1 INFO-VAX	Thu, 27 Sep 2001	Volume 2001 : Issue 537       Contents: Re: 2nd CD drive on a PWS500au1 Re: A restore problem with VMS 7.1 Backup Utility 1 Re: A restore problem with VMS 7.1 Backup Utility & Re: Alias problem Backup image restore& Re: Alias problem Backup image restore& Re: Alias problem Backup image restore% Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS ) Re: Announcing ht://Dig 3.1.5 for OpenVMS * Re: Anyone found a use for LNM$SYSCLUSTER?* Re: Anyone found a use for LNM$SYSCLUSTER? Re: Apache/VMS Question  Apache/VMS Question  Re: Apache/VMS Question  Re: Apache/VMS Question  Re: Apache/VMS Question  Apcahe on VAX OpenVMS 6.1  Re: Apcahe on VAX OpenVMS 6.1  Available Qbus SCSI Controllers  Backing up an HSZ PCM + Re: c$main undefined in hello-world program + Re: c$main undefined in hello-world program + Re: c$main undefined in hello-world program  Cannot detect display device Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations Re: CETS2001 presentations DECTalk help needed  Re: DECTalk help needed  Re: DECTalk help needed  ECP data collector Re: EDT for Linux  Re: FORTRAN terminal IO  Re: FORTRAN terminal IO  Re: FORTRAN terminal IO  Re: FORTRAN terminal IO  Help with UNZIP.EXE  Re: Help with UNZIP.EXE  Re: Help with UNZIP.EXE  Re: Help with UNZIP.EXE  Re: Help with UNZIP.EXE ? Re: HOW DO I change FAB file attributes in C code?, $FAB_STORE? ? Re: HOW DO I change FAB file attributes in C code?, $FAB_STORE? ? Re: HOW DO I change FAB file attributes in C code?, $FAB_STORE? D Re: In times like these ... OpenVMS now more than ever for security! inserting module Re: inserting module( Is it possible to set system-wide flags?, Re: Is it possible to set system-wide flags?& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy" Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy& Re: Major VMS reseller says VMS legacy Need help with RMS Re: Need help with RMS$ Re: New..Wiring in terminals...help? News Group Observation' Re: Pathworks, Advanced Server issue...  Re: Really Unhackable...???? Re: Really Unhackable...???? Router load host or local   Re: Samba on OpenVMS protection." select() returning errno=1 (EPERM)' Re: Set Command for SUBPROCESS Question  SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMS Re: SMTP MAIL FROM VMSA Re: SNMP client software for OpenVMS (Alpha or VAX, Motif or CLI) = SNMP client software for OpenVMS (Alpha or VAX, Motif or CLI) A Re: SNMP client software for OpenVMS (Alpha or VAX, Motif or CLI)  Re: Sophos and PMDF ! Re: SuSE Linux Desupports ALPHA ? ! Re: SuSE Linux Desupports ALPHA ? ! Re: SuSE Linux Desupports ALPHA ? ! Re: SuSE Linux Desupports ALPHA ? ! Re: SuSE Linux Desupports ALPHA ? ! Re: SuSE Linux Desupports ALPHA ?  Re: UNIX Daemons -> OpenVMS ?  UNIX Daemons -> OpenVMS ?  Re: UNIX Daemons -> OpenVMS ? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away? # Re: VAX 7000 hw support going away?  Re: VAX disaster tolerance Re: VAX disaster tolerance Re: VAX disaster tolerance Re: VAX disaster tolerance Re: VAX disaster tolerance Re: VAX disaster tolerance Re: VAX disaster tolerance VMS V7.2-2 vs V7.3 Re: VMS V7.2-2 vs V7.3 Re: VMS V7.2-2 vs V7.3 Re: VMS V7.2-2 vs V7.35 Re: VMScluster using fibrechannel and memory channel? 6 RE:  VMScluster using fibrechannel and memory channel?5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel? 5 Re: VMScluster using fibrechannel and memory channel?  Re: We are back from CETS  weird MRU/scsi robot problems ! Re: weird MRU/scsi robot problems & Re: Whose Running The Oldest Software?& Re: Whose Running The Oldest Software? Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center Re: World Trade Center0 Re: [jfc] Seen in one of my older crash dumps...0 Re: [jfc] Seen in one of my older crash dumps...0 Re: [jfc] Seen in one of my older crash dumps...C Re: [OpenVMS] V7.2 VAX satellite doesn't find V7.3 Alpha bootserver 8 {HP,Compaq} user groups (was Re: CETS2001 presentations)< Re: {HP,Compaq} user groups (was Re: CETS2001 presentations)  F ----------------------------------------------------------------------  % Date: Wed, 26 Sep 2001 14:44:36 -0400 2 From: Atlant Schmidt <atlantnospam@mindspring.com>' Subject: Re: 2nd CD drive on a PWS500au . Message-ID: <3BB22214.CDB688A1@mindspring.com>   Dirk Munk wrote:  U > Every CD-Rom drive on a VMS system should be able to use 512 byte blocks instead of Q > the standard 4k blocks. Check if there is a jumper setting for 512 bytes on the  > Acer.   U That's a non-sequitor. The IDE/ATA/ATAPI driver knows how to operate CD-ROMs in their R native 2K-byte-block mode; I know -- I wrote it. Only the SCSI CD-ROM driver cared about 512-byte-block mode.  P This stuff is supposed to work and certainly has worked in the past, on both theT Cypress and Acer IDE chips. During testing, I ran CD-ROMs simultaneously on all four
 IDE ports.  U "%MOUNT-F-IVADDR, invalid media address" rings a bell with me, but I can't recall the V details; I'm think that it was a known error that occurred with certain early versionsM of the DQDRIVER with certain sized CD-ROM volumes. Make sure your DQDRIVER is V up-to-date and if it is, SPR the problem. By the way, SYS$EXAMPLES had the source codeV for the driver and there were extensive comments about the bugs fixed in each version.   AtlantC (Who is no longer professionally associated with any of this stuff)    ------------------------------  % Date: Wed, 26 Sep 2001 08:55:56 -0400 0 From: "Syltrem" <syltrem@videotron.spammenot.ca>: Subject: Re: A restore problem with VMS 7.1 Backup Utility5 Message-ID: <5cks7.62653$TW.340605@tor-nn1.netcom.ca>   K The request will go at the operator console (OPA0:)  if you use /ASSIST and . for that to work OPCOM must be up and running.K Otherwise the message will be output at your terminal (or .LOG wherever the  BACKUP command is running from)   G In the process that executes the backup, what is the last thing you see  (message) before it froze?- Do you have OPCOM running (SH SYS/PROC=OPCOM)  --   Syltrem I http://pages.infinit.net/syltrem (OpenVMS related web site - en franais) > To reply to myself directly, remove .spammenot from my address  I "Andrew Aylward" <Andrew.Aylward@ast.co.za> a crit dans le message news: = EE170DD1AC76FC4D8A71DA5912DBBF6E050C05@nnc-exc01.ast.co.za...  Good Day Everyone   H I have a problem with a restore that I am doing.  The backup was made ofH all the data on a device "DRA1".  The save set ran along two tapes, bothE with proper ANSI labels.  The problem I have is that when the restore E gets to the end of tape1 it spits out the tape (from device "MKA500") H and then hangs.  I was told by another of the more experienced guys hereB that it is meant to prompt you for the next tape saying, "OperatorH Intervention required... type YES..."  But what it does is just hang and does no prompting at all. H Should I use /ASSIST, /LABEL, or even /EXACT_ORDER to help me with this," or am I barking up the wrong tree?   Yours faithfully   Andrew Aylward AST Group - Newcastle Branch Private Bag X6613  Newcastle, 2940  Republic of South Africa   W:        +2734 314 8153 Cell:      +2783 652 6708  ____________________E Any views or opinions presented are solely those of the author and do ? not necessarily represent those of the AST Group or any of it's C affiliates. This message is for the named person's use only. It may G contain confidential, proprietary or legally privileged information. No G confidentiality or privilege is waived or lost by any mis-transmission. F If you receive this message in error, please immediately delete it andD all copies of it from your system, destroy any hard copies of it andG notify the sender. You must not, directly or indirectly, use, disclose, F distribute, print, or copy any part of this message if you are not theF intended recipient. AST GROUP reserves the right to monitor all e-mailH communications through its networks. Any views expressed in this messageC are those of the individual sender, except where the message states G otherwise and the sender is authorised to state them to be the views of G any such entity. Unless otherwise stated, any pricing information given E in this message is indicative only, is subject to change and does not E constitute an offer to deal at any price quoted. Any reference to the H terms of executed transactions should be treated as preliminary only and+ subject to our formal written confirmation.    ------------------------------  % Date: Wed, 26 Sep 2001 14:54:37 +0200 2 From: "Ren Schelbaum" <rene.schelbaum@datakom.at>: Subject: Re: A restore problem with VMS 7.1 Backup UtilityG Message-ID: <3bb1cf4a$0$31972$5039e797@newsreader01.highway.telekom.at>   B "Andrew Aylward" <Andrew.Aylward@ast.co.za> schrieb im NewsbeitragB news:EE170DD1AC76FC4D8A71DA5912DBBF6E050C05@nnc-exc01.ast.co.za... Good Day Everyone   H I have a problem with a restore that I am doing.  The backup was made ofH all the data on a device "DRA1".  The save set ran along two tapes, bothE with proper ANSI labels.  The problem I have is that when the restore E gets to the end of tape1 it spits out the tape (from device "MKA500") H and then hangs.  I was told by another of the more experienced guys hereB that it is meant to prompt you for the next tape saying, "OperatorH Intervention required... type YES..."  But what it does is just hang and does no prompting at all. H Should I use /ASSIST, /LABEL, or even /EXACT_ORDER to help me with this," or am I barking up the wrong tree?   Yours faithfully   Andrew Aylward AST Group - Newcastle Branch Private Bag X6613  Newcastle, 2940  Republic of South Africa   W:        +2734 314 8153 Cell:      +2783 652 6708  ____________________E Any views or opinions presented are solely those of the author and do ? not necessarily represent those of the AST Group or any of it's C affiliates. This message is for the named person's use only. It may G contain confidential, proprietary or legally privileged information. No G confidentiality or privilege is waived or lost by any mis-transmission. F If you receive this message in error, please immediately delete it andD all copies of it from your system, destroy any hard copies of it andG notify the sender. You must not, directly or indirectly, use, disclose, F distribute, print, or copy any part of this message if you are not theF intended recipient. AST GROUP reserves the right to monitor all e-mailH communications through its networks. Any views expressed in this messageC are those of the individual sender, except where the message states G otherwise and the sender is authorised to state them to be the views of G any such entity. Unless otherwise stated, any pricing information given E in this message is indicative only, is subject to change and does not E constitute an offer to deal at any price quoted. Any reference to the H terms of executed transactions should be treated as preliminary only and+ subject to our formal written confirmation.      Hi!    /assist is the default!   J Guess, if you log in as an operator, type 'reply /enable' and 'reply/stat'E you will find an outstanding operator request for mounting the second  volume. G What you have to do is either mount the second volume and reply to that ; request, or abort the backup and restart it with /noassist. L This way it will ask you for the second tape on the terminal you started the backup from.   Regards    Ren Schelbaum   ------------------------------    Date: 26 Sep 2001 10:14:04 -0700- From: afeldman@gfigroup.com (Alan E. Feldman) / Subject: Re: Alias problem Backup image restore = Message-ID: <af1e4ce6.0109260914.74e2f4b6@posting.google.com>    "Hansford, Paul (ELS)" <P.Hansford@elsevier.co.uk> wrote in message news:<FFDBA0B630B5D211954E0008C70D240D05B6CA4E@elsoxfs12304.elsevier.co.uk>...1 > Hi this is my first ever submission to info-vax  > OS: Vms Version 6.2 onwards N > Problem:  VMS backup processes alias file (set file/ent)and makes two copies
 > on the tape B > If I were performing an image backup restore I would have to use/ > backup/image/noalias <tapedrive>:saveset/save M > because if you use /alias (which is the default) it would process the alias J > file rather than just copy the link in indexf.sys, so doubling up on theN > number of files on the tape if there was one alias per file.  When I came to1 > restore the disk I would run out of disk space.  > H > What I would like to know is, does anyone know of a utility which willL > display all known file aliases,  it is not shown in dir/full.  If there isN > no util around has anyone else encountered this problem, and did you resolve > it?   4 DFU from the OpenVMS Freeware CD (you can find it on3 www.openvms.compaq.com) will show all file aliases.   < Try using BACKUP/IMAGE/NOALIAS when backing up your disk andA BACKUP/IMAGE when restoring your disk. Next time, please post the   exact commands you used. Thanks.  8 Disclaimer: JMHO. This is not a statement from gfigroup. Alan E. Feldman  afeldman&gfigroup.com    ------------------------------   Date: 26 Sep 2001 10:37:08 GMT3 From: gartmann@immunbio.mpg.de (Christoph Gartmann) / Subject: Re: Alias problem Backup image restore 0 Message-ID: <9osb4k$5i2$1@n.ruf.uni-freiburg.de>   In article <FFDBA0B630B5D211954E0008C70D240D05B6CA4E@elsoxfs12304.elsevier.co.uk>, "Hansford, Paul (ELS)" <P.Hansford@elsevier.co.uk> writes: 0 >Hi this is my first ever submission to info-vax >OS: Vms Version 6.2 onwardsM >Problem:  VMS backup processes alias file (set file/ent)and makes two copiesb >on the tapeA >If I were performing an image backup restore I would have to uset. >backup/image/noalias <tapedrive>:saveset/saveL >because if you use /alias (which is the default) it would process the aliasI >file rather than just copy the link in indexf.sys, so doubling up on thenM >number of files on the tape if there was one alias per file.  When I came to 0 >restore the disk I would run out of disk space. >MG >What I would like to know is, does anyone know of a utility which will.K >display all known file aliases,  it is not shown in dir/full.  If there ishM >no util around has anyone else encountered this problem, and did you resolveo >it?  M Using /IMAGE when creating the backup should resolve your problem. I remembero/ that there was a backup patch as well for 6.2 .u   Regards,    Christoph Gartmannp  H -- --------------------------------------------------------------------+H | Max-Planck-Institut fuer      Phone   : +49-761-5108-464   Fax: -452 |H | Immunbiologie                                                        |H | Postfach 1169                 Internet: gartmann@immunbio.mpg.de     |H | D-79011  Freiburg, FRG                                               |H +--------- http://www.immunbio.mpg.de/home/english/menue.html ---------+   ------------------------------  # Date: Wed, 26 Sep 2001 23:52:20 GMTe- From: "Dave Pampreen" <davepampreen@home.com>t/ Subject: Re: Alias problem Backup image restoree, Message-ID: <USts7.130925$K6.59860084@news2>  3 Alan - Will those commands restore the alias links?E   In other words If> DIR1.DIR and DIR2.DIR were aliases, would DIR2.DIR come back ?   Dave   Dave: "Alan E. Feldman" <afeldman@gfigroup.com> wrote in message7 news:af1e4ce6.0109260914.74e2f4b6@posting.google.com...eE > "Hansford, Paul (ELS)" <P.Hansford@elsevier.co.uk> wrote in messageeL news:<FFDBA0B630B5D211954E0008C70D240D05B6CA4E@elsoxfs12304.elsevier.co.uk>. ..3 > > Hi this is my first ever submission to info-vaxe > > OS: Vms Version 6.2 onwardsII > > Problem:  VMS backup processes alias file (set file/ent)and makes two  copies > > on the tape D > > If I were performing an image backup restore I would have to use1 > > backup/image/noalias <tapedrive>:saveset/saveeI > > because if you use /alias (which is the default) it would process theI alias L > > file rather than just copy the link in indexf.sys, so doubling up on theH > > number of files on the tape if there was one alias per file.  When I came to 3 > > restore the disk I would run out of disk space.e > >eJ > > What I would like to know is, does anyone know of a utility which willK > > display all known file aliases,  it is not shown in dir/full.  If theree isH > > no util around has anyone else encountered this problem, and did you resolves > > it?e >y6 > DFU from the OpenVMS Freeware CD (you can find it on5 > www.openvms.compaq.com) will show all file aliases.M >e> > Try using BACKUP/IMAGE/NOALIAS when backing up your disk andC > BACKUP/IMAGE when restoring your disk. Next time, please post theG" > exact commands you used. Thanks. >e: > Disclaimer: JMHO. This is not a statement from gfigroup. > Alan E. Feldman  > afeldman&gfigroup.com    ------------------------------  % Date: Wed, 26 Sep 2001 09:47:03 +0200 < From: "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de>. Subject: Announcing ht://Dig 3.1.5 for OpenVMS4 Message-ID: <9os15r$epg1o$1@ID-56200.news.dfncis.de>  J ht://Dig is "a system for indexing and searching a finite (not necessarilyK small) set of sites or intranet. It is not meant to replace any of the manyO internet-wide search engines."7 (Quoting from the FAQ at http://www.htdig.org/FAQ.html)   J An early adopters kit of my port of ht://Dig 3.1.5 to OpenVMS is availableL at ftp://ftp.pdv-systeme.de/vms/htdig-3_1_5-vms.zip (5.7 MB, includes sourceH and documentation). An installation guideline is in the README.VMS file.E It has been tested with WASD HTTPd 7.1.1 and cursorily with CSWS 1.1.   I While this kit already is fully functional, I'd like some feedback beforey I go on working on it.  I The port was developed on OpenVMS Alpha V7.2-1H1 using Compaq C V6.4-008,rE Compaq C++ V6.2-016, and MMS V3.4-3. On this occasion: Many thanks tou< DECpaq and DECUS/Encompass for the OpenVMS Hobbyist program.   Enjoy,   Martin  C P.S.: Is there a way to set the LRL attribute *after* writing (i.e.pG       just before closing) the file? A file gets created by one programuC       (Stream_LF, LRL 32767), and then SOR$ted by another program -0D       which devours time and disk space. As a workaround, I insertedE       a call to CONVERT which renders the file as (LRL less than 100) #       which helps the SOR$t (much).t --J One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer7 One OS to find them           | work: mv@pdv-systeme.de J One OS to bring them all      |   http://www.pdv-systeme.de/users/martinv/> And in the Darkness bind them.| home: martin@radiogaga.harz.de   ------------------------------  % Date: Wed, 26 Sep 2001 22:24:23 +0200l  From: Paul Sture <paul@sture.ch>2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS+ Message-ID: <VA.00000462.16e8c6ef@sture.ch>u  N In article <3bb219c8.524144494f47414741@radiogaga.harz.de>, Martin Vorlaender  wrote:# > Paul Sture (paul@sture.ch) wrote:e > > Robert Deininger wrote:MK > > > In C, you can fopen with "rfm=var", "rat=cr".  I recently noticed thehN > > > miserable sort performance on C-created stream files, and this change to > > > fopen cured it.g > >eK > > I've seen this too. The sort workfiles were using several GB of space, aO > > blowing up before the sort was complete, even on a surprisingly small file.i > " > Yup, that's what I noticed, too. > J > > I didn't know which sources to hit so put in the following workaround: > >eK > > --------------------start of DCL -------------------------------------- 4 > > $ CONVERT /FDL=SYS$INPUT  infile.dat outfile.dat > > FILE+ > >      ORGANIZATION            sequentiall > 
 > > RECORD0 > >      CARRIAGE_CONTROL        carriage_return) > >      FORMAT                  variable # > > $ ! Convert Job done, now sort:s6 > > $ SORT infile.dat outfile.dat        ! do the sortL > > ----------------------end of DCL --------------------------------------- > F > I'll use the CONV$ interface (as the SOR$t also gets executed in theE > program), if I find that I can't use the fopen parameters - I'm not C > sure whether I already know all the execution paths in the rathere > largish programs.  > ; Firstly, in my example above the last line should have readn     $ SORT outfile.dat *  M Secondly, someone else suggested using /FDL=NLA0:, though I haven't tried it p yet. ___o
 Paul Sture Switzerlands   ------------------------------  % Date: Wed, 26 Sep 2001 11:16:31 -0400 2 From: rdeininger@mindspring.com (Robert Deininger)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMSL Message-ID: <rdeininger-2609011116310001@user-2ivec5b.dialup.mindspring.com>  H In article <9os15r$epg1o$1@ID-56200.news.dfncis.de>, "Martin Vorlaender") <martin.vorlaender@pdv-systeme.de> wrote:     E > P.S.: Is there a way to set the LRL attribute *after* writing (i.e.aI >       just before closing) the file? A file gets created by one programcE >       (Stream_LF, LRL 32767), and then SOR$ted by another program -eF >       which devours time and disk space. As a workaround, I insertedG >       a call to CONVERT which renders the file as (LRL less than 100)u% >       which helps the SOR$t (much).s  B This begs the question:  why not create the file with the sensible attributes in the first place?  E In C, you can fopen with "rfm=var", "rat=cr".  I recently noticed therH miserable sort performance on C-created stream files, and this change to fopen cured it.t   -- u Robert Deininger rdeininger@mindspring.comI   ------------------------------  % Date: Wed, 26 Sep 2001 17:31:42 +0200a  From: Paul Sture <paul@sture.ch>2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS+ Message-ID: <VA.00000461.15dcd29a@sture.ch>o   In article eI <rdeininger-2609011116310001@user-2ivec5b.dialup.mindspring.com>, Robert i Deininger wrote:J > In article <9os15r$epg1o$1@ID-56200.news.dfncis.de>, "Martin Vorlaender"+ > <martin.vorlaender@pdv-systeme.de> wrote:m > G > > P.S.: Is there a way to set the LRL attribute *after* writing (i.e.-K > >       just before closing) the file? A file gets created by one programoG > >       (Stream_LF, LRL 32767), and then SOR$ted by another program -mH > >       which devours time and disk space. As a workaround, I insertedI > >       a call to CONVERT which renders the file as (LRL less than 100)e' > >       which helps the SOR$t (much).l > D > This begs the question:  why not create the file with the sensible  > attributes in the first place? > G > In C, you can fopen with "rfm=var", "rat=cr".  I recently noticed the J > miserable sort performance on C-created stream files, and this change to > fopen cured it.l >dG I've seen this too. The sort workfiles were using several GB of space, FL blowing up before the sort was complete, even on a surprisingly small file. F I didn't know which sources to hit so put in the following workaround:  G --------------------start of DCL --------------------------------------l0 $ CONVERT /FDL=SYS$INPUT  infile.dat outfile.dat FILE'      ORGANIZATION            sequential(   RECORD,      CARRIAGE_CONTROL        carriage_return%      FORMAT                  variablee $ ! Convert Job done, now sort:f2 $ SORT infile.dat outfile.dat        ! do the sortH ----------------------end of DCL --------------------------------------- ___a
 Paul Sture SwitzerlandT   ------------------------------  % Date: Wed, 26 Sep 2001 19:28:30 +0200eI From: =?iso-8859-1?Q?Jean=2DFran=E7ois=20PI=C9RONNE?= <jfp@altavista.net>r2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS- Message-ID: <3BB2103E.27F43C83@altavista.net>e  P I have indexed the same pages (7000 files, 300000 blocks) using htdig and swish:  5 - swish: less than 4 minutes, index size 73000 blockss/ - htdig: near 4 hours, index size 380000 blocksi  L one of the major difference is that swish can index directly file, and htdig  retrieve file using http access.I I have not found how to access file directly with htdig, may be it is notn
 possible ?  7 Using htsearch.exe interactively, i have two questions:d Enter value for words: Enter value for format:s  " which generate a correct HTML file  O but if i try it as a CGI (activation from the search.html, WASD 7.2) i have thei following error:  B ERROR 502  -  External agent did not respond (or not acceptably). . Script did not provide an acceptable response.   From the help: 502 Bad Gateway L The server, while acting as a gateway or proxy, received an invalid responseK from the upstream server it accessed in attempting to fulfill the request. t      N tracing the dialog between WASD and htsearch show the same string (Enter value@ for words:) send back to WASD which lead to the describe problem  
 Any idea ?     Jean-Fancois Pieronne    ------------------------------  % Date: Wed, 26 Sep 2001 20:09:12 +0200a2 From: martin@radiogaga.harz.de (Martin Vorlaender)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS; Message-ID: <3bb219c8.524144494f47414741@radiogaga.harz.de>0  ! Paul Sture (paul@sture.ch) wrote:h > Robert Deininger wrote::I > > In C, you can fopen with "rfm=var", "rat=cr".  I recently noticed the$L > > miserable sort performance on C-created stream files, and this change to > > fopen cured it.r > I > I've seen this too. The sort workfiles were using several GB of space, oM > blowing up before the sort was complete, even on a surprisingly small file.     Yup, that's what I noticed, too.  H > I didn't know which sources to hit so put in the following workaround: >lI > --------------------start of DCL --------------------------------------p2 > $ CONVERT /FDL=SYS$INPUT  infile.dat outfile.dat > FILE) >      ORGANIZATION            sequentialp   > RECORD. >      CARRIAGE_CONTROL        carriage_return' >      FORMAT                  variabler! > $ ! Convert Job done, now sort:t4 > $ SORT infile.dat outfile.dat        ! do the sortJ > ----------------------end of DCL ---------------------------------------  D I'll use the CONV$ interface (as the SOR$t also gets executed in theC program), if I find that I can't use the fopen parameters - I'm notuA sure whether I already know all the execution paths in the rather6 largish programs._   Thanks,    Martin -- iD                     | Martin Vorlaender    |    VMS & WNT programmer-   Smiert Spamionem  | work: mv@pdv-systeme.de'D                     |       http://www.pdv-systeme.de/users/martinv/4                     | home: martin@radiogaga.harz.de   ------------------------------  % Date: Wed, 26 Sep 2001 20:03:38 +0200n2 From: martin@radiogaga.harz.de (Martin Vorlaender)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS; Message-ID: <3bb2187a.524144494f47414741@radiogaga.harz.de>o  3 Robert Deininger (rdeininger@mindspring.com) wrote:u? > "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de> wrote:gG > > P.S.: Is there a way to set the LRL attribute *after* writing (i.e. K > >       just before closing) the file? A file gets created by one programiG > >       (Stream_LF, LRL 32767), and then SOR$ted by another program -aH > >       which devours time and disk space. As a workaround, I insertedI > >       a call to CONVERT which renders the file as (LRL less than 100)o' > >       which helps the SOR$t (much).n > D > This begs the question:  why not create the file with the sensible  > attributes in the first place?  E Because I cannot know beforehand how long the records written will beaB (and BTW, I haven't yet found - in the pile of classes - the place the file gets opened ;-)  G > In C, you can fopen with "rfm=var", "rat=cr".  I recently noticed the J > miserable sort performance on C-created stream files, and this change to > fopen cured it.d  5 I'll try that - as soon as I find the fopen, that is.e   Thanks,b   Martin -- aF                           | Martin Vorlaender  |  VMS & WNT programmer3  Cetero censeo            | work: mv@pdv-systeme.de F  Redmondem delendam esse. |   http://www.pdv-systeme.de/users/martinv/:                           | home: martin@radiogaga.harz.de   ------------------------------  % Date: Wed, 26 Sep 2001 21:26:52 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS; Message-ID: <3bb22bfc.524144494f47414741@radiogaga.harz.de>A  J =?iso-8859-1?Q?Jean=2DFran=E7ois=20PI=C9RONNE?= (jfp@altavista.net) wrote:G > I have indexed the same pages (7000 files, 300000 blocks) using htdigs > and swish: >f7 > - swish: less than 4 minutes, index size 73000 blocks 1 > - htdig: near 4 hours, index size 380000 blocksu  ? Can't comment on that, as I haven't tried swish in a long time.y  N > one of the major difference is that swish can index directly file, and htdig" > retrieve file using http access.  K Which bears the possibility to index a site not residing on the local host.r@ And all the documents are indexed exactly as the user sees them.  K > I have not found how to access file directly with htdig, may be it is notc > possible ?  I Please look up the "local_urls" configuration file keyword. I haven't yetuC touched the "local_user_urls" portion, so that probably won't work.l  9 > Using htsearch.exe interactively, i have two questions:i > Enter value for words: > Enter value for format:s > $ > which generate a correct HTML file >aH > but if i try it as a CGI (activation from the search.html, WASD 7.2) i > have the following error:n >yD > ERROR 502  -  External agent did not respond (or not acceptably). 0 > Script did not provide an acceptable response.  I I've seen this error if htsearch doesn't have access to any of the commont< or database files. Make sure these are owned by HTTP$SERVER.  E HTTP status codes are a really *bad* indicator as to what went wrong.u$ Error logs are much better for that.  J > tracing the dialog between WASD and htsearch show the same string (EnterH > value for words:) send back to WASD which lead to the describe problem  C The interactive portion only is activated if htsearch doesn't see asH REQUEST_METHOD. I suggest to deactivate the "WWW_" CGI variables' prefix@ on the directory htsearch is in. I'll put that on the ToDo list.   cu,/   Martin -- iG  Your mouse has moved.     | Martin Vorlaender  |  VMS & WNT programmeru4  Windows must be restarted | work: mv@pdv-systeme.deH  for the change to take    |    http://www.pdv-systeme.de/users/martinv/;  effect. Reboot now? [OK]  | home: martin@radiogaga.harz.dee   ------------------------------   Date: 26 Sep 2001 16:36 PDTn+ From: rankin@eql14.caltech.edu (Pat Rankin)l2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS1 Message-ID: <26SEP200116362501@eql14.caltech.edu>n  5 In article <9os15r$epg1o$1@ID-56200.news.dfncis.de>,\ A  "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de> writes...rE > P.S.: Is there a way to set the LRL attribute *after* writing (i.e. & >       just before closing) the file?  :      If you write the file in record mode instead of using0 stream mode then RMS will do that automatically.  * >       A file gets created by one programE >       (Stream_LF, LRL 32767), and then SOR$ted by another program -1F >       which devours time and disk space. As a workaround, I insertedG >       a call to CONVERT which renders the file as (LRL less than 100)-% >       which helps the SOR$t (much).h  8      I don't know how to change the attributes when C or7 C++ still has the file open, but if you get the routinen: LYVMS_FixedLengthRecords() out of the lynx sources, you'll8 find something which does the type of operation you want7 after closing the file.  It's setting a couple of others9 attributes in addition of the LRL and MRS fields, but thes8 other stuff could be removed easily.  I'm sure there are8 umpteen other examples of using the XQP interface to set4 file attributes.  Now that I think about it, the RMS8 routine SYS$MODIFY() might be able to handle this simple	 task too.s  :      If you're processing the data with multiple programs,8 why not just stick a SET FILE /ATTRIBUTE=LRL=xxx command8 in between them?  That's effectively what the code above7 would be doing anyway, but with SET FILE you won't need  to do any extra coding....  2                 Pat Rankin, rankin@eql.caltech.edu   ------------------------------  % Date: Thu, 27 Sep 2001 05:07:59 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS; Message-ID: <3bb2980f.524144494f47414741@radiogaga.harz.de>b  ! Paul Sture (paul@sture.ch) wrote:o6 > > > $ CONVERT /FDL=SYS$INPUT  infile.dat outfile.dat
 > > > FILE- > > >      ORGANIZATION            sequentiall > >  > > > RECORD2 > > >      CARRIAGE_CONTROL        carriage_return+ > > >      FORMAT                  variable/% > > > $ ! Convert Job done, now sort:d8 > > > $ SORT infile.dat outfile.dat        ! do the sort >v= > Firstly, in my example above the last line should have readi >     $ SORT outfile.dat *  B I wasn't aware that you can wildcard the output file - this really@ fascinates me about this OS: after 10+ years of VMS, I still see something new every day.  E > Secondly, someone else suggested using /FDL=NLA0:, though I haven't  > tried it yet.t  D Yup, I guess that's what I do now (by leaving out the /FDL qualifier
