1 INFO-VAX	Mon, 23 Jul 2001	Volume 2001 : Issue 406       Contents:( Re: Alpha:  an invitation to communicate( Re: Alpha:  an invitation to communicate1 Re: Alpha:  ostriches and sheep (financials part) ' Re: Alpha: an invitation to communicate - Another reason to READ THOSE RELEASE NOTES... 1 Re: Another reason to READ THOSE RELEASE NOTES...  Basic VMS Question Re: Basic VMS Question Re: Basic VMS Question Re: Basic VMS Question RE: Basic VMS Question RE: Basic VMS Question* Re: Best file spec for incremental backup?* Re: Best file spec for incremental backup?  Can a "White Alpha" box run VMS?& Re: Compaq FUD and lack of information% Does my file has an extension header? ) Re: Does my file has an extension header? ) Re: Does my file has an extension header? $ RE: Golden Opportunity for Microsoft) Re: help mounting drives 3000/300 vms 5.5 ? Re: IA-64 has Little-Endian (was: Re: No chance for OpenVMS...)  Re: IA64 Discussion stuff  Re: IA64 Rocks My World  Re: IA64 Rocks My World  RE: IA64 Rocks My World  Re: IA64 Rocks My World  Ignore this " Re: LPs on the Web (Was: Re: DCPS) MDB Systems Inc % Re: Missing XDPS$DPSLIBSHR.EXE on 7.3  Re: No chance for OpenVMS  Re: No chance for OpenVMS  Re: No chance for OpenVMS  Re: No chance for OpenVMS ? Re: Now we're cooking with gas. (was:  Wailing and moaning....)  Re: OT: Dr Who.  Re: Process adopts Itanium6 Re: Reference for VMS anti-hacker credentials required6 Re: Reference for VMS anti-hacker credentials required6 Re: Reference for VMS anti-hacker credentials required: RESULTS:  Re: Poll: Move OpenVMS to IPF: Agree or DisagreeD RE: Security-Enhancing Products for Compaq OpenVMS from PointSecu	re$ Re: Selling VMS to another company ?$ Re: Selling VMS to another company ?$ Re: Selling VMS to another company ?5 Re: SMS on VMS [was RE: Full port of VMS to Itanium.] 5 RE: SMS on VMS [was RE: Full port of VMS to Itanium.] 1 Re: The death of VMS has been greatly exaggerated 1 Re: The death of VMS has been greatly exaggerated 1 Re: The death of VMS has been greatly exaggerated 1 RE: The death of VMS has been greatly exaggerated F Triggering tasks from network file arrival - Was Re: Basic VMS Questio+ Re: Trying to load VMS on MicroVAX 3100....  Re: V7.3 XFCACHE-W-DATALOSS ? 
 vms backup Re: vms backup Re: Re: vms backup Re: vms backup Re: vms backup Re: Re: vms backup Re: Re: vms backup, Re: VMS remains secure at DEFCON hacker fest Re: VMS Trivia Question  Re: VMS Trivia Question  VS3100 & (dead?) ST51080N  Re: VS3100 & (dead?) ST51080N  Re: VS3100 & (dead?) ST51080N  Re: VS3100 & (dead?) ST51080N P Re: Wailing and moaning.... (was: Reward for the first of the next 50...) 50...) Re: Zero Quadword Time Poll  Re: Zero Quadword Time Poll  Re: Zero Quadword Time Poll  Re: Zero Quadword Time Poll - Re: [DWMOTIF V1.2-6] VAX EURO patch missing ? - Re: [DWMOTIF V1.2-6] VAX EURO patch missing ? - Re: [DWMOTIF V1.2-6] VAX EURO patch missing ? 1 [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ? 5 RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ? 5 Re: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ? 5 RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ? 5 RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ? 5 RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?   F ----------------------------------------------------------------------    Date: 23 Jul 2001 10:31:10 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> 1 Subject: Re: Alpha:  an invitation to communicate H Message-ID: <y4r8v8ujw1.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  / "Richard D. Piccard" <piccard@ohio.edu> writes:   F > > Didn't the US patent law had a clause that independent reinvention, > > allows you to use your invention though?N > No; the standard example that I used to use, to emphasize to my students theO > importance of keeping an accurate lab notebook, was that "Edison invented the N > light bulb" -- known to all because the patent court found in his favor:  heO > could document the critical step (if I recall correctly, the [2.] evacuation, O > after briefly powering up the filament, to outgas adsorbed molecules from its O > surface) five days before Hiram Percy Maxim (who also invented a machine gun) $ > did the same thing, independently.  L However, this principle of "first to invent" has lead to lurid court battlesM and is deemed unworkable in the rest of the world, where "first to patent" is H the giuding principle. It seems likely that his feature of the US patentE system will change in the (near?) future in a realignment with global 	 practice.    	Jan   ------------------------------  % Date: Mon, 23 Jul 2001 15:12:31 +0200 F From: Martin =?iso-8859-1?Q?H=F8yer?= Kristiansen <martin@netimage.dk>1 Subject: Re: Alpha:  an invitation to communicate + Message-ID: <3B5C22BF.516BEB5E@netimage.dk>    Dave Hansen wrote: > 4 > On Fri, 20 Jul 2001 21:36:47 +0200, Terje Mathisen' > <terje.mathisen@hda.hydro.com> wrote:  >  > >Russell Wallace wrote:  > [...]  > >>2 > >> Ah well. (Cue song from 'The Life of Brian'.) > >  > >I assume you're thinking of > > 4 > >       "Always look on the bright side of life" ? > > K > >Not too easy if you're an Alpha architect/supporter/(bigot?) these days.  > >:-( > F > I don't imagine it was too easy for Brian (the character) at the end > of the movie either.  ;-)   B I suppose that Compaq is the 'Judean People's Front, Crack suicide, squad', ie did the Alpha absolutely no good.   Cheers Martin   ------------------------------  % Date: Mon, 23 Jul 2001 19:42:07 +0200 & From: John McLean <mcleanj@dplanet.ch>: Subject: Re: Alpha:  ostriches and sheep (financials part)* Message-ID: <3B5C61EF.A0852006@dplanet.ch>   Hi Bill,  @ Interesting comments you made about Compaq financials.  They canH sometimes be a goldmine of information. You are quite right about the PCC market (personal and commercial) being a consistent money loser and  dragging Compaq down.    A couple of brief comments:   E - In a GOOD quarter PC's seem to return a profit about 1 or 2% of the G money spent on them; in most quarters they lose money and have done for A the last few years.  By contrast Dell are returning about 6-10%.  9 Enterprise stuff from Compaq regularly returns about 14%.   @ - In the Annual Report for 2000 (ie. December) Compaq says aboutE Commercial Personal Computing, "Costs for processors, memory and hard G drives for desktops and portables declined during the year.", and about ? Consumer PC's "Higher component costs also contributed to lower C operating income."  The contradiction is not impossible but it does  deserve an explanation.   D - The April 2001 figures are interesting because Commercial PC's andD Consumer PC's are now under the same umbrella, called "Access".  TheF report says  "The following restated prior period data is provided forH informational purposes as a result of the segment reporting changes madeC in the first quarter of 2001".  (ie. they are posting the last five F quarters of data according to this new classification.)  Summing the 4C quarters of 2000 and comparing it to the stated Annual figures, the H interesting thing is that the Revenue for "Access" is simply an addition< of the old 2 categories but somehow the income to all of theH classifications has been shuffled around.  In particular the income fromF "Enterprise Computing" is down by $400 million (revenue unchanged) andG income from Global Services has risen from $884 million to $944 million = and revenue has changed to $6993 million from $7483 million.  B Interestingly revenue for all PC products (now called "Access") isA unchanged but the income is up from $243 million to $459 million.   D Most importantly the TOTAL INCOME for the four quarters is now $2803F million compared to $3570 million shown in the annual statement.  WhatD has happened to the other $750 million - more than 20% of the profitH that was stated in the annual report ?  If anyone sees it, please let meG know.  (If anything is seriously wrong with my calculations please also  let me know !)  H -  For anyone who thinks that shareholders might be worried about CompaqG it is a realtively easy business to use www.yahoo.com or www.nasdaq.com E to discover who the institutional shareholders are.  Using the Nasdaq B site you can click on their names to see what other holdings theseG people have.  I suspect that Barclays (the largest) is like many of the D others.  Barclays have holdings in IBM, Sun, Dell, HP, Cisco, OracleD ....etc.  Realistically they probably don't care much about which ITD company wins or loses because they are "betting" on all the players.    D Okay, enough of these interesting snippets and back to the matter atH hand.  Yes, the decision to move to Alpha appears to have been primarily motivated by the bean-counters.   1 But it is time for a non-aggressive, "So what ?".   G The decision has been made by Compaq - for better or for worse - and we H just have to try to make the best of it.   Can they back out of the dealF ?  Probably only with extreme difficulty.  Do they want to back out of it ?  Probably very unlikely.   D What we need now to move forward are some reassuring statements from= Compaq and some commitments about how they plan to handle the H transition.  It is not just this newsgroup that wants and/or needs theseG statements, there are a lot of VMS sites who would like to know what is 	 going on.   ? It is now almost a month since the big statement from Compaq, a G statement with little detail and absolutely no follow-up.  I really did " expect something better from them.        John McLean        Bill Todd wrote: > K > It seems to be time to tie together a few issues, and Robert's post below  > touches on some of them. >   C (The full text snipped for brevity, not because it was irrelevant.  - Please find it in the thread if you want it.)    ------------------------------  % Date: Mon, 23 Jul 2001 19:09:33 +0010 % From: paddy.o'brien@zzz.tg.nsw.gov.au 0 Subject: Re: Alpha: an invitation to communicate5 Message-ID: <01K6A4UTG2PE002ZY7@tgmail.tg.nsw.gov.au>    Larry Bohan wrote:  D >On Sat, 21 Jul 2001 19:36:44 +0010, paddy.o'brien@zzz.tg.nsw.gov.au >wrote:   J [And I'm still stuck with this ten minutes -- currently 9 hours from GMT.]  H >>When the company was taken over by Compaq, these key people gradually 
 >disappeared  R >>and not replaced by any equivalent knowledgeable people.  The turnover of their P >>contactors is horrendous, and few have a chance to learn our corporate set-up  >>before they leave. >>N >>Every change scenario that has happened on our corporate machines has meant  >that K >>nothing works correctly for several days.  Their normal solutions to any  
 >problems O >>are at high cost to our company.  One of the few things our corporate IT has   >done B >>right is apparently to vow that never again will Compaq be used. >>P >>Luckily, I manage our departmental machines and am only beholden on corporate ) >>(hence Compaq contractors) for network.  > < >to satisfy my own curiosity,  you refer to both contactors 3 >and contRactors;  are you referring in general to  3 >what the CSC calls Technical Acount Mgrs (tam's),  ; >or higher tier consultants from DEC/CPQ who work onsite?    > 6 >What level of service did your company subscribe to ?  Q The variant without the "R" was a typo.  To the second part, our company took on  I Compaq as contractors with their OMC bid.  Anything extra is billed as a  R "special" -- I forget the terminology, but there are two rates for small or large  additional projects.  P The only people we have on site are the plethora of engineers to look after the N day to day failures of the Mickey$oft machines.  An Operations Manager and an > operator are on site where the corporate machines are located.  P The machines I look after are on a 9-5 for 5 days with about an hour turnaround Q for hardware and software support.  And not accessible by the Compaq contractors.   ' I trust that assuages your curiosity:-)   N O.K., now mine :-)  Every other mail I receive from Info-VAX has the sender's N address (possibly forged) including my own when I send to Info-VAX, but yours L gets to me with the Info-VAX address.  How?  Or is that your forged address?   Regards, Paddy   ------------------------------  # Date: Mon, 23 Jul 2001 14:03:07 GMT 2 From: stephend@pdms99.ksc.nasa.gov (Dave Stephens)6 Subject: Another reason to READ THOSE RELEASE NOTES...2 Message-ID: <3b5c2c3c.323896812@news.ksc.nasa.gov>  C Looks like I just upgraded myself out of viewing Polycenter graphs.   C This one slipped in under the radar. In my exuberance to test drive ? the new VMS version, I installed it on my DEC 3000 workstation. F Oopsie! Now I can't view our CPU reports. A quick Google check of thisF group shows me why - no more support for Display PostScript with MotifA v1.2-6. And do I understand correctly that I can't revert back to A Motif v1.2-5 with VMS 7.3? Ouch! Anyone else out there viewing PS ( files via DECwindows Mail? I am annoyed. David Stephens VMS Systems Administration Payload Data Management Systems  Boeing/Kennedy Space Center ! stephend@xch-bsco-02.ksc.nasa.gov    ------------------------------    Date: 23 Jul 2001 17:51:33 +0200* From: eplan@kapsch.net (Peter LANGSTOEGER): Subject: Re: Another reason to READ THOSE RELEASE NOTES...* Message-ID: <3b5c4805$1@news.kapsch.co.at>  g In article <3b5c2c3c.323896812@news.ksc.nasa.gov>, stephend@pdms99.ksc.nasa.gov (Dave Stephens) writes: D >Looks like I just upgraded myself out of viewing Polycenter graphs.  
