or directory lookup failed ? > -SYSTEM-F-UNREACHABLE, remote node is not currently reachablei >t& > All other nodes can reach this node.: > From node NODE2 to the node with teh problem works fine. >e< > We are running OpenVMS V7.2-1 and decnet DECNET_OSI V7.2-1 >b > Any idea?a >  > Greetings, >  > Piet Timmers   ------------------------------  # Date: Thu, 27 Jun 2002 05:39:59 GMT * From: "Bill Todd" <billtodd@metrocast.net>< Subject: Re: Reuters test - Itanium II blows away Sparky IIIB Message-ID: <OyxS8.418542$%y.30564687@bin4.nnrp.aus1.giganews.com>  @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message, news:6HoS8.49$Bn6.663686@news.cpqcorp.net...J > Bill read your own posts.  Ignoring Alpha (which is the real bee in your? > bonnet), you have consistantly predicted the failure of IA64.?  L From the available evidence I'd say that I may well both read and write withH more proficiency than you do, Fred.  And while I have certainly laid outJ many reasons why Itanic may, and quite likely should, sink forever beneathK the waves (and hence should be considered at least somewhat risky to commitvH to), I am well aware that products do not always get the fates that theyF deserve and have refrained from anything even resembling a categorical; prediction that such complete failure would actually occur.s     But as ispL > highlighted in your frothing at the mouth reply, the core of your issue isF > that you want IA64 (and in fact VMS)  to fail as validation that the) > retirement of Alpha was a bad decision.a  H Fred, if you believe that I require *any* additional validation that theK retirement of Alpha was a bad decision then you are even more clueless thano
 I thought.  E I want Itanic to fail because  a) it deserves to technically (i.e., I L believe that natural selection is in general a good thing where products areG concerned) and  b) those who have attempted to muscle it into dominancesI despite its inferiority should pay dearly enough for their arrogance that F any others who might be similarly inclined will remember the potentialL consequences long into the future (i.e., I think that vendors should be moreL focused on trying to satisfy the needs of their customers than on internal -E or external - politics, and that such a monumental failure would be atH decidedly healthy object lesson to this effect).  Both points reflect myH appreciation of the capitalist competitive ideal and my revulsion at the6 ways in which Intel and its cronies have distorted it.   >lL > Should IA64 not fail, and VMS to successfuly move forward to an IA64 base, IhD > expect fully that you will be the last to acknowledge your lack of > infallibility.  L Since I have not been so rash as to predict their failure rather than simplyH the very real risk of it, the issue does not arise.  I realize that suchK distinctions may be a bit subtle for you, but that's not my problem (except-8 for the time I waste responding to your misconceptions).  J As for VMS 'successfully' moving forward, we'll just have to wait two moreJ years to see:  I've never suggested that VMS couldn't be ported, just thatJ VMS was already stagnating competitively before the Alphacide and that theL Alphacide was likely to make things significantly worse *at least* until theG port was complete (which certainly seems to be the effect it has had son far).b   >OB > So keep shouting at the top of your lungs, and attack anyone who
 disagrees.D > The sorry thing is that it's driven you into the loving embrace of	 Andy-Boy.?  H Not really:  mostly, it's just made me far less tolerant of incompetence( and/or sleaze, regardless of the source.  G Both Andrew and Kerry have agendas, the main difference (aside from theaF different products they support) being that Andrew makes at least someL attempt to argue on merits rather than concentrate on pure spin.  That makesI real information flow (as contrasted with tit-for-tat sniping) at least anI possibility, so when people seem just to be talking past each other I mayiL attempt to improve their focus on the off chance that someone might actuallyJ learn something.  But I'd be more engaged with his discussions if the hardK information content in both his posts and the responses to them was higher: B haranguing without the discipline to back up one's statements with6 verifiable, quantitative references is simply tedious.  F When the cHumPaq apologists attempt to build up Itanic by running downI USIII, it seems appropriate (whether Andrew is involved or not) to don mylJ Truth Squad uniform just as I did when they attempted to run down Alpha inJ much the same way:  after all, the only Itanic that there's yet any publicL performance information available for from credible (i.e., unbiased) sourcesI is Merced, and to suggest that it competes with *anything* effectively isa simply ludicrous..  L Back when Paul DeMone was predicting that McKinley would hit 800 SPECint2K IL suggested that 600 - 700 seemed far more realistic (and in fact Paul droppedL his prediction to 700 in a discussion post).  The few figures that Intel hasL since released seem consistent with my estimate, but only when third partiesK have shipping systems to test will we really know whether McKinley performs I any better than USIII in any area save FP performance (and if it does, in  which areas and by how much).   J At this point, one can't help wondering whether there's a reason Intel hasH been so shy about providing, or allowing other parties to provide, solidL performance data for McKinley (since IIRC SPEC allows figures to be releasedJ 3 months before a product ships, and vendors have usually been eager to doH so).  Given historical precedent, changing their tune yet again to "JustJ wait for Madison!" could be rather dangerous unless Madison was truly justH around the next corner, so minimizing the interval between disappointingH performance revelations about McKinley and the release of something more  respectable could explain a lot.  G Or not.  But their record of performance predictions hardly errs on thesL conservative side, so a degree of skepticism seems indicated until such time as the veil is removed.r   >oF > So why don't you stick to ruminations about file system performance.  L Because that's not the only area in which I'm both competent and interested.   - bill   ------------------------------  % Date: Thu, 27 Jun 2002 13:40:29 +0100 U From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>S< Subject: Re: Reuters test - Itanium II blows away Sparky III0 Message-ID: <aff125$7e5$1@new-usenet.uk.sun.com>   Fred Kleinsorge wrote:  K > Well, I gotta give you high marks on FUDmenship.  Especially since in theiK > not-to-distant future I expect to see Sun fare poorly against IA64 as the L > real systems get into real peoples hands - and it goes beyond your abilityL > to spin things.  You of course can hope that I'm wrong.  But I'm not goingF > to argue with you over it, I can wait to see the actual performance. >  >     7 Just as a matter of interest what are your criteria for 1 something you would consider to be a valid publicm benchmark ?   % Mine not in any particular order are.   / Published specification for the benchmark or atd" least available at a minimum cost.2 Published set of benchmark run rules and reporting rules.- Agreed workload that is preferably usefull to2 someone somewhere. Agreed set of metrics.3 Audit or external scrutiny of the indevidual tests.g  1 This is a set of rules/principles that is broadlyc4 accepted by most people doing comparative benchmarks5 and the bodys or companies that create them from SPEC"3 to NotesBench. Now how many of these tests does the  Reuters "Benchmark" meet ?  5 I find it highly revealing that you try to ignore ther0 results that your own company has published that1 do meet these criteria, SAP, TPC, etc but instead . seem to be fixeted on the results if thats the0 right word of something which if you can call it/ a benchmark at all does not apparentyl meet anyC of the criteria.   Regards  Andrew Harrisont   ------------------------------  # Date: Thu, 27 Jun 2002 14:26:52 GMTT# From: "John Smith" <a@nonymous.com>i< Subject: Re: Reuters test - Itanium II blows away Sparky IIIG Message-ID: <MgFS8.10979$MIK.1969@news01.bloor.is.net.cable.rogers.com>   5 "Nick Maclaren" <nmm1@cus.cam.ac.uk> wrote in messages* news:afd92r$bmv$1@pegasus.csx.cam.ac.uk... > > K > >> Of course, might does not always prevail, and the fact that POWER4 and J > >> Hammer - and on the low end IA32, and at least for a while EV7 - will stillA@ > >> be around to make it clear that Itanic offers best-of-breed capabilities inrG > >> no market segment whatsoever may yet cause it to flounder into thel gravepI > >> that even HP's engineers recognized it belonged in years ago.  We'llt just > >> have to wait and see. > >dI > >Do you have a citation about this last point regarding HP's engineers?. >yB > I do, but am not posting - it was private Email.  I would phraseA > it as "SOME of HP's engineers", perhaps even "MOST of ...", but 5 > wouldn't go as far as saying just "HP's engineers".r >    </flogging a dead horse on>hK Which begs the question, just what the hell was Capellas thinking last Juned( 25th and in the months leading up to it?  K I guess, like most CEO's, he just didn't know what was going on, or perhapspA just didn't care enough to dig deep enough to find out the truth.H  B Does anyone here have any real sense how deep down into the CompaqK organization the decision to kill Alpha went, ie. was Engineering consulted:E at all in any meaningful way, or was it just Capellas and a few otherrK stratospheric 'thinkers' than made the decision on their own (with coachingf from Intel and Carly)?   </flogging a dead horse off>   ------------------------------  # Date: Thu, 27 Jun 2002 14:35:22 GMTe# From: "John Smith" <a@nonymous.com>e< Subject: Re: Reuters test - Itanium II blows away Sparky IIIE Message-ID: <KoFS8.1791$2er.593@news01.bloor.is.net.cable.rogers.com>   < "Terry C. Shannon" <terryshannon@attbi.com> wrote in message$ news:a_pS8.3973$Uu2.295@sccrnsc03... >r > > B > > The first cracks appeared a year ago - i.e. the first negativeB > > reports in the Wall Street Journal.  If the McKinley flops, in0 > > any visible way, expect them to take notice. > >u > I > Yep, especially in light of INTC's performance in the stock market overc theeL > past three weeks. Ah well, we should have more info on McKinley within the > next three weeks.   J So in the light of all the different disclosures about all sorts of thingsI these days, could Intel be hit with a double-dose of trouble - on the cpueK and accounting performance fronts? It seems that they may on the cpu score.hJ I have no idea how clean their accounting is (I suspect it's probably ok),K but if they get caught up in the popular exercise of 'throwing out the baby1I with the bath water' syndrome that often accompanies any whiff of troubleRJ (real or imagined), I wonder which development efforts would be candidatesG to trash if they usher thousands of employees out the door? It might be#K interesting to think this through as a hypothetical exercise. Ideas anyone?    ------------------------------    Date: 26 Jun 2002 23:04:42 -0700# From: dooleys@snowy.net.au (dooley)i2 Subject: Re: RMS : preserving a RAB across $OPEN ?= Message-ID: <1ca82fc6.0206262204.69f9e2a6@posting.google.com>-  a JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message news:<3D1A1DCA.D21F3E6B@videotron.ca>...e > Scenario::E > I have multiple streams accessing a file that was opened read-only.u > F > A new stream is created but it needs read-write access to that file. > N > Can I $CLOSE the file, then $OPEN it with read-write without having to "tear: > down" all the existing RABs and rebuilding  them later ? >  > I.E. $OPEN (read only) > 	$CONNECT with RAB12 > 	$GET with RAB1 by key > 	 > 	$CLOSEd- > 	$OPEN (read-write, same FAB block address) J > 	$GET with RAB1 sequentially (continuing from the previous read by key). > O > If I $CONNECT RAB1 back to the FAB after the second $OPEN, does this wipe oute9 > the context and current position contained in the RAB ?b > O > It would be VERY NICE if I could reopen the file allowing me to add a new rab0N > for read write without affecting the existing RABs and whatever they were in# > the process of doing at the time.  I don't think so rA however one way to do this is to have two different logical nameso% that refer to the same physical file i& eg. define filein disk1:[data]file.dat&     define fileio disk1:[data]file.dat+ each logical file keeps its record pointersC