 alltogether).-   cu,-   Martin -- eF                           | Martin Vorlaender  |  VMS & WNT programmer3  Cetero censeo            | work: mv@pdv-systeme.de1F  Redmondem delendam esse. |   http://www.pdv-systeme.de/users/martinv/:                           | home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 27 Sep 2001 05:00:01 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS; Message-ID: <3bb29631.524144494f47414741@radiogaga.harz.de>p  , Pat Rankin (rankin@eql14.caltech.edu) wrote:7 > In article <9os15r$epg1o$1@ID-56200.news.dfncis.de>,\uC >  "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de> writes...hG > > P.S.: Is there a way to set the LRL attribute *after* writing (i.e.l( > >       just before closing) the file? >i< >      If you write the file in record mode instead of using2 > stream mode then RMS will do that automatically.  E Problem is, I'll have to find all subsequent fopen()s then and modify : them accordingly. But that's probably what I'll do anyway.  , > >       A file gets created by one programG > >       (Stream_LF, LRL 32767), and then SOR$ted by another program -rH > >       which devours time and disk space. As a workaround, I insertedI > >       a call to CONVERT which renders the file as (LRL less than 100)i' > >       which helps the SOR$t (much).e >d: >      I don't know how to change the attributes when C or9 > C++ still has the file open, but if you get the routine < > LYVMS_FixedLengthRecords() out of the lynx sources, you'll: > find something which does the type of operation you want > after closing the file. [...]   D Which, of course, is as good. Thanks for the hint. I'll have a look.   > I'm sure there are: > umpteen other examples of using the XQP interface to set > file attributes. 4  9 That's the primary reason why I was asking. I'm not (yet)n6 very comfortable with handling files at the XQP level.  $ > Now that I think about it, the RMS: > routine SYS$MODIFY() might be able to handle this simple > task too.:  9 Errr, beg your pardon? There's no such routine in my copyw of the RMS Services Reference?!   < >      If you're processing the data with multiple programs,: > why not just stick a SET FILE /ATTRIBUTE=LRL=xxx command: > in between them?  That's effectively what the code above9 > would be doing anyway, but with SET FILE you won't need  > to do any extra coding....  > I'd have to code it as a system() call then, because after the@ end of the first program, I don't have the correct LRL any more.  C But that's effectively what I do now, with a call to CONVERT (plain  simple, without an /FDL).-   Thanks,-   Martin -- OD                        |  Martin Vorlaender  |  VMS & WNT programmer1   OpenVMS: When you    |  work: mv@pdv-systeme.de E   KNOW where you want  |     http://www.pdv-systeme.de/users/martinv/ 8   to go today.         |  home: martin@radiogaga.harz.de   ------------------------------  % Date: Thu, 27 Sep 2001 05:25:49 +0200-2 From: martin@radiogaga.harz.de (Martin Vorlaender)2 Subject: Re: Announcing ht://Dig 3.1.5 for OpenVMS; Message-ID: <3bb29c3d.524144494f47414741@radiogaga.harz.de>t  3 Martin Vorlaender (martin@radiogaga.harz.de) wrote:aL > =?iso-8859-1?Q?Jean=2DFran=E7ois=20PI=C9RONNE?= (jfp@altavista.net) wrote:F > > one of the major difference is that swish can index directly file,. > > and htdig retrieve file using http access. >uM > Which bears the possibility to index a site not residing on the local host.nB > And all the documents are indexed exactly as the user sees them.  K A related thought: Using ht://Dig with WASD HTTPd's Conan the librarian andnH HyperShelf, you can index the names in your libraries (not too useful, IB guess), and all of your online documentation. Try that with SWISH.   cu,0   Martin -- eD                        |  Martin Vorlaender  |  VMS & WNT programmer1   OpenVMS: When you    |  work: mv@pdv-systeme.deaE   KNOW where you want  |     http://www.pdv-systeme.de/users/martinv/s8   to go today.         |  home: martin@radiogaga.harz.de   ------------------------------  % Date: Wed, 26 Sep 2001 07:11:02 -0700 * From: James Gessling <jgessling@yahoo.com>3 Subject: Re: Anyone found a use for LNM$SYSCLUSTER? ) Message-ID: <3BB1E1F6.D4C50778@yahoo.com>m   Arthur Cochrane wrote:  K >     Has anyone found a good use for the VMS 7.x cluster wide logical namec >     table, LNM$SYSCLUSTER.  K I'll leave it to you to judge wether this is a "good use" or not, but I can-N describe how we us it.  On each node in our cluster (12 nodes) we run the WASDM web server.  A batch job on one of the nodes collects various data concerning,J the application configuration and performance.  This job sets cluster wideK logical names with this data.  The web server scripts format this data intor> nice looking displays, tables and such, for the users browser.  H Things to note:  Having a web server on every node doesn't cost much andL allows for using the (IP) cluster alias to be used in the URL.  Only one jobK is required to collect the data, not one on each node.  Reading the logicalaK names and formating for display is a a pretty lightweight task, compared toe* the web script collecting the data itself.   Jim6  N p.s. I know that if the job that collects the data goes down, the updates willK stop.  That's why there's a timestamp logical name too, that's displayed at K the bottom of each page.  If it's noticed to be out of data, the collection # job can be released on another nodec   ------------------------------  % Date: Tue, 25 Sep 2001 15:36:46 -0400n- From: "Richard D. Piccard" <piccard@ohio.edu>r3 Subject: Re: Anyone found a use for LNM$SYSCLUSTER? ( Message-ID: <3BB0DCC2.CD4800E2@ohio.edu>  M One possibility is to use the lexical F$TRNLNM, specifying the table, and seea< if the logical name(s) in question is (are) already defined.  #                                 RDPn     Arthur Cochrane wrote:  K >     Has anyone found a good use for the VMS 7.x cluster wide logical name-L >     table, LNM$SYSCLUSTER. It sounds like a good idea but how often do youJ >     need to redefine a system wide logical. Also, when a system boots isL >     there some way to see if this is the first system to boot and needs toH >     define the cluster wide logicals or if not the first then skip theJ >     definition of the cluster wide logicals? Or just define them anyway. >-I >     I wish there was a LNM_SYSTEM logical name table for system manager K >     logicals names in the search list between LNM$GROUP and LNM$SYSTEM. I=A >     have done this myself but wish it was supported officially.1 >fJ >     I also wish there was a VMS_COMMON logical in the SYS$SYSROOT searchI >     list. This would be empty on a fresh install and the system managersI >     would use to create the necessary procedures to boot and manage thedK >     system. (Also, a SYS_MANAGER, SYS_SYSTEM, SYS_HELP, etc.) I have also < >     done this myself but wish it was supported officially.   --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------    Date: 26 Sep 2001 09:00:40 -07001 From: nothome@spammers.are.scum (Malcolm Dunnett)l  Subject: Re: Apache/VMS Question, Message-ID: <Ea1$55VL6R+9@malvm6.mala.bc.ca>  . In article <3BB1C83C.C3B3798B@bellsouth.net>, 3     Michael Austin <miaustin@bellsouth.net> writes:a  O > First, your program would have to understand how to "send" the information tol? > your webserver as this is generally an interactive interface.o > F    The HTTP protocol is well documented in the relevant RFCs and isn'tJ that hard to implement. I've written programs that do something along thisF line in a hundred or so lines of DEC BASIC. It's also possible to findK prebuilt modules in languages like Perl which implement this functionality.b  Q    Once you have the HTTP communication part written it's quite a simple process, I you just connect to the server, send your form parameters to a cgi script < running on the Apache/VMS server and read back the response.   ------------------------------  % Date: Tue, 25 Sep 2001 20:16:32 -0400s  From: Kuff@Tessco.Com (Hal Kuff) Subject: Apache/VMS QuestionO Message-ID: <70B7D1B67CAAEC8A.13701515DDF7570F.34BB7A2A6DF39B20@lp.airnews.net>   L    Does anyone know if it is possible to use Apache in the following manner.  H    Apache web server on OpeVMS, remote client (actually a program ratherC than a browser) does a post of 4096 bytes to a form, the form callsiJ foo.exe on the alpha... changes the contents of the 4096 byte argument andC returns it to the remote client ... how would one orchestrate that?    ------------------------------  % Date: Wed, 26 Sep 2001 05:41:11 +0200f2 From: martin@radiogaga.harz.de (Martin Vorlaender)  Subject: Re: Apache/VMS Question; Message-ID: <3bb14e57.524144494f47414741@radiogaga.harz.de>e  ! Hal Kuff (Kuff@Tessco.Com) wrote:tG > Apache web server on OpeVMS, remote client (actually a program rather E > than a browser) does a post of 4096 bytes to a form, the form callsgL > foo.exe on the alpha... changes the contents of the 4096 byte argument andE > returns it to the remote client ... how would one orchestrate that?s  F If the client produces a valid POST request to, say, /cgi-bin/foo.exe,G foo.exe (which must be located in a directory marked +ExecCGI) can read-% it in and return whatever it desires.e   What's the problem?    cu,a   Martin -- hG So long, and thanks        | Martin Vorlaender  |  VMS & WNT programmere4 for all the books...       | work: mv@pdv-systeme.deK In Memoriam Douglas Adams  |       http://www.pdv-systeme.de/users/martinv/i;             1952-2001      | home: martin@radiogaga.harz.de    ------------------------------  % Date: Wed, 26 Sep 2001 08:21:17 -0400 - From: Michael Austin <miaustin@bellsouth.net>c  Subject: Re: Apache/VMS Question- Message-ID: <3BB1C83C.C3B3798B@bellsouth.net>t   Hal,  M First, your program would have to understand how to "send" the information to = your webserver as this is generally an interactive interface.   I Can you explain  the program?  I do something similar using Rdb with somesJ external procedures, althugh mine is interactive.  I have a form that getsO updated by a user, it goes to the database, updated and to process it ,  I have 6 a "c" program that prints it out to a network printer.  P You would need to code using an API into lynx or some other browser (not sure ifL an API exists) to use it as the transport.  Then once on your server, can be* changed, updated and returned via the API.   Michael Austin First DBA Source, Inc. 704-947-1089   Hal Kuff wrote:   N >    Does anyone know if it is possible to use Apache in the following manner. > J >    Apache web server on OpeVMS, remote client (actually a program ratherE > than a browser) does a post of 4096 bytes to a form, the form calls L > foo.exe on the alpha... changes the contents of the 4096 byte argument andE > returns it to the remote client ... how would one orchestrate that?t   ------------------------------  % Date: Wed, 26 Sep 2001 20:32:01 +0200-2 From: martin@radiogaga.harz.de (Martin Vorlaender)  Subject: Re: Apache/VMS Question; Message-ID: <3bb21f21.524144494f47414741@radiogaga.harz.de>d  3 Martin Vorlaender (martin@radiogaga.harz.de) wrote:sH > If the client produces a valid POST request to, say, /cgi-bin/foo.exe,I > foo.exe (which must be located in a directory marked +ExecCGI) can readt' > it in and return whatever it desires.i >y > What's the problem?d   OK, details:  : The client has to send a request that looks something like   POST /cgi-bin/foo.exe HTTP/1.0 Host: www.the.target.hostt& Content-Type: application/octet-stream Content-Length: nnn  (blank line) (body)  F where nnn is the message (body)'s length in bytes (or octets, like the RFC calls them).  $ foo.exe must answer the request like   Status: 200 OK& Content-Type: application/octet-stream Content-Length: nnne (blank line) (body)  B The status line can be 'HTTP/1.0 200 OK' if foo.exe is marked as aB Non-Parsed-Header (NPH) script (i.e. the web server just sends outD everything foo.exe prints). It can also contain some 4xx or 5xx codeD in case there's an error on the client or server side, respectively.  G The header lines (i.e. everything up to and including the (blank line))c must be terminated by CR LF.  I I suggest you read (at least parts of) RFC 2616 which deals with HTTP/1.1v: Please find it e.g. at http://www.ietf.org/rfc/rfc2616.txt   Hope it helps,   Martin --  F                           | Martin Vorlaender  |  VMS & WNT programmer3  Cetero censeo            | work: mv@pdv-systeme.dehF  Redmondem delendam esse. |   http://www.pdv-systeme.de/users/martinv/:                           | home: martin@radiogaga.harz.de   ------------------------------  % Date: Wed, 26 Sep 2001 14:22:43 -1300s* From: Shri Kanekal <kanekal@surya.umd.edu>" Subject: Apcahe on VAX OpenVMS 6.1- Message-ID: <3BB1E4B3.32B17B8E@surya.umd.edu>h   Folks,C         This may be old hat but please help. Is APACHE running on  e@  VAX OpenVMS ? If so where could download binaries/source from ?*  I am on a VAX 4000/90 running OpenVMS 6.1    thanks,  Shrip  G  ======================================================================iH  Shri Kanekal                                      phone: (301)405-6237 G  LASP, Univ of Colorado, Boulder                   FAX  : (301)314-9547c  Current Mailing Address:V  Dept of Physics,S  University of Marylande  College Park, MD 20742  1G  ======================================================================a   ------------------------------  % Date: Wed, 26 Sep 2001 21:38:18 +0200r2 From: martin@radiogaga.harz.de (Martin Vorlaender)& Subject: Re: Apcahe on VAX OpenVMS 6.1; Message-ID: <3bb22eaa.524144494f47414741@radiogaga.harz.de>w  + Shri Kanekal (kanekal@surya.umd.edu) wrote:t$ > Is APACHE running on VAX OpenVMS ? ...F+ > I am on a VAX 4000/90 running OpenVMS 6.1g  M Nope. Apache/CSWS needs Alpha and at least VMS 7.1-2 (7.2 if you want to use   mod_perl). SeeK http://www.openvms.compaq.com/openvms/products/ips/apache/csws_install.htmlh   cu,    Martin --  G So long, and thanks        | Martin Vorlaender  |  VMS & WNT programmeri4 for all the books...       | work: mv@pdv-systeme.deK In Memoriam Douglas Adams  |       http://www.pdv-systeme.de/users/martinv/e;             1952-2001      | home: martin@radiogaga.harz.de    ------------------------------  # Date: Thu, 27 Sep 2001 00:06:34 GMT  From: vze2mj6m@mail.verizon.netd( Subject: Available Qbus SCSI Controllers0 Message-ID: <3BB26D8C.2325539B@mail.verizon.net>  4 We know that there are a bunch of Qbus system owners; that would like a SCSI controller for their systems, so ...s  5 The following Qbus SCSI Controllers are available forr6 sale or trade. Sale prices are at lower end commercial9 market, but if you're "just a hobbyist", the trade optiona6 might be your preferred method. If you wish to propose3 a trade, first ensure you're committed to following 5 through. Second, contact one of the major dealers and-5 find out from them what they'll give you "cash price" 7 for your parts. Third, ask them what their price is forc6 the SCSI controller you're interested in. If the "cash6 price" they offer you for your parts is lower than the8 SCSI controller's price, we won't be able to approve the8 proposal. Also, we can only approve a proposal for parts5 that we don't already have plenty of, and that we can $ already find a buyer for in advance.  0 Pictures of these controllers can be seen at ...  , SCSI Disk and Tape - Aviv_QSA_50-1276_Rev_F1J http://members.bellatlantic.net/~vze2mj6m/dec/Aviv_QSA_50-1276_Rev_F1_.jpg   SCSI Disk - Dilog_SQ706_Rev_ClD http://members.bellatlantic.net/~vze2mj6m/dec/Dilog_SQ706_Rev_C_.jpg   SCSI Disk - Dilog_SQ706A_Rev_DE http://members.bellatlantic.net/~vze2mj6m/dec/Dilog_SQ706A_Rev_D_.jpge  6 SCSI Disk - Emulex_UC07_10401-01_Rev_F - Two AvailableM http://members.bellatlantic.net/~vze2mj6m/dec/Emulex_UC07_10401-01_Rev_F_.jpga   SCSI Tape - MTI_QTS25_Rev_DuB http://members.bellatlantic.net/~vze2mj6m/dec/MTI_QTS25_Rev_D_.jpg  & The list of parts already in stock ...  6 http://members.bellatlantic.net/~vze2mj6m/web/home.htm    )                     dec.parts@verizon.net    ------------------------------  # Date: Tue, 25 Sep 2001 19:13:03 GMTo From: dittman@dittman.nete Subject: Backing up an HSZ PCM> Message-ID: <3H4s7.272$ab2.172729@e420r-atl2.usenetserver.com>  ; I have an HSZ40 PCM that I'd like to back up just in case ID; have a problem later.  Does anyone know how I can backup up1 the PCM? -- . Eric Dittman dittman@dittman.netl< Check out the DEC Ethusiasts Club at http://www.dittman.net/   ------------------------------    Date: 26 Sep 2001 09:56:49 -0700/ From: chris@applied-synergy.com (Chris Scheers) 4 Subject: Re: c$main undefined in hello-world program< Message-ID: <754a27c1.0109260856.929e899@posting.google.com>  j 0vazan@st.fmph.uniba.sk (Bubak) wrote in message news:<87266f0d.0109251047.531b1a57@posting.google.com>...B > Thanks, it worked. Now just one question. How is it that command > cc/ver or cc /verIA > returns syntax error? I have discovered I am using OpenVMS V7.1    Use the commands        $ CC/LIST TESTt      $ TYPE TEST  B The version number can be found in the header of the listing file.   ------------------------------  % Date: Tue, 25 Sep 2001 22:07:03 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)4 Subject: Re: c$main undefined in hello-world program; Message-ID: <3bb0e3e7.524144494f47414741@radiogaga.harz.de>   & Bubak (0vazan@st.fmph.uniba.sk) wrote:B > Thanks, it worked. Now just one question. How is it that command > cc/ver or cc /ver A > returns syntax error? I have discovered I am using OpenVMS V7.1h  = It seems like you're not using DEC C (which has this switch).m   cu,t   Martin --  G                            | Martin Vorlaender  |  VMS & WNT programmery4  UNIX is user friendly.    | work: mv@pdv-systeme.deG  It's just selective about |   http://www.pdv-systeme.de/users/martinv/,;  who its friends are.      | home: martin@radiogaga.harz.de    ------------------------------   Date: 25 Sep 2001 22:56 PDTu+ From: rankin@eql14.caltech.edu (Pat Rankin)c4 Subject: Re: c$main undefined in hello-world program1 Message-ID: <25SEP200122564100@eql14.caltech.edu>T  4 In article <$HsjcCxGzDat@eisner.encompasserve.org>,\2  koehler@encompasserve.org (Bob Koehler) writes...1 >   IIRC there are no C$ prefixes in the VAXCRTL.r  =      Yes there are:  C$MAIN and C$MAIN_ARGS (and also a bunchu? of C$_xxx symbols corresponding to errno values).  Those two doL? the startup code that initializes stdio and getenv and a couple = of other things.  The second one also splits the command linea> string into C's traditional argc,argv[] tokens.  The first was= used in this case because the user's definition of main() hase omitted the optional arguments.   ;      DEC C (aka Compaq C) works differently; it uses jacket ? code to first call the startup code and then to call the user's-@ main() as a subroutine.  But VAX C inserts a call to the library= startup code at the beginning of the user's main routine; theDA program needs to be linked with VAXCRTL in order to resolve that.o  2                 Pat Rankin, rankin@eql.caltech.edu   ------------------------------  % Date: Thu, 27 Sep 2001 08:03:34 +0800t' From: "Kenneth" <chehon@netvigator.com>d% Subject: Cannot detect display devicet0 Message-ID: <9otq4j$qgo1@imsp212.netvigator.com>  L I have downgrade my machine by restore the previous system image. When I tryK to start the DEC Window, it will return cannot detect graphical device. How.  can I config the graphic device?   ------------------------------  % Date: Tue, 25 Sep 2001 23:19:06 -0400h  From: John Santos <JOHN@egh.com># Subject: Re: CETS2001 presentationsn4 Message-ID: <1010925225809.882A-100000@Ives.egh.com>  - On Tue, 25 Sep 2001, David J. Dachtera wrote:a   > David Spencer wrote: > > = > > In article <3BAEAD0A.F52209A2@fsi.net>, David J. Dachterae" > > <djesys.nospam@fsi.net> wrote: > >  > > > Larry Kilgallen wrote: > > > > C > > > > In article <9oj8tt$q6p$1@bob.news.rcn.net>, "warren sander" ( > > > > <sander@ma.ultranet.com> writes:R > > > > > I didn't have time to 'htmlize' the cets presentations on friday. I will > > > > > do? > > > > > that next week so you can view them with any browser.t > > > > " > > > > Thanks.  Much appreciated. > > >.L > > > Second Larry's motion, but still have an issue with the attendees-only- > > > thing. Not nice. Really sucks, in fact.t > > E > > Years ago I was fairly active in DECUS. During that time I made awI > > number of contributions to the library and gave several presentationsuG > > at the conference. The time I spent and the software I produced wasjE > > offered voluntarily. I spent money out of my own pocket to attendiI > > these conferences. It was all well worth it because of the atmosphere 3 > > of common interests and sharing of information.r > > J > > Apparently I've been out of the loop too long. What has happened here?I > > Did I read this thread correctly - people have volunteered their timeoF > > and spent their own money to make presentations at the conference.I > > And unless you've had the time and paid your money to attend - you'reaH > > not allowed to share?? What the hell is going on here? I'm outraged! > > E > > If there are the new terms for DECUS or CETS or whatever the damnt* > > thing is called, you can count me out. > A > Knowing the kind of response such comments draw these days, I'mrJ > expecting some one to pipe up and say that since you presented your wereG > given the symposium gratis, and were, therefore, compensated. (That's 3 > lame, but not at all unexpected - by me, anyway.)   F Dave D: Speakers used to have to pay the regular symposium fees, IIRC.  B Dave S: I also think that the ENCOMPASS policy only applies to theE Encompass web site - you have to have been a registered CETS attendee C to browse their CETS web site, if I understand it correctly.  There G is no policy that prevents a speaker from posting his/her session notesEF on his/her own web site.  Keith Parris and Hoff Hoffman have both done this.s  D Also, nothing I know of prevents any attendee from taking notes at aE session, and posting their summary and commentary right here, or on acD web site. (I recall mention of NDA sessions, but in all the DECUS'esD I've attended over the years, I never encountered one.)  There mightF be copyright problems with transcribing a session verbatim and postingB the text, but absent a contract to the contrary, the 1st amendment# should prevail on reporting events.   A I think it is extremely counterproductive for either Encompass oraA Compaq to do otherwise.  The whole purpose of the symposium is tow' spread information, not to restrain it.s  = Speaking of which, does anyone have any summaries to post, ors> comments to make?  I'm particulary interested in the Q&A stuff> in the Alpha->IPF sessions.  (I've viewed several of the PPT's@ of Hoff's sessions so far.  It was all pretty much as I expectedC with no stunningly good or bad news, but there might have been some > seemingly minor bullets on some of the slides which went right? past me.  Hmm - I didn't mean that last clause to come out likeh that.  Duck and cover!!)   -- O John Santosy Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Wed, 26 Sep 2001 15:17:09 -0400j- From: JF Mezei <jfmezei.spamnot@videotron.ca>5# Subject: Re: CETS2001 presentationsc, Message-ID: <3BB229B1.C426D84B@videotron.ca>   Jeff Killeen wrote:bM > to Encompass members sometime in the future for this year only.  For future L > years there is a good chance they will be posted with timed audio and onlyF > available for a fee or to folks who already attended the conference.  N I am curious. Considering that Compaq will cease to exist in a few months, howA can what-used-to-be-DECUS discuss  the future with such details ?   N Have there been any official statements from Carly about support for what used
 to be DECUS ?   L > I am sure there will be folks who believe everything should be offered forM > free.  There are real costs involved in setting up a web site to distribute J > these.  There are significant costs to doing the timed audio if we do it > next year.  M Under the original DECUS, all materials submitted at DECUS was public domain,  wasn't it ?h  L As far as costs for maintaing the web site, let me give you an example: WithN ALL-IN-1, all you'd have to do is drop the presentation in a folder, give it aL good (long) name and voila, it is indexed and available to the world the theL web interface. Surely, whoevere/whatever runs your web site could organise a? similar setup where materials can easily be added without cost.b   ------------------------------  % Date: Wed, 26 Sep 2001 15:53:18 -0400r- From: JF Mezei <jfmezei.spamnot@videotron.ca> # Subject: Re: CETS2001 presentationsf, Message-ID: <3BB23226.BF689132@videotron.ca>   Gary McCready wrote:G > - If some folks could get all the content on-line, they might not go..  M Organisations such as what-used-to-be-DECUS have to adapt to the (not so new) K reality of the Internet. The purpose of user groups is twofold: disseminateeM information (best done on Internet) and promote peer to peer contacts as well 8 as vendor-customer contacts (best done in a conference).  J With budget cutbacks, conferences are (and have been in many countries forK long time) in jeopardy. Many chapters of what used to be DECUS died/peteredeN out because they insisted on focusing on the conference instead of developpingO internet services to disseminate information and promote peer-to-peer contacts.l  H And what happened on June 25 showed perhaps the shape of things to come:K Compaq vising those customers it cares about, and leaving all others in thesM dark. And the "customers it cares about" are in small enough numbers now that-M Compaq doesn't need a conference to meet them, it just goes to their premisesh to visit them.  J I'd say that now is not the time for what-used-to-be-DECUS to piss off theH community by restricting information flow. Those customers not worthy ofI getting a visit from Compaq on June 25 are starved for information and iflI what-used-to-be-DECUS doesn't provide information, it becomes irrelevant.n  L Of course, it may be part of Compaq's plan to piss off the smaller customersM whom it judges are not profitable, in which case making what-used-to-be-DECUS % irrelevant would be part of the plan.g   ------------------------------  % Date: Tue, 25 Sep 2001 13:56:31 -0400g- From: "Peter Weaver" <peter.weaver@stelco.ca>r# Subject: Re: CETS2001 presentations 2 Message-ID: <xz3s7.31784$Z2.453046@nnrp1.uunet.ca>  D Thank you Warren and Hoff and everyone else involved in getting this information out.  I Is there any chance that someone can find the *%#$&*^ reponsible for CETShH and hit them with something large and heavy until they release the other" presentations for general viewing?   ------------------------------  % Date: Tue, 25 Sep 2001 10:17:05 -0700.3 From: David Spencer <spencer@spaamfree.recneps.com>b# Subject: Re: CETS2001 presentations > Message-ID: <250920011017051599%spencer@spaamfree.recneps.com>  9 In article <3BAEAD0A.F52209A2@fsi.net>, David J. Dachtera- <djesys.nospam@fsi.net> wrote:   > Larry Kilgallen wrote: > > ? > > In article <9oj8tt$q6p$1@bob.news.rcn.net>, "warren sander"d$ > > <sander@ma.ultranet.com> writes:N > > > I didn't have time to 'htmlize' the cets presentations on friday. I will > > > do; > > > that next week so you can view them with any browser.e > >  > > Thanks.  Much appreciated. > H > Second Larry's motion, but still have an issue with the attendees-only) > thing. Not nice. Really sucks, in fact.   A Years ago I was fairly active in DECUS. During that time I made a.E number of contributions to the library and gave several presentationsnC at the conference. The time I spent and the software I produced wasiA offered voluntarily. I spent money out of my own pocket to attendeE these conferences. It was all well worth it because of the atmospheret/ of common interests and sharing of information.   F Apparently I've been out of the loop too long. What has happened here?E Did I read this thread correctly - people have volunteered their timeiB and spent their own money to make presentations at the conference.E And unless you've had the time and paid your money to attend - you'reeD not allowed to share?? What the hell is going on here? I'm outraged!  A If there are the new terms for DECUS or CETS or whatever the damnR& thing is called, you can count me out.   -- Dave Spencerh   ------------------------------  # Date: Wed, 26 Sep 2001 20:54:46 GMT & From: "Jeff Killeen" <Jeff@IDM-IO.com># Subject: Re: CETS2001 presentationse6 Message-ID: <qgrs7.163$bY4.23842@typhoon2.gnilink.net>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3BB229B1.C426D84B@videotron.ca... > Jeff Killeen wrote:-H > > to Encompass members sometime in the future for this year only.  For futureI > > years there is a good chance they will be posted with timed audio and- onlyH > > available for a fee or to folks who already attended the conference. >lL > I am curious. Considering that Compaq will cease to exist in a few months, howiC > can what-used-to-be-DECUS discuss  the future with such details ?k  / Encompass is legally independent from Compaq...=   ------------------------------  % Date: Wed, 26 Sep 2001 17:23:55 -0400 ' From: Jim Becker <jbecker@ui.urban.org>U# Subject: Re: CETS2001 presentations , Message-ID: <3BB2476B.CA9462F3@ui.urban.org>   JF Mezei wrote:P [snip]N > As far as costs for maintaing the web site, let me give you an example: WithP > ALL-IN-1, all you'd have to do is drop the presentation in a folder, give it aN > good (long) name and voila, it is indexed and available to the world the theN > web interface. Surely, whoevere/whatever runs your web site could organise aA > similar setup where materials can easily be added without cost..  ? "Without cost" is a myth. If you consider any active, thriving,mF Internet-connected server, it has costs associated with it. Who coversF those costs will vary from site to site, but chances are that somebody? somewhere is bearing the cost of having the material available.h  C Even in your example, someone would have paid for the software, theeF hardware, the storage space, the Internet connection, and the facilityA where the server resides. Someone may well be paying for on-goingtA software and hardware support, or there's the "cost" of running aaF system that's not under any support agreements. There's the staff costC of on-going system management, on-going application management, andbA on-going data management. Not all costs are monetary. Even if theoE server is run entirely by volunteers, there is a "cost" in recruitinguC and supporting the volunteers, who are covering the cost of running E the system by donating their personal time. Someone is always bearinge	 the cost.n  F The idea of making CETS materials available to people who weren't ableA to attend has some definite attractions, but keep these points inr mind:r7 1) There is a non-zero cost associated with doing this.tD 2) While there's a value proposition to those who want access to theC materials, any proposal should also spell out the value proposition # for those who would bear the costs.rE 3) There's also the issue of "appropriate use" -- using the materialsh@ only in ways consistent with agreements and commitments (between2 speaker & CETS, between Compaq & Encompass, etc.).   --
 Jim Becker+ The Urban Institute (http://www.urban.org/)l' Encompass (http://www.encompassus.org/)t. ESILUG (http://encompasserve.org/lugs/esilug/)   ------------------------------  % Date: Tue, 25 Sep 2001 21:50:45 -0500e1 From: "David J. Dachtera" <djesys.nospam@fsi.net>I# Subject: Re: CETS2001 presentationsm' Message-ID: <3BB14285.3F2D3352@fsi.net>_   David Spencer wrote: > ; > In article <3BAEAD0A.F52209A2@fsi.net>, David J. Dachteram  > <djesys.nospam@fsi.net> wrote: >  > > Larry Kilgallen wrote: > > >FA > > > In article <9oj8tt$q6p$1@bob.news.rcn.net>, "warren sander"l& > > > <sander@ma.ultranet.com> writes:P > > > > I didn't have time to 'htmlize' the cets presentations on friday. I will
 > > > > do= > > > > that next week so you can view them with any browser.n > > >/  > > > Thanks.  Much appreciated. > > J > > Second Larry's motion, but still have an issue with the attendees-only+ > > thing. Not nice. Really sucks, in fact.i > C > Years ago I was fairly active in DECUS. During that time I made a2G > number of contributions to the library and gave several presentationsbE > at the conference. The time I spent and the software I produced wasaC > offered voluntarily. I spent money out of my own pocket to attendcG > these conferences. It was all well worth it because of the atmosphere-1 > of common interests and sharing of information.0 > H > Apparently I've been out of the loop too long. What has happened here?G > Did I read this thread correctly - people have volunteered their timeiD > and spent their own money to make presentations at the conference.G > And unless you've had the time and paid your money to attend - you'reeF > not allowed to share?? What the hell is going on here? I'm outraged! > C > If there are the new terms for DECUS or CETS or whatever the damn ( > thing is called, you can count me out.  ? Knowing the kind of response such comments draw these days, I'meH expecting some one to pipe up and say that since you presented your wereE given the symposium gratis, and were, therefore, compensated. (That'su1 lame, but not at all unexpected - by me, anyway.)    -- - David J. Dachtera4 dba DJE Systemso http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Wed, 26 Sep 2001 12:13:47 GMTa& From: "Jeff Killeen" <Jeff@IDM-IO.com># Subject: Re: CETS2001 presentationsC7 Message-ID: <%Djs7.2116$M7.131593@typhoon2.gnilink.net>   < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3BAEAD0A.F52209A2@fsi.net...rH > Second Larry's motion, but still have an issue with the attendees-only) > thing. Not nice. Really sucks, in fact.   G FYI - The sessions have never been available on a web site before - Any K suggestions that they have been offered in the past and are now not offereduL is a false one.  This is a new service and there are costs involved with it.  L There is a fairly good chance the CETS-2001 will become generally accessibleK to Encompass members sometime in the future for this year only.  For future$J years there is a good chance they will be posted with timed audio and onlyD available for a fee or to folks who already attended the conference.  J I am sure there will be folks who believe everything should be offered forK free.  There are real costs involved in setting up a web site to distribute H these.  There are significant costs to doing the timed audio if we do it
 next year.   ------------------------------  # Date: Wed, 26 Sep 2001 14:52:47 GMTC  From: jlsue <jlsuexxxz@home.com># Subject: Re: CETS2001 presentationsl8 Message-ID: <o6q3rtg6rie4q741rjdfu66ejpl5051vsp@4ax.com>  @ Now Jeff, I see no reason that we can't distribute the symposiumF presentations just like we do for software.  Just create a CD with theD presentations.  Possibly, even put it online on Eisner, but at leastD allow us to purchase it for minor charge (just like what we have for the software distribution).     B On Wed, 26 Sep 2001 12:13:47 GMT, "Jeff Killeen" <Jeff@IDM-IO.com> wrote:  = >"David J. Dachtera" <djesys.nospam@fsi.net> wrote in messagea" >news:3BAEAD0A.F52209A2@fsi.net...I >> Second Larry's motion, but still have an issue with the attendees-onlyr* >> thing. Not nice. Really sucks, in fact. >oH >FYI - The sessions have never been available on a web site before - AnyL >suggestions that they have been offered in the past and are now not offeredM >is a false one.  This is a new service and there are costs involved with it.n >eM >There is a fairly good chance the CETS-2001 will become generally accessiblenL >to Encompass members sometime in the future for this year only.  For futureK >years there is a good chance they will be posted with timed audio and onlycE >available for a fee or to folks who already attended the conference.a >tK >I am sure there will be folks who believe everything should be offered fortL >free.  There are real costs involved in setting up a web site to distributeI >these.  There are significant costs to doing the timed audio if we do it  >next year.e >d >r >s   ------------------------------    Date: 26 Sep 2001 08:00:24 -0700) From: google@mccready.com (Gary McCready) # Subject: Re: CETS2001 presentationsC= Message-ID: <6e64ea70.0109260700.1d0726bf@posting.google.com>t  g "Peter Weaver" <peter.weaver@stelco.ca> wrote in message news:<xz3s7.31784$Z2.453046@nnrp1.uunet.ca>...uF > Thank you Warren and Hoff and everyone else involved in getting this > information out. > K > Is there any chance that someone can find the *%#$&*^ reponsible for CETSnJ > and hit them with something large and heavy until they release the other$ > presentations for general viewing?  & It wasn't me...but having said that...   (Sigh) Consider the following:A - Of course, Compaq wants to attract as many people to attend then conference as possibleE - If some folks could get all the content on-line, they might not go.lD - If not enough folks go, Compaq(HP, etc) might decide not to have a, confernce with a Compaq (Digital, VMS) focus- - Thus, no central place for content. Period.y  ? Now, what would be a good solution? Only having a free, on-linen@ conference? Allowing folks to buy all or part on a CD or via webE download? Are there any good examples of traditional conferences thate? have survived for several years with full release of all of thepC technical content on-line? Perhaps that could be used as example tor! CETS as a good model to follow...e   --Gary McCreadya1 -My opinions have nothing to do with my employer.s   ------------------------------  % Date: Wed, 26 Sep 2001 12:44:56 -0400h- From: "Richard D. Piccard" <piccard@ohio.edu>.# Subject: Re: CETS2001 presentationsa( Message-ID: <3BB20607.F8FF3A90@ohio.edu>   Jeff,s  N     Please consider a business model that includes non-attendees being able toN access the information.  Even if that means we have to pay for it, or can onlyM get to after a delay, it would create a much more positive response to CETS-nt+ and Encompass than the current exclusivity.   #                                 RDPn     Jeff Killeen wrote:g  > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message# > news:3BAEAD0A.F52209A2@fsi.net...rJ > > Second Larry's motion, but still have an issue with the attendees-only+ > > thing. Not nice. Really sucks, in fact.  > I > FYI - The sessions have never been available on a web site before - AnypM > suggestions that they have been offered in the past and are now not offeredeN > is a false one.  This is a new service and there are costs involved with it. >sN > There is a fairly good chance the CETS-2001 will become generally accessibleM > to Encompass members sometime in the future for this year only.  For future L > years there is a good chance they will be posted with timed audio and onlyF > available for a fee or to folks who already attended the conference. >aL > I am sure there will be folks who believe everything should be offered forM > free.  There are real costs involved in setting up a web site to distribute-J > these.  There are significant costs to doing the timed audio if we do it > next year.   --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------    Date: 27 Sep 2001 00:50:47 +0800, From: Paul Repacholi <prep@prep.synonet.com># Subject: Re: CETS2001 presentationsw- Message-ID: <87lmj128wo.fsf@prep.synonet.com>   + google@mccready.com (Gary McCready) writes:i  ( > It wasn't me...but having said that... >   > (Sigh) Consider the following:C > - Of course, Compaq wants to attract as many people to attend the  > conference as possibleG > - If some folks could get all the content on-line, they might not go. F > - If not enough folks go, Compaq(HP, etc) might decide not to have a. > confernce with a Compaq (Digital, VMS) focus/ > - Thus, no central place for content. Period.s  E Sigh, you obviously have totally missed the reason for DECUS/CETS/...h. It has nothing to do with what you are saying.  @ BTW, people  who have presented sessions have been putting theirB matterial on the SIGtapes etc forever. Be nice if it was supported