 Not really...   D >This one slipped in under the radar. In my exuberance to test drive@ >the new VMS version, I installed it on my DEC 3000 workstation.G >Oopsie! Now I can't view our CPU reports. A quick Google check of this G >group shows me why - no more support for Display PostScript with Motif B >v1.2-6. And do I understand correctly that I can't revert back toB >Motif v1.2-5 with VMS 7.3? Ouch! Anyone else out there viewing PS) >files via DECwindows Mail? I am annoyed.   ? Install GHOSTSCRIPT and display postscript on every X11 display D (like X-terminals, PC with X11-Software) and not only on x11 servers? with DPS (like Q's workstations up to DECwindows-MOTIF V1.2-5).   J In other words, don't rant for DPS, rant for Q prepackaged GHOSTSCRIPT kit on the VMS CD instead...   --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888 < <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------  % Date: Mon, 23 Jul 2001 09:37:37 +0100 * From: Andrew Robinson <arobinson@hspg.com> Subject: Basic VMS Question M Message-ID: <CDA4BAD1E10ED41181AC00508B6051D3C3E2E7@grumpy.internal.hspg.com>   , Please could anyone help with the following:  E I have a procedure running on one of my VMS boxes which is constantly G checking a directory for a file which can be FTPed into it. If the file I appears it then runs another routine to process and remove the file ready J for the next one. I have a simple routine at the moment written in Diabold" which runs as a detached process. L The problem I have is that to make this process work, it is a simple looping2 check, which eats CPU and slaughters the disk IO. K Is there anything in VMS which you can utilize which you can use to trigger 2 an event if a file appears in a specific location.3 I using an Alpha 1200 running OVMS 7.2-1 + patches     Thank you in advance   Andrew Robinson    ------------------------------  # Date: Mon, 23 Jul 2001 10:02:23 GMT . From: Burnie M <burniem.NOSPAM@ozemail.com.au> Subject: Re: Basic VMS Question 8 Message-ID: <ghtnltknqqpmfcscukpdp6i797s9s331ea@4ax.com>  3 On Mon, 23 Jul 2001 09:37:37 +0100, Andrew Robinson  <arobinson@hspg.com> wrote:   - >Please could anyone help with the following:  > F >I have a procedure running on one of my VMS boxes which is constantlyH >checking a directory for a file which can be FTPed into it. If the fileJ >appears it then runs another routine to process and remove the file readyK >for the next one. I have a simple routine at the moment written in Diabold # >which runs as a detached process.  M >The problem I have is that to make this process work, it is a simple looping:3 >check, which eats CPU and slaughters the disk IO. :L >Is there anything in VMS which you can utilize which you can use to trigger3 >an event if a file appears in a specific location.e4 >I using an Alpha 1200 running OVMS 7.2-1 + patches  >s >Thank you in advanceE >T >Andrew Robinson     Nothing springs to mind.0 I'm assuming that you have a wait in this loop ?   $ WAIT 00:03   Cheers,E	 	Burnie M    ------------------------------  % Date: Mon, 23 Jul 2001 08:25:38 -0300r) From: fabio_compaq@ep-bc.petrobras.com.bro Subject: Re: Basic VMS QuestioncL Message-ID: <OF552FADA4.A7A8DC05-ON03256A92.003EB356@ep-bc.petrobras.com.br>   Hellos  K Why dont you use the "WAIT" command  to run the procedure from time to timey ????   Regardse   FC        ; Andrew Robinson <arobinson@hspg.com> em 23/07/2001 05:37:37   6 Favor responder a Andrew Robinson <arobinson@hspg.com>             Info-VAX@Mvb.Saic.Com4       Assunto: Basic VMS Question     , Please could anyone help with the following:  E I have a procedure running on one of my VMS boxes which is constantlyIG checking a directory for a file which can be FTPed into it. If the file I appears it then runs another routine to process and remove the file readynJ for the next one. I have a simple routine at the moment written in Diabold! which runs as a detached process.MD The problem I have is that to make this process work, it is a simple loopingl1 check, which eats CPU and slaughters the disk IO.nK Is there anything in VMS which you can utilize which you can use to triggerE2 an event if a file appears in a specific location.2 I using an Alpha 1200 running OVMS 7.2-1 + patches   Thank you in advance   Andrew Robinson    ------------------------------  # Date: Mon, 23 Jul 2001 11:24:28 GMTS From: LBohan@dbc.spam_less..coma Subject: Re: Basic VMS Question.8 Message-ID: <p22oltsb9sfb2vo9665k016mp80i5j5r19@4ax.com>  3 On Mon, 23 Jul 2001 09:37:37 +0100, Andrew Robinsonh <arobinson@hspg.com> wrote:l  - >Please could anyone help with the following:a > F >I have a procedure running on one of my VMS boxes which is constantlyH >checking a directory for a file which can be FTPed into it. If the fileJ >appears it then runs another routine to process and remove the file readyK >for the next one. I have a simple routine at the moment written in Diabold # >which runs as a detached process.  M >The problem I have is that to make this process work, it is a simple looping13 >check, which eats CPU and slaughters the disk IO.  L >Is there anything in VMS which you can utilize which you can use to trigger3 >an event if a file appears in a specific location.r4 >I using an Alpha 1200 running OVMS 7.2-1 + patches  >w >Thank you in advanceo >a >Andrew Robinson  # ran across this a few years back;  s% (tip of the hat/thanks to James Duff)a' may need some fixups due to line wraps,E     /* **++ ** **  MODULE DESCRIPTION:t **C **      Did you ever write a program that would periodically poll a 	 directory F **      to see if there are files to process and wonder if there was a betterF **      way?  Well, this is it!  This program demonstrates how to take outlC **      what is known as the "Serialization lock" for a file.  This- lock is-F **      described in the "VMS File System Interals" manual.  This lock is takenE **      out before the file system attempts to write to the directoryu file,uB **      and, as this program has already taken it out in exclusive mode,rD **      a blocking ast fires which drops the lock (allowing the file system/ **      to get it) and awakens the main module.U ** **  AUTHORS: ** **      James F. Duffi( **      JDT Computer Systems Pty Limited8 **      P.O. Box R1561 Royal Exchange NSW 2000 Australia **      JDTCOM@DECUS.ORG.AUs" **      Phone: +61 (0) 419 802 714 ** **  CREATION DATE: ** **      05-Sep-1994  ** ** **  ENVIRONMENT: **F **      User and Kernel modes.  SYSLCK and CMKRNL privileges required. ** ** **  MODIFICATION HISTORY:l **/ **      X01-00      Jim Duff        05-Sep-1994 " **      Original version of module ** **-- */   #include <stdio.h> #include <stsdef.h>  #include <ssdef.h> #include <lckdef.h>  #include <descrip.h> #include <dvidef.h>e #include <lib$routines.h>s #include <starlet.h> #include <rms.h>   #define EFN 20   /*C ** Declare a global structure to hold vaiables that the kernel moded ** routines use. */ static struct {      struct {         short int status;n         short int unused;          long int lock_id;      } plock_sb;n     struct {         short int status;e         short int unused;          long int lock_id;8     } clock_sb;      char logname[12];i     short int fid[3];A } kmp;  P /******************************************************************************/ void drop_lock (void) {i /*3 ** This routine is called as an AST in kernel mode.t */   static long int r0_status;       /*"     ** Drop the serialization lock     */8     r0_status = sys$deq (kmp.clock_sb.lock_id, 0, 0, 0);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {o         lib$signal (r0_status);      }      /*     ** Drop the volume locki     */8     r0_status = sys$deq (kmp.plock_sb.lock_id, 0, 0, 0);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {t         lib$signal (r0_status);-     }      /*9     ** Set the event flag the main routine is waiting on.q     */      r0_status = sys$setef (EFN);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {-         lib$signal (r0_status);:     }0 }   P /******************************************************************************/ void get_lock (void) { /** ** This routine is executed in kernel mode */   static long int r0_status;   static struct {i     char prefix[6];i     char logname[12];f! } parent_resource = { "F11B$v" };    static struct {s     char prefix[6];b     short int fid;     char unused;
     char rvn;o  } child_resource = { "F11B$s" };  0 static struct dsc$descriptor parent_resource_d =;                                 { sizeof (parent_resource),r0                                   DSC$K_DTYPE_T,0                                   DSC$K_CLASS_S,5                                   &parent_resource };o  / static struct dsc$descriptor child_resource_d =c:                                 { sizeof (child_resource),0                                   DSC$K_DTYPE_T,0                                   DSC$K_CLASS_S,4                                   &child_resource };       /*0     ** Construct the parent and child resources.     */6     memcpy (parent_resource.logname, kmp.logname, 12);$     child_resource.fid = kmp.fid[0];$     child_resource.rvn = kmp.fid[2];       /*6     ** Obtain a null mode lock on the volume resource.     */     r0_status = sys$enqw (0,'                           LCK$K_NLMODE,a(                           &kmp.plock_sb,'                           LCK$M_SYSTEM,o-                           &parent_resource_d,                            0,                           0,                           0,                           0,                           0,                           0);o-     if (!($VMS_STATUS_SUCCESS (r0_status))) {c         lib$signal (r0_status);t     }f       /*D     ** Obtain the serialization lock for the file in exclusive mode. Note wesC     ** specify the blocking ast parameter as our drop_lock routine.f     */     r0_status = sys$enqw (0,'                           LCK$K_EXMODE,a(                           &kmp.clock_sb,'                           LCK$M_SYSTEM,p,                           &child_resource_d,/                           kmp.plock_sb.lock_id,i                           0,                           0,%                           &drop_lock,3                           0,                           0);d-     if (!($VMS_STATUS_SUCCESS (r0_status))) {h         lib$signal (r0_status);a     }e }t  P /******************************************************************************/