 so  you couldl% $OPEN filein              (read only)a $CONNECT with RAB1 $GET with RAB1 by key   1& $OPEN fileio              (read-write) $FIND (using key from above) $GET   sequentially[ $UPDATE or $DELETE if you want   $CLOSE   from the rms manual:  B The Find service gives you the following functional capabilities: B You can skip records when you are accessing a file sequentially by2 making successive invocations of the Find service.A You can establish the current record context prior to invoking an $ Update, Delete, or Truncate service.> You can establish a random access starting point in a file for@ subsequent sequential access operations such as the Get service.D When you follow the Find service with a sequential access operation,E such as the Get service, the current record context is established byeF the Find service and the sequential access operation establishes a new? sequential access context. Conversely, when you follow the FindmF service with a nonsequential access operation such as a Delete serviceD or an Update service, the sequential access context remains the same$ as it was prior to the Find service.   All clear now? Phil   ------------------------------  # Date: Thu, 27 Jun 2002 08:08:43 GMT ? From: Jim.Johnson@software-exploration.nospam.com (Jim Johnson)t2 Subject: Re: RMS : preserving a RAB across $OPEN ?/ Message-ID: <3d1abe7c.2988186@news.demon.co.uk>t  D As was said elsewhere, the issue isn't what you do with your RAB, itB is all about what RMS does with its internal RAB equivalent.  ThatE structure is created at a $CONNECT, is torn down at a $DISCONNECT (or': $CLOSE), and is inherently tied to an owning internal FAB.  E The best you're going to be able to do with the sort of scheme you're-C describing is to save away the positioning context (such as the keyi3 value, or an RFA), and reposition after re-opening.   A I'm not quite sure why you don't want to just have the files opena= readonly.  I can think of three or four reasons, centering oneF performance, but I'm not sure how strong your performance requirementsC are.  Furthermore, if the readonly accessor allows writers, I'm nottB sure how much performance you gain in any event.  Just opening the- file read-write is something I'd think about.    Jim.    , On Wed, 26 Jun 2002 16:02:30 -0400, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:   
 >Scenario:D >I have multiple streams accessing a file that was opened read-only. > E >A new stream is created but it needs read-write access to that file.A >,M >Can I $CLOSE the file, then $OPEN it with read-write without having to "tear 9 >down" all the existing RABs and rebuilding  them later ?  >t >I.E. $OPEN (read only)f >	$CONNECT with RAB1 >	$GET with RAB1 by key  >  >	$CLOSE, >	$OPEN (read-write, same FAB block address)I >	$GET with RAB1 sequentially (continuing from the previous read by key).- >-N >If I $CONNECT RAB1 back to the FAB after the second $OPEN, does this wipe out8 >the context and current position contained in the RAB ? >rN >It would be VERY NICE if I could reopen the file allowing me to add a new rabM >for read write without affecting the existing RABs and whatever they were ino" >the process of doing at the time.   Jim Johnsoni Software Exploration, Ltd.) (remove '.nospam' from the reply address)t   ------------------------------  % Date: Thu, 27 Jun 2002 06:10:27 -0400u- From: JF Mezei <jfmezei.spamnot@videotron.ca> 2 Subject: Re: RMS : preserving a RAB across $OPEN ?, Message-ID: <3D1AE490.A1AF4F2C@videotron.ca>   Jim Johnson wrote:F > As was said elsewhere, the issue isn't what you do with your RAB, itD > is all about what RMS does with its internal RAB equivalent.  ThatG > structure is created at a $CONNECT, is torn down at a $DISCONNECT (ors< > $CLOSE), and is inherently tied to an owning internal FAB.  . Are there really internal duplicates of RABs ?  L If, after re-opening the file, I $CONNECT with the old rab, and then issue aJ $FIND using the last RFA that had been obtained, would this bring back the context exactly the same way ?    C > I'm not quite sure why you don't want to just have the files opens > readonly.   L Trying to emulate what ALL-IN-1 does when it accesses rms files. Some of theH files are created read-only. If I try to open read-write a file which isN read-only, it generates a security error which I would like to avoid since the@ user never tries to actually write to the file. (for instance, aG behind-the-scenes validation of a province code against a list of valid . provinces stored in a read only indexed file).  M Problem is that when I open the file initially, it may be fo a read operationlG with no knowedge that later on the user will want to write to the file.    Consider the following  A 	for clients with .area_code = "514" and .postal_code:1 != "H" dot2 		write_change clients %key=.%key, area_code="450"  M This would go through the clients indexed file and update records for clientsu> whose areas code is 514 but postal code doesn't begin with H.   M The first invocation of "clients" is in read only mode with sequential access)K starting at first record. So it would trigger opening the file in read-only K mode. Once it gets to the first record that matches the selection, it wouldeL then execute the write_change which would then want to update the $OPEN to aJ read-write, but after that record, it would need to continue with the next record in the FOR loop.r   ------------------------------  # Date: Thu, 27 Jun 2002 10:26:12 GMTd? From: Jim.Johnson@software-exploration.nospam.com (Jim Johnson)t2 Subject: Re: RMS : preserving a RAB across $OPEN ?0 Message-ID: <3d1ae5e0.13072497@news.demon.co.uk>  , On Thu, 27 Jun 2002 06:10:27 -0400, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:k   >Jim Johnson wrote: G >> As was said elsewhere, the issue isn't what you do with your RAB, itnE >> is all about what RMS does with its internal RAB equivalent.  ThatuH >> structure is created at a $CONNECT, is torn down at a $DISCONNECT (or= >> $CLOSE), and is inherently tied to an owning internal FAB.m >e/ >Are there really internal duplicates of RABs ?e  D Ayup, although the information is generally much more extensive than@ in the user structure.  Try an ANALYZE/SYSTEM, then SHOW PROCESSC /RMS=(IFB,IRB,FWA) for interest.  Fwiw, this will show the internal F FABs, the internal RABs, and the Filename Work Areas (the internal NAM* blocks) for the open files in the process.   > M >If, after re-opening the file, I $CONNECT with the old rab, and then issue amK >$FIND using the last RFA that had been obtained, would this bring back the  >context exactly the same way ?g  C To be honest, I can't remember off the top of my head.  It will getuC you back to the same record, so I think so, but I can't remember ifaE there is any other state that needs to be restored as well.  Bill, doe