 and extended.r   -- o< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda.-@                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  # Date: Wed, 26 Sep 2001 18:00:30 GMT,& From: "Jeff Killeen" <Jeff@IDM-IO.com># Subject: Re: CETS2001 presentationsd5 Message-ID: <2Jos7.80$bY4.11696@typhoon2.gnilink.net>p  8 "Richard D. Piccard" <piccard@ohio.edu> wrote in message" news:3BB20607.F8FF3A90@ohio.edu... >r > Jeff,c >aH >     Please consider a business model that includes non-attendees being able toeK > access the information.  Even if that means we have to pay for it, or cani onlyH > get to after a delay, it would create a much more positive response to CETS-n- > and Encompass than the current exclusivity.v >p% >                                 RDPi  2 That model is under consideration for next year...   ------------------------------  % Date: Wed, 26 Sep 2001 20:40:41 -0500q1 From: "David J. Dachtera" <djesys.nospam@fsi.net> # Subject: Re: CETS2001 presentationss' Message-ID: <3BB28399.6421C894@fsi.net>h   Jeff Killeen wrote:p > > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message# > news:3BAEAD0A.F52209A2@fsi.net...tJ > > Second Larry's motion, but still have an issue with the attendees-only+ > > thing. Not nice. Really sucks, in fact.s > I > FYI - The sessions have never been available on a web site before - AnyiM > suggestions that they have been offered in the past and are now not offeredeN > is a false one.  This is a new service and there are costs involved with it.  = Well, that's fine - everything costs, and we understand this.h  E The point we're trying (desperately!) to drive home is the "attendees-D only" thing. For pay is one thing - restricted audience is something else again."  E I also do not see why it should be limited to Encompass members only,-H unless Encompass is prepared to subsume HP people as well as Tandemites, DECies and Compaqers...i   -- n David J. Dachtera  dba DJE Systemsa http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/n   ------------------------------  % Date: Wed, 26 Sep 2001 21:52:35 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> # Subject: Re: CETS2001 presentations , Message-ID: <3BB28658.4D418ECA@videotron.ca>  0 re: the costs of making the materials available.  K Humm, me thinks that it would cost far more to restrict access only to CETSsT attendees for certain documents/pages than to just make them available to the world.   ------------------------------  % Date: Wed, 26 Sep 2001 20:52:40 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>h# Subject: Re: CETS2001 presentationst' Message-ID: <3BB28668.85622F8A@fsi.net>    Jeff Killeen wrote:a > : > "Richard D. Piccard" <piccard@ohio.edu> wrote in message$ > news:3BB20607.F8FF3A90@ohio.edu... > >l	 > > Jeff,s > >tJ > >     Please consider a business model that includes non-attendees being	 > able torM > > access the information.  Even if that means we have to pay for it, or canl > onlyJ > > get to after a delay, it would create a much more positive response to > CETS-n/ > > and Encompass than the current exclusivity.y > >o' > >                                 RDPn > 4 > That model is under consideration for next year...  H I would *STRONGLY* recommend implementing it ex post facto for this yearF - don't be afraid to admit a mistake. That's what happens what we rushF around like chickens with our heads cut off trying to do everything atA the 11th hour that should have been done months before, trying tohD remember all the "oops!" items from -2000, then get a tripple-whammyH with a terrorist sneak-attack on top of the existing chaos... (D'ya ever# hear the expression: "plan ahead"?)d  @ That's a hint by the way... My recommendation: Let's get busy onF HPETS-2002 *BEFORE* the end of Oct-2001 (I originally wrote "Sep-2001"F and thought better of it - lotta folks (including me) still in shock),H and see if we can ... well, I guess I better not go there. I'd likely be# branded an ingrate again this year.    -- a David J. Dachtera4 dba DJE Systemsc http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/r   ------------------------------  % Date: Wed, 26 Sep 2001 21:06:00 -0500u1 From: "David J. Dachtera" <djesys.nospam@fsi.net> # Subject: Re: CETS2001 presentations ' Message-ID: <3BB28988.3F0C9068@fsi.net>a   Gary McCready wrote: > [snip]G > - If some folks could get all the content on-line, they might not go.a   Speaking from experience:y  D There tends to be a *LARGE* difference between the presentation (theH actual session) and the "presentation" (.PPT, etc. file). In general theG material as presented tends to be considerably expanded from the actual. slides.>  H Many presenters don't prepare much more than just "slides". Check out myG Fall '99 presentations at http://www.djesys.com/vms/support/ and you'llpB notice that I made use of the "(speaker's) notes" view (PowerpointC terminology - not sure what its called in Koffice or StarOffice) to.2 render more of a "textbook" than a "presentation".  H So, just having the "presentation" tends to be of somewhat limited value, without actual (hand-written) session notes.   -- c David J. Dachteran dba DJE Systemss http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/c   ------------------------------  # Date: Thu, 27 Sep 2001 02:19:52 GMTl& From: "Jeff Killeen" <Jeff@IDM-IO.com># Subject: Re: CETS2001 presentationse6 Message-ID: <c1ws7.484$jW2.33679@typhoon1.gnilink.net>  < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3BB28668.85622F8A@fsi.net...eJ > I would *STRONGLY* recommend implementing it ex post facto for this year  @ If we do something for this year it will likely be at no cost...   ------------------------------  # Date: Thu, 27 Sep 2001 02:19:33 GMTa& From: "Jeff Killeen" <Jeff@IDM-IO.com># Subject: Re: CETS2001 presentations 6 Message-ID: <V0ws7.483$jW2.33679@typhoon1.gnilink.net>  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3BB28658.4D418ECA@videotron.ca...2 > re: the costs of making the materials available. > H > Humm, me thinks that it would cost far more to restrict access only to CETSK > attendees for certain documents/pages than to just make them available too
 the world.  8 CETS web site is developed to provide this functionality   ------------------------------   Date: 27 Sep 2001 05:48:17 GMT3 From: vance@alumni.caltech.edu (Vance R. Haemmerle)l# Subject: Re: CETS2001 presentationsh, Message-ID: <9ouej1$3cn@gap.cco.caltech.edu>  , In article <3BB2476B.CA9462F3@ui.urban.org>,) Jim Becker  <jbecker@ui.urban.org> wrote:e >eG >The idea of making CETS materials available to people who weren't able:B >to attend has some definite attractions, but keep these points in >mind:8 >1) There is a non-zero cost associated with doing this.E >2) While there's a value proposition to those who want access to thehD >materials, any proposal should also spell out the value proposition$ >for those who would bear the costs.F >3) There's also the issue of "appropriate use" -- using the materialsA >only in ways consistent with agreements and commitments (betweenr3 >speaker & CETS, between Compaq & Encompass, etc.).q  K   At the DECUS convention in San Diego in '99, I only had a free Trade show:L pass but didn't register or attend the sessions and I was able to order the E Session Notes CDrom for $85 at the booth.  All I needed was my Decus bH membership number.  What's different about CETS? Can't anyone order the  Session Notes CDrom?     -- Vance Haemmerlee vance@alumni.caltech.edu   ------------------------------   Date: 26 Sep 2001 20:24:43 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)o Subject: DECTalk help needed+ Message-ID: <9otdib$3oc$1@info.cs.uofs.edu>e  D While tis is not directly VMS related, it is closer than many of our topics, so here goes.t  ; Is there anywhere I can find ROM images for the EPROMs in ae DECTalk DTC01??s  C The one I have and was hoping to use for a demonstration project is0D not working and I have had at least onperson who used to work with aD large number of them tell me that he saw a number of cases where theD EPROMs became corupt and had to be erased and re-programmed.  I haveD not been able to determine any specific reason for mine not to work,, so I would like to give this solution a try.  @ I guess this is addressed particularly to our COMPAQ audience asE I can't imagine anyone else who is likely to have access to somethingu like EPROM images.   Any help greatly appreciated.c   bill   --  J Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   4   ------------------------------  # Date: Thu, 27 Sep 2001 01:14:15 GMTM= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)-  Subject: Re: DECTalk help needed0 Message-ID: <00A02A58.A8D0DF32@SendSpamHere.ORG>  _ In article <9otdib$3oc$1@info.cs.uofs.edu>, bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:iE >While tis is not directly VMS related, it is closer than many of oure >topics, so here goes. >i< >Is there anywhere I can find ROM images for the EPROMs in a >DECTalk DTC01?? > D >The one I have and was hoping to use for a demonstration project isE >not working and I have had at least onperson who used to work with a E >large number of them tell me that he saw a number of cases where the E >EPROMs became corupt and had to be erased and re-programmed.  I havesE >not been able to determine any specific reason for mine not to work,o- >so I would like to give this solution a try.h >yA >I guess this is addressed particularly to our COMPAQ audience asoF >I can't imagine anyone else who is likely to have access to something >like EPROM images.e >c >Any help greatly appreciated. >  >billi >  >-- K >Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvespE >bill@cs.scranton.edu     |  and a sheep voting on what's for dinner.g >University of Scranton   | B >Scranton, Pennsylvania   |         #include <std.disclaimer.h>       G If you've got a way to read the EPROMs (non-destructively) to reprogramsH yours, I might be convinced to let you have a go with my DTC01's EPROMs. --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMk            oJ   "And of course, I'm a genius, so people are naturally drawn to my fiery I   intellect.  Their admiration overwhelms their envy!" -- Calvin & Hobbest   ------------------------------  % Date: Wed, 26 Sep 2001 21:45:44 -04000* From: "Stanley F. Quayle" <stan@stanq.com>  Subject: Re: DECTalk help needed. Message-ID: <3BB24C88.22714.F7B3391@localhost>  = On 27 Sep 2001, at 1:14, Brian Schenkenberger, VAXman- wrote:t > a > In article <9otdib$3oc$1@info.cs.uofs.edu>, bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:s> > >Is there anywhere I can find ROM images for the EPROMs in a > >DECTalk DTC01?? >vI > If you've got a way to read the EPROMs (non-destructively) to reprogramiJ > yours, I might be convinced to let you have a go with my DTC01's EPROMs.  E I have a programmer that can read EPROM's, if they are in DIP format pF (pins, 0.1 inch on center, as opposed to surface-mount or edge-pin).   I'd be glad to help.     --Stan  
 ----------G Stanley F. Quayle, P.E.   N8SQ   +1 614-868-1363   Fax: +1 614 868-1671t1 8572 North Spring Ct. NW, Pickerington, OH  43147e= Preferred address:  stan@stanq.com       http://www.stanq.comh   ------------------------------  % Date: Wed, 26 Sep 2001 09:22:01 -0400 3 From: "Pohl, Kathy" <Kathy.Pohl@itec.mail.suny.edu>  Subject: ECP data collectordI Message-ID: <0B41ACBECF83734D98ABC83DD8E097E15DE739@cipher.itec.suny.edu>7  J This message is in MIME format. Since your mail reader does not understand< this format, some or all of this message may not be legible.  ' ------_=_NextPart_001_01C1468E.39F655B0e Content-Type: text/plain;w 	charset="iso-8859-1"6     	Hi,9 	I am using ECP data collector for Open VMS V 5.4 and then+ documentation I have is somewhat limited.   D 	I was wondering if anyone can tell where I can find the definitions$ for some of the metrics such as Lock 	Enqueues and Dequeues?  y 	l 	Thanks, 	Kathy    ' ------_=_NextPart_001_01C1468E.39F655B0' Content-Type: text/html; 	charset="iso-8859-1" + Content-Transfer-Encoding: quoted-printablea  1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">o <HTML> <HEAD>9 <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =o charset=3Diso-8859-1">@ <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
 5.5.2653.12">l" <TITLE> ECP data collector</TITLE> </HEAD>n <BODY> <BR>  > <P>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <FONT SIZE=3D2 = FACE=3D"Arial">Hi,</FONT>l? <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <FONT SIZE=3D2 = I FACE=3D"Arial">I am using ECP data collector for Open VMS V 5.4 and the =e7 documentation I have is somewhat limited.&nbsp; </FONT>n? <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <FONT SIZE=3D2 =mH FACE=3D"Arial">I was wondering if anyone can tell where I can find the =7 definitions for some of the metrics such as Lock</FONT>e? <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <FONT SIZE=3D2 =i3 FACE=3D"Arial">Enqueues and Dequeues?&nbsp; </FONT>a1 <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20s? <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <FONT SIZE=3D2 =e FACE=3D"Arial">Thanks,</FONT>f? <BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <FONT SIZE=3D2 =  FACE=3D"Arial">Kathy</FONT>d </P>   </BODY>r </HTML>e) ------_=_NextPart_001_01C1468E.39F655B0--e   ------------------------------  % Date: Wed, 26 Sep 2001 14:30:45 -0400l2 From: Atlant Schmidt <atlantnospam@mindspring.com> Subject: Re: EDT for Linux. Message-ID: <3BB21ED4.637DA8DD@mindspring.com>  I While it's true that most Emacsen claim to have an EDT mode, some of themND also explicitly claim that EDT mode is broken, and I find that to be2 generally true whether they make the claim or not.  = Ultimately, it's a question of "What do you want/need from aneJ EDT replacement?" If all you want is that the keypad work (so that you canM edit with fewer than nineteen simultaneous key-presses), and you can get usedlL to the missing DELW /UNDELW or DEL/UNDEL key on the keypad, then it's prettyM easy to coerce Emacs or XEmacs to do EDT emulation yourself. On some systems,uM it also requires doing somethiing like xmodmap so that the "Num Lock" key caneL act as the "Gold" key, but after that, it's pretty straight-forward.;  a fewK dozen defiitions in your Emacs resource file will do the job. Looking atthedL EDT emulation Lisp module will give you a good idea of how to go about this.I Plus you get to keep all the nice Emacs features (such as Syntax Coloring=4 that almost works), rectangular-region editing, etc.  I However, if you want EDT's extended features (such as command definition, L line mode, etc.), then you're definitely going to be SOL with Emacs or doing a *LOT* of programming.-  : I've adapted very well to being Solaris-based, but EDT and7 $ SEARCH/HIGHLIGHT are the two things I miss about VMS.C   Atlant   ------------------------------  % Date: Tue, 25 Sep 2001 08:34:12 -0700e' From: David Mathog <mathog@caltech.edu>w  Subject: Re: FORTRAN terminal IO+ Message-ID: <3BB0A3F4.B525963F@caltech.edu>f   Michael Austin wrote:a   >tL > > >         OPEN(UNIT=4,NAME='PA0',TYPE='NEW')  !  This should open a port > > > OR a file... Why not?n > >eF > > How does it fail?  Do you get an error message?  Do you get a file > > named PA0.DAT or MG1.DAT?i > >C >a1 > No files generated.  No error message. No Data.f   Try instead:  8          OPEN(UNIT=4,NAME='PA0',TYPE='NEW',IOSTAT=istat))          write(6,*)'open status = ',istatt  B Which should at least tell you what it did.  Comments on the code:  ] 1.  Unit 4 may be special ( I don't recall that it is, and don't have a fortran manual handy,e but 5 and 6 are).lX 2.  Shouldn't that be PA0: (with a colon)?  Otherwise it's probably going to open a file calledZ      "pao.dat" in the current working directory.  Bet there's one there now.  Or rather, a
 bunch of theme\      if you've run it several times as written - it would have created a newer version  each time.-8 3.  TYPE='NEW' seems inappropriate for an existing port.   Regardsr   David Mathog mathog@caltech.edu   ------------------------------    Date: 25 Sep 2001 11:44:11 -0500 From: briggs@encompasserve.org  Subject: Re: FORTRAN terminal IO3 Message-ID: <Ch5o1G4O3F4H@eisner.encompasserve.org>.  W In article <3BB08993.9AA6EBAE@att.net>, Michael Austin <firstdbasource@att.net> writes:g] > I have no idea why they did anything the way they did it. I believe at one time, the serial R > port was connected directly to a serial port on a plasma metal cutter. (OriginalY > PDP11->>VAX/VMS5.1->> OpenVMS7.3(Alpha)  I am just trying to make it compile and/or run  > VESTED on an Alpha.s >  >>K >> >         OPEN(UNIT=4,NAME='PA0',TYPE='NEW')  !  This should open a port1 >> > OR a file... Why not? >>E >> How does it fail?  Do you get an error message?  Do you get a filen >> named PA0.DAT or MG1.DAT? >> > 1 > No files generated.  No error message. No Data.h  ; Your original report indicated that it also failed when the.> MG1 logical name pointed to a file name.  Did that also result! in no files and no error message?   ? Is it possible that you have error messages suppressed somehow?s   >>M >> >         WRITE(4,4500) CLEARI,CLEARO,XON         !  writing XON to a fileb" >> > should also not be a problem. >>5 >> Does this line fail?  Do you get an error message?t > K > Line does not fail when using OS tools.  "Copy login.com lta600: " works.u' > REPLY/TERM=LTA600: "HELLO"  ... worksg  = So... Does that line fail?  Does it produce an error message?r   >> Did you use LATCPC >> to establish a default association between the LTAxxx device andt) >> a particular terminal server and port?y > 7 > The above would not work if this had not been done...   F There are ways and ways.  It could indeed be made to work even without LATCP.  D My best guess at this point is that you have either suppressed error messages with something like:o  > $ set message /nofacility /noseverity /noidentifcation /notext  ' or redirected them with something like:o   $ DEFINE SYS$OUTPUT NL:R $ DEFINE SYS$ERROR NL:    = I would recommend inserting debugging code into your program.f   	OPEN ( unit=4, file=... 	type *, 'the file was opened' 	WRITE ( 4, 4500...o' 	type *, 'a line of output was written'  	...  : If you don't see the debugging output, you'll have a clue.   	John Briggs   ------------------------------  % Date: Tue, 25 Sep 2001 15:27:16 -0400D- From: Michael Austin <miaustin@bellsouth.net>j  Subject: Re: FORTRAN terminal IO- Message-ID: <3BB0DA93.BBCA5DA2@bellsouth.net>f  ` After another compile run of this application, it is amazing to me that this thing ever compiledR correctly and runs properly.  There are missing variable definitions resulting in:  E "Variable <name> is used before its value has been defined" warnings.n  G THIS ADDRESS IS TEMPORARY AND WILL CHANGE within the next day or so....t: see: http://adsl-214-17-83.clt.bellsouth.net/build_new.txtG THIS ADDRESS IS TEMPORARY AND WILL CHANGE within the next day or so....    briggs@encompasserve.org wrote:   Y > In article <3BB08993.9AA6EBAE@att.net>, Michael Austin <firstdbasource@att.net> writes:e_ > > I have no idea why they did anything the way they did it. I believe at one time, the serial T > > port was connected directly to a serial port on a plasma metal cutter. (Original[ > > PDP11->>VAX/VMS5.1->> OpenVMS7.3(Alpha)  I am just trying to make it compile and/or runi > > VESTED on an Alpha.  > >p > >>M > >> >         OPEN(UNIT=4,NAME='PA0',TYPE='NEW')  !  This should open a port  > >> > OR a file... Why not? > >>G > >> How does it fail?  Do you get an error message?  Do you get a fileT > >> named PA0.DAT or MG1.DAT? > >> > >i3 > > No files generated.  No error message. No Data.t >f= > Your original report indicated that it also failed when thel@ > MG1 logical name pointed to a file name.  Did that also result# > in no files and no error message?D       Correct.   >t >cA > Is it possible that you have error messages suppressed somehow?:  # No.  All other errors get reported.-  O > >> >         WRITE(4,4500) CLEARI,CLEARO,XON         !  writing XON to a filed$ > >> > should also not be a problem. > >>7 > >> Does this line fail?  Do you get an error message?F > >8M > > Line does not fail when using OS tools.  "Copy login.com lta600: " works.h  1 > REPLY/TERM=LTA600: "HELLO"  ... works correctlyo3 > REPLY/TERM=MG1 "HELLO"        ... works correctlyn  3 > REPLY/TERM=PA0 "HELLO"        ... works correctlyh  ? > So... Does that line fail?  Does it produce an error message?s >e   No it works as expected.   >e > >> Did you use LATCPE > >> to establish a default association between the LTAxxx device andc+ > >> a particular terminal server and port?w > >y9 > > The above would not work if this had not been done...h >(H > There are ways and ways.  It could indeed be made to work even without > LATCP. >rF > My best guess at this point is that you have either suppressed error > messages with something like:" >n@ > $ set message /nofacility /noseverity /noidentifcation /notext >s) > or redirected them with something like:  >  > $ DEFINE SYS$OUTPUT NL:. > $ DEFINE SYS$ERROR NL: >    Nope.o   >g? > I would recommend inserting debugging code into your program.- >-! >         OPEN ( unit=4, file=... ' >         type *, 'the file was opened'o >         WRITE ( 4, 4500...0 >         type *, 'a line of output was written'
 >         ...> > < > If you don't see the debugging output, you'll have a clue. >s >         John Briggsn   ------------------------------   Date: 26 Sep 2001 06:11:20 GMT- From: djweath@attglobal.net (Dave Weatherall)e  Subject: Re: FORTRAN terminal IO5 Message-ID: <DTiotGxQ0bj6-pn2-HCW9VElXWxqd@localhost>   F On Tue, 25 Sep 2001 13:41:39, Michael Austin <firstdbasource@att.net>  wrote:   >  > > K > Line does not fail when using OS tools.  "Copy login.com lta600: " works.e' > REPLY/TERM=LTA600: "HELLO"  ... workse  @ Seemingly stupid question but does 'Copy login.com PA0:'  work?   . And what are the logical names PA* defined as?  C The times I've come into contact with using OPEN/READ/WRITE to LAT  B devices I've always had problems. Quite often it came down to the D CLOSE not doing the LAT$IO_DISCONNECT, which meant the LAT port was E 'LAT$LINKed' on one node and thus could not be reached from another. mF This meant that subsequent OPENs on another node could fail.. The onlyE way to resolve the situation was to delete and recreate the LAT port cE on the node that was 'linked'. I mention this because you're testing > using VEST and e> native, which implies different nodes talking to one LAT port.  : I now do all my I/O to LAT ports using QIO and do the LAT F Connect/Disconnect explicitly. However, if someone CTRL-y's out before@ the Disconnect routine is run you have to delete/recreate again.  C I endorse checking that the characteristics of the server port are oE coirrect (i.e. remote). It took me literally a year to work out that sC was why I could no longer reach LAT port that always been working. eA Someone had changed the physical server to which my LAT port was eE assigned. Also ensure the characteristics are DEFINEd as well as SET o :-)9   -- l Cheers - Dave.   ------------------------------   Date: 26 Sep 2001 13:33:49 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)t Subject: Help with UNZIP.EXE, Message-ID: <9oslft$2pql$1@info.cs.uofs.edu>  K I am having a problem using the UNZIP.EXE utility that I got with Multinet.e; Anytime I try to unzip something with subdirectories I get:-  = [ Cannot create output file [.pine-3-91-2.pico]pine_win.lnk ].# [ %RMS-E-DNF, directory not found ]a& [ %SYSTEM-W-NOSUCHFILE, no such file ]9 [ Cannot create output file [.pine-3-91-2.pico]random.c ]e# [ %RMS-E-DNF, directory not found ]i& [ %SYSTEM-W-NOSUCHFILE, no such file ]9 [ Cannot create output file [.pine-3-91-2.pico]region.c ]s# [ %RMS-E-DNF, directory not found ]u& [ %SYSTEM-W-NOSUCHFILE, no such file ]  @ The new subdirectory was successfully created, but it then seems incapable of writting into it.  ' Any quick suggestions for fixing this??b   bill   -- dJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   g   ------------------------------  % Date: Wed, 26 Sep 2001 13:14:49 -0400e' From: Howard S Shubs <howard@shubs.net>i  Subject: Re: Help with UNZIP.EXE< Message-ID: <howard-3C689C.13144926092001@enews.newsguy.com>  , In article <9oslft$2pql$1@info.cs.uofs.edu>,3  bill@triangle.cs.uofs.edu (Bill Gunshannon) wrote:e  B > The new subdirectory was successfully created, but it then seems  > incapable of writting into it. > ) > Any quick suggestions for fixing this??H  H What are the protections on that directory?  What are your default file  protections set to?. -- F Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------   Date: 26 Sep 2001 17:58:10 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)   Subject: Re: Help with UNZIP.EXE, Message-ID: <9ot4vi$30ha$1@info.cs.uofs.edu>  < In article <howard-3C689C.13144926092001@enews.newsguy.com>,*  Howard S Shubs <howard@shubs.net> writes:/ |> In article <9oslft$2pql$1@info.cs.uofs.edu>, 6 |>  bill@triangle.cs.uofs.edu (Bill Gunshannon) wrote: |> eE |> > The new subdirectory was successfully created, but it then seemsm# |> > incapable of writting into it.m |> >  , |> > Any quick suggestions for fixing this?? |>  / |> What are the protections on that directory? t   $ dir /full *.diri  + Directory SYS$COMMON:[000000.HOBBYIST.PINE]a  > PINE-3-91-2.DIR;1             File ID:  (2437,2,0)            0 Size:            1/3          Owner:    [SYSTEM]" Created:   26-SEP-2001 09:29:15.63& Revised:   26-SEP-2001 09:29:15.63 (0) Expires:   <None specified>l Backup:    <No backup recorded>e Effective: <None specified>r Recording: <None specified>  File organization:  Sequential Shelved state:      Online mD File attributes:    Allocation: 3, Extend: 0, Global buffer count: 0K                     No default version limit, Contiguous, MoveFile disabledC"                     Directory fileI Record format:      Variable length, maximum 512 bytes, longest 512 byteso4 Record attributes:  No carriage control, Non-spanned RMS attributes:     None Journaling enabled: None; File protection:    System:RWE, Owner:RWE, Group:RE, World:- Access Cntrl List:  None   Total of 1 file, 1/3 blocks.   J |>                                             What are your default file  |> protections set to?  H Whatever VMS sets them to when it creates the SYSTEM account on install.G I wouldn't know how to change them and have never had reason to before.iB (Which is probably why I have never had a reason to look up how!!)  @ Interestingly enough, all of the subsequent sub-directories wereC created, but none of the files were written into these directories.a8 The permissions on all the sub-directories are the same.   bill   -- RJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   6   ------------------------------  % Date: Wed, 26 Sep 2001 22:14:48 -0400 ' From: Howard S Shubs <howard@shubs.net>m  Subject: Re: Help with UNZIP.EXE< Message-ID: <howard-4E022B.22144826092001@enews.newsguy.com>  , In article <9ot4vi$30ha$1@info.cs.uofs.edu>,3  bill@triangle.cs.uofs.edu (Bill Gunshannon) wrote:-  = > File protection:    System:RWE, Owner:RWE, Group:RE, World:7 > Access Cntrl List:  None  L > |>                                             What are your default file  > |> protections set to? > J > Whatever VMS sets them to when it creates the SYSTEM account on install.I > I wouldn't know how to change them and have never had reason to before.MD > (Which is probably why I have never had a reason to look up how!!)  = IIRC, try SHOW PROTECTION to get the default protection mask.o   Other than that, I'm puzzled.k -- q Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------  % Date: Wed, 26 Sep 2001 22:33:33 -0500c1 From: "David J. Dachtera" <djesys.nospam@fsi.net>-  Subject: Re: Help with UNZIP.EXE' Message-ID: <3BB29E0D.E366999C@fsi.net>e   Howard S Shubs wrote:o > . > In article <9ot4vi$30ha$1@info.cs.uofs.edu>,5 >  bill@triangle.cs.uofs.edu (Bill Gunshannon) wrote:  > ? > > File protection:    System:RWE, Owner:RWE, Group:RE, World:p > > Access Cntrl List:  None > M > > |>                                             What are your default filen > > |> protections set to? > > L > > Whatever VMS sets them to when it creates the SYSTEM account on install.K > > I wouldn't know how to change them and have never had reason to before.sF > > (Which is probably why I have never had a reason to look up how!!) > ? > IIRC, try SHOW PROTECTION to get the default protection mask.o >  > Other than that, I'm puzzled.B  B Hhmmm... I'd need to compare the directory protections against theF current process UIC - SYSTEM owns the dir.'s, but both Group and WorldA are denied write access. Something tells me it may be a privilege; problem in disguise.   -- E David J. Dachtera/ dba DJE Systems; http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/r   ------------------------------  % Date: Wed, 26 Sep 2001 14:28:08 -0400-- From: JF Mezei <jfmezei.spamnot@videotron.ca> H Subject: Re: HOW DO I change FAB file attributes in C code?, $FAB_STORE?, Message-ID: <3BB21E37.BC31D288@videotron.ca>  N > > format=undefined, cc=none.  I saw a MACRO in the VMS documentation, calledQ > > $FAB_STORE...  and I know you can do it using DCL routines on a command line,e# > > but I need to  do it in code...a  I From what I read in the on-line help, $FA_STORE simply moves values to an R existing FAB structure in memory. In C, you can simply use the = to achieve this..  O eg: myFAB.fab$b_org = FAB$C_SEQ ;  /* to set file organisation as sequential */i  H I have not tried this, but isn't the magic done by $FLUSH ?  Reading theJ on-line documentation, it says that it flushes buffers as well as modified9 file information to disk. If so, then the trick would be:o   SYS$OPEN  to open the file& SYS$DISPLAY  to populate the FAB block change the fab attributes 4 SYS$FLUSH  to write the changed buffers back to disk SYS$CLOSE to close the file.   ------------------------------  + Date: Tue, 25 Sep 2001 17:17:52 +0000 (UTC) % From: "Mark Hei" <markhei@iceweb.net>aH Subject: Re: HOW DO I change FAB file attributes in C code?, $FAB_STORE?H Message-ID: <8e9c72d331db102ec83d17a1c490d713.27904@mygate.mailgate.org>   Adrian,d  H I use this type of code...  what is missing is (and I have no real idea M what to do about it), is, after opening the file...  resetting the FAB to be dJ format=undefined, cc=none.  I saw a MACRO in the VMS documentation, calledM $FAB_STORE...  and I know you can do it using DCL routines on a command line,i but I need to  do it in code...u  1 Does anyone know how to use $FAB_STORE in C-code?.   thanks,l   Mark H.      --  = Posted from me-bangor2c-43.bngrme.adelphia.net [24.50.42.43] =1 via Mailgate.ORG Server - http://www.Mailgate.ORG.   ------------------------------   Date: 26 Sep 2001 06:11:18 GMT- From: djweath@attglobal.net (Dave Weatherall)1H Subject: Re: HOW DO I change FAB file attributes in C code?, $FAB_STORE?5 Message-ID: <DTiotGxQ0bj6-pn2-cGZ7OHPpcoZL@localhost>'  D On Tue, 25 Sep 2001 17:17:52, "Mark Hei" <markhei@iceweb.net> wrote:  	 > Adrian,e > J > I use this type of code...  what is missing is (and I have no real idea O > what to do about it), is, after opening the file...  resetting the FAB to be  L > format=undefined, cc=none.  I saw a MACRO in the VMS documentation, calledO > $FAB_STORE...  and I know you can do it using DCL routines on a command line, ! > but I need to  do it in code...  > 3 > Does anyone know how to use $FAB_STORE in C-code?- > 	 > thanks,c > 	 > Mark H.v >    Hi MarkiD                and there's me thinking 'why isn't there a FAB_STORE > macro for C'. Short answer, 'cos no-ones done it (AFAIK). The F $xxx_STORE macros were the way to setup/modify FAB, RAB and NAM blocks@ in Macro-32 and very easy to use, as long as you remembered the / principle that r0/r1 are scratch registers :-).9  F The first (only!) time I did something using FAB/RAB's in C, I cursed ? the fact that I had to write longish statements to do what the v $xxx_STORE macros could do.   C Did anybody ever roll their own versions of xxx_STORE macsos for C?(   Cheers - Dave.  @ PS can you really do FAB_STOREs in DCL? I must read Hoff's book.   ------------------------------  % Date: Wed, 26 Sep 2001 13:48:23 +0200M  From: Paul Sture <paul@sture.ch>M Subject: Re: In times like these ... OpenVMS now more than ever for security!t+ Message-ID: <VA.0000045e.1510607e@sture.ch>m  = In article <3BAF975B.D7370A8@rdrop.com>, Dean Woodward wrote:e > Paul Sture wrote:o > > B > > I'm still running Windows in parallel, but am now anticipatingF > > being able to can it completely in the very near future. Yessssss! > C > http://www.vmware.com/ . I'm about this -> <- far from convertingmA > one of my machines to Linux & Star Office, and using VMWare for!, > the few things that I can't do from Linux. > C Assuming I don't come across any seriously hiccupping Office files tH (which I could take to work to read), I'm down to some language courses  I have (mainly German).A  F I see that vmware Express at $49 only supports Win9x, jumping to $299 H ($329 with the docs and CD) for the full version with NT support. Nah - C I'll stick to dual booting for my purposes (please don't ask me to  E revisit that pit of hell known as Win98:-) - it's where this journey   started). I have 2 PCs anyway.  H I see that Linux clustering is in with this package - I've just _gotta_  have a look at that :-)t  D > The primary machine that I use, though, will have to stay Windoze,G > as I'm doing gui programming there (To connect to a VMS/RDB backend).  >e- Ayup. No sense introducing possible unknowns.I  E (other replies coming privately, since I don't want to create a long r Linux thread). ___ 
 Paul Sture Switzerland,   ------------------------------  % Date: Sat, 29 Sep 2001 12:11:56 +1200s& From: A Bonaveidogo <Asena@fsc.com.fj> Subject: inserting modulea2 Message-ID: <01C148DF.EFC8B6A0@PATRICK.FSC.COM.FJ>  n I wanted to insert a module into the system default device control library. When I apply the following command  D library/insert/text sys$library:sysdevctl.tlb sys$library:xxxxx.txt   p I got the following error message "  LIBRAR-F-OPENIN, error opening SYS$SYSROOT:[SYSLIB]SYSDEVCTL.TLB;2 as input1 -RMS-E-FLK, file currently locked by another users   what shall i doa   ab   ------------------------------  % Date: Wed, 26 Sep 2001 20:32:16 -0400I- From: JF Mezei <jfmezei.spamnot@videotron.ca>  Subject: Re: inserting module4, Message-ID: <3BB2738C.7F307586@videotron.ca>   A Bonaveidogo wrote:r > I got the following error message "  LIBRAR-F-OPENIN, error opening SYS$SYSROOT:[SYSLIB]SYSDEVCTL.TLB;2 as input3 > -RMS-E-FLK, file currently locked by another users  K You have to stop all print queues that use this file. Once stopped, you cano+ update the file, and then restart the file.d  ( Another option is to copy the .TLB file:  M copy SYSDEVCTL.TLB;2 ;3  to create a new version, update the new version, andoD then stop queues and then restart them so the new version gets used.   ------------------------------  % Date: Wed, 26 Sep 2001 18:22:20 +0200 9 From: "Groen, N. (sl)" <Norbert.Groen@soz.pinkroccade.nl>?1 Subject: Is it possible to set system-wide flags? @ Message-ID: <7DE31FAF0D4FD211A4460000F87A853B026AEE9A@ASZMSG001>   Hello,  @ I am trying hard to give processes acces to a global section.=20* These processes are not in the same group.G Now I have to setup flags (or semaphores) and events to lock acces to =s theD section.  E I tried everything but I couldn't get it to work system-wide with thed system-function as there are:r sys$clref(), sys$waitfr() ....  " These events occur only groupwise.  " Is there a change to get it right?  C (I also wrote a version with pp$, but this doesn't seem to work OK)o  * Oh yes, it has got to work on VAX and AXP.     Met vriendelijke groet,    Norbert Groen.   --=20w PinkRoccade Sociale Zekerheid"# Framework Development & Support/=20/ Sofware Levering" Postbus 57005, 1040 CG   Amsterdam% tel: 020-8514154    fax : 020-8514195r  Norbert.Groen@soz.pinkroccade.nl priv=E9 mobiel: 06-41220650r nok@nok.xs4all.nl=20   ------------------------------  % Date: Wed, 26 Sep 2001 19:51:37 -0400a* From: "Stanley F. Quayle" <stan@stanq.com>5 Subject: Re: Is it possible to set system-wide flags? - Message-ID: <3BB231C9.2282.F1301CA@localhost>e  K > Now I have to setup flags (or semaphores) and events to lock acces to then
 > section.  E Use the SYS$ENQ system service.  Before writing, convert the lock to s4 exclusive; when done, convert to null.  Works great.  E You might want to include the node name in the name of the resource, rC just in case you run the same programs on two nodes -- their locks  @ would collide (lock on node A would lock the section on node B).     --Stan  
 ----------G Stanley F. Quayle, P.E.   N8SQ   +1 614-868-1363   Fax: +1 614 868-1671 1 8572 North Spring Ct. NW, Pickerington, OH  43147.= Preferred address:  stan@stanq.com       http://www.stanq.com.   ------------------------------  % Date: Wed, 26 Sep 2001 15:07:31 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>-/ Subject: Re: Major VMS reseller says VMS legacyj, Message-ID: <3BB2276F.7E7E2E6E@videotron.ca>   Alan Greig wrote:aF > There's more like this. So a major reseller has decided to scare VMS2 > customers further for their own commercial gain.  M The remaining VMS customers that matter to Compaq-HP (those worthy of gettinglH a visit from Compaq) are probably perfectly aware that VMS has a limited5 lifetime but are happy and comfortable with its fate.   J If you're confident that VMS, while remaining obscure and unmarketed, willM remain for another 10 years, even if only on the dead-end Alphas, and you arebM confident that the remaining Alpha performance improvements will allow you tow: last 6-7 years, then there is no big hurry to get off VMS.  K But it does mean that new applications are going to different platforms andnK you know that you will eventually migrate from VMS to another platform. AndaI the longer you wait, the less you will lose when to migrate to Unix or NTe4 because the later will have closed the gap with VMS.  K What is left of the VMS customer base has survived over 10 years of "VMS isTN dead", including a certain number of years of Digital telling customers to get3 off VMS. What is left is pretty robust I would say.u   ------------------------------  % Date: Wed, 26 Sep 2001 15:22:28 -0400e- From: JF Mezei <jfmezei.spamnot@videotron.ca>./ Subject: Re: Major VMS reseller says VMS legacy , Message-ID: <3BB22AEF.1F44E5CF@videotron.ca>   Syltrem wrote: > K > Why don't you post the email address of this guy so that we SPAM him with  > our complaints?r    M Why ? The guy is right.  Those still on VMS accept that it is a legacy system L with a not-so-bright future, but figured that it was still better to stay on? VMS and wait for an acceptable replacement to become available.E   ------------------------------    Date: 26 Sep 2001 14:47:40 -0500- From: koehler@encompasserve.org (Bob Koehler) / Subject: Re: Major VMS reseller says VMS legacy 3 Message-ID: <9fzawouRXuDk@eisner.encompasserve.org>   \ In article <3BB2276F.7E7E2E6E@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > Alan Greig wrote:aG >> There's more like this. So a major reseller has decided to scare VMSe3 >> customers further for their own commercial gain.i [...]  > L > If you're confident that VMS, while remaining obscure and unmarketed, willO > remain for another 10 years, even if only on the dead-end Alphas, and you aresO > confident that the remaining Alpha performance improvements will allow you toc< > last 6-7 years, then there is no big hurry to get off VMS.  A    So you don't believe in iVMS on $8K (maybe less) workstations?c   ------------------------------  % Date: Wed, 26 Sep 2001 12:26:48 +0100o% From: Alan Greig <a.greig@virgin.net> + Subject: Major VMS reseller says VMS legacy 8 Message-ID: <jqd3rtseiknbhdgs9dt6fafjuajf8rh7av@4ax.com>  ; A major UK Compaq ISV and VMS specialist has just set up ancD outsourcing subsidiary XAMAN who are continually on the phone to me.C Constant phone calls prompted me to read the brochure and found theb