 main (void) {l /*D ** Main routine.  Obtain the name of the directory the user wants to watch.B ** Find the label of the device the file is on, and the FID of the file.  TheseC ** pieces of info are required to format the resource names for thei locks.F ** Get to kernel to obtain the locks and then wait for the AST to fire to ** tell us to go ahead.h */   static struct FAB fab; static struct NAM nam;   static long int r0_status; static long int i;   static char file[255]; static char es[255]; static char rs[255]; static char lockname[13];o   static struct {y     short int length;h     short int code;      void *bufadr;e     void *retlen;d     long int end_list;6 } dviitms = { 12, DVI$_VOLNAM, kmp.logname, NULL, 0 };   static struct {o     long int status;     long int unused; } iosb;a  6 static struct dsc$descriptor file_d = { sizeof (file),6                                         DSC$K_DTYPE_T,6                                         DSC$K_CLASS_S,/                                         file };l  6 static $DESCRIPTOR (prompt_d, "Directory to watch: ");       /*<     ** Get the name of the directory the user wants to watch     */<     r0_status = lib$get_input (&file_d, &prompt_d, &file_d);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {a         lib$signal (r0_status);m     }h       /*?     ** Set up a RMS file access block specifying the file name.t Attach a     ** Name attributes block.e     */     fab = cc$rms_fab;      fab.fab$l_nam = &nam;l     fab.fab$l_fop = FAB$M_NAM;     fab.fab$l_fna = file; (     fab.fab$b_fns = file_d.dsc$w_length;       /*%     ** Set up a name attributes blockt     */     nam = cc$rms_nam;t     nam.nam$l_esa = es;      nam.nam$b_ess = sizeof(es);      nam.nam$l_rsa = rs;      nam.nam$b_rss = sizeof(rs);        /*F     ** Use RMS to parse the file to pick up any fields in the filename"     ** the user failed to specify.     */!     r0_status = sys$parse (&fab); ,     if (!($VMS_STATUS_SUCCESS(r0_status))) {         lib$signal (r0_status);2     }a       /*D     ** Use RMS to search for the file.  This fills in the NAM block,	 includingi     ** the file id.      */"     r0_status = sys$search (&fab);,     if (!($VMS_STATUS_SUCCESS(r0_status))) {         lib$signal (r0_status);      }a       /*+     ** Reuse our filename string descriptoro     */     strcpy (file, rs);(     file_d.dsc$w_length = strlen (file);       /*E     ** Call SYS$GETDVI to obtain the 12 character volume label.  This  is usedeB     ** to construct the resource for the volume lock, which is the parent"     ** of all serialization locks.     */     r0_status = sys$getdviw (0,e                              0,o%                              &file_d,n&                              &dviitms,#                              &iosb,l                              0,m                              0,n                               0);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {          lib$signal (r0_status);t     }f/     if (!($VMS_STATUS_SUCCESS (iosb.status))) { !         lib$signal (iosb.status);i     }e       /*D     ** SYS$GETJPI returns a null filled string in the last call, but we needaD     ** space filled, so loop through the string replacing nulls with spaces.n     */     for (i = 11; i >= 0; i--) {h%         if (kmp.logname[i] != '\0') {o             break;         } else {!             kmp.logname[i] = ' '; 	         }l     }a       /*1     ** Store the file id in the global structure.r     */"     kmp.fid[0] = nam.nam$w_fid[0];"     kmp.fid[1] = nam.nam$w_fid[1];"     kmp.fid[2] = nam.nam$w_fid[2];       /*,     ** Get to kernel and lock the resources.     */*     r0_status = sys$cmkrnl (&get_lock, 0);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {e         lib$signal (r0_status);r     }9       /*E     ** Wait on an event flag.  The blocking AST which drops the locks  will*     ** set this flag so the program exits.     */      r0_status = sys$clref (EFN);-     if (!($VMS_STATUS_SUCCESS (r0_status))) {e         lib$signal (r0_status);      } !     r0_status = sys$waitfr (EFN);w-     if (!($VMS_STATUS_SUCCESS (r0_status))) {w         lib$signal (r0_status);O     }a }o   ------------------------------  % Date: Mon, 23 Jul 2001 13:59:06 +0100a* From: Andrew Robinson <arobinson@hspg.com> Subject: RE: Basic VMS QuestionSM Message-ID: <CDA4BAD1E10ED41181AC00508B6051D3C3E2E8@grumpy.internal.hspg.com>3   > -----Original Message-----+ > From: fabio_compaq@ep-bc.petrobras.com.br . > [mailto:fabio_compaq@ep-bc.petrobras.com.br] > Sent: 23 July 2001 12:26 > To: Info-VAX@Mvb.Saic.Comh! > Subject: Re: Basic VMS Questionn >  >  > Hellos > < > Why dont you use the "WAIT" command  to run the procedure  > from time to time  > ????  H I have a little wait, but the speed that the file requires processing isH critical. Its there to allow an old legacy system to process an internet ordering system. e    > 	 > RegardsP >  > FC > G > I have a procedure running on one of my VMS boxes which is constantlyn> > checking a directory for a file which can be FTPed into it.   F Because the file doesn't exist already, I can't monitor locks etc on aG specific file, but I'm just checking for a file of a certain extension.E   Regards:   Andrew Robinson    ------------------------------  # Date: Mon, 23 Jul 2001 14:10:14 GMT> From: noone@ishere.com (Bob) Subject: RE: Basic VMS Question@: Message-ID: <Xns90E768AB450C8gneissmailroom@63.210.103.21>  I Would the following DCL procedure work for you?  This was posted here in T$ response to my own similar question:   $ fn = "foo.bar" $TRY:eB $ if F$SEARCH(fn) .EQS. "" then goto LOCKED     !has file arrived?D $ open/read/write/error=LOCKED channel 'fn'     !is it done writing? $ close/nolog channel H $ goto RUNJOB                                   !if so, proceed with job $LOCKED: $ wait 00:00:10a5 $ write sys$output "waiting for input file to arrive"n
 $ goto TRY $RUNJOB: $ ......       >> D= >> Why dont you use the "WAIT" command  to run the procedure a >> from time to time >> ????a > I >I have a little wait, but the speed that the file requires processing isDI >critical. Its there to allow an old legacy system to process an internets >ordering system.  >  t >> Q
 >> Regards >> T >> FCk >> nH >> I have a procedure running on one of my VMS boxes which is constantly? >> checking a directory for a file which can be FTPed into it. 4 > G >Because the file doesn't exist already, I can't monitor locks etc on a H >specific file, but I'm just checking for a file of a certain extension. >- >Regards >- >Andrew Robinson >e   ------------------------------    Date: 23 Jul 2001 08:18:08 -0700- From: afeldman@gfigroup.com (Alan E. Feldman)-3 Subject: Re: Best file spec for incremental backup?t= Message-ID: <af1e4ce6.0107230718.6cd34537@posting.google.com>:  G zessin@decus.de wrote in message news:<009FF59D.0FE66841.1@decus.de>...  > Tom Adams wrote:= > > We had been using [*...] as the file spec for incrementalt1 > > backups that might be used to recover a disk.e > > ( > > Is [000000...] a better choice? Why? > ? > I beleive it is a better choice, because this will find fileso: > in the master file directory (eg. [000000]ABC.DAT) which9 > [*...] does not (at least not the last time I checked).e >   D You may also want to add /FAST to your incremental backup command. IA have twice inherited incremental backup scripts that did not takeoB advantage of this qualifier. At one job the daily incremental wentC from about 45 to 50 minutes without /FAST to about 15 to 20 minutes F with. I don't remember the amount of performance increase at the otherF job, but it was also very good. Therefore, your incremental backup may2 run 2, 3, or more times faster with this. O, YMMV.   Disclaimer: JMHO   &-) afeldman@gfigroup.comh   ------------------------------  # Date: Mon, 23 Jul 2001 15:28:22 GMTI2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)3 Subject: Re: Best file spec for incremental backup?i3 Message-ID: <qoX67.1099$rc5.72576@news.cpqcorp.net>n  c In article <793af3df.0107200921.72861d7f@posting.google.com>, tadamsmar@aol.com (Tom Adams) writes:a: :We had been using [*...] as the file spec for incremental. :backups that might be used to recover a disk. :s% :Is [000000...] a better choice? Why?   J http://www.openvms.compaq.com/doc/73final/6017/6017pro_044.html#6017backupL http://www.openvms.compaq.com/doc/73final/6017/6017pro_048.html#inc_disk_sec    N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  % Date: Mon, 23 Jul 2001 12:10:22 +0200Q" From: "Hans Vlems" <hvlems@iae.nl>) Subject: Can a "White Alpha" box run VMS?c( Message-ID: <9jgsv2$bko$1@news.IAEhv.nl>  8 A friend of mine has an unused Alpha sitting on a shelf.8 It's a white box Alpha, and IIRC it was sold to run WNT.> Is it possible to boot VMS on it, and if so what modifications are needed?n  
 Hans Vlems   ------------------------------  % Date: Mon, 23 Jul 2001 17:40:27 -0000l- From: wspencer@ap.nospam.org (Warren Spencer) / Subject: Re: Compaq FUD and lack of informationh/ Message-ID: <tloocb3mag0e09@news.supernews.com>l  I mcleanj@dplanet.ch (John McLean) wrote in <3B587F23.8120049C@dplanet.ch>:i  
 -- snip --   >eE >Can you please urge the people responsible in Compaq to start makingm% >some confidence-building statements?  >e  H I can do without the hummer, but like you, I'm anxiously awaiting facts.   ws   -- o   Warren Spencer Senior Software Engineer The Associated Press  L ** My employer does not necessarily agree with my statements - neither do I  **   ------------------------------   Date: 23 Jul 2001 13:21:19 GMT@ From: matthias.maisenbacher@de.bosch.com (Matthias Maisenbacher). Subject: Does my file has an extension header?6 Message-ID: <9jh8cf$a7j$1@proxy.fe.internet.bosch.com>   Hello,  6 our disk ran into the problem that we still had enough9 file entries "Maximum # files" in indexf.sys, but already7, filled up the space with additional headers.   Read:   6 Maximum # files: 733725 (according DFU report command)6 Free headers:  1000 (yes, we already was at zero here)/ Number of files on disk is no more than 530000.h  6 The documentation says that this extension headers are used if " a) the disk is quite fragmented or b) ACLs are heavily used.I  5 Fragmentation is at 1,005 (according to DFU) and, yes + we do use many ACLs (in average 8 per file)o  ? While we have to wait for a timeslot to increase the indexf.sysn' (which cost 3 days of backup and back)  : we have to live with this free blocks as long as possible.  4 So we try to remove some ACLs. But for testing this,6 we need to know, WHEN our ALCs are less enough to fit  in one header.  3 So, is there a way to find out whether a given fileo= uses just one header or jumps over the boundary and uses two./   As usual, thank you in advance         Matthias Maisenbacher         ------------------------------  % Date: Mon, 23 Jul 2001 10:50:15 -0400r- From: "Peter Weaver" <peter.weaver@stelco.ca> 2 Subject: Re: Does my file has an extension header?4 Message-ID: <HQW67.267332$Z2.3247350@nnrp1.uunet.ca>  E "Matthias Maisenbacher" <matthias.maisenbacher@de.bosch.com> wrote in 8 message news:9jh8cf$a7j$1@proxy.fe.internet.bosch.com... >...5 > So, is there a way to find out whether a given fileT? > uses just one header or jumps over the boundary and uses two.o >...  D I believe "DFU SEARCH device /MULTIPLE" will give you what you want.  K "DFU SEARCH device /FRAG=MIN=2" will show you all files that are fragmentedcJ and if the fragmented file has multiple headers then you will also see theJ number of headers. But if the file has multiple headers but it is actuallyK not fragmented (i.e. a .DIR in a Pathworks share) then the /FRAG=MIN=2 will K not report it. This is a file with multiple headers and multiple fragments;NH          DRA0:[SYSTEM_SUPPORT.DSN30]DSNLINK030.C;1      10278/10287 3/15J The 15 is the number of fragments and the 3 is the number of headers. "DFUL DEFRAG" will try to reduce both the fragmentation and the number of headers.  