 you remember?g   >n >cD >> I'm not quite sure why you don't want to just have the files open
 >> readonly. e >cM >Trying to emulate what ALL-IN-1 does when it accesses rms files. Some of thedI >files are created read-only. If I try to open read-write a file which isnO >read-only, it generates a security error which I would like to avoid since therA >user never tries to actually write to the file. (for instance, auH >behind-the-scenes validation of a province code against a list of valid/ >provinces stored in a read only indexed file).a >nN >Problem is that when I open the file initially, it may be fo a read operationH >with no knowedge that later on the user will want to write to the file. >  >Consider the followings >lB >	for clients with .area_code = "514" and .postal_code:1 != "H" do3 >		write_change clients %key=.%key, area_code="450"  >eN >This would go through the clients indexed file and update records for clients? >whose areas code is 514 but postal code doesn't begin with H. p >eN >The first invocation of "clients" is in read only mode with sequential accessL >starting at first record. So it would trigger opening the file in read-onlyL >mode. Once it gets to the first record that matches the selection, it wouldM >then execute the write_change which would then want to update the $OPEN to arK >read-write, but after that record, it would need to continue with the next  >record in the FOR loop.  E Yeah, ok.  That makes sense -- if I understand correctly, you need tocE support files that you don't have write access to, and you don't wantpE to generate failure audits unless the user really did try to write toaE a file they weren't allowed to (otherwise, you could try to open readsD write, and drop back to readonly if you get an access denied error).   Jim.   Jim Johnson  Software Exploration, Ltd.) (remove '.nospam' from the reply address)s   ------------------------------  # Date: Thu, 27 Jun 2002 11:39:11 GMTn From: system@SendSpamHere.ORGy2 Subject: Re: RMS : preserving a RAB across $OPEN ?0 Message-ID: <00A10136.3107CC48@SendSpamHere.ORG>  F In article <slrnahka01.du4.danco@pebble.org>, danco@pebble.org writes:= >In article <3D1A1DCA.D21F3E6B@videotron.ca>, JF Mezei wrote:r >dP >> If I $CONNECT RAB1 back to the FAB after the second $OPEN, does this wipe out: >> the context and current position contained in the RAB ? > @ >Well, even if it didn't wipe out context in the "external" RAB,B >I doubt the context in the "internal" RAB would be the same after? >a close->open.  In real life the record context can be complexlA >enough that just saving the RFA of the current record before thed@ >close and doing a FIND by RFA after the open isn't good enough.@ >But perhaps might your application be simple enough for that to >work for you? >f6 >(Have you seen the undocumented SYS$MODIFY trick that8 >LIB$FIND_IMAGE_SYMBOL uses to position by RFA and start: >reading the symbol table within an image file in variable >length record access mode?) >s >- Dan  I $MODIFY is coded to allow modification of the file's record formats (RFM)dG but I see no way to use $MODIFY to alter the file's access modes (FAC).      --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM_            u5   "Well my son, life is like a beanstalk, isn't it?" a   ------------------------------  # Date: Thu, 27 Jun 2002 15:10:41 GMTK From: danco@pebble.org2 Subject: Re: RMS : preserving a RAB across $OPEN ?- Message-ID: <slrnahmbn1.egg.danco@pebble.org>   < In article <3D1AE490.A1AF4F2C@videotron.ca>, JF Mezei wrote:  N > If, after re-opening the file, I $CONNECT with the old rab, and then issue aL > $FIND using the last RFA that had been obtained, would this bring back the  > context exactly the same way ?  B If, before you closed the file, you had done a FIND by KEY using a= secondary key to establish a key of reference, then after youd@ close, reopen, reconnect, and do your FIND by RFA, I don't thinkC your current key of reference will not have been restored.  I thinkt2 your key of reference will now be the primary key.  A Lets say that your program does a FIND by KEY on a secondary key,t> then does several sequential gets (using the order established: by that key of reference) with a limit in order to get theD record with that key and all it's duplicates (or something similar).? While it's doing this, it does $UPDATEs on some of the records.SC If you interrupt this with your CLOSE/OPEN/RECONNECT, I don't thinkB@ you will be able to restore the exact record context with just a simple FIND by RFA.t  > Someone please correct me if I'm wrong, but FIND by RFA always8 resets your current key of reference to the primary key.   - Dant   ------------------------------  # Date: Thu, 27 Jun 2002 15:19:25 GMTd From: danco@pebble.org2 Subject: Re: RMS : preserving a RAB across $OPEN ?- Message-ID: <slrnahmc7j.egg.danco@pebble.org>,  E In article <slrnahmbn1.egg.danco@pebble.org>, danco@pebble.org wrote::  B > close, reopen, reconnect, and do your FIND by RFA, I don't thinkE > your current key of reference will not have been restored.  I thinkr  + Oops, typo in the above.  Stike that "not".s   - Danm   ------------------------------  # Date: Thu, 27 Jun 2002 16:53:16 GMT * From: "Bill Todd" <billtodd@metrocast.net>2 Subject: Re: RMS : preserving a RAB across $OPEN ?C Message-ID: <0qHS8.468803$Oa1.32350850@bin8.nnrp.aus1.giganews.com>i  L "Jim Johnson" <Jim.Johnson@software-exploration.nospam.com> wrote in message* news:3d1ae5e0.13072497@news.demon.co.uk.... > On Thu, 27 Jun 2002 06:10:27 -0400, JF Mezei' > <jfmezei.spamnot@videotron.ca> wrote:v >o > >Jim Johnson wrote: I > >> As was said elsewhere, the issue isn't what you do with your RAB, ityG > >> is all about what RMS does with its internal RAB equivalent.  ThattJ > >> structure is created at a $CONNECT, is torn down at a $DISCONNECT (or? > >> $CLOSE), and is inherently tied to an owning internal FAB.a > >s1 > >Are there really internal duplicates of RABs ?   L More that there is internal context maintained for the record access stream,J so that the program can't screw it up within operations (for the executionH context established at the operation's start) or between operations (for5 context carried over from one operation to the next).    >qF > Ayup, although the information is generally much more extensive thanB > in the user structure.  Try an ANALYZE/SYSTEM, then SHOW PROCESSE > /RMS=(IFB,IRB,FWA) for interest.  Fwiw, this will show the internalnH > FABs, the internal RABs, and the Filename Work Areas (the internal NAM, > blocks) for the open files in the process. >  > >0G > >If, after re-opening the file, I $CONNECT with the old rab, and thene issue a I > >$FIND using the last RFA that had been obtained, would this bring backo the.! > >context exactly the same way ?  >nE > To be honest, I can't remember off the top of my head.  It will get@E > you back to the same record, so I think so, but I can't remember if G > there is any other state that needs to be restored as well.  Bill, dol > you remember?   L As Dan said, you won't get the index context you had previously if it was inH an alternate index.  But since you will get the same record, perhaps youK could just perform a second file open (requesting write access), modify theeF record there (you might have to issue an unlock for the record in yourK read-only access stream), and then return to the read-only access stream to . continue processing in the context you wanted.   - bill   ------------------------------  % Date: Thu, 27 Jun 2002 14:00:33 -0400(- From: JF Mezei <jfmezei.spamnot@videotron.ca>a2 Subject: Re: RMS : preserving a RAB across $OPEN ?, Message-ID: <3D1B52C0.5619AF08@videotron.ca>   danco@pebble.org wrote: D > If, before you closed the file, you had done a FIND by KEY using a? > secondary key to establish a key of reference, then after you B > close, reopen, reconnect, and do your FIND by RFA, I don't think= > your current key of reference will not have been restored. -  L But if my RAB contains the former key of reference, wouldn't the first $FIND: after I have reopened the file set that key of reference ?  I When I do a sequential $GET after a $FIND, doesn't RMS look at the key ofmF reference inside the RAB or does it use only the internal hidden one ?   ------------------------------  % Date: Thu, 27 Jun 2002 16:02:06 +0100nU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>l+ Subject: Re: Sun benchmarketeering campaignt0 Message-ID: <aff9bn$a8r$1@new-usenet.uk.sun.com>   Rob Young wrote:  C >         There are several issues here.  Sun in a real sense needsCD >         to abandon benchmarks altogether and highlight application >         availability.d > F >         We have learned recently that Sun has essentially abandondedN >         TPC-C, abandoning SPEC2000 or in the process thereof.  That decisionL >         appears to be because senior management in part doesn't understandJ >         how the benchmark works and why it is one of the best benchmarks >         there is.e > P >         How long before they abandon SPECjbb?  How long before TPC-H suddenly H >         isn't of great importance?  What do they do if large customers7 >         formulate RFPs that require these benchmarks?i > M >         Cases in point.  First, David Yen, VP for the processor and networkiK >         products at Sun, makes the following claims about SPECfp/SPECint:i >  > 4 > http://www.eweek.com/article2/0,3959,277705,00.asp > P > "In our opinion, many of the processor-oriented benchmarks are being outmoded"Q > by new chip designs, and therefore don't accurately reflect system performance,n4 > Yen said.  [The CPU is part of the system Mr. Yen] > P > He cited two popular benchmarks as "misleading," SPECfp and SPECint, which areH > used to measure floating point and integer performance, respectively.  > N > The problem, he said, is that the larger on-die memory caches on some chips,P > such as Itanium 2's 3MB cache, skew the results since they can accommodate theL > entire benchmark program, and thus don't have to off-die to access memory, > which is unrealistic." >  > ---t > J >         Specint/Specfp are not "misleading", quite revealing actually.  P >         Perhaps Mr. Yen isn't very "technical" or doesn't understand "techie"  >         issues?e >     F David is very technical, he was Mr SPARCenter2000 arguably Sun's first
 large server.e  C And Devid understands the issues rather well its a shame you don't.hD Take the IBM P690 for example. The minium number of CPU's in a P690 H Turbo actually installed is 8, on one 8 way MCM module. For the purposesB of SPECint and SPECfp IBM turn off all but one of these CPU's, but@ they don't turn off the half of the L2 cache shared by the other? core or 7/8ths of the L3 cache. So you get a SPECint and SPECfps> number being run on 1 CPU with 2 CPU's worth of L2 and 8 CPU's> worth of L3 cache. None of the SPECint or fp benchmarks run in@ more than 200 MB and the majority now fit nicely into the 128 MB of L3 cache on the system.  ? Is this against the rules of SPEC, no, they were drafted before ? this kind of configuration was possible on micorprocessor basedg systems.  C But it means that it is now virtually impossible to use SPECint and D SPECfp as a usefully measure ot even raw CPU int and fp performance.  C No customer is going to turn off 7 of the 8 CPU's they get in theiraB P690 and 8 is the minimum and very few people will buy a P690 just
 to use 1 CPU.i  @ By saying that SPECint/fp are revealing all you have revealed is@ that you don't really understand systems architecture and how it; might influence the results in a way that isn't helpfull to0 actual apps.  > In addition you really need to discuss SPECfp and the validity= of SPECfp results with Freddy Boy who you may remember binnedd? 10 years of Compaq SPECfp marketing BS by admitting or claimings< that SPECfp wasn't representative of customers applications.    H >         Cache can accomadate the entire benchmark program?  Don't have, >         to off-die access memory?  Sheesh.K >         Really a shame there.  He obviously hasn't a clue about Spec2000.rB >         Perhaps someone could shoot the following Mr. Yen's way: >     A I guess you didn't understand how a P690 is architected did you ?   < While the cache used for the IBM P690 SPECint/fp numbers was> 128 MB which is less than the 199 MB maximum required for gzip? 14 of the 26 SPECint/fp benchmarks now have a resident set size A that fits entirely in cache while an additional 2 spend more thane@ half of the run-time with their resident set fitting into cache,* the rest get a much better cache hit rate.  ; If the P690 had all 8 CPU's active then the available cachea: per CPU would be 16 MB. For a workload like SPECint/SPECfp: this would reduce the number of apps fitting entirely into? cache to 4 (what the designers of CPU 2000 envisaged) with none : that have larger set size spending any significant part of: their run in cache the rest would have a reduced cache hit= rate, what that would do for the SPECint number who knows butd then thats the point isn't it.  > How about displaying your grasp of another subject rather than not doing so in this one.      Regardsa   Andrew Harrisons    7 > http://www.specbench.org/osg/cpu2000/analysis/memory/r > Q > The SPEC CPU2000 benchmarks are intended to exercise the CPU itself, the memory E > hierarcy, and the compilers. How much memory do they actually use? y > Q > The data collected here show that SPEC met its goals for memory footprint: mostIP > benchmarks are larger than common cache sizes, many are larger than 100MB, and > none are larger than 200MB.  > M > It is useful to have benchmarks that are larger than common caches, becauseeP > SPEC would like to differentiate its benchmarks from "toy benchmarks" that are. > too easy to run or that simply reflect MHz. L > It is useful to keep the benchmarks under 200MB so that the suite leaves aL > reasonable margin on a 256MB machine. The other 56MB are available for theP > operating system, graphics system, network daemons, etc, without using 'singleP > user mode' on Unix systems, or killing processes on NT systems. (Such measuresC > may not be representative of how most people use their systems.) g >  > ---t > B >         Secondly, we read about SPECjbb and how StarCat fares at4 >         that benchmark in comparison to SuperDome: > 4 > http://www.eweek.com/article2/0,3959,278481,00.asp > F > In the Java application space, the new Superdome scored 594,161 JavaP > operations-per-second on the SPECjbb2000 benchmark, a result 75 percent higherO > than IBM's 32-way p690 server, code-named Regatta, and 37 percent better thanoG > Sun Microsystems Inc.'s 72-way Sun Fire 15K server, known as Starcat.  > J >         Perhaps they could trumpet the fact that if you bought a StarCatI >         with 104 processors (versus SuperDome's 64) the end-user can do A >         1.4% more jbbs than a SuperDome (602270 versus 594161).  >  >         Thirdly... > 7 > http://www.tpc.org/tpch/results/tpch_perf_results.asp  > 4 > http://www.eweek.com/article2/0,3959,278481,00.asp > N > For example, in testing targeted at measuring data warehousing, the upgradedP > Superdome running Oracle 9i Database Release 2 set a world-record one terabyteQ > benchmark result of 25,805 queries-per-hour, according to HP, a mark that bestsnF > all single-system or cluster servers on the market at one terabyte.  > Q >         TPC-H at 1000 GB?  HP has better Price/QphH and 37% better performance  1 >         than StarCat.  25805 QphH versus 18802.Y > I >         Yeah... you can see Sun's benchmarketeers have a great campaign N >         ahead of them.  Hoping no one notices must be part of that campaign. > % >                                 Robh >  >    ------------------------------    Date: 27 Jun 2002 11:13:03 -0600+ From: young_r@encompasserve.org (Rob Young) + Subject: Re: Sun benchmarketeering campaign 3 Message-ID: <yOI3v0udcYde@eisner.encompasserve.org>t   In article <aff9bn$a8r$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com> writes: >  >  > Rob Young wrote: > D >>         There are several issues here.  Sun in a real sense needsE >>         to abandon benchmarks altogether and highlight applicationc >>         availability. >> NG >>         We have learned recently that Sun has essentially abandonded-O >>         TPC-C, abandoning SPEC2000 or in the process thereof.  That decisioncM >>         appears to be because senior management in part doesn't understandlK >>         how the benchmark works and why it is one of the best benchmarkso >>         there is. >> fQ >>         How long before they abandon SPECjbb?  How long before TPC-H suddenly  I >>         isn't of great importance?  What do they do if large customers 8 >>         formulate RFPs that require these benchmarks? >> nN >>         Cases in point.  First, David Yen, VP for the processor and networkL >>         products at Sun, makes the following claims about SPECfp/SPECint: >>   >> r5 >> http://www.eweek.com/article2/0,3959,277705,00.aspi >>  Q >> "In our opinion, many of the processor-oriented benchmarks are being outmoded"vR >> by new chip designs, and therefore don't accurately reflect system performance,5 >> Yen said.  [The CPU is part of the system Mr. Yen]n >> mQ >> He cited two popular benchmarks as "misleading," SPECfp and SPECint, which arexI >> used to measure floating point and integer performance, respectively. a >> t  O >> The problem, he said, is that the larger on-die memory caches on some chips,FQ >> such as Itanium 2's 3MB cache, skew the results since they can accommodate the M >> entire benchmark program, and thus don't have to off-die to access memory,r >> which is unrealistic."h >> r >> --- >> rK >>         Specint/Specfp are not "misleading", quite revealing actually.  pQ >>         Perhaps Mr. Yen isn't very "technical" or doesn't understand "techie" i >>         issues? >> a >  > H > David is very technical, he was Mr SPARCenter2000 arguably Sun's first > large server.  > E > And Devid understands the issues rather well its a shame you don't.jF > Take the IBM P690 for example. The minium number of CPU's in a P690  > Turbo actually installed is 8t  ? 	He does understand the issues?  That's a surprising statement.I  5 	In another round of missing a point (once again) ...e  C 	Do you ever stop your campaign of leaving a great deal of what is oE 	written in a reply and taking out the key segments and/or even most RI 	annoyingly ignoring the key segments???????  Like you are doing in this  	 	case!!!!e  