 following:  E "Such legacy systems comprise, in the main, a large number of DigitaleF VMS environments. In its time, VMS dominated manufacturing, and in theE right hands is still capable of running large,, modern workloads. ButsF it is now functionally stabilized and the tide is running out for much5 of the UK's installed base of older Digital systems."!  D "Our skills with VMS run deep. To this day we continue to wring more: performance efficiencies from the final generation of this@ technology,... And we yield to nobody in our ability to plan andE implement effective transitional and migration strategies to UNIX and9 Microsoft environments."  D There's more like this. So a major reseller has decided to scare VMS0 customers further for their own commercial gain.  : Richard George - Compaq UK (to whom this is also emailed),  E You will recall I spoke to you at the CSF (the company involved here)g> Alpha=:>Itanium issues day. How do you feel about them sendingA customers this sort of garbage? "VMS legacy", "final generation",f "migrate to UNIX and WIndows"M     -- Alan   ------------------------------  % Date: Wed, 26 Sep 2001 09:02:09 -0400 0 From: "Syltrem" <syltrem@videotron.spammenot.ca>/ Subject: Re: Major VMS reseller says VMS legacy 5 Message-ID: <Qhks7.62655$TW.340674@tor-nn1.netcom.ca>d  I Why don't you post the email address of this guy so that we SPAM him with: our complaints?   H A thousand emails from the subscribers of cov may get him to understand?   --   SyltremnI http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)i> To reply to myself directly, remove .spammenot from my address  ? "Alan Greig" <a.greig@virgin.net> a crit dans le message news:s- jqd3rtseiknbhdgs9dt6fafjuajf8rh7av@4ax.com...r >g= > A major UK Compaq ISV and VMS specialist has just set up anaF > outsourcing subsidiary XAMAN who are continually on the phone to me.E > Constant phone calls prompted me to read the brochure and found the  > following: > G > "Such legacy systems comprise, in the main, a large number of Digital H > VMS environments. In its time, VMS dominated manufacturing, and in theG > right hands is still capable of running large,, modern workloads. ButiH > it is now functionally stabilized and the tide is running out for much7 > of the UK's installed base of older Digital systems."  >eF > "Our skills with VMS run deep. To this day we continue to wring more< > performance efficiencies from the final generation of thisB > technology,... And we yield to nobody in our ability to plan andG > implement effective transitional and migration strategies to UNIX andt > Microsoft environments." >cF > There's more like this. So a major reseller has decided to scare VMS2 > customers further for their own commercial gain. >2< > Richard George - Compaq UK (to whom this is also emailed), >tG > You will recall I spoke to you at the CSF (the company involved here)0@ > Alpha=:>Itanium issues day. How do you feel about them sendingC > customers this sort of garbage? "VMS legacy", "final generation",e > "migrate to UNIX and WIndows": >, >o > -- > Alan   ------------------------------    Date: 26 Sep 2001 09:20:42 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)o/ Subject: Re: Major VMS reseller says VMS legacy 3 Message-ID: <RVXLg+x$1Mw6@eisner.encompasserve.org>o  h In article <Qhks7.62655$TW.340674@tor-nn1.netcom.ca>, "Syltrem" <syltrem@videotron.spammenot.ca> writes:K > Why don't you post the email address of this guy so that we SPAM him witht > our complaints?    Because Alan has more sense ?   N ==============================================================================I The Boulder Pledge: "Under no circumstances will I ever purchase anything3J      offered to me as the result of an unsolicited email message. Nor willI      I forward chain letters, petitions, mass mailings, or virus warningsuH      to large numbers of others. This is my contribution to the survival      of the online community."N ==============================================================================   ------------------------------  % Date: Wed, 26 Sep 2001 10:29:37 -0400 % From: "John Vottero" <John@mvpsi.com>n/ Subject: Re: Major VMS reseller says VMS legacyV/ Message-ID: <tr3pij36oi5e0d@news.supernews.com>t  2 "Alan Greig" <a.greig@virgin.net> wrote in message2 news:jqd3rtseiknbhdgs9dt6fafjuajf8rh7av@4ax.com... >b= > A major UK Compaq ISV and VMS specialist has just set up anCF > outsourcing subsidiary XAMAN who are continually on the phone to me.E > Constant phone calls prompted me to read the brochure and found thep > following: >tG > "Such legacy systems comprise, in the main, a large number of DigitalrH > VMS environments. In its time, VMS dominated manufacturing, and in theG > right hands is still capable of running large,, modern workloads. ButwH > it is now functionally stabilized and the tide is running out for much7 > of the UK's installed base of older Digital systems."a > F > "Our skills with VMS run deep. To this day we continue to wring more< > performance efficiencies from the final generation of thisB > technology,... And we yield to nobody in our ability to plan andG > implement effective transitional and migration strategies to UNIX and  > Microsoft environments." >I  2 These dolts don't seem to know anything about VMS.  . Makes you wonder if they know anything at all.   ------------------------------  % Date: Wed, 26 Sep 2001 16:31:41 +0200s  From: Paul Sture <paul@sture.ch>/ Subject: Re: Major VMS reseller says VMS legacyg+ Message-ID: <VA.00000460.15a5e035@sture.ch>l  D In article <jqd3rtseiknbhdgs9dt6fafjuajf8rh7av@4ax.com>, Alan Greig  wrote:= > A major UK Compaq ISV and VMS specialist has just set up anaF > outsourcing subsidiary XAMAN who are continually on the phone to me.E > Constant phone calls prompted me to read the brochure and found the  > following: > G > "Such legacy systems comprise, in the main, a large number of DigitalfH > VMS environments. In its time, VMS dominated manufacturing, and in theG > right hands is still capable of running large,, modern workloads. But<H > it is now functionally stabilized and the tide is running out for much7 > of the UK's installed base of older Digital systems."a > F > "Our skills with VMS run deep. To this day we continue to wring more< > performance efficiencies from the final generation of thisB > technology,... And we yield to nobody in our ability to plan andG > implement effective transitional and migration strategies to UNIX andd > Microsoft environments." > F > There's more like this. So a major reseller has decided to scare VMS2 > customers further for their own commercial gain. > < > Richard George - Compaq UK (to whom this is also emailed), > G > You will recall I spoke to you at the CSF (the company involved here)/@ > Alpha=:>Itanium issues day. How do you feel about them sendingC > customers this sort of garbage? "VMS legacy", "final generation",  > "migrate to UNIX and WIndows"f > I http://www.XAMAN.CO.UK/indexmain.html is full of the stuff (incidentally oI it caused me considerable browser problems, and won't work at all if you n+ don't use a frames enabled browser - grrrr)p   ___d
 Paul Sture Switzerland(   ------------------------------  % Date: Wed, 26 Sep 2001 16:01:14 +0100e% From: Alan Greig <a.greig@virgin.net>t/ Subject: Re: Major VMS reseller says VMS legacyc8 Message-ID: <62r3rt4ve32kak8igrsbm8hi6tkrsbakgf@4ax.com>  E On Wed, 26 Sep 2001 16:31:41 +0200, Paul Sture <paul@sture.ch> wrote:e    J >http://www.XAMAN.CO.UK/indexmain.html is full of the stuff (incidentally J >it caused me considerable browser problems, and won't work at all if you , >don't use a frames enabled browser - grrrr)  B It appears their web site is being/has been edited. I just clickedF "VMS and manufacturing" and went to a page with no mention of VMS. TheF "Managing IT for Value" printed brochure (the one with most of the VMSF dead content) has been truncated virtually out of existence on the web6 copy. Looks like someone might be getting the message.   -- Alan   ------------------------------  % Date: Wed, 26 Sep 2001 20:23:39 +0100i% From: Alan Greig <a.greig@virgin.net>C/ Subject: Re: Major VMS reseller says VMS legacyS* Message-ID: <3BB22B3B.5FA13147@virgin.net>   JF Mezei wrote:-  M > What is left of the VMS customer base has survived over 10 years of "VMS iscP > dead", including a certain number of years of Digital telling customers to get5 > off VMS. What is left is pretty robust I would say.t  Q But a lot of it isn't. Take manufacturing for example. There's a lot of VMS still R in these sites - including my own. However there is a constant pressure to move toN SAP and it is sites which fit this profile XAMAN are targeting. If they didn'tQ think there were still a significant number of sites likely to outsource prior totQ migration or just migrate straight away they wouldn't be targeting them. And they Q should know - they sold the systems to many of these companies in the first placedL and process the VAX/Alpha maintenance contracts for many others since Compaq! outsourced this function to them.    --
 Alan Greig   ------------------------------    Date: 26 Sep 2001 12:07:43 -0700+ From: hobbesnet@hotmail.com (Scott Squires)  Subject: Need help with RMS(= Message-ID: <ff921edf.0109261107.5e861be7@posting.google.com>9  ; I am working on a database program for keeping track of our.8 HobbesNet users and aiding in the process of registering< accounts, and have run into a problem either with RMS or the< software package I am using to access RMS.  At least I think< it is a problem involving RMS; I can't make any sense of it.: So I'm here to beg for help hoping someone will be able to figure this out.  < I'm using the RMS Interface package by Mark S. Frank which I? found in Hunter Goatley's fileserv.  Calvin is running VMS 7.2.-  ? The file is indexed with fixed length records.  The primary key E is the 'id' field in struct acnt_data.  There are no alternate keys. F  ? The problem is in updating a record.  When the program loads its? reads all of the records into memory.  When you change a recordcA and save it, update_record() is called.  update_record() receives:A the record which is to be updated, acnt_upd.  It calls RMS_FIND() 9 to find the record by the primary key, id.  Then it calls @ RMS_UPDATE() to update the record.  RMS_UPDATE returns RMS$_CHG,@ saying that I tried to change a key.  But I don't understand how? that can happen because I used the same data to find the recorda as I am using to update it.d  ? There is a FIND and a GET at the beginning which I put there to B make sure that it was trying to update the correct record, which I@ have commented out because when they execute, the UPDATE returns= SS$_NORMAL but appearantly does not actually update the file.a0 (when I read the data again, it is not changed).  " Here is the routine update_record:  % ---------- update_record() ----------c   #include <rmsdef.h>o) #include "infodefs.h" /* RMS_INFO defs */i #include "defs.h"v #include "structs.h"  P int update_record(struct program_data * prgrm_data, struct acnt_data * acnt_upd) {-   struct acnt_data acnt_test;1
   int ioflag;    long fabstv;   long rabstv;   Clear_Screen();c/ printf("Finding record id %d\n", acnt_upd->id);i  /   ioflag = RMS_FINDKEY ( prgrm_data->acnt_file,74                          acnt_upd, sizeof(long), 0);   if (ioflag != RMS$_NORMAL)4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }   /* printf("Getting record...\n");  -   ioflag = RMS_GETSEQ( prgrm_data->acnt_file,D>                        &acnt_test, sizeof(struct acnt_data) );;   if ( (ioflag != RMS$_NORMAL) && (ioflag != RMS$_OK_ALK) )e4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }  A printf("  fname: %s\n  id: %d\n", acnt_test.fname, acnt_test.id);o  /   ioflag = RMS_FINDKEY ( prgrm_data->acnt_file,h4                          acnt_upd, sizeof(long), 0);;   if ( (ioflag != RMS$_NORMAL) && (ioflag != RMS$_OK_ALK) )i4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); } */  3 printf("Updating record id %d...\n", acnt_upd->id);0  9   ioflag = RMS_UPDATE ( IFI_( &(prgrm_data->acnt_file) ),d9                         RSZ_( sizeof(struct acnt_data) ), )                         UBF_( acnt_upd ),,8                         USZ_( sizeof(struct acnt_data) )                         );   if (ioflag != RMS$_NORMAL):     { RMS_PUTERRMSG(ioflag); /* return(ADM$_ERR_RMS); */ }  
 printf("\n");t)   PRINT_FILE_INFO(prgrm_data->acnt_file);   "   RMS_INFO( prgrm_data->acnt_file,!             RMS__FABSTV, &fabstv, #             RMS__RABSTV, &rabstv );s   printf("FAB STV:\n");o   RMS_PUTERRMSG(fabstv); printf("RAB STV:\n");    RMS_PUTERRMSG(rabstv);  2 printf("Finding record id %d...\n", acnt_upd->id);  /   ioflag = RMS_FINDKEY ( prgrm_data->acnt_file,l;                          &(acnt_upd->id), sizeof(long), 0);    if (ioflag != RMS$_NORMAL)4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }  2 printf("Getting record id %d...\n", acnt_upd->id);  -   ioflag = RMS_GETSEQ( prgrm_data->acnt_file,/>                        &acnt_test, sizeof(struct acnt_data) );;   if ( (ioflag != RMS$_NORMAL) && (ioflag != RMS$_OK_ALK) )i4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }  C printf("  id: %d\n  fname: %s\n\n", acnt_test.id, acnt_test.fname);o  # printf("Press enter to continue.");j
   getch();     return(ADM$_NORMAL); }u   ---------- end ----------   E When I attempt to save a record with first name Rudy and id 10000001,(2 this is the output which the above code generates:   ---------- output ----------   Finding record id 10000001 Updating record id 10000001...7 %RMS-F-CHG, invalid key change in $UPDATE (CHG not set)   M -------------  F I L E   I N F O R M A T I O N   D I S P L A Y  -------------a3                             File Name: ACCOUNTS.DATiI  File access: <UPD,DEL,GET,PUT>                      Last IO op: <UPDATE>tF      Sharing: <GET,PUT>                               File Type: <IDX>E File options: <CBT>                                 Rec. attrib: <CR>eF    Name spec: <CNCL_DEV,EXP_NAME,EXP_TYPE>          Rec. format: <FIX>F Rec. options: <>                                    Rec. access: <KEY>D            Full Spec: DISK$USER:[SQUIRES.PROJECTS.ADM]ACCOUNTS.DAT;1J       Created: 20-SEP-2001 18:03:33.45    Revised: 26-SEP-2001 14:37:47.79C              FAB status word: 00010001   FAB status value: 000000E0l=                   %RMS-S-NORMAL, normal successful completion-C              RAB status word: 0001849C   RAB status value: 00000000eC             %RMS-F-CHG, invalid key change in $UPDATE (CHG not set)gL     EndFile Blk: 862      MaxRecNum: 0      MaxRecSize: 1026      Keysize: 4O  ------------------------------------------------------------------------------o FAB STV:. %NONAME-S-NORMAL, normal successful completion RAB STV:( %NONAME-W-NOMSG, Message number 00000000 Finding record id 10000001...  Getting record id 10000001...    id: 10000001
   fname: Rudyr     ---------- end ----------E  C When the FIND and GET in the beginning are not commented out, and I E change the fname field from Rudy to anything else, this is the outputrD (fname doesn't actually get updated, though update returns a success code):   ---------- output ----------   Finding record id 10000001 Getting record...r
   fname: Rudy    id: 10000001 Updating record id 10000001...  M -------------  F I L E   I N F O R M A T I O N   D I S P L A Y  -------------n3                             File Name: ACCOUNTS.DAT I  File access: <UPD,DEL,GET,PUT>                      Last IO op: <UPDATE>rF      Sharing: <GET,PUT>                               File Type: <IDX>E File options: <CBT>                                 Rec. attrib: <CR> F    Name spec: <CNCL_DEV,EXP_NAME,EXP_TYPE>          Rec. format: <FIX>F Rec. options: <>                                    Rec. access: <KEY>D            Full Spec: DISK$USER:[SQUIRES.PROJECTS.ADM]ACCOUNTS.DAT;1J       Created: 20-SEP-2001 18:03:33.45    Revised: 26-SEP-2001 14:31:01.36C              FAB status word: 00010001   FAB status value: 000000E0o=                   %RMS-S-NORMAL, normal successful completioncC              RAB status word: 00010001   RAB status value: 00000000t=                   %RMS-S-NORMAL, normal successful completionoL     EndFile Blk: 862      MaxRecNum: 0      MaxRecSize: 1026      Keysize: 4O  ------------------------------------------------------------------------------h FAB STV:. %NONAME-S-NORMAL, normal successful completion RAB STV:( %NONAME-W-NOMSG, Message number 00000000 Finding record id 10000001...f Getting record id 10000001...e   id: 10000001
   fname: Rudyi   ---------- end ----------s  D Thinking that perhaps it was a problem with RMS Interface, I wrote aF small program that writes a couple records to a file, and then updatesC one of them and reads in the updated record.  I tried to make is ase% close to update_record() as possible."  E The FAB, RAB, and KEY defs are the same as are used in update_record.    Here is rms_test.c:h   ---------- rms_test ----------   #include <stdio.h> #include <stdlib.h>e #include <rms.h> #include <rmsdef.h>y) #include "infodefs.h" /* RMS_INFO defs */    #define ADM$_ERR_RMS 0   struct rms_datab {    unsigned long id;r   char data[32]; };  
 int main() {f   struct rms_data * data;e   struct rms_data * new_data;    int file_id = 0;   int ioflag = 0;    int count = 0;   short key_flg = XAB$V_CHG;  @   data     = (struct rms_data *)malloc(sizeof(struct rms_data));@   new_data = (struct rms_data *)malloc(sizeof(struct rms_data));  .   RMS_FABDEF( ifi_(&file_id), fnm_("RMS.DAT"),N               str_("FOP=CBT; ORG=IDX; RFM=FIX; FAC=GET,PUT,UPD; SHR=GET,PUT"),.               mrs_(sizeof(struct rms_data)) );     RMS_RABDEF( ifi_(&file_id),/.               rsz_( sizeof(struct rms_data) ),'               str_("RAC=KEY; KRF=0") );d     RMS_KEYDEF( ifi_(&file_id),y.               str_("DTP=IN4; POS=0; REF=0") );  )   ioflag = RMS_CREATE( file_(&file_id) );s   if (ioflag != RMS$_NORMAL)5     { RMS_PUTERRMSG(ioflag); return (ADM$_ERR_RMS); }s     data->id = 711;u   strcpy(data->data, "Narf!");  '   printf("Writing data record 1...\n");p  @   ioflag = RMS_PUTKEY( file_id, data, sizeof(struct rms_data) );   if (ioflag != RMS$_NORMAL)5     { RMS_PUTERRMSG(ioflag); return (ADM$_ERR_RMS); }   D   printf("Data:\n  id:   %d\n  data: %s\n\n", data->id, data->data);     data->id = 712;l   strcpy(data->data,"Pinky");e  '   printf("Writing data record 2...\n");y  @   ioflag = RMS_PUTKEY( file_id, data, sizeof(struct rms_data) );   if (ioflag != RMS$_NORMAL)5     { RMS_PUTERRMSG(ioflag); return (ADM$_ERR_RMS); }d  D   printf("Data:\n  id:   %d\n  data: %s\n\n", data->id, data->data);  =   printf("Changing record:\n  id:   711\n  data: Poit!\n\n");r     data->id = 711;    strcpy(data->data, "Poit!");  0   printf("Finding record id %d...\n", data->id);  @   ioflag = RMS_FINDKEY ( file_id, &(data->id), sizeof(long), 0);   if (ioflag != RMS$_NORMAL)4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }  1   printf("Updating record id %d...\n", data->id);l  '   ioflag = RMS_UPDATE ( IFI_(&file_id),a. 		UBF_(data), USZ_(sizeof(struct rms_data)) );   if (ioflag != RMS$_NORMAL):     { RMS_PUTERRMSG(ioflag); /* return(ADM$_ERR_RMS); */ }  0   printf("Finding record id %d...\n", data->id);  @   ioflag = RMS_FINDKEY ( file_id, &(data->id), sizeof(long), 0);   if (ioflag != RMS$_NORMAL)4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }  0   printf("Getting record id %d...\n", data->id);  G   ioflag = RMS_GETSEQ ( file_id, new_data, sizeof(struct rms_data), 0);h;   if ( (ioflag != RMS$_NORMAL) && (ioflag != RMS$_OK_ALK) )d4     { RMS_PUTERRMSG(ioflag); return(ADM$_ERR_RMS); }  :   printf("Data after update:\n  id:   %d\n  data: %s\n\n",  		new_data->id, new_data->data);  !   RMS_CLOSEOUT( ifi_(&file_id) );=   }=   ---------- end ----------=  * Here is the output from the above program:   ---------- output ----------   Writing data record 1... Data:e   id:   711t
   data: Narf!t   Writing data record 2... Data:f   id:   712l
   data: Pinkyn   Changing record:   id:   711g
   data: Poit!g   Finding record id 711... Updating record id 711...v Finding record id 711... Getting record id 711... Data after update:   id:   711=
   data: Poit!=   ---------- end -----------  & Any help would be greatly appreciated.   Regards,
 Scott SquiresJ   hobbesnet@hotmail.comc   ------------------------------  % Date: Wed, 26 Sep 2001 18:40:35 -04003- From: JF Mezei <jfmezei.spamnot@videotron.ca>. Subject: Re: Need help with RMSe+ Message-ID: <3BB25951.43F294F@videotron.ca>.   Scott Squires wrote:A > The file is indexed with fixed length records.  The primary keyuF > is the 'id' field in struct acnt_data.  There are no alternate keys.  A > The problem is in updating a record.  When the program loads it A > reads all of the records into memory.  When you change a record ) > and save it, update_record() is called.V  N  I suspect you may have sequential access set on that RAB and you might need aM second RAB for direct/indexed access to the file. Does the update routine sete3 the RAC field of the RAB to RAB$C_KEY or RAB$C_SEQ.e  L For instance, when you initially load all records, you probably have the SEQL option in the RAB to indicate that each $GET is sequential. But once that isL done, you need to set it to KEY to indicate that from now on, $FIND and $GET work in an indexed fashion.i  M From the code snippets you gave, it looks like you have some wrapper routinessM that give you access to RMS services. You will need to look into those to see 6 what sort of record processing options are being used.  M Look that the VMS documentation, in particular the RMS manual for the $UPDATERK and $FIND routines , notably which fields are read (input) and set (output)r for each of those routines.e   ------------------------------  % Date: Tue, 25 Sep 2001 12:53:27 -0500a. From: Duncan Brown <brown_du@eisner.decus.org>- Subject: Re: New..Wiring in terminals...help?y0 Message-ID: <3BB0C497.2BDABEA9@eisner.decus.org>   Phil Mendelsohn wrote: > * > On Tue, 25 Sep 2001, Duncan Brown wrote: > J > > When you ask about configuration, if you mean DEC VT-series terminals,4 > > then just press the F2 key and follow the menus. > L > On my VT-320 and VT-220 it's F3.  I think it's the same on the 100 I don't > have anymore.y  5 You are correct, sir.  That was a (badly timed) typo.e F1 - Hold screen
 F2 - Print
 F3 - Setup   Duncan   ------------------------------  % Date: Wed, 26 Sep 2001 06:46:28 -0400 ) From: "Neil Rieck" <n.rieck@sympatico.ca>e Subject: News Group ObservationA; Message-ID: <Amis7.16265$Df4.1696240@news20.bellglobal.com>   E About a month ago I decided to highlight posts from Compaq employees. G (OUTLOOK EXPRESS> TOOLS> MESSAGE RULES> NEWS: Apply this rule after theiK message arrives Where the From line contains 'compaq.' or 'dec.'  Highlighti
 it with Red )C  G I was surprised to see that at least 12 different employees are postingyK here. This is probably not the type of VMS promotion that we've been hoping " for, but it's better than nothing.  ; ps. I guess I should start looking for posts from "hp." :-)   
 Neil Rieck Kitchener/Waterloo/Cambridge,/ Ontario, Canada.! http://www3.sympatico.ca/n.rieck/(   ------------------------------  % Date: Tue, 25 Sep 2001 20:13:48 -0400a3 From: "Brad McCusker" <Brad.McCuskerSP@Mcompaq.com>g0 Subject: Re: Pathworks, Advanced Server issue...1 Message-ID: <l99s7.628$YP.21310@news.cpqcorp.net>1  / Yep, your print job database is likely corrupt.t  I Delete the job_data_base and queue_data_base files in PWRK$LMROOT:[PRINT]2H and restart the server.  Both files will get recreated and you should be good to go.l   Regards,   Brad     --( The opinions expressed herein are my own' and do not reflect those of my employerm or anyone else.o   Brad  0 "CasinoOp2" <casinoop2@aol.com> wrote in message3 news:20010912194125.28363.00000245@mb-md.aol.com...bL > There must be someone out there that can answer this question. On my AlphaI > OpenVMS system, Advanced Server process aborts. The following is in ther loge > file pwrk$lmsrv_vnc2.log...  >d& > PPS_jdb_open: Could not read header!G > 12-SEP-2001 10:25:35.02 00000237:007325C0 PPS Initialization failure:  Cannot > read job database!K > PANIC: aborting from module AS$BLD_ROOT:[AS.UTIL.SRC]PPS_UTIL.C;1 at linen 105! > %SYSTEM-F-ABORT, abort >     .........V >r >lK > Any idea what's going wrong here? I am at my wits end.  Any help would bew > appreciated! >  >    Clark Calkins, programmer >    Schafer Corpy >    ccalkins@schaferLabs.comu >g   ------------------------------  # Date: Tue, 25 Sep 2001 20:19:38 GMT.* From: Steven Whatley <swhatley@blkbox.com>% Subject: Re: Really Unhackable...????tC Message-ID: <uF5s7.659428$ai2.51333892@bin2.nnrp.aus1.giganews.com>u  . Jack Patteeuw <jjpatteeuw@peoplepc.com> wrote:H > Read the book "The Kukoo's (sp?) Egg" about the "early days"  hacking.  F Read the book years ago.  I though it was very good and interesting.  F There is also a "Hackers Guide to VMS" floating around.  But it prettyC much requires the hacker to already have an account with the CMKRNLy priviledge.    Later, Steven   ------------------------------  % Date: Wed, 26 Sep 2001 10:41:51 -0400a# From: Jim Agnew <agnew@hsc.vcu.edu>i% Subject: Re: Really Unhackable...???? + Message-ID: <3BB1E92F.FE56C09E@hsc.vcu.edu>S  E i have that upon request, and there are 2, i think.. i'll have to digj thru my emails...    jime   Steven Whatley wrote:6 > 0 > Jack Patteeuw <jjpatteeuw@peoplepc.com> wrote:J > > Read the book "The Kukoo's (sp?) Egg" about the "early days"  hacking. > F > Read the book years ago.  I though it was very good and interesting.H > There is also a "Hackers Guide to VMS" floating around.  But it prettyE > much requires the hacker to already have an account with the CMKRNLh
 > priviledge.h >  > Later, > Steven   ------------------------------  % Date: Thu, 27 Sep 2001 08:04:31 +0800 ' From: "Kenneth" <chehon@netvigator.com>'" Subject: Router load host or local0 Message-ID: <9otq6d$qgr2@imsp212.netvigator.com>  I How can I check if the router config is load for itself or from the host?a   ------------------------------  % Date: Wed, 26 Sep 2001 14:08:06 -0400l4 From: John Malmberg <Malmberg@dskwld.zko.dec.compaq>) Subject: Re: Samba on OpenVMS protection.e4 Message-ID: <3BB21986.9040102@dskwld.zko.dec.compaq>  	 pj wrote:o6 > I installed the latest version of Samba for openvms.  6 The latest binaries I am aware of available by ftp are2 for SAMBA 2.0.6.  They are on the "latest" OpenVMS Freeware CD-ROM.   > But i am getting errors. > " > Checking with testparms unveals: <snip>  ; > WARNING: lock directory /samba_root/var/locks should havet' > permissions 0755 for browsing to worke  8 See the SAMBA-VMS FAQ with the SAMBA 2.0.6 distribution. The warning is bogus.t  H > I already changed permissions on the directory , even with (W:RWED) no	 > result._ > I > When I want to open a share logon goes OK but the a message apears witht >  > $ smbclient //utopjr/kitse > E > Added interface ip=16.197.80.8 bcast=16.255.255.255 nmask=255.0.0.0  > C > Password: Domain=[DIGITALWSUTO] OS=[OpenVMS] Server=[Samba 2.0.3]u >  > tree connect failed: code 0   # I am not sure.  I never used 2.0.3.   I It looks like it did not accept the username/password for the connection.1  1 A higher debug level might give more information.l   > What is wrong here?i  I Try the tests in diagnosis.txt.  If it is not in your kit, IIRC it is on -E the documentation page of http://www.samba.org mirror closest to you.f   -John! Personal Opinion OnlyM malmberg@dskwld.zko.dec.compaq   ------------------------------  % Date: Tue, 25 Sep 2001 17:55:18 +0100(- From: "Martin J. Evans" <martin@easysoft.com> + Subject: select() returning errno=1 (EPERM)f, Message-ID: <3BB0B6F6.FF922F3D@easysoft.com>  B I have a problem using select() with UCX 5.0A on OpenVMS 7.2/alpha% I was hoping someone might have seen.s  > My server process listens on a port accepting new connections.8 After the listen() call the server calls select() on the8 listening socket fd with a time out of 5 seconds so that; if a connection does not arrive in that time it can do some; house keeping.  8 When a new connection comes in it vfork()s and the child8 immediately execs the same executable as the main server< but with different arguments. As I was not aware VMS vforked> processes inherit file descriptors (when the code was written); the child gets the network device as an argument, assigns a(@ channel to it and calls decc$get_sdc to get the file descriptor.> When the client closes the socket the child closes its socket,  deassigns the channel and exits.  = At this point the main server's select() returns -1 and errno @ is set to EPERM - this does not always happen but about half the child exits cause this.   = If the server ignores the EPERM and simply calls select again 8 there is no problem and the server appears to work fine.  < select() does not appear to be documented as returning EPERM< and this only appears to happen when the child process exits( when the parent is blocking in select().   Any ideas why?   Thanks   Martin -- Martin J. Evansl Easysoft Ltd, UK Developmentn   ------------------------------  # Date: Tue, 25 Sep 2001 19:50:57 GMT = From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) 0 Subject: Re: Set Command for SUBPROCESS Question0 Message-ID: <00A02962.543F5BCC@SendSpamHere.ORG>  W In article <C2256AD2.0065A9E4.00@jklh21.valmet.com>, norm.raphael@jamesbury.com writes:_ >M >; >u > B >I have an app that in due course spawns a subprocess to execute a >t6 >BACKUP/VERIFY/DELETE files tape:saveset command, then >a& >deletes the subprocess and continues. >-? >Of course, it keeps the tape allocated during the delete pass.0 >.@ >I want to modify the BACKUP verb for the subprocess so that the >-9 >/RELEASE_TAPE qualifier is defaulted for that one BACKUP- >- >invokation. > C >I cannot change the application code (I do not own it or have it).T >T< >Can this be done by some form of SET COMMAND for the tables >P* >for a subprocess that does not exist yet?  F Is this DCL code that SPAWNs the process?  If so, try /TABLE= with the SPAWN command.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM              J   "And of course, I'm a genius, so people are naturally drawn to my fiery I   intellect.  Their admiration overwhelms their envy!" -- Calvin & Hobbes:   ------------------------------    Date: 26 Sep 2001 03:40:09 -0700' From: postmaster@roblock.com (Rob Lock)  Subject: SMTP MAIL FROM VMSs= Message-ID: <8b0a76c5.0109260240.22fe05ff@posting.google.com>s  N I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP.  C but I want to sent smtp mail , but I keep getting the same problem:6  
 MAIL> send To:     user@tst.nhs.uk-2 %MAIL-E-ERRACTRNS, error activating transport SMTPP %LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MA
 ILSHR.EXE; -RMS-E-FNF, file not found   MAIL>   & What do I need to do to resolve this ?   Thanks in advance1   Robm   ------------------------------  % Date: Wed, 26 Sep 2001 14:53:57 -0400g- From: JF Mezei <jfmezei.spamnot@videotron.ca>e Subject: Re: SMTP MAIL FROM VMSg, Message-ID: <3BB22441.5552B2CF@videotron.ca>   Rob Lock wrote:t > P > I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP. > To:     user@tst.nhs.uko4 > %MAIL-E-ERRACTRNS, error activating transport SMTPR > %LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MA > ILSHR.EXE; > -RMS-E-FNF, file not found  < On my system, the file is SYS$LIBRARY:TCPIP$SMTP_MAILSHR.EXE  L So if VMSmail is looking for SMTP_MAILSHR, it means that there is no logicalL name defined to redirect the request to the right file (TCPIP$SMTP_MAILSHR).: The simple answer is that the SMTP service is not started.  % look for a logical MAIL$PROTOCOL_SMTPo   Another way is to : F $TCPIP SHOW SERVICES and you will see if SMTP has been enabled or not.   You could try:  & to:   other_node::smtp%user@tst.nhs.uk  / In case the service is started on another node..      L SYS$MANAGER:TCPIP$SMTP_STARTUP.COM is the procedure to start the service (orS you can use @TCPIP$CONFIG to ensure your smto service is properly configured first)0   ------------------------------    Date: 26 Sep 2001 14:01:40 -0700( From: bob@instantwhip.com (Bob Ceculski) Subject: Re: SMTP MAIL FROM VMS = Message-ID: <d7791aa1.0109261301.16cf732b@posting.google.com>-  d norm lastovica <norman.lastovica@oracle.com> wrote in message news:<3BB1D5FA.BE8F6DCD@oracle.com>...I > make sure that you're running the current version of TCP/IP (V5.1 ECO 1 G > I believe) and make sure that you've configured and started SMTP (viad > TCPIP$CONFIGURE).- >  > Rob Lock wrote:t > > R > > I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP. > > G > > but I want to sent smtp mail , but I keep getting the same problem:r > >  > > MAIL> send > > To:     user@tst.nhs.uko6 > > %MAIL-E-ERRACTRNS, error activating transport SMTPT > > %LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MA > > ILSHR.EXE; > > -RMS-E-FNF, file not found > > 	 > > MAIL>m > > * > > What do I need to do to resolve this ? > >  > > Thanks in advancen > >  > > Rob0    = why are you running ucx smtp ... tcpware has a wonderful smtp; package w/alot more features...d   ------------------------------  % Date: Wed, 26 Sep 2001 08:39:27 -0400l- From: Michael Austin <miaustin@bellsouth.net>A Subject: Re: SMTP MAIL FROM VMSr- Message-ID: <3BB1CC7F.FEC9A143@bellsouth.net>e   Which version/vendor for TCPIP?s  O Normally in VMSMail to use smtp, the syntax is smtp%"user@domain.com".  In your O tcpip package, you usually configure the SMTP FORWARDER to be your primary mail ? server.  Have you done this?  Did you enable the SMTP protocol?i   Michael Austin DBA Consultant   Rob Lock wrote:   P > I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP. > E > but I want to sent smtp mail , but I keep getting the same problem:e >) > MAIL> send > To:     user@tst.nhs.uk 4 > %MAIL-E-ERRACTRNS, error activating transport SMTPR > %LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MA > ILSHR.EXE; > -RMS-E-FNF, file not found >a > MAIL>s >c( > What do I need to do to resolve this ? >  > Thanks in advanceo >g > Robr   ------------------------------  % Date: Wed, 26 Sep 2001 08:31:36 -0400>5 From: David Beatty <David.Beatty@qwertysasasdfgh.com>d Subject: Re: SMTP MAIL FROM VMSn2 Message-ID: <b8qxO+KdnB3R8CZfLGLLcUFVZa+Q@4ax.com>  =     What value is logical name MAIL$PROTOCOL_SMTP?  It shouldR7 be UCX$SMTP_MAILSHR for UCX V4.X and TCPIP$SMTP_MAILSHR ! for Digital TCP/IP Services V5.X.    David R. Beattya  @ On 26 Sep 2001 03:40:09 -0700, postmaster@roblock.com (Rob Lock) wrote:  O >I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP.  >oD >but I want to sent smtp mail , but I keep getting the same problem: >d >MAIL> sendg >To:     user@tst.nhs.uk3 >%MAIL-E-ERRACTRNS, error activating transport SMTPEQ >%LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MAU >ILSHR.EXE;& >-RMS-E-FNF, file not found_ >z >MAIL> t >s' >What do I need to do to resolve this ?  >  >Thanks in advance >; >Rob   ------------------------------  % Date: Wed, 26 Sep 2001 09:19:54 -0400;2 From: norm lastovica <norman.lastovica@oracle.com> Subject: Re: SMTP MAIL FROM VMS * Message-ID: <3BB1D5FA.BE8F6DCD@oracle.com>  G make sure that you're running the current version of TCP/IP (V5.1 ECO 1 E I believe) and make sure that you've configured and started SMTP (via) TCPIP$CONFIGURE).!   Rob Lock wrote:i > P > I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP. > E > but I want to sent smtp mail , but I keep getting the same problem:_ >  > MAIL> send > To:     user@tst.nhs.uk)4 > %MAIL-E-ERRACTRNS, error activating transport SMTPR > %LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MA > ILSHR.EXE; > -RMS-E-FNF, file not found >  > MAIL>  > ( > What do I need to do to resolve this ? >  > Thanks in advancer >  > Rob    --  > norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  # Date: Wed, 26 Sep 2001 15:31:45 GMTt' From: "Hans Vlems" <hvlems@zfree.co.nz>- Subject: Re: SMTP MAIL FROM VMSw" Message-ID: <3bb1dfc2@zfree.co.nz>  @ Depending on whether you run UCX or TCPIP, you need to configure@ SMTP: @sys$manager:UCX$CONFIGURE (or TCPIP$CONFIGURE of course).8 There may be a separate DCL file to configure SMTP with.  