L "DUMP/HEAD/RECORD=COUNT=0 filename" will also show the number of headers andJ will show you the actual ACE that is in each header. BTW: I just looked atL one particular file here and each ACE took one header, I do not know if that is standard or not.      -- Peter WeaverJ Using a WIN NT/WIN 2000 box to manage your VMS systems is like towing your7 mechanic in a 5th wheel motor home behind your Porsche.    ------------------------------  % Date: Mon, 23 Jul 2001 10:59:05 -0400c2 From: rdeininger@mindspring.com (Robert Deininger)2 Subject: Re: Does my file has an extension header?L Message-ID: <rdeininger-2307011059060001@user-2ive6lk.dialup.mindspring.com>  6 In article <9jh8cf$a7j$1@proxy.fe.internet.bosch.com>,A matthias.maisenbacher@de.bosch.com (Matthias Maisenbacher) wrote:n      5 > So, is there a way to find out whether a given file ? > uses just one header or jumps over the boundary and uses two.  >   > As usual, thank you in advance   $ DUMP/HEADER/BLOCK=COUNT:0V  + ... will show you the header(s) for a file.$   -- ( Robert Deininger rdeininger@mindspring.com    ------------------------------  % Date: Mon, 23 Jul 2001 09:15:58 -05000+ From: Christopher Smith <csmith@amdocs.com>)- Subject: RE: Golden Opportunity for Microsoft L Message-ID: <3B55D7F383B0D31197D9009027541CBF1170DA4F@cmiexch1.cmi.itds.com>   > -----Original Message-----2 > From: Christof Brass [mailto:brass@infopuls.com]  8 > Which ones? Are there any? Also with Solaris are there# > C-libraries without memory leaks?*  L I'm pretty sure there are some functions -- whole libraries, though, I can't	 say... :)e   Regards,   Chrisl    ! Christopher Smith, Perl Developer  Amdocs - Champaign, IL   /usr/bin/perl -e '? print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");p 'i  ]   ------------------------------  # Date: Mon, 23 Jul 2001 05:00:45 GMTe From: adroso@home.com (ADR)r2 Subject: Re: help mounting drives 3000/300 vms 5.59 Message-ID: <3b5b63fb.61572603@news.jamison1.pa.home.com>   % First, determine the volume-label ...   - $ MOUNT R1SQUA$DIA0: /OVERRIDE=IDENTIFICATION  $ SHOW DEVICE DIA0:   A Write down the volume-label to reference later.  Then, use a textr9 editor to open up either SYS$MANAGER:SYSTARTUP_VMS.COM orl= SYS$MANAGER:SYSTARTUP_V5.COM (whichever you have) and add the F following line at the end (if you desire a logical-name [such as DISK0C or KYLE, etc.], include that too).  Be very careful of typos in the & device and volume-label parameters ...  6 $ MOUNT /SYSTEM R1SQUA$DIA0: volume-label logical-name  B The logical-name is optional, but it may make referencing the diskD easier in COMs, etc.  The /SYSTEM qualifier makes the mounted device7 public (that is, available to all users on the system).$  : Once you're done editing SYS$MANAGER:SYSTARTUP_VMS.COM (orD SYS$MANAGER:SYSTARTUP_V5.COM, as may be the case), save the file and reboot.   % See $ HELP MOUNT for additional info.     ( cassidy@netaxs.com (Kyle Cassidy) wrote:  E >howdy kids -- i just got vms 5.5 installed on a vax 3000/400 and i'm K >having a dilly of a time mounting the drives .... when i try (for example)  >$ >mount r1squa$dia0:  > K >i'm asked for a label and a log file, i make some up, get a request on themJ >screen to physically mount the disk (whih isn't what i'm looking for) andH >nothing .... i'm assuming i need to mount the disks before i can access
 >them... yes?  > $ >thanks for your time and expertise, >  >kc  >    ------------------------------  # Date: Mon, 23 Jul 2001 15:41:41 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)H Subject: Re: IA-64 has Little-Endian (was: Re: No chance for OpenVMS...)3 Message-ID: <VAX67.1101$rc5.72742@news.cpqcorp.net>   S In article <3B5A89E0.6C170D2B@dplanet.ch>, John McLean <mcleanj@dplanet.ch> writes: ? :No not at all.  I am talking about using the recovered files. *  3   BACKUP will be available, and will be compatible.   H :Exectuable images for VMS on Alpha may be unusable on IPF and trying to) :rebuild a scenario will be impossible.  t  E   That is nothing different than the present VAX to Alpha operations.i  H   I know that we have been receiving at least a little feedback and someF   questions around any plans for binary translation and/or for binary I   emulation, but I'll have to defer an answer to those questions for now.t  ! :IIRC, we also have an issue with H :big-endian and little-endian data with Alpha and IPF.  Does this go all" :the way through to files on disk?  E   IPF operates little-endian.  (And IPF, like Alpha microprocessors, u'   can also operate on big-endian data.)e  B   IPF also has capabilities for the requsite four processor modes.  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  , Date: Mon, 23 Jul 2001 12:11:29 +0200 (CEST): From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>" Subject: Re: IA64 Discussion stuffI Message-ID: <Pine.LNX.4.21.0107231154310.5962-100000@irys.stanpol.com.pl>     Hello !  , On Tue, 17 Jul 2001, Freddy Meerwaldt wrote:   >+Hi there!a >++ >+I keep on receiving 100s of E-Mails [...] 5 >+Would it be possible to get back to on-topic stuff?l' >+OK - IA64 is not completely off-topic.    Really ! :)  % >+ but it is only a Hardware Platformt* >+where Compaq said they will port VMS to.    "Only" ? Great...  J >+Sad, that they're replacing the Alphas with that, but as this is a final2 >+decision, much talking doesn't help any further.  <  Personally will disagree. The paging protection scheme (and9 limitation from IA doc pointed here) is definitely in thef
 proper class. =  Agree, some "political only" opinion (may be: *most* in some0< subject) are found. But... "dosn't help" may be IMHO wrong !9  May *not change* the COMPAQ decision, but *may help* theh2 peoples here re-think what may expect in future...  7  Will not reject you the right having separate opinion.U8  Will not say, that you can filter anything uninteresant6 (because that may be problem for you, if you b.ex. pay6 for internet-connection-time) but will apeal, that you6 accept the fact that *multiple* peoples from the group8 participate in the discuission. At least some of it you 7 must know as competent in VMS at all. The point may be k- a reason to classify the subject as "proper".t5 (och, really, any something-war also produces loooong   subject, I am aware... !)  J >+Please: Let IA64 be IA64 and Alpha be Alpha and VAX be VAX and let's getE >+back to the Operating System VMS, for which this group was created.   9  Och, from time to time we will go *FORWARD* to O.S.VMS !   -:)  ( >+I expect _no_ replies on this posting.    Well.7  We all - at least from time to time - make mistakes ;)S/ (please don't flame me for the little malice !)n   >+Greetings - Freddy    Best regards - GotfrydT8 PS: Sure, I must self-check and self-limit participation;  of subject like the "politics" and "war", because in otherc&  way will end with time-debt :( But... -- =E ===================================================================== F $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=ME . $!                        GS@stanpol.zabrze.plE =====================================================================f   ------------------------------  % Date: Mon, 23 Jul 2001 13:00:45 +0100a0 From: andrew harrison <andrew.nospam@uk.sun.com>  Subject: Re: IA64 Rocks My World* Message-ID: <3B5C11ED.2BCDDF5E@uk.sun.com>   "Main, Kerry" wrote: >  > Chris, > E > >>> People do use Oracle's OPS, but they use it for reliability notoK > performance.  Typically the other "node" is just a hot standby incase the G > main system goes down.  This allows the downtime to be minimal and is 5 > absolutely needed technology to run a 24x7 shop.<<<$ > K > The Oracle hot standby option is a totally different config than OPS. HotlN > Standby is a replication scheme that does not do load balancing and requiresM > server HW to be standing by waiting for something to happen on the primary.: >   G Yes and no, you can of course be servicing other DBMS's on the standby y/ node and this is quite a common configuration. -  M > The following is feedback from an Oracle Consultant on where you might wantt > to use Oracle OPS: > N > 1)      Load Balancing - Balancing the application requests across availableL > servers and database instances, increased scalability and use of available# > hardware resources is obtainable.? >   7 But Oracle generally only recommend this when you have p7 exhausted the performance you can get from a single boxs8 because scalability is not guaranteed and it costs more.  L > 2)      Parallel Query Support - OPS allows the customer to take advantageJ > of Oracle's Parallel Query feature.  In a parallel server environment, aN > single parallel query can spread the workload across all instances connected > to the database.  9 This is a self fufilling prophecy. The fact that you are a6 using OPS means that Parallel Query has to be able to 9 use multiple instances otherwise you don't have a chance e< of getting any scaling. If you don't need to use OPS because< you do have a scalable server (not WildFire) then you don't  need the feature.    > G > 3)      Decision Support - OPS is very well suited to query intensive L > applications that only occasionally modify data. An example is a financialJ > transaction that is queried most of the time and has new data added only > occasionally.s > L > As can be expected, there are down sides as well and there are cases where' > Oracle OPS is not a good strategy ie.r > L > 1)      OPS is not advantageous when there is contention between instancesC > on a single block in the database as performance can be impacted.e > L > 2)      OPS is not advantageous when there is a significant amount of dataG > synchronization overhead (network traffic and CPU processing) betweendI > systems with hardware that is inadequate to meeting the traffic and CPUR > requirements.F  = 3) OPS isn't advantageous when the apps vendor that is using i9 Oracle as a storage subsystem does not support their app c7 running on OPS, as was the case with Oracle themselves.s  B 4) OPS isn't advantageous if you are only using it for performance< reasons and by choosing a more scalable server platform you  could avoid using it.e     > J > Now, if you want to talk real games, look at the 400K-700K clustered TPCM > numbers on NT (both IBM and Compaq) - over 6,000 disk drives with up to 250 K > local scsi adapters and NO RAID .. cost $15M - now there is an example ofe > real life config's.m >   7 Dangerous territory, these cluster numbers exploit the t6 same warehouse tuning method that the cluster in a box5 scheme used on the WildFire does. Who is to say that  5 their results are any less relevant than the GS320's.   5 If you beleive as you obviously do that the GS TPC-C f2 result is in any way predictive of real customers 5 application performance, then you will have a really  + hard time attacking the MS cluster results.n   RegardsA Andrew Harrisone Enterprise IT Architecta   ------------------------------  % Date: Mon, 23 Jul 2001 13:19:03 +0100 0 From: andrew harrison <andrew.nospam@uk.sun.com>  Subject: Re: IA64 Rocks My World* Message-ID: <3B5C1637.B5489DE3@uk.sun.com>   Fred Kleinsorge wrote: > I > OK.  I'm, not a database expert, but why is this a bad thing?  If TPC-ChH > truly measures relative performance that could be expected in the realI > world - and if partitioning the system gets you this performance - then-H > what's the big deal?  Tru64 clusters do provide a single system image,J > making it relatively easy to administer.  On the other hand, if TPC-C isM > just another meaningless benchmark number - then lets stop talking about ita+ > altogether - or define a meaningful test.g >   1 What would you say if a vendor did the following a with a standard benchmark.  7 1. Pretend that a clustered system with multiple nodes c= and multiple database instances is equivalent to a similarly p. performing single node/single instance system.  < 2. Pretend that they have non clustered system when in fact < they have a single system image but with multiple instances 7 because without multiple instances the system does not - perform.  = 3. Compare favourably this system with an equally performant m9 system which used one system image and one DBMS instance.-  ? The fact that your employer has done all three of these things e7 tends to inject a note of unintentional irony into your-; "just another meaningless benchmark number" comment. CompaqH2 has hardly done TPC-C any favours in this respect.   regardsnB > Chris Ruemmler wrote in message <9j00of$c44@cello.hpl.hp.com>...- > >In article <3B45E1CD.FBDBC266@us.ibm.com>,a, > >Greg Pfister  <pfister@us.ibm.com> wrote: > >>andrew harrison wrote:
 > >>[snip]< > >>> The problem for WildFire is that it was late but  when5 > >>> it was announced it wasn't competitive with the1> > >>> previous generation of servers from Compaqs competitors. > >>>:9 > >>> It recently reached the dizzy heights of 230,000 onc6 > >>> TPC-C, using the latest 1001 Mhz CPU's, a result9 > >>> puts it just ahead of the IBM P680, a machine thatss+ > >>> just about to be replaced by Regatta.  > >>>i5 > >>> When Rob Young started his WildFire boosting inn6 > >>> the dark and dim past initial numbers of 200,0006 > >>> were being bandied about, I can only assume that6 > >>> this number was a design goal that was leaked to3 > >>> Rob (unless he made it up). The fact that thei2 > >>> initial numbers came out way lower than that3 > >>> and then needed OPS in a box tends to suggest * > >>> that they missed their design goals. > >>E > >>Do you know for a fact that the OPS-in-a-box solution was used in-E > >>Wildfire? That's often rather difficult to tell from official TPCn@ > >>disclosures. If it was used, it of course makes the high TPC8 > >>results rather meaningless (in my personal opinion). > >>C > >Yes, all of the GS320 TPC-C numbers have been a cluster in a canUJ > >with OPS except the very first one they published and withdrew (becauseD > >it was so poor in terms of performance).  If you look at the full > disclosureK > >reports you'll see 8 configuration files for Oracle in the back, one forrH > >each instance.  In addition, their "clients" are always configured inC > >multiples of 8 which is another hint that partitioning was used.i > > G > >It is sad that Compaq had to resort to TPC-C tricks like this to trybD > >and show performance on the box.  I imagine the "real" single boxE > >performance is not terrible, just not leadership.  The NUMA hit onWD > >that box is really much too large to run TPC-C really well with a > >single instance.h > >g
 > >--Chris > >My own viewsd   -- u Andrew Harrisond Enterprise IT Architectu   ------------------------------  % Date: Mon, 23 Jul 2001 09:38:20 -0400-+ From: "Main, Kerry" <Kerry.Main@compaq.com>0  Subject: RE: IA64 Rocks My WorldR Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4AD7F25@kaoexc01.americas.cpqcorp.net>   Andrew,D  J >>> Yes and no, you can of course be servicing other DBMS's on the standby2 node and this is quite a common configuration. <<<  L Sure, however, assuming you have other apps and assuming want fast fail-overJ times, then all the system and database parameters on the secondary serverB must be setup to handle the mission critical database when it doesK fail-over. This may not be the best thing for these other non-critical DB'sl
 performance. e  H Of course, these DB's would likely not be available during the fail-overG period if they in any way impact the mission critical applications. The J users on that secondary server would likely be kicked off very abruptly as well.N  H Then there are the issues of fail-back ie. when to do it since down time will be required etc.A  H >>> But Oracle generally only recommend this when you have exhausted theD performance you can get from a single box because scalability is not  guaranteed and it costs more.<<<  I OPS can (depends on application, as previously noted) partition the loadslK and when one server goes down, the other server automatically keeps handing J any requests (no fail-over, no fail-back, no restarting of applications.)   < Does it cost more - you bet (I really agree with you here).   L Bottom line is that with many applications in todays world,  availability isJ just as important as performance. Imho, what most Customers want today areK solutions that can easily handle peak loads but have zero downtime when one < node fails or needs to be taken offline for whatever reason.  I >>> If you beleive as you obviously do that the GS TPC-C result is in anyu! way predictive of real customers  L application performance, then you will have a really hard time attacking the MS cluster results.<<<  H My point is that everyone (all vendors) take advantage of the latest TPCF "techniques" they can to wring the most performance out of whatever HW config they have.   J The GS320 was designed to be a partitioned or large instance single serverH and Oracle OPS is designed for multiple partions / OS instances for loadL balancing, scalability and availability. This TPC benchmark simply used what both were designed for.   E And before you start with the but, but, but ... The single server TPCs numbers are NO different.   J Take 1,400-1,700 disk drives with no RAID (6,000+ drives on NT benchmarks)H and then HAND partition the TPC database across all these drives and, inI some cases, 100+ local scsi adapters (250 on the NT TPC numbers), so that F there is absolutely minimal database contention on any drive/adapter.   G And this is supposed to be a real world application representation? AlloG vendors do it (including Compaq), but please - most med-large Customers"= understand the vendor games that go on with these benchmarks.-  E So, while your thoughts on what a real mans performance benchmark areoD interesting, remember that what Customers want with their real world+ solutions are performance AND availability.p   Regards   
 Kerry Main Senior Consultanta Compaq Canada Inc. Professional ServicesI Voice: 613-592-4660W Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----7 From: andrew harrison [mailto:andrew.nospam@uk.sun.com]o Sent: July 23, 2001 8:01 AMC To: Info-VAX@Mvb.Saic.Com   Subject: Re: IA64 Rocks My World     "Main, Kerry" wrote: >  > Chris, > E > >>> People do use Oracle's OPS, but they use it for reliability not K > performance.  Typically the other "node" is just a hot standby incase themG > main system goes down.  This allows the downtime to be minimal and ise5 > absolutely needed technology to run a 24x7 shop.<<<l > K > The Oracle hot standby option is a totally different config than OPS. HotsE > Standby is a replication scheme that does not do load balancing andY requiresD > server HW to be standing by waiting for something to happen on the primary. >   G Yes and no, you can of course be servicing other DBMS's on the standby h/ node and this is quite a common configuration. x  H > The following is feedback from an Oracle Consultant on where you might want > to use Oracle OPS: > D > 1)      Load Balancing - Balancing the application requests across	 availablerL > servers and database instances, increased scalability and use of available# > hardware resources is obtainable.i >   7 But Oracle generally only recommend this when you have  7 exhausted the performance you can get from a single box 8 because scalability is not guaranteed and it costs more.  L > 2)      Parallel Query Support - OPS allows the customer to take advantageJ > of Oracle's Parallel Query feature.  In a parallel server environment, aD > single parallel query can spread the workload across all instances	 connectedh > to the database.  9 This is a self fufilling prophecy. The fact that you are  6 using OPS means that Parallel Query has to be able to 9 use multiple instances otherwise you don't have a chance o< of getting any scaling. If you don't need to use OPS because< you do have a scalable server (not WildFire) then you don't  need the feature.    > G > 3)      Decision Support - OPS is very well suited to query intensiveeL > applications that only occasionally modify data. An example is a financialJ > transaction that is queried most of the time and has new data added only > occasionally., > L > As can be expected, there are down sides as well and there are cases where' > Oracle OPS is not a good strategy ie.  > L > 1)      OPS is not advantageous when there is contention between instancesC > on a single block in the database as performance can be impacted.p > L > 2)      OPS is not advantageous when there is a significant amount of dataG > synchronization overhead (network traffic and CPU processing) betweenlI > systems with hardware that is inadequate to meeting the traffic and CPU- > requirements.s  = 3) OPS isn't advantageous when the apps vendor that is using a9 Oracle as a storage subsystem does not support their app -7 running on OPS, as was the case with Oracle themselves.1  B 4) OPS isn't advantageous if you are only using it for performance< reasons and by choosing a more scalable server platform you  could avoid using it.5     > J > Now, if you want to talk real games, look at the 400K-700K clustered TPCI > numbers on NT (both IBM and Compaq) - over 6,000 disk drives with up to  250lK > local scsi adapters and NO RAID .. cost $15M - now there is an example ofl > real life config's.  >   7 Dangerous territory, these cluster numbers exploit the s6 same warehouse tuning method that the cluster in a box5 scheme used on the WildFire does. Who is to say that t5 their results are any less relevant than the GS320's.n  5 If you beleive as you obviously do that the GS TPC-C c2 result is in any way predictive of real customers 5 application performance, then you will have a really "+ hard time attacking the MS cluster results.m   Regards) Andrew Harrison. Enterprise IT ArchitectO   ------------------------------  % Date: Mon, 23 Jul 2001 18:30:38 +0100e0 From: andrew harrison <andrew.nospam@uk.sun.com>  Subject: Re: IA64 Rocks My World* Message-ID: <3B5C5F3E.9FBE7852@uk.sun.com>   "Main, Kerry" wrote: > 	 > Andrew,c > L > >>> Yes and no, you can of course be servicing other DBMS's on the standby4 > node and this is quite a common configuration. <<< > N > Sure, however, assuming you have other apps and assuming want fast fail-overL > times, then all the system and database parameters on the secondary serverD > must be setup to handle the mission critical database when it doesM > fail-over. This may not be the best thing for these other non-critical DB'sp > performance. >   G Why would it effect the other DB's running on the system. They are all oD Oracle DBMS's, what specifically at the system level would you tune " for one Oracle DBMS over another.   