 	Mr Yen says:i  L The problem, he said, is that the larger on-die memory caches on some chips,N such as Itanium 2's 3MB cache, skew the results since they can accommodate theJ entire benchmark program, and thus don't have to off-die to access memory, which is unrealistic."  ; 	He states: "Itanium 2's 3 MB cache, skew the results since @ 	they can accomadate the entire benchmark program and thus don't( 	have to [go] off-die to access memory"   3 	That is totally incorrect.  Spin Spin Spin Andrew.   F 	No where does he mention IBM.  In that context he is totally wrong.  $ 	Laughable actually.  Why?  Because:  6  http://www.specbench.org/osg/cpu2000/analysis/memory/  eP  The SPEC CPU2000 benchmarks are intended to exercise the CPU itself, the memoryD  hierarcy, and the compilers. How much memory do they actually use?   iP  The data collected here show that SPEC met its goals for memory footprint: mostO  benchmarks are larger than common cache sizes, many are larger than 100MB, and   none are larger than 200MB.    rL  It is useful to have benchmarks that are larger than common caches, becauseO  SPEC would like to differentiate its benchmarks from "toy benchmarks" that are -  too easy to run or that simply reflect MHz. aK  It is useful to keep the benchmarks under 200MB so that the suite leaves ayK  reasonable margin on a 256MB machine. The other 56MB are available for theoO  operating system, graphics system, network daemons, etc, without using 'singleeO  user mode' on Unix systems, or killing processes on NT systems. (Such measurestB  may not be representative of how most people use their systems.)   r  9 	IBM was NOT THE POINT AT ALL and a casual observer wouldn> 	recognize that fact.  SPEC2000 running on Itanium 2 does NOT ) 	(I repeat NOT) live in Itanium 2 caches.e  @ 	Besides that Andrew, prattling on about IBM ignores a couple ofC 	key facts.  First, is that given single CPU benchmarks IBM is near$G 	the top of SPECint but not leading.  Secondly, SPECint/SPECfp measuresPF 	three key components, CPU, memory subsystem and compilers.  It is one( 	of THE BEST system benchmarks there is.  G 	Face it, UltraSparc is weak and you can try to wiggle out of that but s 	facts are facts.f   	Good day Andrew.m   				Rob    ------------------------------  # Date: Thu, 27 Jun 2002 17:05:51 GMT25 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>e+ Subject: Re: Sun benchmarketeering campaigna2 Message-ID: <PBHS8.28$1U6.518303@news.cpqcorp.net>  6 Andrew Harrison SUNUK Consultancy wrote in message ... >, >o >,D >And Devid understands the issues rather well its a shame you don't.D >Take the IBM P690 for example. The minium number of CPU's in a P690I >Turbo actually installed is 8, on one 8 way MCM module. For the purposes C >of SPECint and SPECfp IBM turn off all but one of these CPU's, buteA >they don't turn off the half of the L2 cache shared by the othere@ >core or 7/8ths of the L3 cache. So you get a SPECint and SPECfp? >number being run on 1 CPU with 2 CPU's worth of L2 and 8 CPU'sr? >worth of L3 cache. None of the SPECint or fp benchmarks run inTA >more than 200 MB and the majority now fit nicely into the 128 MBr >of L3 cache on the system.h >n@ >Is this against the rules of SPEC, no, they were drafted before@ >this kind of configuration was possible on micorprocessor based	 >systems.i >e  J You know, I understand your complaint, but it is IMHO really just a way ofH trying to hide your own poor performance numbers.  IBM doesn't sell thisI system as a single CPU configuration, so the only way to get a single CPUrH number is to turn the others off.  ***This doesn't invalidate the singleK stream numbers, it just means you should by no means try to extrapolate theaI single stream number to the performance of a CPU in an MP configuration -eF which you *can't* do anyway.***  If you really need that single streamJ performance, you can always turn off the other CPU.  Heck, they could turnL off every other CPU.  Does this make the *cost* to get that performance bad?L Yup.  But that's a different question.  It doesn't mean that you *can't* getL that performance if you want to (and if IBM will allow you to configure it -9 which I assume they must - to conform to the SPEC rules).h  L There are *other* benchmarks that show the MP performance characteristics of	 a system.h  H The Sparc problem is that even playing compiler optimization tricks, theG SPEC numbers for them are not good, getting worse, and unlikely to everhH catch up to the competition.  So you need to find some *other* benchmarkB that you can, of only while nobody else has numbers, look good on.  H It's all a game.  Just admit it.  If a miracle happened, and Sparc 9 wasG launched with SPEC numbers 100x better than the competition - your tuner would change in an instant.a   ------------------------------  % Date: Thu, 27 Jun 2002 13:25:24 +0100 U From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>n Subject: Re: TCO study0 Message-ID: <aff05s$75s$1@new-usenet.uk.sun.com>   Main, Kerry wrote:  E >>>>Of course they will, but to customers buying your systems now its  >>>>3 > what they are capable of now thats important ..<<j > 5 > Yep, and that's why Customers continue to buy them.  >  >     4 Really, so declining from 11% of the Market to 5% is3 a victory because 5% still means that customers areK still buying systems !  4 I guess if you think that having the worst benchmark4 numbers for TPC-C when compared with current systems7 that you categorise as being the same size as the GS320 6 is acceptable then a mere halving of your market share is also great.    H >>>>No they don't sorry, the GS320's numbers are quite frankly terrible. >>>>G > Worse than Fujitsu and more expensive, worse than IBM (P690) and moreyJ > expensive, worse than SuperDome and more expensive, almost the same as a! > IBM P680 and more expensive.<<<  > ? > Yada, yada, yada - let me know when Sun does a TPC benchmark.  >   9 I don't need to I have 3 very acceptable benchmarks and 6c  C very acceptable benchmark results, so why would Sun need to provides> a result for a benchmark that everyone accepts is very flawed.   Regardsh Andrew Harrisona   ------------------------------  % Date: Thu, 27 Jun 2002 13:55:40 +0100pU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>  Subject: Re: TCO study0 Message-ID: <aff1ul$7oi$1@new-usenet.uk.sun.com>   Michael Austin wrote:i  * > Andrew Harrison SUNUK Consultancy wrote: >  >>Michael Austin wrote:s >> >>+ >>>Andrew Harrison SUNUK Consultancy wrote:  >>>f >>>b >>>>Main, Kerry wrote: >>>> >>>> >>>> >>>>>Gerard, >>>>>f  >>>>>Re: TCO Study .. Check out:J >>>>>http://www.openvms.compaq.com/openvms/whitepapers/enterprise_tco.html >>>>>l >>>>>s >>>>> G >>>>Sorry its horribly flawed and you know it, putting the GS320 in theaJ >>>>same class as a bunch of servers that at a minium have 2x its capacity4 >>>>is hardly going to generate like for like costs. >>>>D >>>>Bang the GS320 back into the mid-range section where it belongs,@ >>>>re-do the study and hey presto you find that the TCO numbers, >>>>don't look any thing like as attractive. >>>> >>>> >>>> >>>rK >>>So what you are saying is that while the GS320 has half the capacity, iteJ >>>does the same amount of work and has better TCO than Sun with twice the: >>>capacity?  