 Hans Vlems  ( postmaster@roblock.com (Rob Lock) wrote:H >I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP. >rD >but I want to sent smtp mail , but I keep getting the same problem: >k >MAIL> send  >To:     user@tst.nhs.uk3 >%MAIL-E-ERRACTRNS, error activating transport SMTPeQ >%LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MA  >ILSHR.EXE;I >-RMS-E-FNF, file not foundi >a >MAIL>   >t' >What do I need to do to resolve this ?f >d >Thanks in advance >  >Rob       http://www.zfree.co.nz   ------------------------------  # Date: Wed, 26 Sep 2001 15:31:46 GMT ' From: "Hans Vlems" <hvlems@zfree.co.nz>p Subject: Re: SMTP MAIL FROM VMSh$ Message-ID: <3bb1dfc2$1@zfree.co.nz>  @ Depending on whether you run UCX or TCPIP, you need to configure@ SMTP: @sys$manager:UCX$CONFIGURE (or TCPIP$CONFIGURE of course).8 There may be a separate DCL file to configure SMTP with.  
 Hans Vlems  ( postmaster@roblock.com (Rob Lock) wrote:H >I have two clusters, one alpha's, one vaxes all running oVMS 7.1-2 with TCPIP. >sD >but I want to sent smtp mail , but I keep getting the same problem: >S >MAIL> sendh >To:     user@tst.nhs.uk3 >%MAIL-E-ERRACTRNS, error activating transport SMTP(Q >%LIB-E-ACTIMAGE, error activating image $6$DKA0:[SYS0.SYSCOMMON.][SYSLIB]SMTP_MAu >ILSHR.EXE;1 >-RMS-E-FNF, file not foundn >o >MAIL> _ >e' >What do I need to do to resolve this ?g >  >Thanks in advance >s >Rob       http://www.zfree.co.nz   ------------------------------  # Date: Wed, 26 Sep 2001 20:31:28 GMTw= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-) J Subject: Re: SNMP client software for OpenVMS (Alpha or VAX, Motif or CLI)0 Message-ID: <00A02A31.27EFEB19@SendSpamHere.ORG>  f In article <cc5619f2.0109260937.16ac5223@posting.google.com>, jordan@ccs4vms.com (Rich Jordan) writes:A >I've been doing a lot of searching to no avail; I hope I haven't-G >missed something obvious and posted a faq as a result.  Nothing in theNF >VMS FAQ, freeware lists, google searches, or the various VMS software@ >sites I can get to (an abbreviated list today; can't get to the >European sites at all). >g; >Is there available SNMP _client_ software for VMS?  Either>D >graphical/Motif or CLI-based would be useful, and Alpha more usefulF >than VAX; I can deal with compiling source, but do not have time (forF >this requirement) to attempt porting one of the Linux/Unix packages. G >Similarly, I did see mention of various PERL modules (though they seemnF >to require underlying code that would probably need to be ported) but; >really don't have time to try and get up to speed on Perl./ >eG >Multinet, TCPware, and UCX/TCPIP all have SNMP servers, but outside of E >one mention of an old CMU-SNMP package that a poster talked about inaG >1996 or so (and considered obsolete then), no word at all on a client.w >sB >Thanks in advance.  An email response would be great, but I'll beF >following up here as well (so you don't have to say 'ask here, get an >answer here, Hoff; I know :)t >a >Rich Jordan  J TCP/IP services has SNMP_REQUEST, SNMP_TRAPRCV, and SNMP_TRAPSND commands H which work quite well for my purposes.  I've been using them to test the. SNMP stuff I am incorporating into a product.  --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM             eJ   "And of course, I'm a genius, so people are naturally drawn to my fiery I   intellect.  Their admiration overwhelms their envy!" -- Calvin & Hobbesa   ------------------------------    Date: 26 Sep 2001 10:37:04 -0700& From: jordan@ccs4vms.com (Rich Jordan)F Subject: SNMP client software for OpenVMS (Alpha or VAX, Motif or CLI)= Message-ID: <cc5619f2.0109260937.16ac5223@posting.google.com>I  @ I've been doing a lot of searching to no avail; I hope I haven'tF missed something obvious and posted a faq as a result.  Nothing in theE VMS FAQ, freeware lists, google searches, or the various VMS softwarep? sites I can get to (an abbreviated list today; can't get to they European sites at all).b  : Is there available SNMP _client_ software for VMS?  EitherC graphical/Motif or CLI-based would be useful, and Alpha more usefuluE than VAX; I can deal with compiling source, but do not have time (for3E this requirement) to attempt porting one of the Linux/Unix packages. cF Similarly, I did see mention of various PERL modules (though they seemE to require underlying code that would probably need to be ported) butn: really don't have time to try and get up to speed on Perl.  F Multinet, TCPware, and UCX/TCPIP all have SNMP servers, but outside ofD one mention of an old CMU-SNMP package that a poster talked about inF 1996 or so (and considered obsolete then), no word at all on a client.  A Thanks in advance.  An email response would be great, but I'll be E following up here as well (so you don't have to say 'ask here, get ant answer here, Hoff; I know :)   Rich Jordanj   ------------------------------  % Date: Wed, 26 Sep 2001 22:46:51 -0500 , From: "Rich Jordan" <rjordan@mindspring.com>J Subject: Re: SNMP client software for OpenVMS (Alpha or VAX, Motif or CLI)2 Message-ID: <9ou7dl$151$1@slb5.atl.mindspring.net>  . Brian Schenkenberger, VAXman- wrote in message' <00A02A31.27EFEB19@SendSpamHere.ORG>...h> >In article <cc5619f2.0109260937.16ac5223@posting.google.com>,( jordan@ccs4vms.com (Rich Jordan) writes: >>< >>Is there available SNMP _client_ software for VMS?  EitherE >>graphical/Motif or CLI-based would be useful, and Alpha more usefuld >>than VAX; .....s >>H >>Multinet, TCPware, and UCX/TCPIP all have SNMP servers, but outside ofF >>one mention of an old CMU-SNMP package that a poster talked about inH >>1996 or so (and considered obsolete then), no word at all on a client. >rJ >TCP/IP services has SNMP_REQUEST, SNMP_TRAPRCV, and SNMP_TRAPSND commandsI >which work quite well for my purposes.  I've been using them to test the . >SNMP stuff I am incorporating into a product. >--t     Brian,K      thanks.  I don't know how I missed those; the SPD only mentions serverOJ (agents) and I must have glossed over those in the docs.  The site in needC is running TCPIP V5.1 on an Alpha so I'll hit those books tomorrow.a   Rich Jordanh   ------------------------------  % Date: Tue, 25 Sep 2001 12:37:25 -0400h0 From: "Mike Zarudzki" <mike.zarudzki@compaq.com> Subject: Re: Sophos and PMDF1 Message-ID: <cp2s7.612$YP.20628@news.cpqcorp.net>   ' The newsgroup has lots of info on this:r   vmsnet.mail.pmdf  A One can customize the behaviour to suit a particular requirement.d   -Mike Z.  / "John Nebel" <nebel@csdco.com> wrote in message.@ news:Pine.OSF.4.21.0109240847250.6556-100000@athena.csdco.com... >2E > It works fine - the below is right out of Process web site with thes4 > exception that the virii and headers are archived. >e
 > pmdf.cnf >h > ! Rewrite rule for testing > !v< > testing.csdco.com                        $U%$D@tcp-testing > + > tcp_testing smtp single_sys queue stoppedy
 > TCP-TESTINGt >  > conversions. >t( > in-channel=*; out-channel=tcp_testing;% >  in-type=application; in-subtype=*;i/ >  parameter-symbol-0=NAME; parameter-copy-0=*;-5 >  dparameter-symbol-0=FILENAME; dparameter-copy-0=*;51 >  message-header-file=2; original-header-file=1;t >  override-header-file=1;# >  command="@pmdf_table:vsweep.com"  >B > in-channel=*; out-channel=*;% >  in-type=application; in-subtype=*;X/ >  parameter-symbol-0=NAME; parameter-copy-0=*;m5 >  dparameter-symbol-0=FILENAME; dparameter-copy-0=*; 1 >  message-header-file=2; original-header-file=1;  >  override-header-file=1;# >  command="@pmdf_table:vsweep.com"  >  > vsweep.com >h% > $ vsweep:==$pmdf:[vsweep]vsweep_axpt > $ run pmdf_exe:unique_id? > $ vsweep 'INPUT_FILE'/ff/ns/il/macv/output=vs-'unique_id'.lise9 > $ if sweep$_status .nes. "SWEEP$_VIRUS" then goto cleane) > $ copy 'INPUT_FILE' pmdf:[vsweep.virii]s. > $ copy 'INPUT_HEADERS' pmdf:[vsweep.headers]0 > $ copy 'MESSAGE_HEADERS' pmdf:[vsweep.headers]2 > $ copy pmdf_table:delete_virus.txt 'OUTPUT_FILE' > $ define/job OUTPUT_TYPE TEXTs# > $ define/job OUTPUT_SUBTYPE PLAINt > $ define/job OUTPUT_MODE TEXTh# > $ define/job OUTPUT_ENCODING NONE  > $ delete vs-'unique_id'.lis; > $ exit > $!
 > $ clean:% > $ rename 'INPUT_FILE' 'OUTPUT_FILE'R > $ delete vs-'unique_id'.lis; >  > John Nebel >h > " > On 24 Sep 2001, manrubble wrote: >o > > Alpha 1200, 512mb, > > Openvms 7.2i > > PMDF 5.2.x > >eI > > We are testing Sophos Vsweep for thirty day's.  Currently it is setupcB > > to run as a batch.    To date it has eradicated approx. 60,000I > > infected e-mails (we have over 1500 students so this doesn't surprise G > > me).   My question to the tech's at Sophos was, "How do I integratemF > > PMDF with VSWEEP so files are scanned before the reach the user(s)2 > > mailbox."  They told me to check with Process. > >eJ > > Has anyone integrated Sophos and PMDF successfully?  I assume it's viaH > > the conversion channel, but I was wondering if there is another way. > >. > > best regards,a > >@ > > russ > >r >s   ------------------------------   Date: 26 Sep 2001 11:46:22 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon) * Subject: Re: SuSE Linux Desupports ALPHA ?, Message-ID: <9osf6e$2lqr$1@info.cs.uofs.edu>  0 In article <3BB12FC4.9EF82881@yahoo.commercial>,4  Ed Wensell III <ewensell3@yahoo.commercial> writes: |> |> You want commercial support?a |> r |> http://www.redhat.com |> http://www.wasabisystems.comI |> oI |> I realize the second one is NetBSD not Linux... But it boots Alpha and G |> VAX right from the CDROM.... Among many other platforms. Support and " |> development services available.  D Actually, I have yet to find any Linux CD (including the REDHat fromD CETS) that will boot right out of the box on my Alpha.  FreeBSD does% and installs flawlessly, including X.    bill   -- eJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   o   ------------------------------    Date: 26 Sep 2001 10:18:48 -05001 From: Graham Allan <allan@physics.umn.edu.nospam>,* Subject: Re: SuSE Linux Desupports ALPHA ?/ Message-ID: <w53ofnyou93.fsf@crail.spa.umn.edu>b  3 bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:   2 > In article <3BB12FC4.9EF82881@yahoo.commercial>,6 >  Ed Wensell III <ewensell3@yahoo.commercial> writes: > |>! > |> You want commercial support?_ > |> H > |> http://www.redhat.com! > |> http://www.wasabisystems.comt > |>  K > |> I realize the second one is NetBSD not Linux... But it boots Alpha and-I > |> VAX right from the CDROM.... Among many other platforms. Support and@$ > |> development services available. > F > Actually, I have yet to find any Linux CD (including the REDHat fromF > CETS) that will boot right out of the box on my Alpha.  FreeBSD does' > and installs flawlessly, including X.o  K Does the regular FreeBSD CD set contain the Alpha version as well as Intel?rD That would be nice... I've been meaning to take a look at FreeBSD onG Alpha but haven't had much time so far... and the FreeBSD web site is a:% bit quiet about Alpha-specific stuff.    Graham -- SI -------------------------------------------------------------------------e Graham Allan9 School of Physics and Astronomy - University of MinnesotaaI -------------------------------------------------------------------------o   ------------------------------   Date: 26 Sep 2001 15:37:57 GMT1 From: bill@triangle.cs.uofs.edu (Bill Gunshannon)o* Subject: Re: SuSE Linux Desupports ALPHA ?, Message-ID: <9ossol$2t8n$1@info.cs.uofs.edu>  / In article <w53ofnyou93.fsf@crail.spa.umn.edu>,c4  Graham Allan <allan@physics.umn.edu.nospam> writes:6 |> bill@triangle.cs.uofs.edu (Bill Gunshannon) writes: |> 0I |> > Actually, I have yet to find any Linux CD (including the REDHat frommI |> > CETS) that will boot right out of the box on my Alpha.  FreeBSD doess* |> > and installs flawlessly, including X. |> gN |> Does the regular FreeBSD CD set contain the Alpha version as well as Intel?G |> That would be nice... I've been meaning to take a look at FreeBSD onrJ |> Alpha but haven't had much time so far... and the FreeBSD web site is a( |> bit quiet about Alpha-specific stuff.  C No.  When you go for the ISO image you grab the one for the desiredeB architecture.  Somehow, I doubt that three copies of everything onC the CD now would fit on a single CD. (I say three because they also 
 support PPC.)e   bill    -- sJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------   Date: 26 Sep 2001 19:07:01 GMT) From: wkb@freebie.xs4all.nl (Wilko Bulte).* Subject: Re: SuSE Linux Desupports ALPHA ?; Message-ID: <3bb22755$0$87446$e4fe514c@newszilla.xs4all.nl>a  Z In <w53ofnyou93.fsf@crail.spa.umn.edu> Graham Allan <allan@physics.umn.edu.nospam> writes:  4 >bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:  3 >> In article <3BB12FC4.9EF82881@yahoo.commercial>, 7 >>  Ed Wensell III <ewensell3@yahoo.commercial> writes:@ >> |>c" >> |> You want commercial support? >> |>  >> |> http://www.redhat.como" >> |> http://www.wasabisystems.com >> |> L >> |> I realize the second one is NetBSD not Linux... But it boots Alpha andJ >> |> VAX right from the CDROM.... Among many other platforms. Support and% >> |> development services available.  >> tG >> Actually, I have yet to find any Linux CD (including the REDHat fromeG >> CETS) that will boot right out of the box on my Alpha.  FreeBSD doesn( >> and installs flawlessly, including X.  L >Does the regular FreeBSD CD set contain the Alpha version as well as Intel?E >That would be nice... I've been meaning to take a look at FreeBSD onhH >Alpha but haven't had much time so far... and the FreeBSD web site is a& >bit quiet about Alpha-specific stuff.  G There is a seperate 4-CD FreeBSD/alpha CD set. I'd suggest downloading S7 the FreeBSD 4.4 install .iso image and give it a spin..Z   W/   --- |   / o / /_  _   		email: 	wilko@FreeBSD.orgL1 |/|/ / / /(  (_)  Bulte		Arnhem, The Netherlands	X   ------------------------------   Date: 26 Sep 2001 19:07:36 GMT) From: wkb@freebie.xs4all.nl (Wilko Bulte)0* Subject: Re: SuSE Linux Desupports ALPHA ?; Message-ID: <3bb22777$0$87451$e4fe514c@newszilla.xs4all.nl>o  W In <9ossol$2t8n$1@info.cs.uofs.edu> bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:b  0 >In article <w53ofnyou93.fsf@crail.spa.umn.edu>,5 > Graham Allan <allan@physics.umn.edu.nospam> writes:I7 >|> bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:M >|> J >|> > Actually, I have yet to find any Linux CD (including the REDHat fromJ >|> > CETS) that will boot right out of the box on my Alpha.  FreeBSD does+ >|> > and installs flawlessly, including X.e >|> O >|> Does the regular FreeBSD CD set contain the Alpha version as well as Intel?jH >|> That would be nice... I've been meaning to take a look at FreeBSD onK >|> Alpha but haven't had much time so far... and the FreeBSD web site is a ) >|> bit quiet about Alpha-specific stuff.e  D >No.  When you go for the ISO image you grab the one for the desiredC >architecture.  Somehow, I doubt that three copies of everything ontD >the CD now would fit on a single CD. (I say three because they also >support PPC.)  G Make that: a PPC port is in progress. It is not ready by any stretch oft imagination.   W/ --- |   / o / /_  _   		email: 	wilko@FreeBSD.orgA1 |/|/ / / /(  (_)  Bulte		Arnhem, The Netherlands	    ------------------------------  % Date: Wed, 26 Sep 2001 20:28:49 -0500o1 From: "David J. Dachtera" <djesys.nospam@fsi.net> * Subject: Re: SuSE Linux Desupports ALPHA ?' Message-ID: <3BB280D1.A6A3FC81@fsi.net>-   Wilko Bulte wrote: > \ > In <w53ofnyou93.fsf@crail.spa.umn.edu> Graham Allan <allan@physics.umn.edu.nospam> writes: > 6 > >bill@triangle.cs.uofs.edu (Bill Gunshannon) writes: > 5 > >> In article <3BB12FC4.9EF82881@yahoo.commercial>,M9 > >>  Ed Wensell III <ewensell3@yahoo.commercial> writes:e > >> |> $ > >> |> You want commercial support? > >> |>t > >> |> http://www.redhat.comr$ > >> |> http://www.wasabisystems.com > >> |>eN > >> |> I realize the second one is NetBSD not Linux... But it boots Alpha andL > >> |> VAX right from the CDROM.... Among many other platforms. Support and' > >> |> development services available.S > >>I > >> Actually, I have yet to find any Linux CD (including the REDHat fromKI > >> CETS) that will boot right out of the box on my Alpha.  FreeBSD doeso* > >> and installs flawlessly, including X. > N > >Does the regular FreeBSD CD set contain the Alpha version as well as Intel?G > >That would be nice... I've been meaning to take a look at FreeBSD on J > >Alpha but haven't had much time so far... and the FreeBSD web site is a( > >bit quiet about Alpha-specific stuff. > H > There is a seperate 4-CD FreeBSD/alpha CD set. I'd suggest downloading9 > the FreeBSD 4.4 install .iso image and give it a spin..   G I'd love to ... as soon as someone provides wireless broadband out heretE in the boonies (no cable internet, too far from teh CO so no DSL, tooeH far from Sears tower and/or no line of sight for either Sprint BroadBandF Direct or GL National, IOLS isn't here yet). Funny - the "frontier" of= suburban sprawl is already a good 60 miles west of here; yet,S) technologically, we're still in the Styx.a   -- K David J. Dachtera] dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/g   ------------------------------    Date: 26 Sep 2001 07:32:08 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen)e& Subject: Re: UNIX Daemons -> OpenVMS ?3 Message-ID: <zyu6lRbVz+l4@eisner.encompasserve.org>:  [ In article <9osct0$8vr$1@neptunium.btinternet.com>, "Gary" <someone@btinternet.com> writes:B  L > What I didn't want to have to do, was get the child to run another copy ofJ > the application code, but if that is the only way (i.e. exec the APP.EXEK > from the child) is there a way to tell in the new app that it was startedaJ > from the old one, rather than having to try and get round this by having% > some command line param to tell it?I  = You can check to see whether it is a detached process or not.a  B But why is it necessary to have the "parent" ?  Why not just start it as detached ?   ------------------------------  % Date: Wed, 26 Sep 2001 12:00:40 +0100;% From: "Gary" <someone@btinternet.com>d" Subject: UNIX Daemons -> OpenVMS ?3 Message-ID: <9osct0$8vr$1@neptunium.btinternet.com>s  	 Hi there,y  G I am porting an application to OpenVMS from UNIX.  It runs on UNIX as a L daemon, so when you start it up it forks, the parent closes and the child is left running in the background.o  H I have seen some talk about the differences between the UNIX and the VMSL fork (->vfork) calls.  How do people generally accomplish the daemonizing ofE an app on OpenVMS?  Is there some way of using vfork to start a childlI process, but then making it continue running in the process it was forked  from, whilst it's parent dies?  J What I didn't want to have to do, was get the child to run another copy ofH the application code, but if that is the only way (i.e. exec the APP.EXEI from the child) is there a way to tell in the new app that it was started H from the old one, rather than having to try and get round this by having# some command line param to tell it?    Cheers   Gary   ------------------------------  % Date: Wed, 26 Sep 2001 09:10:42 -0400l0 From: "Syltrem" <syltrem@videotron.spammenot.ca>& Subject: Re: UNIX Daemons -> OpenVMS ?5 Message-ID: <Rpks7.62656$TW.340677@tor-nn1.netcom.ca>    $ run/detached yourprogram.exe   And if the program needs DCL,J create yourprogram.com $ run yourprogram.exeM ^Z= $ run/detached sys$system:loginout.exe /input=yourprogram.coma) /log=yourprogram.log /err=yourprogram.logn  H No need to fork and stuff. Just start the program as a detached process.   --   Syltrem I http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)t> To reply to myself directly, remove .spammenot from my address  = "Gary" <someone@btinternet.com> a crit dans le message news:a( 9osct0$8vr$1@neptunium.btinternet.com... > Hi there,i >/I > I am porting an application to OpenVMS from UNIX.  It runs on UNIX as a K > daemon, so when you start it up it forks, the parent closes and the childa is! > left running in the background.c >eJ > I have seen some talk about the differences between the UNIX and the VMSK > fork (->vfork) calls.  How do people generally accomplish the daemonizingl ofG > an app on OpenVMS?  Is there some way of using vfork to start a childaK > process, but then making it continue running in the process it was forked-  > from, whilst it's parent dies? >uL > What I didn't want to have to do, was get the child to run another copy ofJ > the application code, but if that is the only way (i.e. exec the APP.EXEK > from the child) is there a way to tell in the new app that it was started,J > from the old one, rather than having to try and get round this by having% > some command line param to tell it?  >c > Cheers >o > Gary >r >.   ------------------------------    Date: 26 Sep 2001 07:59:34 -0500- From: koehler@encompasserve.org (Bob Koehler)m, Subject: Re: VAX 7000 hw support going away?3 Message-ID: <SbN6jkkvrJ+r@eisner.encompasserve.org>R  ` In article <ot53rtgd1ig7hsorqhk4gdhc00crto1qel@4ax.com>, Alan Greig <a.greig@virgin.net> writes:  G > And yes Compaq should restart production of anything it takes to keephG > equipment supported for a minimum of five years after last ship date.dH > That's what they promised after all. Ford still manufactures new spare1 > parts for cars that haven't sold for ten years.M  H    Ford doesn't manufacture many of thier own parts, and those who do doF    so because there is a market.  What market could you have predicted    for VAX 7000 spare parts?   ------------------------------  % Date: Wed, 26 Sep 2001 14:38:50 -0400i- From: "Peter Weaver" <peter.weaver@stelco.ca>a, Subject: Re: VAX 7000 hw support going away?2 Message-ID: <chps7.32288$Z2.462670@nnrp1.uunet.ca>  2 "Alan Greig" <a.greig@virgin.net> wrote in message2 news:ot53rtgd1ig7hsorqhk4gdhc00crto1qel@4ax.com...4 > On Tue, 25 Sep 2001 17:41:34 GMT, "Mark D. Jilson"# > <jilly@clarityconnect.com> wrote:- >- >...G > Long before September 11th Compaq had a *large* list of products withmE > internal retirement dates already set, I kicked up a fuss about the2> > Storageworks desupport and suggested that all customers withG > non-current hardware on maint, call Compaq and ask them the followingsF > question: "Will you guarantee that no date for de-support of XYZ has: > been set internally within Compaq" and see what happens. >...  4 This is what Compaq Canada said about HSZ70 support;G # The HSZ70 will be at it's end of life on the 30th of January 2002 fordK # i t's software. We will still be issuing patches for it and will continuevJ # to support it as a best effort. The product itself is not at it's end ofH # life and hardware controllers are still available. We will continue to< # provide telephone support through your contract agreement.   ------------------------------  % Date: Wed, 26 Sep 2001 14:57:33 -0500 1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>5, Subject: Re: VAX 7000 hw support going away?8 Message-ID: <9otbv7$69b$1@fizban.fizban.pprd.abbott.com>  L Just got a call from Compaq Services.  We can renew our hw service contracts for the VAX 7000.m   Dave...-  < "Dave Gudewicz" <david.gudewicz@abbott.com> wrote in message2 news:9oq6ot$kpc$1@fizban.fizban.pprd.abbott.com...K > Wondered if anyone here has received news that their hw service contractse= > for VAX 7000 will not be renewed after current ones expire.e >xI > We have 2 such systems, one contract expires at the end of 2001 and thet > other in 2002. > E > We were "told" that yesterday.  I haven't seen anything in writing.N >_1 > I'm wondering if this was some sort of mistake.' >d >v   ------------------------------  % Date: Tue, 25 Sep 2001 12:14:54 -0500p1 From: "Dave Gudewicz" <david.gudewicz@abbott.com>e, Subject: Re: VAX 7000 hw support going away?8 Message-ID: <9oqe27$loo$1@fizban.fizban.pprd.abbott.com>  L If what you state is true, then I would like Compaq to advise customers well1 in advance of the unavailability of "pixie dust".   0 3 months notice is not what I call "ample time".  I And I'll restate that perhaps we missed some communication on this issue.m9 That's our fault.  If we didn't, then its Compaq's fault.e   Dave...t  < "Mark D. Jilson" <jilly@clarityconnect.com> wrote in message, news:3BB0AC80.1EE72E67@clarityconnect.com...G > Given that the chips needed for some/most/all of the modules in theseeH > systems have not been produced for years and that supplies of them areF > close to or indeed non-existent and that the cost to restart fabs toJ > produce those chips would cause maintenance support contracts to requireB > pricing that no one would buy, what would you like Compaq to do?5 > Manufacture the necessary spares out of pixie dust?. >  > Dave Gudewicz wrote: > > C > > Wondered if anyone here has received news that their hw service2	 contracts2? > > for VAX 7000 will not be renewed after current ones expire.: > >SK > > We have 2 such systems, one contract expires at the end of 2001 and the  > > other in 2002. > >2G > > We were "told" that yesterday.  I haven't seen anything in writing.e > >>3 > > I'm wondering if this was some sort of mistake.w >e > --F > Jilly - Working from Home in the Chemung River Valley - Lockwood, NY/ > - jilly@clarityconnect.com - Brett Bodine fan>- > - Mark.Jilson@Compaq.com - since 1975 or sop- > - http://www.jilly.baka.com               -    ------------------------------  # Date: Tue, 25 Sep 2001 16:09:24 GMTD1 From: "Mark D. Jilson" <jilly@clarityconnect.com> , Subject: Re: VAX 7000 hw support going away?2 Message-ID: <3BB0AC80.1EE72E67@clarityconnect.com>  E Given that the chips needed for some/most/all of the modules in theseaF systems have not been produced for years and that supplies of them areD close to or indeed non-existent and that the cost to restart fabs toH produce those chips would cause maintenance support contracts to requireA pricing that no one would buy, what would you like Compaq to do? :3 Manufacture the necessary spares out of pixie dust?f   Dave Gudewicz wrote: > K > Wondered if anyone here has received news that their hw service contractsB= > for VAX 7000 will not be renewed after current ones expire.3 > I > We have 2 such systems, one contract expires at the end of 2001 and the  > other in 2002. > E > We were "told" that yesterday.  I haven't seen anything in writing.n > 1 > I'm wondering if this was some sort of mistake.    -- rD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  # Date: Tue, 25 Sep 2001 17:41:34 GMTg1 From: "Mark D. Jilson" <jilly@clarityconnect.com> , Subject: Re: VAX 7000 hw support going away?2 Message-ID: <3BB0C21A.357E1B39@clarityconnect.com>  D Additionally I believe the events on Sept. 11 seriously effected theE worldwide availability of spares for everyone for the VAX 7000 series-E system and that without the Sept. 11 events one would have seen a bithG longer of a notice.  AFAIK from messages posted here and other places a-H large number of VAX 7000 systems were destroyed on Sept. 11 which I hopeB you would give Compaq a bit of slack for not forseeing this event.   Dave Gudewicz wrote: > N > If what you state is true, then I would like Compaq to advise customers well3 > in advance of the unavailability of "pixie dust".o > 2 > 3 months notice is not what I call "ample time". > K > And I'll restate that perhaps we missed some communication on this issue.:; > That's our fault.  If we didn't, then its Compaq's fault.d > 	 > Dave...f > > > "Mark D. Jilson" <jilly@clarityconnect.com> wrote in message. > news:3BB0AC80.1EE72E67@clarityconnect.com...I > > Given that the chips needed for some/most/all of the modules in thesenJ > > systems have not been produced for years and that supplies of them areH > > close to or indeed non-existent and that the cost to restart fabs toL > > produce those chips would cause maintenance support contracts to requireD > > pricing that no one would buy, what would you like Compaq to do?7 > > Manufacture the necessary spares out of pixie dust?  > >w > > Dave Gudewicz wrote: > > > E > > > Wondered if anyone here has received news that their hw servicen > contracts|A > > > for VAX 7000 will not be renewed after current ones expire.e > > >|M > > > We have 2 such systems, one contract expires at the end of 2001 and then > > > other in 2002. > > >nI > > > We were "told" that yesterday.  I haven't seen anything in writing.S > > > 5 > > > I'm wondering if this was some sort of mistake.. > >j > > --H > > Jilly - Working from Home in the Chemung River Valley - Lockwood, NY1 > > - jilly@clarityconnect.com - Brett Bodine fan./ > > - Mark.Jilson@Compaq.com - since 1975 or soa/ > > - http://www.jilly.baka.com               -n   -- )D Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  % Date: Tue, 25 Sep 2001 14:28:58 -0400D5 From: David Beatty <David.Beatty@qwertysasasdfgh.com>r, Subject: Re: VAX 7000 hw support going away?2 Message-ID: <2cywO9cSx4PEANBsrguRhC+gFu6=@4ax.com>  ) Which model do you have?  David R. Beattyu  3 On Tue, 25 Sep 2001 10:10:33 -0500, "Dave Gudewicz"o" <david.gudewicz@abbott.com> wrote:  J >Wondered if anyone here has received news that their hw service contracts< >for VAX 7000 will not be renewed after current ones expire. > H >We have 2 such systems, one contract expires at the end of 2001 and the >other in 2002.c >iD >We were "told" that yesterday.  I haven't seen anything in writing. >i0 >I'm wondering if this was some sort of mistake. >    ------------------------------  % Date: Tue, 25 Sep 2001 19:56:18 +0100 + From: "antonio.carlini" <arcarlini@iee.org>	, Subject: Re: VAX 7000 hw support going away?' Message-ID: <3BB0D352.6EA07CB7@iee.org>-   "Mark D. Jilson" wrote:  > G > Given that the chips needed for some/most/all of the modules in theseuH > systems have not been produced for years and that supplies of them areF > close to or indeed non-existent and that the cost to restart fabs toJ > produce those chips would cause maintenance support contracts to requireB > pricing that no one would buy, what would you like Compaq to do?5 > Manufacture the necessary spares out of pixie dust?a  + When I was with the DECnis group and we didt+ our "end of life" buy of NVAX parts for ther$ MPC3, one of the things we had to do* was make allowances for future maintenance) requirements. EOSL (End Of Service Life)  - was supposed to be at least five years beyondi. end of production. I was not directly involved+ with all of this but I was on the periphery + and I was given to understand that this was  normal practice within Digital.S  % It may be that COMPAQ just do things h& differently and don't quite understand! that these boxes and Wintel boxesf don't work the same way.  # Or it could be that the emails I'ves  seen forwarded to various lists   supposedly from COMPAQ personnel  looking for VAX 7800s to replace the 70-odd allegedly wiped out two weeks ago were real and as a result the various stockrooms/ have just be drained.     I'd prefer to believe the latter" (something noone could reasonably " allow for) rather than the former.   Antonioa    e --     ---------------0- Antonio Carlini             arcarlini@iee.orgf   ------------------------------  % Date: Wed, 26 Sep 2001 22:18:59 +0100.% From: Alan Greig <a.greig@virgin.net>:, Subject: Re: VAX 7000 hw support going away?* Message-ID: <3BB24643.B66DBAB2@virgin.net>   Dave Gudewicz wrote:  N > Just got a call from Compaq Services.  We can renew our hw service contracts > for the VAX 7000.M >   N Good to hear. Hopefully Compaq might review some of the other retirals already communicated to customers. --