J > Of course, these DB's would likely not be available during the fail-overI > period if they in any way impact the mission critical applications. ThegL > users on that secondary server would likely be kicked off very abruptly as > well.i >   F Whats your point exactly you only have so much resource to play with. B Its obvious that if you lose one node in a cluster then the other D nodes are going to be more loaded. OpenVMS/Tru64/Solaris/MS clusters+ all behave in the same way in this respect.   D And in any cluser environment you may choose to kick the user of theE non-critical apps off if you are running with less than full systems fE capacity. Of course you could use the resource managers to wind down 03 their share of the system without kicking them off.>    J > Then there are the issues of fail-back ie. when to do it since down time > will be required etc.  > J > >>> But Oracle generally only recommend this when you have exhausted theF > performance you can get from a single box because scalability is not" > guaranteed and it costs more.<<< > K > OPS can (depends on application, as previously noted) partition the loadsyM > and when one server goes down, the other server automatically keeps handingeK > any requests (no fail-over, no fail-back, no restarting of applications.)o >   @ But you are selling a benefit that TPC-C does not measure. TPC-C measuresB price and performance, it does not measure resiliance. All this HA discussion r? is a smoke screen, Compaq didn't use OPS because they wanted tor demonstrate E HA, they used it because if they hadn't they would not have acheived w a competive throughput.:  ; In addition it bullshit when used to justify your use of ite< for TPC-C because the Compaq configuration wasn't resiliant,= there was only one node, the failure of that node, which OPS b= can mitigate against in an N+1 cluster would cause the whole   DBMS to go offline.   > So in fact you demonstrated the worst of both worlds, you have; a more complex, harder to manage, more costly to own systemQ1 without the benefit of any improved reliability. n  ; If you had wanted to provide HA you would have used 2 nodese9 say 2 x GS160's (does OPS support a mixture of instances r$ in a node and remote instances ???).  G > So, while your thoughts on what a real mans performance benchmark areaF > interesting, remember that what Customers want with their real world- > solutions are performance AND availability.e >   4 So why are you touting the GS320 TPC-C results ?????     Regardsv Andrew Harrisone Enterprise IT Architect    ------------------------------  / Date: Mon, 23 Jul 2001 08:24:45 +0200 (MET DST)d& From: Rudolf Wingert <win@fom.fgan.de> Subject: Ignore this6 Message-ID: <200107230624.IAA03402@sinet1.fom.fgan.de>   Hello,  F i did not get any message since Friday. So I will send this message to& the INFOVAX and waiting for come back.   Regards Rudolf Wingert   ------------------------------  % Date: Mon, 23 Jul 2001 08:32:20 -0700a' From: David Mathog <mathog@caltech.edu>e+ Subject: Re: LPs on the Web (Was: Re: DCPS)b+ Message-ID: <3B5C4384.C0B32C5E@caltech.edu>)   Freddy Meerwaldt wrote:u   > Hi!e >iK > >   Now that DCPS licenses are included with VMS, how about changing thatn > > andcJ > > also including them on the VMS distribution CDs?  The license for Java > > JDKsC > > are also included with VMS and they are distributed on the web.e >rL > Why doesn't Compaq make all Layered Products available for Download on the > web?  " Pick one or more of the following:  9 Because the company is run by a pack of drooling morons?? Q Because the ability to do so is only five or six years old and they need to studyo it for3 another decade before they can make up their minds? L Because the lawyers have not been able to come up with a contract that would allow such a downloadrN but would restrict the use of  the software to a single user machine, and then only for the purpose ofl heating a room?   P They  also charge an awful lot for the CONDIST, so they may consider it a profit center.t   >h5 > Everyone who doesn't have the license can't use it.hG > But there are Hobbyist Users (like me), who do have the PAKs from the 7 > Hobbyist Program - so I may use the software legally.f, > But it's very hard to get to the software.  Q A point which I made emphatically in several messages to the company with regardsh to the farce of new education program.w  L Stop expecting Compaq to do anything that in any way benefits the current or potential OpenVMS:0 customer and you won't be disappointed any more.   Regards,   David Mathog mathog@caltech.eduL **************************************************************************** *I            RIH   Compaq, *tL ****************************************************************************   ------------------------------  % Date: Mon, 23 Jul 2001 10:47:24 +0200g, From: "Dave McDonald" <browser@acenet.co.za> Subject: MDB Systems Inc- Message-ID: <9jgoac$sfv$1@ctb-nnrp2.saix.net>   
 Hi people,? Please excuse the cross-post, not sure where I'll find the mostm knowledgeable on this subject.  I I'm looking for support material - schematics, manuals, diagnostics - for . MDB Systems Inc MLSI-DR11-W, and MDB/MLSI-WLL.  H The first is a Qbus equivalent of DEC Unibus DR11-W (16 bit parallel DMA I/O).t= The second is a long line driver that attaches to the DR11-W.a  K I have a customer still using these in a production environment, so willing  to pay reasonable costs.  % Any help will be greatly appreciated.s   Thanks and regards  
 Dave McDonald 
 Johannesburg.    ------------------------------    Date: 23 Jul 2001 08:24:21 -0700& From: jordan@ccs4vms.com (Rich Jordan). Subject: Re: Missing XDPS$DPSLIBSHR.EXE on 7.3= Message-ID: <cc5619f2.0107230724.4d5b2249@posting.google.com>t  q > In article <FDCKCwGx4xj1@eisner.encompasserve.org>, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) writes: T > >In article <M1wUNFrREJ5I@mcrc16.med.nyu.edu>, smithp01@mcrc16.med.nyu.edu writes:R > >> Don't ask why, but I still like to use DECWRITE to edit old documents on the / > >> workstation rather than MS Word on my Mac.e > >> sS > >> However, after upgrading to 7.3 DECWRITE won't open any more due to a missing   > >> file... > >> aR > >> -CLI-E-IMAGEFNF, image file not found MCRC16$DKA100:[SYS0.SYSCOMMON.][SYSLIB] > >> XDPS$DPSLIBSHR.EXEs > > B > >Display Postscript was dropped from VMS V7.3 due to contractual > >requirements from Adobe.P >   C Does this mean that DECWrite is effectively nonfunctional on a V7.3 E system?  Any possibility of a fix for this barely supported product? tA I use it too (hobbyist, so not exactly paying for support), whichnE means an upgrade to V7.3 at home just got very unlikely.  I knew we'diA lose DPS, but not that layered products would also be effectivelye lost.a   Rich Jordane rjordan@mcs.netu   ------------------------------  # Date: Mon, 23 Jul 2001 07:15:53 GMTw. From: "Duane Sand" <duane.sand@mindspring.com>" Subject: Re: No chance for OpenVMSA Message-ID: <JaQ67.243252$%i7.135959481@news1.rdc1.sfba.home.com>t   > Duane Sand wrote:l> > > ... Alpha and IPF both support both big- and little-endian? > > operating systems and applications.  (Alpha belatedly addedf$ > > this about five years ago.)  ...   John Santos repliedn> > Are you sure?  ISTR that Alpha was both-endian from day one.  ; The first two generations of Alpha chips lacked the current : way to get XOR'ing of low address bits for free.  On those? first chips, "big endian" software could run only by using justo@ one size (eg 64 bits) for all loads and stores, or by explicitly3 adding address XOR steps to the instruction stream.i. Cray must have done one or the other, or both.  9 I'm not sure just when the current big-endian support was 3 added; it was not there when Tandem evaluated usingr: the Alpha chip instead of Mips, in Alpha's early merchant-7 chip-marketing push and again in 1996.  The support wase8 present by 1998, when Pfeiffer forced Tandem to halt the= IPF-targeted work in favor of a Alpha target, to help Alpha'ss marketing msg.   ------------------------------    Date: 23 Jul 2001 09:01:03 -0500- From: koehler@encompasserve.org (Bob Koehler)-" Subject: Re: No chance for OpenVMS3 Message-ID: <NssfXasE4+E2@eisner.encompasserve.org>m  U In article <3B5AD93D.5E808FA6@iee.org>, "antonio.carlini" <arcarlini@iee.org> writes:@ >  >  > John McLean wrote:K >> rebuild a scenario will be impossible.  IIRC, we also have an issue withnJ >> big-endian and little-endian data with Alpha and IPF.  Does this go allJ >> the way through to files on disk?  (I don't know and I ask the question
 >> honestly.)y > ) > Alpha and Intel are both little-endian.e+ > (Alpha can be configured to be big-endiant  > but AFAIK only Cray did that). >   ; The Alpha is bi-endian, whether you use it that way or not.'  B IA-64 can access data in either big-endian or little-endian form. # Instructions must be little-endian.h  A So what do we call a system that's half way between bi-endian andn little-endian?  bill-endian?  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporatione= NASA GSFC Flight Software       | Federal Sector, Civil GroupaE                                 | please remove ".aspm" when replyingl   ------------------------------    Date: 23 Jul 2001 09:02:42 -0500- From: koehler@encompasserve.org (Bob Koehler)h" Subject: Re: No chance for OpenVMS3 Message-ID: <kSRKbXZBmI7$@eisner.encompasserve.org>n  r In article <W2I67.241716$%i7.135039115@news1.rdc1.sfba.home.com>, "Duane Sand" <duane.sand@mindspring.com> writes: > 8 > Alpha and IPF both support both big- and little-endian= > operating systems and applications.  (Alpha belatedly addedo > this about five years ago.)r  B Even the pre=release Alpha architecture manuals specify that it is> bi-endian.  IA-64 is not truley bi-endian, data access is, but- the instruction stream must be little-endian.   F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporationl= NASA GSFC Flight Software       | Federal Sector, Civil GroupiE                                 | please remove ".aspm" when replyingq   ------------------------------  % Date: Mon, 23 Jul 2001 18:58:25 +0200 & From: John McLean <mcleanj@dplanet.ch>" Subject: Re: No chance for OpenVMS* Message-ID: <3B5C57B1.B932C10C@dplanet.ch>   Larry Kilgallen wrote: >  > Just a few snippits: > U > In article <3B5A89E0.6C170D2B@dplanet.ch>, John McLean <mcleanj@dplanet.ch> writes:  > >e > >y > > Larry Kilgallen wrote:     ... snip ...   > >>M > >> I have read nothing to indicate an IA64 processor could ever be inserted6N > >> into a GS320.   Thus one is in the same position one is today -- tried to/ > >> buy any new CPUs for a TurboLaser lately ?s > >dL > > An IA64 could not be inserted into GS320 but since VMS is moving to thatF > > hardware, one hopes that at some time an IPF (ie. same family, newH > > child) will be in a QBB on a GS320.  