Hmmm. I guess buying Sun is better... NOT!!! >>>r >>>o >>>e1 >>No by capacity I mean its ability to do usefulla( >>work, what else did you think I ment ? >> > F > Let' see: a "mid-range" Alpha GS320 that does the equivilant of yourF > "high-end" system and yours is better??? what planet do you live on? >     ; Do you have some sort of massive comprehesnion gap, a GS320p8 can based on the published benchmarks for the box manage7 between 1/2 and 1/3 of the throughput of a Sun F15000K.   7 Its a mid range machine not because of its price or itsa: footprint (its expensive and huge) but because it can only) do 1/3 to 1/2 to work of its competition.s  ? It even gets beaten by Sun F6800's which we and the TCO "Study"P class as a mid range system.  < What is the point of grouping a server like the GS320 in the: same category as a F15000, SuperDome or P690, on the basis9 of all the available usefull benchmark data these are all  much bigger machines than it.a  9 Of course putting the machine into the mid range categorya6 where it in fact resides would totally mess up the TCO9 analysis because it would then be costed against machinese9 with 1/2-1/3 of its cost 1/2-1/3 of its maintenance costsm etc etc.  ; Thank you for posting though I havn't had such a good laught since Kerrys last post.S     Regards    Andrew Harrison    ------------------------------  # Date: Thu, 27 Jun 2002 07:33:50 GMTm+ From: Roland Barmettler <rob@bbp.ch.remove>m' Subject: Re: Trouble with BACKUP/RECORD 7 Message-ID: <20020627093347.0904ed26.rob@bbp.ch.remove>    Hi Jim   Thanks for the reply.7  F > I ran a quick check and had no problems on both versions of OpenVMS.@ > Does your v7.3 computer have the DEC-AXPVMS-VMS73_BACKUP-V0100, > patch applied?  ($ PRODUCT  SHOW  HISTORY)  ) Yes, it's been applied on the 7.3 system.c OK, here's the log (7.3):   / NABOO:$ BACKUP/RECORD/LOG *.TXT;* TEST.BCK/SAVEtA %BACKUP-S-COPIED, copied DSA0:[SYS0.][SYSMGR]ANNOUNCE_ALCOR.TXT;2nB %BACKUP-S-COPIED, copied DSA0:[SYS0.][SYSMGR]ANNOUNCE_NABOO.TXT;12 %BACKUP-S-COPIED, copied7 DSA0:[SYS0.][SYSMGR]DECNET_REGISTER_IMPORT_FILE_NABOO.TS XT;1 %BACKUP-S-COPIED, copied7 DSA0:[SYS0.SYSCOMMON.][SYSMGR]TCPIP$BINDSETUP_HELP.TXT;e 1eD %BACKUP-S-COPIED, copied DSA0:[SYS0.SYSCOMMON.][SYSMGR]WELCOME.TXT;1: %BACKUP-I-STARTRECORD, starting backup date recording pass   NABOO:$ sh time    27-JUN-2002 09:28:18  ? NABOO:$ BACKUP/NORECORD/LOG *.TXT;*/SINCE=BACKUP TEST1.BCK/SAVE A %BACKUP-S-COPIED, copied DSA0:[SYS0.][SYSMGR]ANNOUNCE_ALCOR.TXT;2 B %BACKUP-S-COPIED, copied DSA0:[SYS0.][SYSMGR]ANNOUNCE_NABOO.TXT;12 %BACKUP-S-COPIED, copied7 DSA0:[SYS0.][SYSMGR]DECNET_REGISTER_IMPORT_FILE_NABOO.Te XT;1 %BACKUP-S-COPIED, copied7 DSA0:[SYS0.SYSCOMMON.][SYSMGR]TCPIP$BINDSETUP_HELP.TXT;  1nD %BACKUP-S-COPIED, copied DSA0:[SYS0.SYSCOMMON.][SYSMGR]WELCOME.TXT;1   *** BUT:  > NABOO:$ BACKUP/NORECORD/LOG *.TXT;*/SINCE=27-JUN-2002:09:28:18 TEST1.BCK/SAVEE %BACKUP-W-NOFILES, no files selected from DSA0:[SYS0.][SYSMGR]*.TXT;*C) %BACKUP-W-NOFILES, no files selected from % DSA0:[SYS0.SYSCOMMON.][SYSMGR]*.TXT;*(  0 I still don't get it...  what am I doing wrong ?   Greetings, Rolandt  F --------------- bbp - Biveroni Batschelet Partners AG ----------------:              Bahnhofstrasse 28, CH-5401 Baden, SwitzerlandF ----------------------------------------------------------------------   ------------------------------  / Date: Thu, 27 Jun 2002 10:41:12 +0200 (MET DST) & From: Rudolf Wingert <win@fom.fgan.de>' Subject: Re: Trouble with BACKUP/RECORDi6 Message-ID: <200206270841.KAA13490@sinet1.fom.fgan.de>   Hello,  I the show time command is irrelevant. Please look at the creation, record,oH etc. date and time. me be some in the future. So this won't be backuped.   Regards Rudolf Wingert   ------------------------------  # Date: Thu, 27 Jun 2002 09:19:38 GMT + From: Roland Barmettler <rob@bbp.ch.remove>n' Subject: Re: Trouble with BACKUP/RECORDh7 Message-ID: <20020627111935.54fbd1da.rob@bbp.ch.remove>C   Rudolf Wingert schrieb:e > Hello, > C > the show time command is irrelevant. Please look at the creation,fE > record, etc. date and time. me be some in the future. So this won'th > be backuped.  A The backup date in the file header reads: 27-JUN-2002 09:27:22.22eC which is clearly _before_ 27-JUN-2002 09:28:18 when I did the "SHOWw; TIME" after the backup/record. That's what's puzzling me...g   Greetings, Rolande  F --------------- bbp - Biveroni Batschelet Partners AG ----------------:              Bahnhofstrasse 28, CH-5401 Baden, Switzerland?        Roland Barmettler <rob@bbp.ch>  --  Phone +41-56-2039602iF ----------------------------------------------------------------------   ------------------------------  % Date: Thu, 27 Jun 2002 13:39:19 -0400 - From: "Richard D. Piccard" <piccard@ohio.edu>E' Subject: Re: Trouble with BACKUP/RECORDi( Message-ID: <3D1B4DC7.2AB4376E@ohio.edu>  Y I cannot retrieve the earlier postings, but I would expect the timestamp on the backup ofeW a file to be when the backup OF THAT FILE STARTED, which certainly ought to be prior too2 the time when the whole backup process completed.    				RDP'     Roland Barmettler wrote: >  > Rudolf Wingert schrieb:u
 > > Hello, > >eE > > the show time command is irrelevant. Please look at the creation,/G > > record, etc. date and time. me be some in the future. So this won't9 > > be backuped. > C > The backup date in the file header reads: 27-JUN-2002 09:27:22.22pE > which is clearly _before_ 27-JUN-2002 09:28:18 when I did the "SHOWc= > TIME" after the backup/record. That's what's puzzling me..., >  > Greetings, Rolandt > H > --------------- bbp - Biveroni Batschelet Partners AG ----------------< >              Bahnhofstrasse 28, CH-5401 Baden, SwitzerlandA >        Roland Barmettler <rob@bbp.ch>  --  Phone +41-56-2039602 H > ----------------------------------------------------------------------   -- eB ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------   End of INFO-VAX 2002.351 ************************