 Alan Greig   ------------------------------  % Date: Wed, 26 Sep 2001 10:03:27 +0100|% From: Alan Greig <a.greig@virgin.net>t, Subject: Re: VAX 7000 hw support going away?8 Message-ID: <ot53rtgd1ig7hsorqhk4gdhc00crto1qel@4ax.com>  2 On Tue, 25 Sep 2001 17:41:34 GMT, "Mark D. Jilson"! <jilly@clarityconnect.com> wrote:   E >Additionally I believe the events on Sept. 11 seriously effected the F >worldwide availability of spares for everyone for the VAX 7000 seriesF >system and that without the Sept. 11 events one would have seen a bitH >longer of a notice.  AFAIK from messages posted here and other places aI >large number of VAX 7000 systems were destroyed on Sept. 11 which I hopedC >you would give Compaq a bit of slack for not forseeing this event.   E Long before September 11th Compaq had a *large* list of products withiC internal retirement dates already set, I kicked up a fuss about the < Storageworks desupport and suggested that all customers withE non-current hardware on maint, call Compaq and ask them the followingLD question: "Will you guarantee that no date for de-support of XYZ has8 been set internally within Compaq" and see what happens.  E And yes Compaq should restart production of anything it takes to keep E equipment supported for a minimum of five years after last ship date.iF That's what they promised after all. Ford still manufactures new spare/ parts for cars that haven't sold for ten years.o  > But some people still haven't quite got it yet. Compaq have noE intention of keeping *any* promise they make. That;s not the way theycA work and why should the bother when every time they break anothereC promise apologists pop up and say it was outwith their control or ao8 special case so we should believe their future promises.     -- Alan   ------------------------------  % Date: Wed, 26 Sep 2001 08:29:32 -0400e5 From: David Beatty <David.Beatty@qwertysasasdfgh.com> , Subject: Re: VAX 7000 hw support going away?2 Message-ID: <9MmxO7tjCtixWGDJ8t8v6sCu=Bwb@4ax.com>  8 Thanks, I was curious.  We have a VAX 7860 here and have5 not heard anything regarding the dropping of support.e   David R. Beattyt  3 On Tue, 25 Sep 2001 15:40:36 -0500, "Dave Gudewicz"v" <david.gudewicz@abbott.com> wrote:  
 >VAX 7000-630? > C >"David Beatty" <David.Beatty@qwertysasasdfgh.com> wrote in message0- >news:2cywO9cSx4PEANBsrguRhC+gFu6=@4ax.com...j >>, >> Which model do you have?  David R. Beatty >>6 >> On Tue, 25 Sep 2001 10:10:33 -0500, "Dave Gudewicz"% >> <david.gudewicz@abbott.com> wrote:X >>M >> >Wondered if anyone here has received news that their hw service contractse? >> >for VAX 7000 will not be renewed after current ones expire.e >> >K >> >We have 2 such systems, one contract expires at the end of 2001 and thes >> >other in 2002. >> >G >> >We were "told" that yesterday.  I haven't seen anything in writing.u >> >3 >> >I'm wondering if this was some sort of mistake.o >> > >> >g   ------------------------------  % Date: Wed, 26 Sep 2001 14:51:02 +0100d% From: Alan Greig <a.greig@virgin.net>l, Subject: Re: VAX 7000 hw support going away?8 Message-ID: <lim3rtg4346hs3d45nggjjlt5e0h3di4b4@4ax.com>  F On 26 Sep 2001 07:59:34 -0500, koehler@encompasserve.org (Bob Koehler) wrote:  a >In article <ot53rtgd1ig7hsorqhk4gdhc00crto1qel@4ax.com>, Alan Greig <a.greig@virgin.net> writes:  >hH >> And yes Compaq should restart production of anything it takes to keepH >> equipment supported for a minimum of five years after last ship date.I >> That's what they promised after all. Ford still manufactures new sparen2 >> parts for cars that haven't sold for ten years. >tI >   Ford doesn't manufacture many of thier own parts, and those who do do G >   so because there is a market.  What market could you have predictedp >   for VAX 7000 spare parts?n  @ Whether Ford manufacture or subcontract doesn't really matter...  D It is not a question of what I could have predicted it is a questionB of what Compaq promised. One of the supposed advantages of doing aE deal with DEC/Compaq over a PC supplier was the 'enterprise' culture.aF You knew whatever necessary would be done to meet a commitment. CompaqE shut down VAX manufacture too early in my opinion. The WTC might haveo= speeded up the problem but, as I said, a list of products andaD desupport dates existed within Compaq prior to 11th September. I wasF given information in confidence for our company's internal use only soD I cannot say more other than I would advise *everyone* with criticalD non-current hardware to ask Compaq for an assurance that they expectA to continue support for a minimum of five years after last ship -g# assuming that is what you expected.m  C During the DEC-10/DEC-20 phase out DEC handed over a lot of cash to D purchase these machines back used to maintain their stock of spares.F We were offered virtually nothing for a VAX 7000 series machine just aF year ago. Surely someone must have noticed spares supplies dropping by then?a -- Alan   ------------------------------  % Date: Wed, 26 Sep 2001 20:42:36 +0200 < From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <noone@home.com>, Subject: Re: VAX 7000 hw support going away?( Message-ID: <3BB2219C.DDE425DB@home.com>  : How can the hardware *not* be at it's end-of-life, but the8 software (that's needed to use the hardware, not ?) be ?  & I must have missunderstod something...   Jan-Erik Sderholm.i   Peter Weaver wrote:  > 6 > This is what Compaq Canada said about HSZ70 support;I > # The HSZ70 will be at it's end of life on the 30th of January 2002 foraM > # i t's software. We will still be issuing patches for it and will continue L > # to support it as a best effort. The product itself is not at it's end ofJ > # life and hardware controllers are still available. We will continue to> > # provide telephone support through your contract agreement.   ------------------------------  % Date: Wed, 26 Sep 2001 20:11:29 +01008% From: Alan Greig <a.greig@virgin.net>., Subject: Re: VAX 7000 hw support going away?) Message-ID: <3BB22861.A3BF693@virgin.net>8   Peter Weaver wrote:c   > "A > 6 > This is what Compaq Canada said about HSZ70 support;I > # The HSZ70 will be at it's end of life on the 30th of January 2002 forTM > # i t's software. We will still be issuing patches for it and will continuer  N How do you manage to continue to issue patches for a product beyond the end ofM software support? That just doesn't make sense and does not match what I haved% been told - which is no more patches.o   >rL > # to support it as a best effort. The product itself is not at it's end ofJ > # life and hardware controllers are still available. We will continue to> > # provide telephone support through your contract agreement.  O I have been assured in writing that there will be no software telephone supportt after 31-JAN-20022  H I strongly suggest anyone with HSZ80s ask the same question about end of% software support for that controller.f   --
 Alan Greig   ------------------------------  % Date: Tue, 25 Sep 2001 12:42:09 -0400i5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> # Subject: Re: VAX disaster tolerance-1 Message-ID: <Ot2s7.614$YP.20612@news.cpqcorp.net>0    Marty Kuhrt wrote in message ... >zF >This was one of the many reasons we changed hardware support vendors.G >The one we are currently using has one of each type of machine we havetE >under contract _running_ at their location.  That way when somethingoF >breaks they have a hot spare that was working when they pulled it out >to bring it to our location.o  I And this helps how, when two customers both have their systems destroyed,vI and each need that backup hardware?  Which also leaves the requirement tod( purchase the replacement for the backup.  J At some point, with old hardware, it all becomes impossible to do - unlessH you set it up with money-is-no-object and bought enough spares up front.   ------------------------------  % Date: Tue, 25 Sep 2001 12:58:16 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>4# Subject: Re: VAX disaster tolerancev$ Message-ID: <3bb0b7bc$1@news.si.com>  / >Using a "legacy"  platform as Mission Critical2, >is not what we can say - a safe investment. > . >It is why I ask people here why do they still >using VAXes...e  J How about "because the mission-critical apps we need don't run on an AlphaG or PC"?  How about "it would cost millions of dollars to get the FAA too? recertify the software we have if we were to port it to another, architecture"? --A Brian Tillman                   Internet: tillman_brian at si.comrA Smiths Aerospace                          tillman at swdev.si.come= 3290 Patterson Ave. SE, MS      Addresses modified to prevents< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------    Date: 25 Sep 2001 12:24:18 -0500+ From: kuhrt@encompasserve.org (Marty Kuhrt) # Subject: Re: VAX disaster tolerance 3 Message-ID: <Brx3$Fx5zqRL@eisner.encompasserve.org>s  i In article <Ot2s7.614$YP.20612@news.cpqcorp.net>, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> writes:e" > Marty Kuhrt wrote in message ... >>G >>This was one of the many reasons we changed hardware support vendors. H >>The one we are currently using has one of each type of machine we haveF >>under contract _running_ at their location.  That way when somethingG >>breaks they have a hot spare that was working when they pulled it oute >>to bring it to our location. > K > And this helps how, when two customers both have their systems destroyed,mK > and each need that backup hardware?  Which also leaves the requirement top* > purchase the replacement for the backup. >   B They don't intermingle customers hardware.  It is not, as you are E implying, that they have one 4100 running for all customers that havew 4100s.   ------------------------------  % Date: Tue, 25 Sep 2001 12:47:35 -0400h5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> # Subject: Re: VAX disaster toleranced1 Message-ID: <Ty2s7.615$YP.20624@news.cpqcorp.net>o  E Your problem does not seem to be the cost to "port to Alpha", but the G question is - what happens if something stops working?  Are you forevernD dependent on code that can't be reproduced?  That would make me very nervous.      $ Andrew Robinson wrote in message ...J >We use VAXes in a Business Critical environment - We can't modify some ofG >the legacy code without having to reinvent the wheel, due to a portioneI >having been written in-house without proper control over source code (Wen are  >not unusual in that one!).wK >The cost of porting to Alpha's, when the licenses, programming & man hoursoJ >are involved, just aren't justifiable. We don't need the processing power ofK >Alpha's, so what is our motive to spend X amount for no business benefit ?t >o >Regards >M >Andrew Robinson >	 >-----Original Message-----:6 >From: Fabio Cardoso [mailto:fabiopenvms@yahoo.com.br] >Sent: 24 September 2001 13:56 >To: Info-VAX@Mvb.Saic.Com$ >Subject: Re: VAX disaster tolerance >c > / >Using a "legacy"  platform as Mission Criticale, >is not what we can say - a safe investment. > . >It is why I ask people here why do they still >using VAXes...d >e. >In my opinion the upgrades should be the best >choice for these old systems. >d4 >How can I rebuild a cluster after a disaster if the1 >boards are not available anymore in the market ?s! >Why people still using PDPs ????  >b2 >It is a non responsable act inside the companies. >  >o >Regards >  >FCC >a ><SNIP>f   ------------------------------  % Date: Tue, 25 Sep 2001 20:01:34 +0100 + From: "antonio.carlini" <arcarlini@iee.org>o# Subject: Re: VAX disaster toleranceo' Message-ID: <3BB0D48E.97C09F48@iee.org>    JF Mezei wrote: H > That is the problem isn't it ? You don't want to waste money replacingN > hardware that still does the job just for the sake of having the latest. ButM > then, you also don't want to get a rude awakening because you can no longernO > get replacement machines after a disaster even though you didn't realise thati4 > problem as your configuration was still supported.   Exactly. These days you don'tn need an identical replacementt (unless you are the military ors  run a nuclear power plant etc.).   Most businesses could replace  an Alpha with a newer Alpha -s# there are not that many specialisedq interfaces any more.   However, when you are running % on an architecture that can no longer $ be produced economically (e.g. VAX -" I won't include PDP-11 since there  are compatibles available!) then! you need to at least have a plan!v   Antoniot   --     ---------------w- Antonio Carlini             arcarlini@iee.org    ------------------------------  % Date: Tue, 25 Sep 2001 15:01:10 -05000+ From: "Mike Kier" <michael.kier@compaq.com> # Subject: Re: VAX disaster tolerancea1 Message-ID: <do5s7.623$YP.20931@news.cpqcorp.net>   H Amazing how your name jumps off the screen when you see it in unexpected places :-)  Thanks, Larry.  L Ensuring they had maintainable systems was, indeed, the biggest factor, bothJ hardware (PDPs, DT07 bus switches, synch and asynch multidrop DDCMP, etc.,? as well as 6504 and 8080 ROM-based toll terminals) and softwaree( (spaghetti-coded Fortran 68 on RSX-11M).  J There were other considerations as well, as we've eliminated the magstripeI cards (landfill problem) and added in touch-screen web interfaces on CSWS-F (human maintenance - easier to find people familiar with browsers thanI terminals these days).  The neat part was we did the entire switch-out ofoJ the old technology without every having the entire application down (yeah,$ we wrote a *lot* of throwaway code).  I Now we can pretty much plug in anything that works with TCP/IP and/or BEAtH MessageQ.  In fact we will probably replace all of the Alphaserver 2100sJ with DS10s and a couple of ES40s and fibre-based storage very soon just to6 stay ahead of any future replacement part bottlenecks. --	 Mike Kier1 Compaq Professional Services Cincinnati, OH, USAt michael.kier@compaq.come# Caveat pagina non mutatio quantitase  : "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message- news:VV7qPvFTZnnH@eisner.encompasserve.org...1; > In article <3BB0D48E.97C09F48@iee.org>, "antonio.carlini"c <arcarlini@iee.org> writes:i > >a > >e > > JF Mezei wrote:EK > >> That is the problem isn't it ? You don't want to waste money replacingeE > >> hardware that still does the job just for the sake of having ther latest. ButaI > >> then, you also don't want to get a rude awakening because you can not longerE > >> get replacement machines after a disaster even though you didn'tp realise that7 > >> problem as your configuration was still supported.o > >e! > > Exactly. These days you don'te! > > need an identical replacement # > > (unless you are the military ord$ > > run a nuclear power plant etc.). > >t! > > Most businesses could replace ! > > an Alpha with a newer Alpha -o' > > there are not that many specialised( > > interfaces any more. > >p! > > However, when you are runningn) > > on an architecture that can no longern( > > be produced economically (e.g. VAX -& > > I won't include PDP-11 since there$ > > are compatibles available!) then% > > you need to at least have a plan!u >dC > Mike Kier from DEQ, who upgraded the Indiana Turnpike from PDP-11aD > to Alpha said the major reason was spare parts availability, since> > their "process control" at the toll gates didn't really needA > any more speed or throughput.  Moore's law applies to computersg  > but not to turnpike speeds :-)   ------------------------------  + Date: Wed, 26 Sep 2001 11:29:50 +0000 (UTC)f' From: david20@alpha1.mdx.ac.uk (D.Webb)e# Subject: Re: VAX disaster tolerance + Message-ID: <9ose7a$847$1@aquila.mdx.ac.uk>-  U In article <3BB0D48E.97C09F48@iee.org>, "antonio.carlini" <arcarlini@iee.org> writes:- >- >  >JF Mezei wrote:I >> That is the problem isn't it ? You don't want to waste money replacing:O >> hardware that still does the job just for the sake of having the latest. ButaN >> then, you also don't want to get a rude awakening because you can no longerP >> get replacement machines after a disaster even though you didn't realise that5 >> problem as your configuration was still supported.c >d >Exactly. These days you don't >need an identical replacement  >(unless you are the military or! >run a nuclear power plant etc.).@ >g >Most businesses could replace w >an Alpha with a newer Alpha -$ >there are not that many specialised >interfaces any more.S >2   Not entirely true.0 If your software requires an old version of VMS.N Eg we have Alpha systems running VMS 6.2-1H3 because the cost of rewriting the= application to support a later version of Oracle is too high.sL The Oracle version restricts us to VMS 6.2-1H3 which in turn restricts what 5 models of Alpha could possibly replace these systems.t    
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  # Date: Wed, 26 Sep 2001 21:56:43 GMTn9 From: rickm123@oregon456.uoregon789.edu (Rick Millhollin)l Subject: VMS V7.2-2 vs V7.3 2 Message-ID: <3bb24e89.799537414@news2.uoregon.edu>  C I just saw some info on VMS V7.2-2 that made it sound brank new.  IrC received V7.3 a while back.  I'm confused - what's the deal?  UnderoE what circumstances would I upgrade to V7.2-2 instead of V7.3?  ThanksX for any insight...   ------------------------------  % Date: Wed, 26 Sep 2001 20:53:59 -0400r2 From: norm lastovica <norman.lastovica@oracle.com> Subject: Re: VMS V7.2-2 vs V7.3 * Message-ID: <3BB278A7.1F597B31@oracle.com>  > AFAIK, 7.2-2 is a rollup of existing patches (bug fixes) alongD with support for some new hardware and some performance enhancements@ (aimed at larger SPM environments).  I'd go with V7.3 if I could> (to get all the latest stuff) and V7.2-2 if I could not.  Some> people are more comfortable, I guess, with smaller incrementalD upgrades and fixes.  I believe that support for 7.2-1 (and variants); will be ending before too long now that 7.2-2 is out there.-   Rick Millhollin wrote: > E > I just saw some info on VMS V7.2-2 that made it sound brank new.  IyE > received V7.3 a while back.  I'm confused - what's the deal?  UndereG > what circumstances would I upgrade to V7.2-2 instead of V7.3?  Thanks. > for any insight...   --  > norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  % Date: Wed, 26 Sep 2001 21:00:39 -0400d2 From: norm lastovica <norman.lastovica@oracle.com> Subject: Re: VMS V7.2-2 vs V7.3@* Message-ID: <3BB27A37.7C83A849@oracle.com>   Rick Millhollin wrote:E > I just saw some info on VMS V7.2-2 that made it sound brank new.  I E > received V7.3 a while back.  I'm confused - what's the deal?  UnderiG > what circumstances would I upgrade to V7.2-2 instead of V7.3?  Thanks  > for any insight...  	 checkout:>  E http://www.openvms.compaq.com:8000/722final/6650/6650pro.html#preface>  , 1.1 Overview of OpenVMS Alpha Version 7.2--21 OpenVMS Alpha Version 7.2--2 has been designed as 1 a follow-on release to replace both OpenVMS Alphaa3 Version 7.2--1 and OpenVMS Alpha Version 7.2--1H1.  / The latter was a limited hardware release that -1 introduced support for the Compaq AlphaServer GS 1. series systems. OpenVMS Version 7.2--2 is the / version Compaq recommends for coexistence with a, OpenVMS Version 7.3 in mixed-version OpenVMS Cluster systems. n  . OpenVMS Alpha Version 7.2--2 includes all the 0 capabilities of its predecessors and introduces 5 new features and hardware support. Some new features i6 have already been released in update kits for certain 1 software modules and system integrated products. u/ Other new features have only been available in r OpenVMS Alpha Version 7.3. h  / The major new features and hardware support of  % OpenVMS Alpha Version 7.2--2 follow: n   - SMP performance enhancements d/ - AlphaServer GS80/160/320 systems support for  > using AlphaServer EV68 CPUs and EV67 CPUs in the same system. 2 - Minicopy support in host-based volume shadowing  - Fibre Channel tape support uC - Compatibility with the OpenVMS Alpha Version 7.3 Registry Server  , - OpenVMS e-Business Infrastructure Package / - New SDA commands, qualifiers, and parameters n  - New system dump documentation  - LAN ATM adapter support < - Disaster-tolerant support for host-based volume shadowing G in an OpenVMS Cluster configuration using shared Fibre Channel storage  6 - Remedial fixes for OpenVMS Alpha Version 7.2--1 and # Version 7.2-1H1 through April 2001     -- e> norman lastovica / oracle rdb engineering / usa / 610.696.4685   ------------------------------  % Date: Wed, 26 Sep 2001 23:21:24 -0400 ' From: Howard S Shubs <howard@shubs.net>s Subject: Re: VMS V7.2-2 vs V7.3u< Message-ID: <howard-C8C517.23212126092001@enews.newsguy.com>  * In article <3BB27A37.7C83A849@oracle.com>,4  norm lastovica <norman.lastovica@oracle.com> wrote:  0 > OpenVMS Alpha Version 7.2--2 includes all the 2 > capabilities of its predecessors and introduces 7 > new features and hardware support. Some new features e8 > have already been released in update kits for certain 3 > software modules and system integrated products. u1 > Other new features have only been available in - > OpenVMS Alpha Version 7.3. ,  L Also, I was able to get 7.2-2 to upgrade 7.2-1, whereas 7.3 failed to do so O successfully, no matter what I tried.  I was also able to upgrade a GS80 and a D: cluster of GS160s, w/o trouble.  Can't say that about 7.3. -- l Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------    Date: 25 Sep 2001 22:04:06 -0500+ From: young_r@encompasserve.org (Rob Young)b> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <5m111pxGsoth@eisner.encompasserve.org>e  a In article <4phK+ha$Nnld@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:d  > > 	From a vendor I received NT cards 3 times from two separateC > 	sources... check and recheck your HBA's... trust no one on that:p > F > http://www.compaq.com/products/storageworks/adapters/KGPSA-firm.html > G > 	You want KGPSA-CX... I believe the KGPSA-CAs are older now.  Upgrade < > 	your firmware on the cards, they are no doubt out of rev. >   = 	I have to be careful here... CX may only be a firmware thingsB 	versus CA.  I have a foggy recollection there.  But I do remember@ 	thinking: "CX must be newer" and somehow discovering it doesn't> 	mean that.  But then again, I may be all turned around again.   				Robo   ------------------------------  % Date: Wed, 26 Sep 2001 10:23:52 -0600f$ From: John.Ottosen@fairchildsemi.com? Subject: RE:  VMScluster using fibrechannel and memory channel?nG Message-ID: <OF5AC42174.FA809DBD-ON87256AD3.0059B630@fairchildsemi.com>t  H Yes, we have gone thru that migration.  We had a CI adapter installed inH our ES42's, which really helped.  I tried sending to your email address, but got a delivery failure.e  % John (john.ottosen@fairchildsemi.com)o  H >Is any one currently running a production VMScluster using FibrechannelI >storage and memory channel interconnect? Has any one migrated from CI to I >Fibrechannel / memory channel? Any advice from someone who has done thisu	 >already?0   >Bob Kaplow0   ------------------------------    Date: 26 Sep 2001 07:10:03 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)Y> Subject: Re: VMScluster using fibrechannel and memory channel?= Message-ID: <cf15391e.0109260610.6797d119@posting.google.com>r  f young_r@encompasserve.org (Rob Young) wrote in message news:<4phK+ha$Nnld@eisner.encompasserve.org>... > 	Even though MC has itshF > 	issues, I would still go with it... 100 MBit as Verele? chart showsE > 	does well with lock traffic and Keith points out Gigabit gains yous5 > 	little at this point as it is harder on your CPUs.y  F Uh, Memory Channel actually takes more host CPU than LAN adapters like< Gigabit Ethernet.  (See Verell Boaen's Cluster InterconnectsD presentation from DFW Days 2001 under http://www.dfwcug.org for dataE on both simulated lock throughput and host CPU overhead.)  My concern @ with Gigabit Ethernet was that as Verell's data shows, for smallE packets (like lock requests), the on-board CPU apparently runs out ofu> steam early for the current (first) generation of GbE adapter.C ------------------------------------------------------------------- C Keith Parris | parris at encompasserve dot org | VMS consulting on:eC Clusters, Disaster Tolerance, Internals, Performance, Storage & I/On   ------------------------------    Date: 26 Sep 2001 07:49:07 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)e> Subject: Re: VMScluster using fibrechannel and memory channel?= Message-ID: <cf15391e.0109260649.40446623@posting.google.com>i  E Given that you have 10-CPU SMP boxes, and 4 CI adapters per box, it's E quite possible that you are currently using Fast_Path on CI to spread D CI interrupt workload across non-primary CPUs.  Be aware that Memory@ Channel does not support Fast_Path (and since no new MC hardware= development is planned, and there's thus no incentive for VMS D Engineering to dive back into the driver code, it's unlikely it everC will support Fast_Path).  I worked at a site where we tried to move C from CI to MC for SCS traffic in a large cluster (16 6-CPU GS-140s)kE and our plans failed due to saturation of CPU 0 in interrupt state on.D lock-master nodes for hot application files.  They had to rip the MCD hardware back out and go to 6 CIs instead.  As a test, you might tryE turning off Fast_Path on CI in your current configuration and measureh! peak CPU 0 interrupt-state usage.h  E Considering that you have only two nodes at present, the odds of thislF being a problem are probably actually fairly small.  If you don't haveD CPU 0 saturation problems (as is the case in many smaller clusters),D MC can work very well, with lock latencies 1/4 that of CI, and about) 1/2 that of using a LAN for lock traffic.r  D Since you have 10 CPUs per box now, and plan to go with ES45s, whichE top out at 4 CPUs per box, I assume you'd end up with more individual-B ES45 boxes, say 5 or more.  This has some repercussions.  For one,F more of your lock traffic will be remote; where today perhaps half the@ processes run on the lock master node and get great lock requestE response times (4-6 microseconds local lock request latency).  With 5 B boxes instead of 2, 20% of lock requests would be local instead ofF 50%.  But the remote lock latency would be around 120 microseconds for@ MC versus about 440 microseconds for CI, to compensate somewhat.  D FC itself has Fast_Path support if you go with version 7.3, so you'dD have the benefit of Fast_Path for the I/O operation traffic.  If youA look at LAN adapters as an alternative to MC for lock traffic, be E aware that they do not have Fast_Path support either; we're told this C will definitely not be in 7.3-1, and may not happen until after theiD IA-64 porting work.  And LAN adapters have twice the lock latency of@ Memory Channel (although they are about half the latency of CI).  E Another factor I haven't seen mentioned is shadow merges.  With FibrerF Channel disks, you'll get time-consuming and I/O-intensive full mergesE when a VMS node crashes, not the very-brief mini-merges possible with E HSJ controllers.  Some applications are significantly affected by the @ I/O performance impact of a shadow merge operation.  Support forB mini-merges on HSG80s is under development, but not available yet.  D As another poster pointed out, to get the level of redundancy you're> used to with CI, you'll need dual independent MC and FC rails.  E The maturity level of the FC stuff is not yet at the level we're usedrF to with CI.  I strongly agree with the earlier poster that you'll wantC to go slowly and incrementally with the implementation, doing tests:? and simulating failures along the way, to make sure things workcB correctly.  Starting out by adding FC adapters, switches, and tape= alone on FC is a low-risk path.  I'd also recommend providingdF redundancy between the old and new hardware as much as possible duringC the proving time: for example, shadow HSJ disks with HSG disks, andn2 keep CIs around as a backup to MC for SCS traffic.C -------------------------------------------------------------------rC Keith Parris | parris at encompasserve dot org | VMS consulting on: C Clusters, Disaster Tolerance, Internals, Performance, Storage & I/Os   ------------------------------    Date: 26 Sep 2001 10:52:01 -0500+ From: young_r@encompasserve.org (Rob Young)e> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <3v$7zh09cP$g@eisner.encompasserve.org>t  o In article <ytia1jFr+3zG@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:ek > In article <Pine.LNX.4.31.0109251515210.871-100000@jaipur>, Ryan Moore <rmoore@rmoore.dyndns.org> writes:nD >> We upgrade our CI-based cluster of 8400s to FC and MC.  The extra > L >> Make sure you get redundant MC hubs since MC is an active bus.  We testedM >> MC hub failures on our backup cluster.  The cluster state transition to gocH >> from one MC to the other took less than one second.  Very impressive." >> Even with lots of I/O going on. >  > Everything will be redundant.b >   2 	Why?  But more importantly do you have the slots?E 	What I mean is, I can see the need for duplicate network connectionswD 	as a necessity and that would be your backup for SCS traffic.  TellB 	us what you are thinking (meant to draw you out... not slam you).   				Rob    ------------------------------    Date: 25 Sep 2001 21:58:07 -0500+ From: young_r@encompasserve.org (Rob Young)m> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <4phK+ha$Nnld@eisner.encompasserve.org>e  o In article <F5Xbti9TFkHR@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:oN > We've been running a VMS CI cluster for many years, going from VAX to Alpha.I > The primary nodes are currently an 8400 and a GS140, each with 10 CPUs.pK > We're running 4 CIs with most of the disks and tapes out there, plus some N > direct SCSI devices. Unfortunately, that takes up a lot of space in our data< > center, and we need to significantly reduce the footprint. > J > What we're looking at for a replacement is a rack of ES45s, an EMA12000,L > Fibrechannel disk and tape, and Memory Channel for a cluster Interconnect.L > DLT tapes would be connected via MDRs. We should be able to do all of thisM > in about 25-30% of the space of the existing footprint, counting all of theo > storage arrays.  >   C 	Shame you can't stick a rack somewhere other than that datacenter. A 	If you could, you could easily and fairly cheaply create a fault > 	tolerant cluster (i.e. stay within a single mode fibre run of@ 	say 1500-3000 feet) and run MC over it.  Even though MC has itsD 	issues, I would still go with it... 100 MBit as Verele? chart showsC 	does well with lock traffic and Keith points out Gigabit gains youa3 	little at this point as it is harder on your CPUs.   ? 	I see according to Hoff's slides that 7.2-2 is shipping.  I'velB 	been reading about it lately.  It is a winner.  Shadow mini-copy,9 	hooray!  and a "one off" on performance fix-ups... i.e. o  = http://www.openvms.compaq.com:8000/722final/6650/6650pro.htmll    o  SMP performance enhancements    	which reading here:  L http://www.openvms.compaq.com:8000/722final/6650/6650pro_001.html#nf_v722_ch  = 	Includes improved scheduling, locking, and spinning ... i.e.d) 	less kernel overhead, more for us users.p  K > We're currently at VMS 7.2, but will upgrade to whatever it takes to makecK > this all work. We'll have to migrate to the new configuration a step at a H > time, rather than doing a massive swap, as we can't be down that long. > I > Is any one currently running a production VMScluster using FibrechannelrJ > storage and memory channel interconnect? Has any one migrated from CI toJ > Fibrechannel / memory channel? Any advice from someone who has done this > already?   >   > 	I can answer yes to both those in 2 weeks ;-).  But extensiveB 	testing is going well.  Disclaimer:  not nearly as large as your	 	setup.:  @ 	If I had a bit more confidence in 7.3 , I would push that.. but 	XFC patch is on hold:  Q http://ftp1.support.compaq.com/patches/public/vms/axp/v7.3/vms73_xfc-v0100.READMEe  6 	so it isn't quite ready for prime time in my opinion.  A 	More advice?  How much money is in your budget?  What I would do C 	would be to stick a Fibre card in your existing 8400 and GS140 fore@ 	starters.  I'm assuming you will have at least 2 Brocade/CompaqH 	switches... so plug those 8400 and GS140 in. Bring in the ESA12000 into@ 	the cluster.  Pull one database/volume over from CI to Fibre atA 	a time.  Modest amount of risk now is to bring in an ES45 as thegD 	only way it trades lock traffic is with the network.  However, you B 	may wish to borrow or you may be forunate to have CIPCAs to stick9 	in the ES45s.  Maybe 1 or 2 ES45s are model 10 (ten slot @ 	PCI to allow:  2 Fibre cards, CIPCA, network... but because of G 	multiple ES45s you only are doing one fibre card per.)  This gives youvG 	good old CI to exchange lock traffic and network outages (cough cough 4F 	cough) won't hose up your cluster.  That may be overkill if you have < 	good networking and the ES45<->8400/GS140 is only for a few 	weeks/months anyhow.t  < 	From a vendor I received NT cards 3 times from two separateA 	sources... check and recheck your HBA's... trust no one on that:o  D http://www.compaq.com/products/storageworks/adapters/KGPSA-firm.html  E 	You want KGPSA-CX... I believe the KGPSA-CAs are older now.  Upgradee: 	your firmware on the cards, they are no doubt out of rev.  C 	The fact that you have Alphas makes this migration simple.  Not so.C 	simple is CI VAX to Alpha Fibre.  Tricks there too, though but not(A 	nearly the luxury of moving slowly over maintaining the existingr 	servers for a while.m   				Robr   ------------------------------    Date: 26 Sep 2001 13:14:33 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) > Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <fteh3dZegGg8@eisner.encompasserve.org>   a In article <4phK+ha$Nnld@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:sE > 	Shame you can't stick a rack somewhere other than that datacenter.dC > 	If you could, you could easily and fairly cheaply create a faultc@ > 	tolerant cluster (i.e. stay within a single mode fibre run ofB > 	say 1500-3000 feet) and run MC over it.  Even though MC has its  G I'd love to do that. We do have 2 other data centers. One is 1727 miles = away, the oher 1375. Juat a bit outside the cluster specs :-)a  @ > 	I can answer yes to both those in 2 weeks ;-).  But extensiveD > 	testing is going well.  Disclaimer:  not nearly as large as your		 > 	setup.7  ) Do keep me informed how thigns things go.e  C > 	More advice?  How much money is in your budget?  What I would dotE > 	would be to stick a Fibre card in your existing 8400 and GS140 foroB > 	starters.  I'm assuming you will have at least 2 Brocade/CompaqJ > 	switches... so plug those 8400 and GS140 in. Bring in the ESA12000 intoB > 	the cluster.  Pull one database/volume over from CI to Fibre at > 	a time.    I We may end up going this route anyway, as it now looks like management is-L hesitant to spend the $$$ on the new CPUs. The biggest space reduction comesH from upgrading the IO farm, which goes back as far as HSJ40s with RZ29s.  : > 	Modest amount of risk now is to bring in an ES45 as theF > 	only way it trades lock traffic is with the network.  However, you D > 	may wish to borrow or you may be forunate to have CIPCAs to stick; > 	in the ES45s.  Maybe 1 or 2 ES45s are model 10 (ten slot B > 	PCI to allow:  2 Fibre cards, CIPCA, network... but because of   K If I do get the ES45s, this is a possibility. Iv'e got one spare CIPCA now,yE and could easilly combine 2 of my 4 stars into one, freeing up 2 moret CIPCAs.h  I > 	multiple ES45s you only are doing one fibre card per.)  This gives youaI > 	good old CI to exchange lock traffic and network outages (cough cough  H > 	cough) won't hose up your cluster.  That may be overkill if you have > > 	good networking and the ES45<->8400/GS140 is only for a few > 	weeks/months anyhow.t  I I'd do SCS over a private LAN if I had to, but wouldn't trust our networku farther than I could throw it.   ------------------------------    Date: 26 Sep 2001 13:26:45 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)o> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <M2JOFfre6+EW@eisner.encompasserve.org>s  q In article <cf15391e.0109260649.40446623@posting.google.com>, keithparris_NOSPAM@yahoo.com (Keith Parris) writes:nG > Given that you have 10-CPU SMP boxes, and 4 CI adapters per box, it's G > quite possible that you are currently using Fast_Path on CI to spreadsF > CI interrupt workload across non-primary CPUs.  Be aware that Memory  - We are most definitely using Fastpath for CI.7  B > Channel does not support Fast_Path (and since no new MC hardware? > development is planned, and there's thus no incentive for VMSiF > Engineering to dive back into the driver code, it's unlikely it everE > will support Fast_Path).  I worked at a site where we tried to movetE > from CI to MC for SCS traffic in a large cluster (16 6-CPU GS-140s) G > and our plans failed due to saturation of CPU 0 in interrupt state on F > lock-master nodes for hot application files.  They had to rip the MCF > hardware back out and go to 6 CIs instead.  As a test, you might tryG > turning off Fast_Path on CI in your current configuration and measure # > peak CPU 0 interrupt-state usage.n  I Most of my CI traffic is HSJ to/from host IO. Very little between the two- nodes.   > F > Since you have 10 CPUs per box now, and plan to go with ES45s, whichG > top out at 4 CPUs per box, I assume you'd end up with more individualhD > ES45 boxes, say 5 or more.  This has some repercussions.  For one,  J The plan was 3. The 8400 has got 5/440 CPUs. An ES45 with 4 GHz CPUs wouldI give me twice the "VUPS" of the old 8400, which is sufficient for plannedaJ growth for now. The GS140 has got 6/525s, which is still a bit less than aG single ES45, so we're looking at replacing it with 2 ES45s to allow forl growth there too.G  F > FC itself has Fast_Path support if you go with version 7.3, so you'dF > have the benefit of Fast_Path for the I/O operation traffic.  If you  K It looks like we'll definitely wait for 7.3 to settle out before going thisl route.  G > Another factor I haven't seen mentioned is shadow merges.  With FibrerH > Channel disks, you'll get time-consuming and I/O-intensive full mergesG > when a VMS node crashes, not the very-brief mini-merges possible with0G > HSJ controllers.  Some applications are significantly affected by theMB > I/O performance impact of a shadow merge operation.  Support forD > mini-merges on HSG80s is under development, but not available yet.  J We'll probably do controller based shadowing to avoid this. That's what we do on our UNIX SAN.r  * Other good advice noted and saved. Thanks.   ------------------------------    Date: 26 Sep 2001 13:40:26 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) > Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <X2S4CfgQaRxg@eisner.encompasserve.org>y  o In article <fteh3dZegGg8@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes: c > In article <4phK+ha$Nnld@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:VF >> 	Shame you can't stick a rack somewhere other than that datacenter.D >> 	If you could, you could easily and fairly cheaply create a faultA >> 	tolerant cluster (i.e. stay within a single mode fibre run ofeC >> 	say 1500-3000 feet) and run MC over it.  Even though MC has itsh > I > I'd love to do that. We do have 2 other data centers. One is 1727 milesh? > away, the oher 1375. Juat a bit outside the cluster specs :-)h  ? If DEQ gave you 1727 miles, you'd probably ask for the moon :-)    ------------------------------    Date: 26 Sep 2001 13:56:22 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)t> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <FZDmXMKjvLqD@eisner.encompasserve.org>h  a In article <3v$7zh09cP$g@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:rq > In article <ytia1jFr+3zG@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:-l >> In article <Pine.LNX.4.31.0109251515210.871-100000@jaipur>, Ryan Moore <rmoore@rmoore.dyndns.org> writes:E >>> We upgrade our CI-based cluster of 8400s to FC and MC.  The extra- >> -M >>> Make sure you get redundant MC hubs since MC is an active bus.  We testedON >>> MC hub failures on our backup cluster.  The cluster state transition to goI >>> from one MC to the other took less than one second.  Very impressive.m# >>> Even with lots of I/O going on.b >>    >> Everything will be redundant. >>   > 4 > 	Why?  But more importantly do you have the slots?G > 	What I mean is, I can see the need for duplicate network connectionsoF > 	as a necessity and that would be your backup for SCS traffic.  TellD > 	us what you are thinking (meant to draw you out... not slam you).  I I should have plenty of slots. haven't seen an ES45 yet, but the ES40 hassG 10. 2 FC, 2 MC, and if not integrated on the motherboard 2 NET and I'veC still got 3 left over.  J I'm not sure what you mean by what am I thinking. The target config is 2-3A ES45s clustered via MC, with all storage on a SAN. The goal is tohF significantly reduce the footprint of my existing configuration, whichC includes lots of old HSJ and some HSZ storage in SW800 and ESA10000gJ cabinets. WITHOUT affecting the reliability that we've come to expect from VMS.   ------------------------------    Date: 26 Sep 2001 14:30:18 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)a> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <tq4EXI7Xs1Bd@eisner.encompasserve.org>C  c In article <X2S4CfgQaRxg@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:nq > In article <fteh3dZegGg8@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes: d >> In article <4phK+ha$Nnld@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:G >>> 	Shame you can't stick a rack somewhere other than that datacenter.oE >>> 	If you could, you could easily and fairly cheaply create a fault)B >>> 	tolerant cluster (i.e. stay within a single mode fibre run ofD >>> 	say 1500-3000 feet) and run MC over it.  Even though MC has its >> nJ >> I'd love to do that. We do have 2 other data centers. One is 1727 miles@ >> away, the oher 1375. Juat a bit outside the cluster specs :-) > A > If DEQ gave you 1727 miles, you'd probably ask for the moon :-)u  H You're probably right. I've always wanted to go there! But 2.6+++ secondG latency per message would not be very good at all. It would make the 20fJ millisecond latency from Chicago to Phoenix seem awful good by comparison.  9 And what happens when the moon gets blasted out of orbit?y   ------------------------------  # Date: Tue, 25 Sep 2001 22:34:54 GMTt+ From: Ryan Moore <rmoore@rmoore.dyndns.org> > Subject: Re: VMScluster using fibrechannel and memory channel?; Message-ID: <Pine.LNX.4.31.0109251515210.871-100000@jaipur>r  A We upgrade our CI-based cluster of 8400s to FC and MC.  The extrarI performance has been great!  We had one CI bus in the cluster and we wereoF using CIXCD cards (XMI version).  We were having major I/O performanceE problems... that is why we went to FC.  The performance is FC is muchf& greater than CI (we are using HSJ50s).  I You'll need to be running at least VMS 7.2-1H1.  Plus you'll need a wholeuF bunch of ECOs to make it run right (like the FIBRE_SCSI ECOs).  Or youJ could go to 7.3.  But the talk about XFC problems in 7.3 would make me shyH away from that.  Note that 7.2-1H1 does not support FC tapes if you want" that.  Only 7.3 supports FC tapes.  J You said your machines have 10 CPUs.  That's a lot.  I hope you aren't CPU@ bound.  Or worse, you aren't CPU bound on the primary processor.F Apparently you will take a big primary CPU hit going from CI to MC forJ cluster communications.  You need to be aware of that problem.  We weren'tB CPU limited on our machines, so we didn't experience this problem.  I Make sure you get redundant MC hubs since MC is an active bus.  We testedSJ MC hub failures on our backup cluster.  The cluster state transition to goE from one MC to the other took less than one second.  Very impressive.: Even with lots of I/O going on.e  G I hope you will have significant downtime to do the switch... like at a H backup site or something.  The transition took some time for us.  We hadJ problems with bad GBICs (essentially transceivers) in the FC switchs.  AndG one FC switch had to be replaced because some FC paths kept going down.7G Or they wouldn't come back up if they failed (like a switch rebooting).pG So I'd recommend once the new stuff is installed to do lots of testing. E Try simulating power fails of a switch to watch FC fail over.  Or try>G turning off an MC hub to watch the cluster fall over to the other path,gI etc.  Maybe try simulating disk failures in the HSGs by yanking an active.@ disk.  We found lots of problems by testing fail-over scenarios.  I We got two PCI bus cages for each of our 8400s.  We put one MC and one FCeF in each bus.  I think the MC card has to be in the first slot or stuffG doesn't work right.  We didn't have it configured that way at first anda1 were having trouble with FC not working properly..  G The switchover wasn't painless for us, but once we got things stable...7I the performance gains have been dramatic.  I wrote some I/O test programseH based on the types of I/O we do at our site.  I saw about a 10-15x speedJ improvement on the FC disks compared to the disks on the HSJ50.  Big jump.J We use RMS indexed files, so it was also great to get our RMS lock trafficH off the CI and onto the MC.  But, again, watch for CPU saturation of the primary CPU.  G Overall, the installation, configuration, and testing time went on over J the period of about a month and a half.  I'm sure it could be done in lessH than, if necessary... but don't count on everything working right out of	 the gate.s  ! On 25 Sep 2001, Bob Kaplow wrote:   N > We've been running a VMS CI cluster for many years, going from VAX to Alpha.I > The primary nodes are currently an 8400 and a GS140, each with 10 CPUs. K > We're running 4 CIs with most of the disks and tapes out there, plus some N > direct SCSI devices. Unfortunately, that takes up a lot of space in our data< > center, and we need to significantly reduce the footprint. >tJ > What we're looking at for a replacement is a rack of ES45s, an EMA12000,L > Fibrechannel disk and tape, and Memory Channel for a cluster Interconnect.L > DLT tapes would be connected via MDRs. We should be able to do all of thisM > in about 25-30% of the space of the existing footprint, counting all of the  > storage arrays.i >pK > We're currently at VMS 7.2, but will upgrade to whatever it takes to makePK > this all work. We'll have to migrate to the new configuration a step at aeH > time, rather than doing a massive swap, as we can't be down that long. > I > Is any one currently running a production VMScluster using FibrechanneloJ > storage and memory channel interconnect? Has any one migrated from CI toJ > Fibrechannel / memory channel? Any advice from someone who has done this
 > already? >v > Bob Kaplow >eD >         They that can give up essential liberty to obtain a littleA >         temporary safety deserve neither liberty nor safety. --iC >         Benjamin Franklin Historical Review of Pennsylvania. 1759  >m   ------------------------------  % Date: Tue, 25 Sep 2001 21:56:36 -0400a( From: Hamlyn Mootoo <univms@bigfoot.com>> Subject: Re: VMScluster using fibrechannel and memory channel?+ Message-ID: <3BB135D4.4C62BEF3@bigfoot.com>l  D Unless you have a specific application or another good reason to useC Memory Channel, I wouldn't use it.  The speed gains are realized by 6 fibre channel, without the added complexity of the MC.   HM   Bob Kaplow wrote:  > N > We've been running a VMS CI cluster for many years, going from VAX to Alpha.I > The primary nodes are currently an 8400 and a GS140, each with 10 CPUs./K > We're running 4 CIs with most of the disks and tapes out there, plus some N > direct SCSI devices. Unfortunately, that takes up a lot of space in our data< > center, and we need to significantly reduce the footprint. > J > What we're looking at for a replacement is a rack of ES45s, an EMA12000,L > Fibrechannel disk and tape, and Memory Channel for a cluster Interconnect.L > DLT tapes would be connected via MDRs. We should be able to do all of thisM > in about 25-30% of the space of the existing footprint, counting all of the  > storage arrays.  > K > We're currently at VMS 7.2, but will upgrade to whatever it takes to makeoK > this all work. We'll have to migrate to the new configuration a step at a2H > time, rather than doing a massive swap, as we can't be down that long. > I > Is any one currently running a production VMScluster using Fibrechannel J > storage and memory channel interconnect? Has any one migrated from CI toJ > Fibrechannel / memory channel? Any advice from someone who has done this
 > already? >  > Bob Kaplow > D >         They that can give up essential liberty to obtain a littleA >         temporary safety deserve neither liberty nor safety. --.C >         Benjamin Franklin Historical Review of Pennsylvania. 1759w   ------------------------------    Date: 25 Sep 2001 20:57:02 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) > Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <9C8xcOuRYdQL@eisner.encompasserve.org>i  D CI and Memory Channel can carry Lock Manager Traffic.  Fibre ChannelA cannot.  It depends on what performance characteristics you want.r  V In article <3BB135D4.4C62BEF3@bigfoot.com>, Hamlyn Mootoo <univms@bigfoot.com> writes:F > Unless you have a specific application or another good reason to useE > Memory Channel, I wouldn't use it.  The speed gains are realized byn8 > fibre channel, without the added complexity of the MC. >  > HM >  > Bob Kaplow wrote:p >> $O >> We've been running a VMS CI cluster for many years, going from VAX to Alpha. J >> The primary nodes are currently an 8400 and a GS140, each with 10 CPUs.L >> We're running 4 CIs with most of the disks and tapes out there, plus someO >> direct SCSI devices. Unfortunately, that takes up a lot of space in our datan= >> center, and we need to significantly reduce the footprint.  >> vK >> What we're looking at for a replacement is a rack of ES45s, an EMA12000,uM >> Fibrechannel disk and tape, and Memory Channel for a cluster Interconnect.)M >> DLT tapes would be connected via MDRs. We should be able to do all of thissN >> in about 25-30% of the space of the existing footprint, counting all of the >> storage arrays. >>  L >> We're currently at VMS 7.2, but will upgrade to whatever it takes to makeL >> this all work. We'll have to migrate to the new configuration a step at aI >> time, rather than doing a massive swap, as we can't be down that long.e >> 2J >> Is any one currently running a production VMScluster using FibrechannelK >> storage and memory channel interconnect? Has any one migrated from CI to K >> Fibrechannel / memory channel? Any advice from someone who has done thisr >> already?s >> m
 >> Bob KaplowI >> mE >>         They that can give up essential liberty to obtain a littleiB >>         temporary safety deserve neither liberty nor safety. --D >>         Benjamin Franklin Historical Review of Pennsylvania. 1759 -- bN ==============================================================================I The Boulder Pledge: "Under no circumstances will I ever purchase anythingcJ      offered to me as the result of an unsolicited email message. Nor willI      I forward chain letters, petitions, mass mailings, or virus warningspH      to large numbers of others. This is my contribution to the survival      of the online community."N ==============================================================================   ------------------------------  # Date: Wed, 26 Sep 2001 21:50:10 GMTw9 From: rickm123@oregon456.uoregon789.edu (Rick Millhollin)s> Subject: Re: VMScluster using fibrechannel and memory channel?2 Message-ID: <3bb24566.797198050@news2.uoregon.edu>  E I've been following this thread with a great deal of interest since I5D am trying to decide on the best route to go with in terms of clusterE interconnect.  I recently upgraded a 3-node CI cluster of 2 7000s andDA an 8400 by replacing the 7000s with a GS80, and replacing all the0D HSJ50 based storage with 2 EMA12000 FC arrays.  I put a CIPCA in the; GS80 and stuck with CI for SCS traffic but no storage. ThissE configuration is working very well.  Now I'm getting ready to order anE second GS80 and am trying to decide how to connect it to the cluster.nC My choices are (1) buy another  CIPCA and stick with CI (2) put GigFB enet in all 3 systems  (3) go to MC in all 3 systems.  Option 1 isA fairly expensive and stays with dead-end technology.  Option 2 isnB cheap since I have or am planning Gig in all 3 anyway, but doesn'tD give me the independence from the network I'd like.  Option 3 soundsC possibly best but is the most expensive.  I'll mention also that myuD SCS traffic is relatively light.  I gather from this discussion thatE relative  performance is (low -> high) CI, Gig, MC.  I'm not clear oniC how to judge relative CPU 0 loads.  Any additional words of wizdom?t	 Thanks...t   ------------------------------    Date: 26 Sep 2001 08:40:45 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)n> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <ganTobAaVJW1@eisner.encompasserve.org>s  V In article <3BB135D4.4C62BEF3@bigfoot.com>, Hamlyn Mootoo <univms@bigfoot.com> writes:F > Unless you have a specific application or another good reason to useE > Memory Channel, I wouldn't use it.  The speed gains are realized byC8 > fibre channel, without the added complexity of the MC.  E The gain from going from CI to MC is saving the floor space of 4 Star 	 Couplers.o   ------------------------------    Date: 26 Sep 2001 08:54:08 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)e> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <ytia1jFr+3zG@eisner.encompasserve.org>   i In article <Pine.LNX.4.31.0109251515210.871-100000@jaipur>, Ryan Moore <rmoore@rmoore.dyndns.org> writes:tC > We upgrade our CI-based cluster of 8400s to FC and MC.  The extraeK > performance has been great!  We had one CI bus in the cluster and we wereiH > using CIXCD cards (XMI version).  We were having major I/O performanceG > problems... that is why we went to FC.  The performance is FC is muchf( > greater than CI (we are using HSJ50s).  I I'm fortunate that my company gives me the resources I need when/before IvJ need them, so I don't really have any proformance issues now. The issue weC need to solve is floor space. Our data center is full, we've got nodH immediate hope of getting more space, and we've got lots of new stuff we= need. Growth is one of the better problems tohave these days.y  L > You said your machines have 10 CPUs.  That's a lot.  I hope you aren't CPUB > bound.  Or worse, you aren't CPU bound on the primary processor.H > Apparently you will take a big primary CPU hit going from CI to MC forL > cluster communications.  You need to be aware of that problem.  We weren'tD > CPU limited on our machines, so we didn't experience this problem.  C We're not really CPU bound. The 8400 is a 5/440, so it's pretty oldoL technology. A single ES45 will probably give us double the "VUPS". The GS140I is obviously newer, but stll has the older 6.525 boards, so a single ES45oJ will still be a slight step up. By replacing that box with 2 of them, that' too should give us double the capacity.w  L And our primary CPU is in good shape. We don't see big Interrupt Stack or MPL synch times at all. My previous employer had a cluster where the primary CPUH pegged out with 3 CPUs in the box. But they were so memory starved, theyJ page faulted up the wazoo. I've got enough memory that I NEVER page to the	 pagefile.e  K > Make sure you get redundant MC hubs since MC is an active bus.  We tested9L > MC hub failures on our backup cluster.  The cluster state transition to goG > from one MC to the other took less than one second.  Very impressive.n! > Even with lots of I/O going on.r   Everything will be redundant..  I > I hope you will have significant downtime to do the switch... like at aiJ > backup site or something.  The transition took some time for us.  We had  L I doubt we will. But we may be able to bring in, test, and configure all theL new hardware, then make the switch a step at a time. Rob Young's suggestionsL (next response) were pretty much what we planned on doing anyway, a piece at a time.>   ------------------------------    Date: 26 Sep 2001 10:45:51 -0500+ From: young_r@encompasserve.org (Rob Young) > Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <u2sjiLo5q9aP@eisner.encompasserve.org>n  q In article <cf15391e.0109260610.6797d119@posting.google.com>, keithparris_NOSPAM@yahoo.com (Keith Parris) writes:uh > young_r@encompasserve.org (Rob Young) wrote in message news:<4phK+ha$Nnld@eisner.encompasserve.org>... >> 	Even though MC has itsG >> 	issues, I would still go with it... 100 MBit as Verele? chart showsnF >> 	does well with lock traffic and Keith points out Gigabit gains you6 >> 	little at this point as it is harder on your CPUs. > H > Uh, Memory Channel actually takes more host CPU than LAN adapters like> > Gigabit Ethernet.  (See Verell Boaen's Cluster InterconnectsF > presentation from DFW Days 2001 under http://www.dfwcug.org for dataG > on both simulated lock throughput and host CPU overhead.)  My concernmB > with Gigabit Ethernet was that as Verell's data shows, for smallG > packets (like lock requests), the on-board CPU apparently runs out ofs@ > steam early for the current (first) generation of GbE adapter. >o  < 	Yes... thanks Keith.  Of course MC is hard on the CPU, duh.B 	Thanks with mini-copy explanation over the years too.  I actuallyA 	got talking about that the other day on the phone and the fellowh@ 	on the other end stated "mini-merge" when we were talking about@ 	"mini-copy" and it afforded me the opportunity to describe themF 	both.  That to say, I *hope* I have 100 Mbit versus Gigabit advantage" 	down now.  If not, I search Deja!   				Rob    ------------------------------    Date: 27 Sep 2001 01:30:11 +0800, From: Paul Repacholi <prep@prep.synonet.com>> Subject: Re: VMScluster using fibrechannel and memory channel?- Message-ID: <87d74d2730.fsf@prep.synonet.com>a  ; kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:m  X > In article <3BB135D4.4C62BEF3@bigfoot.com>, Hamlyn Mootoo <univms@bigfoot.com> writes:  D > > Unless you have a specific application or another good reason to? > > use Memory Channel, I wouldn't use it.  The speed gains are F > > realized by fibre channel, without the added complexity of the MC.  tB > The gain from going from CI to MC is saving the floor space of 4 > Star Couplers.  E So spend $50 (tops...) on a 6' rack and shove ALL the SCs into it. OrV? shove them onto a plasic bin, and toss them under the flooring.n   -- .< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.' Now watch some one ask about the bin...    ------------------------------    Date: 26 Sep 2001 13:06:16 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow)g> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <$tqBMXaVeOwB@eisner.encompasserve.org>w  \ In article <87d74d2730.fsf@prep.synonet.com>, Paul Repacholi <prep@prep.synonet.com> writes:= > kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:t > Y >> In article <3BB135D4.4C62BEF3@bigfoot.com>, Hamlyn Mootoo <univms@bigfoot.com> writes:a > E >> > Unless you have a specific application or another good reason to @ >> > use Memory Channel, I wouldn't use it.  The speed gains areG >> > realized by fibre channel, without the added complexity of the MC.a >  rC >> The gain from going from CI to MC is saving the floor space of 4C >> Star Couplers.r > G > So spend $50 (tops...) on a 6' rack and shove ALL the SCs into it. OrrA > shove them onto a plasic bin, and toss them under the flooring./  L I've already got SIX star couplers (2 16-node, 4 8-node) in one rack. But itH still takes up mode floor space than an MC hub. And it fills up a lot of cable trough space.t   ------------------------------    Date: 26 Sep 2001 21:49:40 -0500+ From: young_r@encompasserve.org (Rob Young)t> Subject: Re: VMScluster using fibrechannel and memory channel?3 Message-ID: <HHymAMO5Aeqt@eisner.encompasserve.org>   o In article <FZDmXMKjvLqD@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:Oc > In article <3v$7zh09cP$g@eisner.encompasserve.org>, young_r@encompasserve.org (Rob Young) writes:ur >> In article <ytia1jFr+3zG@eisner.encompasserve.org>, kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) writes:m >>> In article <Pine.LNX.4.31.0109251515210.871-100000@jaipur>, Ryan Moore <rmoore@rmoore.dyndns.org> writes:CF >>>> We upgrade our CI-based cluster of 8400s to FC and MC.  The extra >>> N >>>> Make sure you get redundant MC hubs since MC is an active bus.  We testedO >>>> MC hub failures on our backup cluster.  The cluster state transition to goeJ >>>> from one MC to the other took less than one second.  Very impressive.$ >>>> Even with lots of I/O going on. >>> ! >>> Everything will be redundant.  >>>  >>  5 >> 	Why?  But more importantly do you have the slots?gH >> 	What I mean is, I can see the need for duplicate network connectionsG >> 	as a necessity and that would be your backup for SCS traffic.  TelloE >> 	us what you are thinking (meant to draw you out... not slam you).  > K > I should have plenty of slots. haven't seen an ES45 yet, but the ES40 hascI > 10. 2 FC, 2 MC, and if not integrated on the motherboard 2 NET and I'veo > still got 3 left over. > L > I'm not sure what you mean by what am I thinking. The target config is 2-3C > ES45s clustered via MC, with all storage on a SAN. The goal is touH > significantly reduce the footprint of my existing configuration, whichE > includes lots of old HSJ and some HSZ storage in SW800 and ESA10000iL > cabinets. WITHOUT affecting the reliability that we've come to expect from > VMS.    = 	Two MC is overkill.  If MC fails, you still have two networko@ 	paths.  Did you receive a recommendation on two MC , or are youA 	just being cautious?  And yes with the model 10, you have enough 2 	slots and you will be sure to order the model 10.   				Robo   ------------------------------    Date: 25 Sep 2001 12:07:46 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)u" Subject: Re: We are back from CETS= Message-ID: <cf15391e.0109251107.73012fb8@posting.google.com>m  Y "Tim Jackson" <tim.jackson@amsjv.com> wrote in message news:<3bb06de2$1@pull.gecm.com>...  > Keith, > + > Tried to access your site but it reports:g > . > Sorry, this site is temporarily unavailable!G > The web site you are trying to access has exceeded its allocated datai5 > transfer. Visit our help area for more information.o  E This is my reward for being cheap and using a free website provider.  E Turns out GeoCities measures data transfer rates and if the rate getsyE too high, they shut down access to a page for an hour (unless you payiF them ransom money).  So if you get this error, please try again later,C with my apologies.  And I'll work to provide the material on a moreu dependable website.uC --------------------------------------------------------------------C Keith Parris | parris at encompasserve dot org | VMS consulting on:pC Clusters, Disaster Tolerance, Internals, Performance, Storage & I/Oe   ------------------------------  % Date: Wed, 26 Sep 2001 12:37:42 -0400m From: Kat <kathee@mindiq.com> & Subject: weird MRU/scsi robot problems@ Message-ID: <5.1.0.14.2.20010926123433.03060e68@mail.mindiq.com>   Hello,  L Have a ESL9326 -- and getting the robot working automated with MRU/robot is J not proving fun or easy.  I have seen this error but found no answers. No J matter what I set the GK device to (we have 3 SCSI interfaces active, but 6 moving to FCTC later) they come up as "not responding"  I The silo works manually of course, and the tape library has been working rK for months. No one ever configured the robot for acfess through vms.  On 3   machines I have:   mkb100, mkb200 mkc100, mkc200 mkc300, mkc400   respectiively.  F I have tried on all systems -- gkb101, gkc101 and gkc301 and no luck..  
 Any ideas?   thanks Katn   ------------------------------  % Date: Wed, 26 Sep 2001 13:03:58 -0500w1 From: "Dave Gudewicz" <david.gudewicz@abbott.com> * Subject: Re: weird MRU/scsi robot problems8 Message-ID: <9ot5a4$56q$1@fizban.fizban.pprd.abbott.com>  K Don't have a ESL9326 as you do, but learned this about another tape library 5 just yesterday, so its still fresh.  Its a TL891 btw.C  I If the Library Mode is set to Sequential, then any and all ROBOT commandshB return a Robot not Ready error.  Sequential mode works as follows.  H When a cartridge is unoladed by the host, it loads the next cartridge inB magazine numerial order until all the cartridges have been loaded.  J This mode is set under the Configuration Options menu on the TL891.  Might be entirely different on yours.i   Dave...m  * "Kat" <kathee@mindiq.com> wrote in message: news:5.1.0.14.2.20010926123433.03060e68@mail.mindiq.com... > Hello, >iJ > Have a ESL9326 -- and getting the robot working automated with MRU/robot isK > not proving fun or easy.  I have seen this error but found no answers. No K > matter what I set the GK device to (we have 3 SCSI interfaces active, buti8 > moving to FCTC later) they come up as "not responding" >oJ > The silo works manually of course, and the tape library has been workingL > for months. No one ever configured the robot for acfess through vms.  On 3 > machines I have: >o > mkb100, mkb200 > mkc100, mkc200 > mkc300, mkc400 >u > respectiively. >gH > I have tried on all systems -- gkb101, gkc101 and gkc301 and no luck.. >1 > Any ideas? >C > thanks > Katn >g   ------------------------------  % Date: Tue, 25 Sep 2001 13:18:19 -0400I; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>d/ Subject: Re: Whose Running The Oldest Software?u$ Message-ID: <3bb0beca$1@news.si.com>   >It is running.... >e >VMS V5.5-1k >CMU IP V6.6-5
 >DCPS V1.0 >COBOL V5.2  >PATHWORKS V4.2-1M >rA >All old stuff, but it was up for 482 days and does the job well.c > 2 >So, is anybody running older software than me? :)  I You call that old?  We have one-half dozen VAX-11/725s in our plant doing>) board testing.  They're running VMS V4.3.a --A Brian Tillman                   Internet: tillman_brian at si.comoA Smiths Aerospace                          tillman at swdev.si.comt= 3290 Patterson Ave. SE, MS      Addresses modified to prevent < Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  # Date: Thu, 27 Sep 2001 05:09:19 GMTc- From: "Richard L. Dyson" <rickdyson@home.com>i/ Subject: Re: Whose Running The Oldest Software? ( Message-ID: <3BB2B466.E6DC4431@home.com>   Dean Woodward wrote: >  > Jonathan McCormack wrote:e > >e9 > > While shutting down our systems we came across an oldH4 > > VaxStation 4000 VLC that we had forgotten about.D > > All old stuff, but it was up for 482 days and does the job well. > J > Ironic that the best advertising you can give a system is that it ran so/ > well it's managers forgot about it.  *sigh*.se  L I recently updated the firmware on some DECserver900TM servers.  Not VMS perM say, but related.  One of them had been up for almost 900 days since its last  reboot.v  L This is a TS that runs instruments, printers and dumb terminals 24x7x366, so it is still used routinely.=   rick   ------------------------------  % Date: Wed, 26 Sep 2001 01:31:39 -0400m- From: JF Mezei <jfmezei.spamnot@videotron.ca>o Subject: Re: World Trade Centers, Message-ID: <3BB16832.566F988D@videotron.ca>   David Froble wrote:@I > As long as Saddam is in power, I wonder whether sanctions (which I feeliH > are not working and should be replaced with somnething that will work)$ > have any bearing on those numbers.  M Sadam is exactly where the USA wants him. Weak but still in power. I was toldtM that if he is thrown out, he would likely be replaced by a fundementalist who.G would turn Iraq into a fundamentalist state and make things MUCH worse,rF especially for its neighbour to the south that has so far succeeded in" avoiding fundementalist pressures.  M And if Iraq turns into a fundamentalist state and starts to destabilise SaudiIK Arabia, it threathens the oil supply to the USA and many others, as well aslN cut off military access for the USA to some of the air bases in Saudia Arabia.   ------------------------------  % Date: Wed, 26 Sep 2001 15:07:57 -0400e+ From: "John Saunders" <jws@ma.ultranet.com>r Subject: Re: World Trade Centers+ Message-ID: <9ot93n$cd2$1@bob.news.rcn.net>n  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3BB10205.FBE36BA5@videotron.ca... > John Saunders wrote:J > > I never once heard the Administration say we were going to war against > > Afghanistan. >iJ > They haven't said it, but they did say that they were going after states thatL > harbour terrorists. And they started to move their heavy aircraft carriersI > towards that region. If that isn't a strong hint, I don't know what is.V  L It's a  hint to the Taliban that by the time we're ready to strike, they had1 better not be a country which supports terrorism.y   ------------------------------  % Date: Tue, 25 Sep 2001 15:36:30 -0400n' From: "Bill Todd" <billtodd@foo.mv.com>r Subject: Re: World Trade Centerw( Message-ID: <9oqm96$p30$1@pyrite.mv.net>  6 "John Saunders" <jws@ma.ultranet.com> wrote in message$ news:9oqlcv$ta$1@bob.news.rcn.net...4 > "Bill Todd" <billtodd@foo.mv.com> wrote in message$ > news:9oq0kc$r81$1@pyrite.mv.net... > >.7 > > "ChrisQ" <lightwork@aerosys.co.uk> wrote in messagea+ > > news:3BB06F6C.AEBDA9C3@aerosys.co.uk...F > >@ > > ...c > <snap>( > >  What else could GB say while tryingA > > > to steady the nation and buy time to find a solution to ther unsolveable, > > at > > > least in the short term ?C > >ML > > He could make our resolve to catch and deal with those responsible every > biteL > > as clear without  a) revving up our military (and civilian) expectations > forr > > a major land war,e >nK > Bill, I was watching the Pentagon and White House news conferences today,5C > and found it interesting that it's the Press who is looking for ah
 full-scale > land war.s  L I watched most of Rumsfeld's conference, and agree with your impression.  InK fact, I really liked what I heard (and it's safe to say I'm a more criticaln audience than most).  J That doesn't get Dubya off the hook for a lot of what I've heard *him* sayK for the past 12 days or so, but it is encouraging.  And while the press has H had a feeding frenzy over the bellicose verbiage emanating from D.C., ifK that verbiage becomes more reasonable they'll get used to a healthier diet,C# as will the rest of our population.n  L Good leaders know when to lead their people and when to follow them.  I hope- ours were just a little confused for a while.    - bill   ------------------------------  % Date: Tue, 25 Sep 2001 17:15:58 -0400 + From: "John Saunders" <jws@ma.ultranet.com>G Subject: Re: World Trade Centerl+ Message-ID: <9oqs7k$2rd$1@bob.news.rcn.net>o  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3BB0EC88.679F9129@videotron.ca... > John Saunders wrote:F > > Bill, I was watching the Pentagon and White House news conferences today,E > > and found it interesting that it's the Press who is looking for a 
 full-scale
 > > land war.h >oI > The media were given the impression by the administration that this waso goinghK > to be done ASAP with grandiose war fireworks, live coverage in stereo etc  etc.J > Remember that initially, the rethoric was very strong as it took a whileF > before Bush Jr realised that war in Afghanistan was useless and that nobodyK > really knew how they could get Bin Landen. That is when he started to sayv thatJ > it would be a long term efforts and started to downplay the "WAR" thing, evenL > though, at the same time, the military is gearing up for a full scale war.  F I never once heard the Administration say we were going to war againstB Afghanistan. We were always going to war against terrorists. WhichB terrorists wasn't clearly communicated at first, but it was alwaysE terrorists. Remember that in his first remarks about 911 he said that-. America had been attacked by faceless cowards?  C > Don't underestimate the fact that the USA is using its media as ae
 propagandaJ > machine to scare the folks in the area into cooperating with the USA, so then+ > USA is understandably flexing its muscle.p  K The Press don't seem very pleased with the restrictions they've been under.nJ And someone today actually asked Rumsfeld whether the DOD would lie to the7 Press. I don't think we're "using" the media very much.s  I > Freezing bank accounts just doesn't have the coverage/ratings potentialw that aJ > full scale war with scuds, cruise missiles etc have. It is taking a hell of aL > time for CNN to remove that "AMERICA's WAR" garbage banner on its screens.L > But in the end, this "war" will just peter out. They won't catch Bin Laden andoK > friends, but they will cut off his funding and isolate them. They may geta sentJ > to some small island nation where, while they will still be alive, won't be > able to run their empire.e  I I believe you're mistaken. At the very least we're going to clean out theaH training camps in Afghanistan and assure ourselves that bin Laden reallyF isn't there. Also, during the time we're in there (after defeating theJ Taliban), we'll probably be able to get some good intellegence from people( who are no longer afraid of the Taliban.  F I also suspect that the Taliban is stupid enough to fight, which meansJ they're going to be massively defeated. If we've got any sense, we'll thenL ask the UN to take over in Afghanistan, perhaps after we do some things like  road building and mine clearing. --
 John Saunders  jws@ma.ultranet.com-   ------------------------------  # Date: Wed, 26 Sep 2001 03:49:44 GMTp/ From: "Randall Burlew" <rkburlew@earthlink.net>  Subject: Re: World Trade Center F Message-ID: <sfcs7.14456$W83.1459929@newsread2.prod.itd.earthlink.net>   Bill Todd wrote: > Randall Burlew wrote:n  ; > > Now you know how Fred felt about the discussion and the@= > > frustration he felt. Maybe knowing how someone else feelsi7 > > is the beginnings of an understanding. We can hope.o >n > Not in my case, you can't.   ??  A I can't hope that you will begin understanding the people you are > holding discussions with? Don't you want to understand others?  . > I was thinking about what I had said to FredK > when I wrote the above, and made sure I understood the differences in thenF > two situations - one obvious example being your wild digression intoK > historical actions of Bill Clinton and Lyndon Johnson which was the point  at8 > which I decided there was no hope for this discussion.  E The point under discussion was that there not much difference between C Democrats and Republicans, especially in foreign policy issues. The G best way I know to predict future behavior is based on past and present B behavior. I gave examples of what Democratic leaders are doing nowK and what they did in the past. I don't see how that is a wild digression --rJ it clearly addresses the point under discussion. I pulled my punches, too.' I didn't mention what Harry Truman did.l  3 > That's judgemental, but no more than you've been.e  D I was really hoping that you might begin to see that clubbing peopleE over the head with your version of the truth was not an effective ways to convince them of anything.s   > My guess is that by nowrG > I've gotten through to just about anyone I'm likely to be able to, sogK > continuing for the benefit of others no longer provides enough benefit to J > make the effort worthwhile - and there's nothing coming from you to make itF > appear worthwhile either.  I don't mind discussing details - in fact details G > are often the real meat, since opinions often turn out to be based on  fluff.K > But your ability to address details at the level required is insufficientu to > make interesting.   D Does this mean "Everyone who thinks like I do agrees with me, I haveE driven everyone else so far away with my attitude that I have no hope0I of convincing them of anything, and I'm tired of playing with you becausen you won't play by my rules?"  J > So the only remaining issue is whether we can have any influence on what ourLK > government does, and if you contend that we can't then I suspect there'lll beJ > plenty of people happy to set you straight without any need for me to do so.e  I I hope I didn't say that we cannot influence what our government does. Ife so,,% that was certainly not the intention.    Randyo   ------------------------------  % Date: Tue, 25 Sep 2001 22:40:04 -0400 ( From: David Froble <davef@tsoft-inc.com> Subject: Re: World Trade Centerd, Message-ID: <3BB14004.9070106@tsoft-inc.com>   Paul Repacholi wrote:t  , > David Froble <davef@tsoft-inc.com> writes: >  >  >>David J. Dachtera wrote: >> >  c >  >>>David Froble wrote: >>>t > 
 >>>>[snipped]o >>>> > D >>Bill's right about unintended consequences.  The death toll amountB >>Afgans fleeing the cities and being stopped at the borders couldD >>easily surpass that in NY.  That's appalling!  The US government'sB >>best course would be to place helping save these lives above any
 >>revenge. >> >  > To give a comparison...a > H > Every month, the sanctions on Iraq kill about 6000s women and children6 > in Iraq, mostly in the Kurdish and southern sectors. >  >   H As long as Saddam is in power, I wonder whether sanctions (which I feel G are not working and should be replaced with somnething that will work)  " have any bearing on those numbers.   Dave   ------------------------------  % Date: Tue, 25 Sep 2001 22:35:55 -0400o( From: David Froble <davef@tsoft-inc.com> Subject: Re: World Trade Centern, Message-ID: <3BB13F0B.3070103@tsoft-inc.com>   Jan Vorbrueggen wrote:  1 > koehler@encompasserve.org (Bob Koehler) writes:  >  > G >>> That is the problem. They are just that - criminals that need to be I >>>prosecuted. But there is not coutry that they represent. So you cannote >>>have a war.   >>>aN >>   But we already did.  Therefor it set a precedent, and we can do it again. >> > M > You're not alone in the world, you know. Call it what you please, but don't-D > expect you partners in combat to accept it as a "war" in the sense" > international law uses the word. >  > 	Jan >   G Acknowledged, I think that's pretty clear.  What we can and should ask TI is for the world to join us in building a better world for all, and that g# WILL require more than destruction.o   Dave   ------------------------------  % Date: Tue, 25 Sep 2001 22:59:23 -0400l( From: David Froble <davef@tsoft-inc.com> Subject: Re: World Trade Centeri, Message-ID: <3BB1448B.2010100@tsoft-inc.com>   Bill Todd wrote:  5 > "ChrisQ" <lightwork@aerosys.co.uk> wrote in messagei) > news:3BB06F6C.AEBDA9C3@aerosys.co.uk...r >  > ...s >  > J >>Got really wound up initially about the rhetoric coming from the US, but >>	 > realisem > - >>now that that was for domestic consumption.r >> > N > While that's undoubtedly true, there's still a very real possibility that he > means what he says.- > K > But even leaving that aside, it's already been sufficiently disturbing topL > frighten millions of Afghans to the point that something like a million ofK > them, ranging from very young to very old, have taken off for the borders-L > (which they may or may not know are nominally closed) with only that whichG > they can carry or, for the more fortunate ones, stuff into a vehicle.4 > I > Afghanistan has some pretty rough country.  If 'only' a bit under 1% ofoH > those people die as a result of our rhetoric, it will have killed moreM > innocent people than were killed two weeks ago - before we've actively donee > a damn thing.r    G I have sympathy for the plight of these Afgans.  However, their flight  F was not caused by any military or terrorist act of the US Government. E No where will you find any mention by the government of any proposed .F action against innocent afgan citizens.  Maybe they fled because they C expected the US to do what they would do if the balance of nuclear 0F weapons was reversed.  In that scenario, who's showing more restraint?  @ Get off this topic Bill, it was not caused by the US government.    M > And, of course, fanning the flames domestically has contributed to creatingbL > an environment in which U.S. Arabs now live in fear, a growing number haveL > actually had their persons, businesses, and/or places of worship attacked,N > and at least two, including one person (a Sikh) who just *looked* a bit like > an Arab, have been killed.    G These actions are very sad, and not supported by the US government, or .G US citizens.  A few radical elements exist in the US also, in case you  F weren't aware of that fact.  Had a KKK rally at the county courthouse I this summer.  Ever been to one of those Bill.  Kind of makes the rest of t& us look more reasonable by comparison.    I > Dubya's 'domestic rhetoric' (and that of his Cold Warrior re-treads) is2> > morally reprehensible and possibly criminally irresponsible.    H How far will we need to extend the left field fence to get you into the  stadium?    & >  What else could GB say while trying > J >>to steady the nation and buy time to find a solution to the unsolveable, >> > at >  >>least in the short term ?p >> > N > He could make our resolve to catch and deal with those responsible every bitN > as clear without  a) revving up our military (and civilian) expectations forM > a major land war,  b) capitalizing upon the attack as an excuse to push for0K > significant curtailments of our civil liberties (no, I'm not talking justgD > about things like increased airport security) that 'law-and-order'L > conservatives have been after for a very long time,  c) bullying the otherM > countries of the world (including our friends) with "You're either with us,WL > or with the terrorists" drivel in an attempt to obtain support for his ownI > excessive demands, and  d) specifically bullying the Taliban government M > (which we don't even deign to recognize) with 'non-negotiable' demands thathM > he says we won't even *discuss* in what is essentially a criminal, though alJ > very major criminal, matter where the proper procedures for dealing withA > other nations are clearly laid out (and mechanisms for building K > international coalitions to address nations which fail to respond to such N > procedures are also well-understood).  Just to name a few things off the top
 > of my head.r    F I wonder if the new restraint of some of the Telliban reported in the H news tonight would have happened without such actions?  Maybe so, since H I have a hard time with the notion that even hard-line Muslems would be ) so insensitive to the loss of human life.T    7 >  If the aim of the terrorists was to draw the US into> > D >>a snap response, nuke afghansitan, whatever, then they have failed >> > dismally. In > B >>that, the restraint shown by your government is just the sort of >> > "leadership by > 6 >>example" that the world could do with a lot more of. >> > K > I suggest that the use of the word 'restraint' to dignify mere failure toeN > lash out immediately and blindly stretches its definition severely, and thatL > praise merely for not doing so (yet) is hardly merited (unless lashing out5 > blindly is what you'd normally expect of the U.S.).v > M > The motives likely behind your attempt at reconciliation here are laudable,tI > but I believe you've confused an understanding of the personal feelingseH > generated by the attack with the behavior a responsible country should
 > exhibit. >  > - bill    E Overall, I see more responsibility that you do.  When the government bG doesn't fall to your low expectations, you seem incapable of accepting w that seeming fact.   Dave   ------------------------------  % Date: Tue, 25 Sep 2001 22:33:36 -0400i( From: David Froble <davef@tsoft-inc.com> Subject: Re: World Trade Centers, Message-ID: <3BB13E80.4060005@tsoft-inc.com>   JF Mezei wrote:k    P > But more importantly, he focused his speech on efforts to bring justice to theF > terrorists instead of bringing the USA back to normalcy and economicU > prosperity. As was said during his father's last campaign: IT'S THE ECONOMY STUPID.n >  > O > If you really want to defeat the terrorists, you have to ignore them and just M > rebuild ASAP and get back to normal life.  (Again with covert operations to N > get to them, and celebrate only once you have proven that they have been put > out of action).e >   G Well, I think some of the above is happening, with regard to trying to R8 get people back to 'normal', whatever that is.  However:    H It would be political suicide to try to act like you're not going after G the culprits, and too many US citizens would not accept such a stance. nF They will accept the reality that it cannot be a conventional type of L war, whatever that is, and the necessity of new methods to combat terrorism.    G On tonight's news there was a piece about a suspected terrorist trying rI to get a US Government loan to buy an ag plane.  (Seems like they really  H like the idea of getting us with our own assets.  Cheeky basterds!)  An G ag plane would be a decent platform for spreading various chemical and  A biological weapons.  Things are NOT normal when we're looking at  K possible further actions.  I wonder if there will be a NORMAL anytime soon.n   Dave   ------------------------------  # Date: Wed, 26 Sep 2001 05:10:09 GMTt" From: GreyCloud <mist@cumulus.com> Subject: Re: World Trade Center + Message-ID: <3BB1643E.1D8C42FC@cumulus.com>    David Froble wrote:i >  > Paul Repacholi wrote:s > . > > David Froble <davef@tsoft-inc.com> writes: > >3 > >6 > >>David J. Dachtera wrote: > >> > >. > >  > >>>David Froble wrote: > >>>a > >l > >>>>[snipped]o > >>>> > >aF > >>Bill's right about unintended consequences.  The death toll amountD > >>Afgans fleeing the cities and being stopped at the borders couldF > >>easily surpass that in NY.  That's appalling!  The US government'sD > >>best course would be to place helping save these lives above any > >>revenge. > >> > >  > > To give a comparison...a > >nJ > > Every month, the sanctions on Iraq kill about 6000s women and children8 > > in Iraq, mostly in the Kurdish and southern sectors. > >  > >e > I > As long as Saddam is in power, I wonder whether sanctions (which I feeleH > are not working and should be replaced with somnething that will work)$ > have any bearing on those numbers. >   > Your feelings are correct.  From what I've been able to gleen,G the mid-eastern countries in times of trouble use the Bazaar as a meansdC of acquiring goods.  Sort of like the black market.  And this leadsv4 to some scary situations or at least scary thoughts.   ------------------------------  % Date: Wed, 26 Sep 2001 20:19:11 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: World Trade Centero' Message-ID: <3BB27E8F.D6643D5A@fsi.net>-   David Froble wrote:  >  > Jan Vorbrueggen wrote: > 3 > > koehler@encompasserve.org (Bob Koehler) writes:  > >c > >:I > >>> That is the problem. They are just that - criminals that need to beeK > >>>prosecuted. But there is not coutry that they represent. So you cannotv > >>>have a war. > >>>uP > >>   But we already did.  Therefor it set a precedent, and we can do it again. > >> > > O > > You're not alone in the world, you know. Call it what you please, but don'teF > > expect you partners in combat to accept it as a "war" in the sense$ > > international law uses the word. > >o
 > >       Jana > >e > H > Acknowledged, I think that's pretty clear.  What we can and should askJ > is for the world to join us in building a better world for all, and that% > WILL require more than destruction.   B True; however, it's rather like trying to have an enjoyable picnic around a hive of African bees.   -- b David J. Dachterat dba DJE Systemsd http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/l   ------------------------------  % Date: Wed, 26 Sep 2001 20:23:53 -0500a1 From: "David J. Dachtera" <djesys.nospam@fsi.net>n Subject: Re: World Trade Centerw' Message-ID: <3BB27FA9.4A7A4A9A@fsi.net>i   David Froble wrote:s >  > JF Mezei wrote:f > R > > But more importantly, he focused his speech on efforts to bring justice to theH > > terrorists instead of bringing the USA back to normalcy and economicW > > prosperity. As was said during his father's last campaign: IT'S THE ECONOMY STUPID.  > >  > >dQ > > If you really want to defeat the terrorists, you have to ignore them and justrO > > rebuild ASAP and get back to normal life.  (Again with covert operations tohP > > get to them, and celebrate only once you have proven that they have been put > > out of action).o > >j > H > Well, I think some of the above is happening, with regard to trying to: > get people back to 'normal', whatever that is.  However: > I > It would be political suicide to try to act like you're not going after H > the culprits, and too many US citizens would not accept such a stance.G > They will accept the reality that it cannot be a conventional type ofTN > war, whatever that is, and the necessity of new methods to combat terrorism. > H > On tonight's news there was a piece about a suspected terrorist tryingJ > to get a US Government loan to buy an ag plane.  (Seems like they reallyI > like the idea of getting us with our own assets.  Cheeky basterds!)  AnhH > ag plane would be a decent platform for spreading various chemical andB > biological weapons.  Things are NOT normal when we're looking atM > possible further actions.  I wonder if there will be a NORMAL anytime soon.,  F There was a report on the business news as I was driving home from the? train today. Seems some folks traded heavily before the attack,fF positioning themselves to profit from drops in airline stocks. HoweverE much they may say they hate capitalism, they sure seem to embrace itseE benefits, if there's even a shred of truth to the implications of theh reported findings.   -- r David J. Dachterae dba DJE Systems1 http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Tue, 25 Sep 2001 11:01:34 -0400 - From: "Richard D. Piccard" <piccard@ohio.edu> 9 Subject: Re: [jfc] Seen in one of my older crash dumps...n( Message-ID: <3BB09C48.2BED2686@ohio.edu>   Keith,  @     Thanks for the extended and very understandable explanation.  ]     The conclusion I take from the last paragraph is that even with a quorum disk, you cannotr[ *EXPECT* 100% uptime from a two-node cluster, since there is a deliberately coded path thatt$ takes out both nodes simultaneously.  Z     Likewise, the surviving node of a symmetrical three-node cluster will hang for lack of5 quorum while the two bug-checked nodes are rebooting.d  ]     What, then, is the minimum cluster that can provide "continuous service" (that is, appearn_ to hang, but only for a time short compared to the reboot time)?  Four nodes plus a quorum diski on shared storage?  Five nodes?                            RDPa     Keith Parris wrote:d  ^ > eplan@kapsch.net (Peter LANGSTOEGER) wrote in message news:<3bafc542$1@news.kapsch.co.at>...< > > This one I've just seen in one of my (older) crash dumps > > 8 > >   Time of system crash: 24-MAY-**** 18:15:48.99<NUL> >lG > Apparently the time quadword contains bogus data, which resulted in agC > year value that is greater than 9999 and cannot be displayed in ae > 4-digit field. >e5 > >   VMScluster node: MARS, a AlphaServer 2100 5/250sL > >   CPU 00 reason for Bugcheck: LOCKMGRERR, Error detected by Lock Manager >eD > LOCKMGRERR (and the similar CNXMGRERR) are triggered when the LockC > Manager (or Connection Manager, respectively) code is decoding ancH > incoming packet and finds garbage in a field in the packet.  In such a+ > case, there are several potential causes:eE > 1) Bugs in the VMS code on the other end which assembled the packet F > (this is the least likely cause nowadays, since the checks have been* > in place in the code for many years now)H > 2) Corruption of the packet on the source node after it was assembled,B > but before it was sent (perhaps by a memory hardware problem, orA > errant code overwriting the packet as it sat in non-paged pool)oE > 3) Corruption of the data as it passed across the LAN.  Despite CRCwG > technology, parity or ECC memory within LAN adapters and bridges, andtE > so forth, hardware problems or firmware errors sometimes cause datayC > corruption that is not detected as a CRC error (or sometimes poordE > bridge design re-calculates a new CRC for internally-corrupted data , > when it is transmitted out the other side)B > 4) Corruption of the packet on the target node after it has beenH > received, but before it is examined, again either by a memory hardwareE > problem or by errant code trampling over it as it sits in non-pagedc > pool >tD > #3 seems to be the most common cause in practice, and probably the> > cause in your case, since you had a bridge acting strangely. >>E > This data validation code and the resultant bugchecks were put intoeG > the code by the maintainer of the lock manager and connection managerbB > because he got sick and tired of seeing bugs in new interconnectF > hardware designs causing odd cluster problems that his code ended upE > getting blamed for.  The intent of these bugchecks is to notify you C > that undetected data corruption has apparently occured within SCSnD > packets, a condition that could be quite dangerous to your data if > uncorrected. >eE > Originally, it was only the receiving node which detected the errordH > and which went down with a bugcheck.  Because some of the problems canA > be caused on the source end rather than the receiving end (e.g. D > corruption of the packet in memory on the source node), there wereE > some cases where a corrupted packet might be sent to multiple nodes.F > (consider a corrupted lock request packet that gets sent to the lockB > master node, crashes that, and then gets re-sent to the new lockE > master node elected after the first node leaves the cluster; I once E > had 9 nodes of a 16-node cluster of 8400s crash one right after theeH > other due to this problem), I'm told VMS in recent versions now causesE > a simultaneous bugcheck on both the receiving and sending nodes, too@ > localize the damage and minimize the potential number of nodes > affected by a bugcheck.aE > ------------------------------------------------------------------- E > Keith Parris | parris at encompasserve dot org | VMS consulting on:iE > Clusters, Disaster Tolerance, Internals, Performance, Storage & I/Oc   --B ==================================================================B Dick Piccard                           Academic Technology ManagerB piccard@ohio.edu                                 Computer ServicesB http://oak.cats.ohiou.edu/~piccard/                Ohio University   ------------------------------  % Date: Wed, 26 Sep 2001 09:47:38 +0200 . From: "Gorazd Kikelj" <gorazd.kikelj@aster.si>9 Subject: Re: [jfc] Seen in one of my older crash dumps...t/ Message-ID: <6Hfs7.550$8v1.18527@news.siol.net>h  8 "Richard D. Piccard" <piccard@ohio.edu> wrote in message" news:3BB09C48.2BED2686@ohio.edu... > Keith, > B >     Thanks for the extended and very understandable explanation. > G >     The conclusion I take from the last paragraph is that even with a  quorum disk, you cannotv@ > *EXPECT* 100% uptime from a two-node cluster, since there is a deliberately coded path that& > takes out both nodes simultaneously. >yK >     Likewise, the surviving node of a symmetrical three-node cluster willh hang for lack of7 > quorum while the two bug-checked nodes are rebooting.s >eE >     What, then, is the minimum cluster that can provide "continuouse service" (that is, appeartH > to hang, but only for a time short compared to the reboot time)?  Four nodes plus a quorum disk! > on shared storage?  Five nodes?t >  >                         RDPo  J Three node cluster with quorum disk can survive if it loose two nodes whenK every node has 1 vote and quorum disk have 2 votes. Assuming that all threey1 nodes have direct access to quorum disk naturaly.h   best, Gorazd Kikelj    ------------------------------  % Date: Wed, 26 Sep 2001 09:49:47 +0100g% From: Alan Greig <a.greig@virgin.net>B9 Subject: Re: [jfc] Seen in one of my older crash dumps... 8 Message-ID: <e553rtc0br1o0edt62j93vidu1lvlfh78n@4ax.com>  B On 25 Sep 2001 07:50:53 -0700, keithparris_NOSPAM@yahoo.com (Keith Parris) wrote:   >  >uC >LOCKMGRERR (and the similar CNXMGRERR) are triggered when the Lock,B >Manager (or Connection Manager, respectively) code is decoding anG >incoming packet and finds garbage in a field in the packet.  In such ad* >case, there are several potential causes:D >1) Bugs in the VMS code on the other end which assembled the packetE >(this is the least likely cause nowadays, since the checks have beent) >in place in the code for many years now)h  B I am not sure that is true as I have only ever seen one lockmgrerr? that I can recall and it  was caused by a known VMS bug and the F workaround did fix it for us.  Problem is present in 7.1 through 7.2-1' at least. with no patch kit available. t     -- Alan   ------------------------------    Date: 26 Sep 2001 11:25:19 +0200* From: eplan@kapsch.net (Peter LANGSTOEGER)L Subject: Re: [OpenVMS] V7.2 VAX satellite doesn't find V7.3 Alpha bootserver( Message-ID: <3bb19eff@news.kapsch.co.at>  q In article <cf15391e.0109242110.72f559c2@posting.google.com>, keithparris_NOSPAM@yahoo.com (Keith Parris) writes:(A >Other things I can think of that could cause a failure to form agC >connection would be a shortage of non-paged pool on one end or thesB >other, or a conflict in SCS nodenames.  SCSCONNCNT shouldn't be aE >problem; VMS reserves an extra 200 slots beyond the SYSGEN parametera& >setting in any recent versions, IIRC.  M SCS nodenames do not conflict. Nothing has changed here (and see also below).  Non-Paged Pool, well maybe.hI Because. as I wrote, I did a VMS upgrade from V7.2-1 to V7.3 on the AlphavH bootserver and VMS upgrades tend to destroy SYSGEN parameters learned byE AUTOGEN feedback a (sometimes not so) little bit, this possible cause F could be fixed with the next AUTOGEN runs. It does however not explain= why this should wrong parameter only effect VAX satellites...e  E >The satellite would be sending out a Solicit_Service datagram to thef@ >multicast address for the cluster group number (did that or theE >cluster password change recently?  Is CLUSTER_AUTHORIZE.DAT the same F >on both system disks, and in the common area, with no other copies inE >other roots?), which contains the device name of the system disk (isd? >that correct in the DECnet or LANACP data for the satellite?),h  B Yup. It is the same. Nothing has changed here for the last decade.  G >                                                                and ataD >least one Alpha bootserver should be answering with a VERF (Channel? >Control Verify) packet indicating that it can indeed serve theeF >requested disk.  The satellite will pick the first node which answers@ >and send a VACK (Channel Control Verify-Ack) to it.  (Any other2 >willing servers eventually time out and give up.) >tF >Appendix F in the VMS Cluster Systems manual, on "Troubleshooting theF >NISCA Protocol", may be helpful.  It includes info on using a network >sniffer to diagnose problems.   Thanks for the suggestions.'  ? >Could a bridge or router be passing MOP requests but not NISCAo >(protocol type 60-07)?i  A Nope. A simple DECswitch 900EF without any filters is in between.i  = >Could packets be getting through but the packet loss rate isbE >excessive?  You might want to check the ReXmt counters in SDA on ther >server node(s).  % Packet loss is under investigation...n  H BUT: I did copy the VAX system image from the served HSJ disk to a localE attached SCSI disk and then this VAX (no longer) satellite is able toaC boot and joind the cluster WITHOUT ANY CHANGE of SYSGEN parameters.-   So for me it looks like:D THE PROBLEM IS ONLY IN THE MSCP SERVER ON THE ALPHA. THERE MUST HAVE* BEEN A CHANGE BETWEEN V7.2(-1) AND V7.3...  F I've now a workaround (though I possibly have now to switch to HBVS toN get back the avaibility of the VAX system disk [previously a HSJ mirror set]).  ! Thanks for reading and responding    -- N< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888g< <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------    Date: 26 Sep 2001 14:05:00 -0700/ From: mark@dino.cacr.caltech.edu (Mark Bartelt) A Subject: {HP,Compaq} user groups (was Re: CETS2001 presentations) 0 Message-ID: <9otfts$koh$1@dino.cacr.caltech.edu>  H (I'm adding comp.sys.hp.hpux to the distribution, even though that's notH really the appropriate forum.  But as far as I can tell, there isn't anyG newsgroup analogous to comp.org.decus for Interex folks to talk to each. other.)t   [ JF Mezei ]  D >>  Considering that Compaq will cease to exist in a few months, howD >>  can what-used-to-be-DECUS discuss the future with such details ? >>D >>  Have there been any official statements from Carly about support >>  for what used to be DECUS ?d  E Actually, your mention of the merger with (or perhaps, should we say,e8 the getting-eaten-up-by) HP raises an interesting point:  G This past May, I attended my first-ever InterWorks meeting.  InterWorks-F is a conference sponsored by Interex (an HP-centric user organization)G and my first impression was that it really reminded me of the old DECUS2D (say early 70s through mid 80s) which many of us remember so fondly.  F I had to admit that part of this was a result of some serious dj vu:F The conference was held in the very same hotel (the Hyatt Regency, SanF Francisco) which was the site of the first DECUS symposium (fall 1973) that I attended.  H But there were other things about InterWorks/2001 which were reminiscentH of DECUS/1973 as well:  Rather small in size (~500 attendees, I believe)H and included sit-down lunches.  But it nonetheless had a nicely eclecticI assortment of sessions, ranging from very basic to highly technical.  And H there were some "feedback to HP" sessions, where developers and managersH listened to comments from the attendees, and that also brought back fond/ memories of the "good old days" of early DECUS.a  G It will be interesting to see what sort of user organization eventuallytH evolves to meet the needs of current HP users and current Compaq (and ofD course, former DEC (and other)) users, {if|when} the merger happens.  I Mark Bartelt                                                 626 395 2522sI Center for Advanced Computing Research              mark@cacr.caltech.edusI California Institute of Technology      http://www.cacr.caltech.edu/~mark,  I "Clothes not busy being worn are busy drying."  --  Dylan, on laundry dayt?           [ singing "It's all right, ma (I'm only bleaching)" ]D   ------------------------------  % Date: Wed, 26 Sep 2001 14:48:15 -0700 ( From: "Greg Cagle" <gregc@gregcagle.com>E Subject: Re: {HP,Compaq} user groups (was Re: CETS2001 presentations)m/ Message-ID: <tr4j93nvgtfj14@corp.supernews.com>e  g "Mark Bartelt" <mark@dino.cacr.caltech.edu> wrote in message news:9otfts$koh$1@dino.cacr.caltech.edu...hJ > (I'm adding comp.sys.hp.hpux to the distribution, even though that's notJ > really the appropriate forum.  But as far as I can tell, there isn't anyI > newsgroup analogous to comp.org.decus for Interex folks to talk to eachn	 > other.)   G I'll chime in here. Regarding Interex discussions, they tend to be heldt> in comp.sys.hp.mpe, which is roughly analagous to comp.os.vms.  B Just to set context, I'm the InterWorks conference co-chair and am@ involved in Interex at a high level (ex-Board member, volunteer,! and a number of other positions).u   > [ JF Mezei ] >bF > >>  Considering that Compaq will cease to exist in a few months, howF > >>  can what-used-to-be-DECUS discuss the future with such details ? > >>F > >>  Have there been any official statements from Carly about support! > >>  for what used to be DECUS ?d  * There has been no mention to my knowledge.  G > Actually, your mention of the merger with (or perhaps, should we say,s: > the getting-eaten-up-by) HP raises an interesting point: >sI > This past May, I attended my first-ever InterWorks meeting.  InterWorks H > is a conference sponsored by Interex (an HP-centric user organization)I > and my first impression was that it really reminded me of the old DECUS F > (say early 70s through mid 80s) which many of us remember so fondly. >dH > I had to admit that part of this was a result of some serious dj vu:H > The conference was held in the very same hotel (the Hyatt Regency, SanH > Francisco) which was the site of the first DECUS symposium (fall 1973) > that I attended. >eJ > But there were other things about InterWorks/2001 which were reminiscentJ > of DECUS/1973 as well:  Rather small in size (~500 attendees, I believe)J > and included sit-down lunches.  But it nonetheless had a nicely eclecticK > assortment of sessions, ranging from very basic to highly technical.  AndoJ > there were some "feedback to HP" sessions, where developers and managersJ > listened to comments from the attendees, and that also brought back fond1 > memories of the "good old days" of early DECUS.o >yI > It will be interesting to see what sort of user organization eventuallyoJ > evolves to meet the needs of current HP users and current Compaq (and ofF > course, former DEC (and other)) users, {if|when} the merger happens.  M Mark, thanks for your kind words about InterWorks. Interex and the conferencewH volunteers work hard to put on this kind of conference, and it's nice to be appreciated 8^).   D Regarding the future of DECUS and Interex, I believe discussions are> being held, but that's about all I know. I myself believe thatE preserving this type of conference is critical to serving the membersm( of whatever organization we end up with.   --
 Greg Cagle gregc at gregcagle dot com   ------------------------------   End of INFO-VAX 2001.537 ************************