Given the modular nature of the; > > components, I would expect these things to be possible.n > E > The combination of your insistence on _proof_ that Compaq will portwJ > VMS to IA64 (which they have announced) together with your _supposition_I > that IA64 could be incorporated into Wildfire (at which they have neveri > hinted) is quite strange.a  F I am not insisting on proof that Compaq will port VMS to IA64 because,H as you say, this is what they have said they are going to do.  What I amH seeking are further details about their plans for this, and some kind of; customer commitment so they don't renege on the whole deal..  > As regards IA64 and Wildfire, what I find very strange is your= insistence that we are dealing with IA64 in Wildfire.  I haveeG consistently stated "IPF family", meaning of course the Intel processoru/ with an unknown degree of Alpha-ism's included.d     John McLean   H PS.  My thanks to those who clarified the point about the "endians".  ItH looks like it will be okay to restore and use data files.  It looks likeC it's just the executable VMS-Alpha images that won't be usable, ando: that's a far lower percentage of file recovery operations.   ------------------------------  # Date: Mon, 23 Jul 2001 17:41:55 GMT 0 From: John Santos <john.santos@post.harvard.edu>H Subject: Re: Now we're cooking with gas. (was:  Wailing and moaning....)> Message-ID: <MPG.15c62bf03b2bbd5d98968b@news.bellatlantic.net>  D In article <3B5BF0D0.DEEE63E5@uk.sun.com>, andrew.nospam@uk.sun.com  says...i > Larry Kilgallen wrote: > > a > > In article <3B56A150.8B23833D@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:i > > >n > > > Fred Kleinsorge wrote: > > >>Q > > >> A quick check of the source listings seems to indicate that I am.  And thelJ > > >> hot air you blow seems to be the only indication of what you do ;-) > > >> > > >y: > > > Really, I would suggest that since the 25th your own3 > > > contributions have had an element of warm air 3 > > > polluting what is normally relatively factuale3 > > > content :):):) hence my re-evaluation of yourd > > > role.h > > K > > Andrew, given Fred's position, we look to him for authoritative answerslJ > > on what VMS Development engineers want to do.  That is not necessarilyI > > what the company management will end up doing, but even in the Compaq : > > environment, it is a powerful influence on management. > >  > ; > Really, note that I said since the 25th. Your point aboutl= > "powerful influence" before that date might well have been a" > defensible, since then it isn't.  D The quote is "powerful influence on management."  In another thread,; aren't you currently denying that you tell us what Compaq's : management's plans are without any documentation?  Can you; provide us with documentation for that "fact" that what VMSnB development engineers want to do is no longer a powerful influence on management?   Spin, spin, spin, spin...a  > > From coments made by various Compaq representatives on this ; > newsgroup it is clear that VMS development engineers werer= > either consulted very very close to the 25th or not at all.t > ; > The decision to move OpenVMS to IA-64 is one of the most  ; > important development decisions that has been made about g< > OpenVMS in the last 5 years. A powerfull influencer would ? > have been consulted, there is no evidence that this happened,O? > if it had then why do "powerfull influencers" post responses e@ > like we have only had 13 days to look at the porting issues ??  D Because the (few) engineers with advanced knowledge don't post here,* or are required to keep their mouths shut?  ? Compaq engineers have posted that the decision was made as much @ for engineering reasons as anything else.  Either they are lying? or they were lied to, or they are telling the truth.  Since youe6 can't handle the truth ;-), you assume they are lying.  ; In a few years, someone might write a book about this wholeo= episode.  I hope it is a book about how Compaq pulled its fath< out of the fire and VMS took over the world, but it might be= a book about how Compaq jumped out of the frying pan and into ( the fire, and now their goose is cooked.  > Anyone want to post the next 100 bad cooking puns?  (Remember,> I was the person who originally stated that if Compaq couldn't= pull off the IA-64 port, and do it right, they were toast...)   	 > Regardsh > Andrew Harrisonp > Enterprise IT Architect:   ------------------------------  # Date: Mon, 23 Jul 2001 15:10:16 GMT,F From: lederman@star.enet.dec.DISABLE-JUNK-EMAIL.com (Bart Z. Lederman) Subject: Re: OT: Dr Who.3 Message-ID: <s7X67.1095$rc5.72648@news.cpqcorp.net>f  > It's been said that your favorite Dr. Who is usually the first9 one you saw.  Although Tom Baker was the first one I saw,t; and I thought the writing on many of his episodes was quite 8 good (you do know that Douglas Adams wrote some of them,4 don't you?), I also think Jon Pertwee was very good.  6 I don't think Sylvester McCoy was all that bad: he was7 better than Colin Baker (though I think the scripts ande+ management were part of the problem there).h  8 As for Mr. Bean and Black Adder: he hasn't done anything. really good since "Not the Nine O'Clock News".  6 Personally, I wish they'd bring back David Allen (what is he doing now, anyway?)c  & Is this off-topic enough for everyone?   -- d(  B. Z. Lederman   Personal Opinions Only  8  Posting to a News group does NOT give anyone permission8  to send me advertising by E-mail or put me on a mailing  list of any kind.  5  Please remove the "DISABLE-JUNK-EMAIL" if you have ae5  legitimate reason to E-mail a response to this post.i   ------------------------------  % Date: Mon, 23 Jul 2001 11:03:30 +0100h% From: Alan Greig <a.greig@virgin.net>c# Subject: Re: Process adopts Itaniume8 Message-ID: <susnlt4m1cij8k13uv7mirtic9ijl1dq70@4ax.com>  , On Fri, 20 Jul 2001 13:44:37 -0400, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:f   > M >One tidbit I heard is that the "business critical" group will cease to exist N >and VMS etc will be handled by the wintel server group. Not sure if this is aO >fact or heresay. But assuming it is true, this has interesting implications inb  D Well I asked that question of Richard George (UK Alpha business unitE manager) and he replied that ISSG would be in charge of the design ofoF the new servers and that the business critical group would continue toF develop and support Alpha technology. Engineers would transfer to ISSG
 over time.    N >terms of transfer of technologies. The Wintel crowd will benefit from the VMSM >and NSK knowledge and expertise in building systems and the Wintel crowd maye? >be able to then sell "wildfires" based on IA64 and running NT.t  C Richard also presented some slides of what the 2004 servers will bedE like. They showed one system running NT, Tru64 (or was it Linux?) andsE VMS in partitions. So in four years time Compaq will start selling an E Intel based server which can do what was first demonstrated in publicS  about two years ago on an Alpha.   >tL >On the other hand, by having the wintel crowd responsible for VMS, it givesI >the slight possibility that Compaq may start to view VMS as a mainstreameL >product and market the hell out of it. However, I do not personally beleiveG >that the attitudes at Compaq and Microsoft would allow this to happen.6   -- Alan   ------------------------------  % Date: Mon, 23 Jul 2001 12:13:06 +0200 < From: "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de>? Subject: Re: Reference for VMS anti-hacker credentials required 4 Message-ID: <9jgtc8$nu1f8$1@ID-56200.news.dfncis.de>   Rob Buxton wrote...yG >At a no longer recent VMS Seminar a point was made that VMS was one ofiF >(or the only) OS that had not been hacked by a certain expert hacker. >oE >I haven't been able to find an actual article to back this claim and./ >would like pointers to any WWW pages for info.0  7 Nothing official, but a Google search quickly turned up = http://www.zdnet.com/zdnn/stories/news/0,4586,2454737,00.htmlh   cu,A   Martin -- oJ One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer7 One OS to find them           | work: mv@pdv-systeme.degJ One OS to bring them all      |   http://www.pdv-systeme.de/users/martinv/? And in the Darkness bind them.| home: martin@radiogaga.harz.de u   ------------------------------    Date: 23 Jul 2001 07:15:20 -0700# From: dooleys@snowy.net.au (dooley) ? Subject: Re: Reference for VMS anti-hacker credentials required ; Message-ID: <1ca82fc6.0107230615.f1ef6b@posting.google.com>i  c rob.buxton@wcc.govt.nz (Rob Buxton) wrote in message news:<3b5b78d8.1473410512@news.wcc.govt.nz>...e	 > Hi All,  > H > At a no longer recent VMS Seminar a point was made that VMS was one ofG > (or the only) OS that had not been hacked by a certain expert hacker.s > F > I haven't been able to find an actual article to back this claim and0 > would like pointers to any WWW pages for info. > D > There's a fear here about access to all Computers and I want to beB > able to make a clear point with some good reference material the > inherant security of VMS.l >  > TIA  >  > Rob.  < You must be referring to http://www.underground-book.com/ :) Phil   ------------------------------    Date: 23 Jul 2001 14:29:07 -00004 From: Doc.Cypher <Use-Author-Address-Header@[127.1]>? Subject: Re: Reference for VMS anti-hacker credentials required 6 Message-ID: <20010723142907.29280.qmail@nym.alias.net>  " -----BEGIN PGP SIGNED MESSAGE-----  K On Mon, 23 Jul 2001, "Martin Vorlaender" <martin.vorlaender@pdv-systeme.de>n wrote: >Rob Buxton wrote...H >>At a no longer recent VMS Seminar a point was made that VMS was one ofG >>(or the only) OS that had not been hacked by a certain expert hacker.e >>F >>I haven't been able to find an actual article to back this claim and0 >>would like pointers to any WWW pages for info. > 8 >Nothing official, but a Google search quickly turned up> >http://www.zdnet.com/zdnn/stories/news/0,4586,2454737,00.html  H The followups contains a message from a Mike Foley of Boston claiming to, have managed a VMS system hacked by Mitnick.  B http://www.zdnet.com/tlkbck/comment/22/0,7056,85280-362454,00.html  F No mention if this was through security expoits or (as I would like to think) social engineering.     Doc. - -- p6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                              http://vmsbox.cjb.nete   -----BEGIN PGP SIGNATURE-----a Version: 2.6.2  @ iQEVAwUBO1ta8sriC3SGiziTAQGGYQf9Go1YwTzahoca9WeXOfRvoXCSDySY/DH1@ VZBXIJg0Ap48ZB57dUE9vo2YPkLVbe0MDiGdbCBiDqv2oSyBhMkMRuC/q6EVfn6+@ NMF8G2jjZYZy9PW1Yhf/fARgpvVYuX+EBVeKTDHp5MHGiVdl9IbG7qDcv8Dru6+S@ DyHH/qgPBwoUJep0wdvEqvfdbd57+0Cy2A9jdMN3VRk8ZhEUfD4EpFwpiRdx02sT@ H07d143IVDYVdJUGHFs818S5pXgtQrPk1VhpSm3YVvli8insThnlGaR7sIqCMfHO8 IeAJRl0SxfTUBY8M/QyILL9GDuHrQCD2B7NLogjlyv18Vyln4KNS5w== =DI2Ns -----END PGP SIGNATURE-----3   ------------------------------  % Date: Mon, 23 Jul 2001 14:56:36 -0000t- From: wspencer@ap.nospam.org (Warren Spencer)iC Subject: RESULTS:  Re: Poll: Move OpenVMS to IPF: Agree or Disagreei/ Message-ID: <tloep4603n45df@news.supernews.com>c  J My apologies to anyone who did not appreciate the extra traffic this poll ? caused.  Larry you were right - should have done this by email.o   The results are:       	10 responses totalr     	 7 Agreed with:  6 "Do you agree or disagree with moving OpenVMS to IPF?"  + Of those 7, few agreed with retiring Alpha.M  0     	 2 did not agree with moving OpenVMS to IPF)     	 1 fell into the category of "other"a  9 Terry's survey, at www.tru64.org offers much more detail.r   ws   --     Warren Spencer Senior Software Engineer The Associated Press  L ** My employer does not necessarily agree with my statements - neither do I  **   ------------------------------  % Date: Mon, 23 Jul 2001 09:01:36 -0500g+ From: Christopher Smith <csmith@amdocs.com> M Subject: RE: Security-Enhancing Products for Compaq OpenVMS from PointSecu	reeL Message-ID: <3B55D7F383B0D31197D9009027541CBF1170DA4E@cmiexch1.cmi.itds.com>   > -----Original Message-----@ > From: system@SendSpamHere.ORG [mailto:system@SendSpamHere.ORG]  < > Seeing as how I am a VMS customer and use VMS, wouldn't a  > message about.. > security and VMS be better targetted at VMS  > users/managers/etc. than at- > PeeCee weenies?   K .but peesee weenies are compaq's target market.  Oh, wait, _THAT'S_ why VMSt? isn't selling so well!  Microshaft isn't advertising it enough!I   Regards,  + Chris (who really couldn't resist this one)w    ! Christopher Smith, Perl Developer  Amdocs - Champaign, IL   /usr/bin/perl -e '? print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");r 'o   ------------------------------  , Date: Mon, 23 Jul 2001 11:41:52 +0200 (CEST): From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>- Subject: Re: Selling VMS to another company ? I Message-ID: <Pine.LNX.4.21.0107231111350.5716-100000@irys.stanpol.com.pl>o  $ On Sat, 21 Jul 2001, JF Mezei wrote:  ( >+"Brian Schenkenberger, VAXman-" wrote:P >+> The death of Alpha makes the sale of VMS to any other party quite a bit less >+> attractive.  [...]M >+I disagree. Once VMS is in IA64, the buyer of VMS would not be dependent onlP >+any box maker for systems and could contract with any one to build VMS-quality! >+boxes or build them themselves.e  	 My .02$$:KB  *only* when having hope, that here will be one (or "nearly one") 2 implementation of IA-64 as computer architecture.   E  Unfortunatelly the IBM-PC implementation of x86 (today named IA-32)  A was a good example of poor processor architecture implementation. E  The (discuised here in past) limitation of interrupt subsystem in PCa3 is NOT enforced by the Intel project, but IBM one !r?  And litle change within x86 was enforced later by use of some h( "reserved" interrupt in B.G. software :]  <  We know others "nearly standard" - SCSI, b.ex., (and I have< buy one CD-ROM with hope to read CD-RW... You know - doesn't< work [with VMS], even with all mentioned on c.o.v tricks :()?  And we know - the reason in most cases is NOT in VMS implemen- 5 tation (here: of SCSI) but miss of expected features.s  6  Then - it IS possible that some time may be available4 "mass production" of IA-64 based computers - but the6 computers may not be able run VMS... Even if generally OVMS-IA64 will be created...=  One look in the PC will explain, that the processor (withoutr: proper motherboard chipsets and motherboard at all) is not= enought. Supose VIA will create a great chipset, where unfor-J3 tunatelly will miss one feature expected by OVMS :[r  "What ? That is great hardware !$  works with Windows and Linux..." :(    Regards - Gotfryd7 (not *THAT* pesimistic, but also not *THAT* optimistic)w -- cE ===================================================================== F $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=ME(. $!                        GS@stanpol.zabrze.plE =====================================================================i   ------------------------------  # Date: Mon, 23 Jul 2001 10:06:16 GMT . From: Burnie M <burniem.NOSPAM@ozemail.com.au>- Subject: Re: Selling VMS to another company ?-8 Message-ID: <botnltc04h232o0km2r87oeg6jggcvcr5t@4ax.com>  , On Sun, 22 Jul 2001 23:27:11 -0400, JF Mezei% <jfmezei.spamnot@videotron.ca> wrote:r   > N >The killing of Alpha is simply a corporate decision by Compaq because this is. >not where Compaq wants to go in its strategy.    ( Exactly where does Compaqs strategy go ?F Anybody from Compaq ? or are you guys/gals as much in the dark as us ?   Burnie M   ------------------------------  % Date: Mon, 23 Jul 2001 13:49:50 -0400t- From: JF Mezei <jfmezei.spamnot@videotron.ca>n- Subject: Re: Selling VMS to another company ? , Message-ID: <3B5C63BB.D3DED9A7@videotron.ca>   Burnie M wrote:dP > >The killing of Alpha is simply a corporate decision by Compaq because this is0 > >not where Compaq wants to go in its strategy. > * > Exactly where does Compaqs strategy go ?H > Anybody from Compaq ? or are you guys/gals as much in the dark as us ?    H Compaq did state some time ago that they would move to industry standardG hardware. Shortly after than, they started to call their wintel serverslB "industry standard", meaning that Alpha was not industry standard.  J Their past financial statements indicated that they beleived that NT would rule the enterprise.  G The TV ad campaign that was designed to portray Compaq as an enterprise K company featured Proliant servers. (granted, that was 2 seconds on an alphaqG mainframe shown, but most would think it is another big wintel server).o  K Compaq's killing of Alpha was designed to cut costs and put everything intoWI one industry standard chip. (yeah, as if Compaq is going to drop the 8086k anytime soon).  K In hindsight, Compaq has sent lots of messages that anything non-wintel wasdK not considered a strategic/core product. Doesn't mean that Compaq will killAM anything that isn't industry standard, but it certaintly leads one to beleivehK that Compaq won't be pushing those "weird" platforms such as Alpha and VMS.n   ------------------------------  % Date: Mon, 23 Jul 2001 11:23:45 -0400a% From: "John Vottero" <John@mvpsi.com> > Subject: Re: SMS on VMS [was RE: Full port of VMS to Itanium.]/ Message-ID: <tlogb1hn14tk41@news.supernews.com>   : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3B594BB0.76EA27F7@videotron.ca... > Burnie M wrote::H > > Sorry for the TLAs but yes VMS is at the core of Telecoms especially > > GSM mobile networks. >(J > You make it sound like all GSM networks make extensive use of GSM. Funny howaI > the local GSM network here in canada seems to only advertise for window. weenie > and Sun Solaris jobs.l > I > I keep hearing how lotteries run on VMS, yet the provicial lottery is an tandemL > shop. I keep hearing how telecom runs on VMS, yet I never see adds for VMS) > folks from the local telecom companies.e >   2 I know that the British Columbia Lottery uses VMS.   ------------------------------  % Date: Mon, 23 Jul 2001 13:34:11 -0400 + From: "Main, Kerry" <Kerry.Main@compaq.com> > Subject: RE: SMS on VMS [was RE: Full port of VMS to Itanium.]R Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF48DBFC5@kaoexc01.americas.cpqcorp.net>  8 >>> I know that the British Columbia Lottery uses VMS.<<   As does Ontario.   Regards,  
 Kerry Main Senior Consultant  Compaq Canada Inc. Professional Serviceso Voice: 613-592-4660a Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----* From: John Vottero [mailto:John@mvpsi.com] Sent: July 23, 2001 11:24 AM To: Info-VAX@Mvb.Saic.Comt> Subject: Re: SMS on VMS [was RE: Full port of VMS to Itanium.]    : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3B594BB0.76EA27F7@videotron.ca... > Burnie M wrote:hH > > Sorry for the TLAs but yes VMS is at the core of Telecoms especially > > GSM mobile networks. >tJ > You make it sound like all GSM networks make extensive use of GSM. Funny howwI > the local GSM network here in canada seems to only advertise for windowe weenie > and Sun Solaris jobs.p > I > I keep hearing how lotteries run on VMS, yet the provicial lottery is a  tandemL > shop. I keep hearing how telecom runs on VMS, yet I never see adds for VMS) > folks from the local telecom companies.  >h  2 I know that the British Columbia Lottery uses VMS.   ------------------------------  % Date: Mon, 23 Jul 2001 11:14:15 +0100 0 From: andrew harrison <andrew.nospam@uk.sun.com>: Subject: Re: The death of VMS has been greatly exaggerated* Message-ID: <3B5BF8F7.C8E29A57@uk.sun.com>   jlsue wrote: > 5 > On Fri, 13 Jul 2001 10:52:19 +0100, andrew harrisone# > <andrew.nospam@uk.sun.com> wrote:w >  > >jlsue wrote:n >  > >>K > >> Hey Arne, Andrew's been doing that for years.  He's repeatedly told usgI > >> what Compaq Management's plans were, when there was no communications+ > >> affect anywhere to back up his claims.  > >> > >e? > >Wrong go and search deja, I havn't ever made any suggestionse5 > >on what Compaqs management plans are or have been.y > >  >  > Bwahahahahahahahaha!!!!!!! > F > You're wrong-o buddy.  I don't have to do a deja search - I wouldn'tH > waste my time.  Even when faced with certain facts you weasel and spin > your way around them.  >   8 Stop wasting time, go search Deja and come back when you have.n  6 Our conversations in the past have been characterised 2 by a lack of research, lack of factual content and4 lack of manners on your part, your response in this , thread is only following the normal pattern.  4 or did you think that Bwahahahahahahahaha!!!!!!! was an adult response.      E > Anyone who's been reading in here for a year or more will know thatJE > you will very often "predict" what Compaq (or DEC) will (would) do.  >   7 No, go and read Deja. I have made suggestions for what  9 I think Compaqs management should do, marketing, businessl9 development etc. I have never offered an opinion on what h@ Compaqs management will do. Do you understand the difference ???  E > It's your reputation man.  Live with it, or do something to improvee > it.e  8 Really and what would you like your reputation to be ???   Regards  Andrew Harrisone Enterprise IT Architecti   ------------------------------  % Date: Mon, 23 Jul 2001 11:17:11 +0100n0 From: andrew harrison <andrew.nospam@uk.sun.com>: Subject: Re: The death of VMS has been greatly exaggerated* Message-ID: <3B5BF9A7.A6B115B0@uk.sun.com>   "Main, Kerry" wrote: > 	 > Andrew,  > N > >> Kerry your choice of words makes it sound as if all the vendors that haveN > their own architecture and IA-64 have the IA-64 platform to ensure that theyA > will have a backup platform if their own architecture fails.<<<" > M > Not at all. I stated "or alternate 64 bit plan" very clearly. I did not say J > IBM's Power strategy was a backup for IA64. You chose to misintepret the" > words to suit your own purposes. >   7 Bullshit if you didn't mean backup why did you say it ?s     Here is you post.cI > In addition, with the exception of MS, each OS provider has a backup oriK > alternate 64bit HW plan (Power4, Alpha EV7, PA, MIPS) that will allow thesM > Customer to migrate when they feel the timing is right - whether it is in 3t > years or 5+ years. >    Is you memory so short ???????   Regardsi Andrew Harrisona Enterprise IT Architects   ------------------------------    Date: 23 Jul 2001 07:07:56 -07005 From: Gerry.Czadowski@NAV-International.com (YYYGAC2)h: Subject: Re: The death of VMS has been greatly exaggerated= Message-ID: <37a44c33.0107230607.1882a643@posting.google.com>e  b andrew harrison <andrew.nospam@uk.sun.com> wrote in message news:<3B5BF9A7.A6B115B0@uk.sun.com>... > "Main, Kerry" wrote: > >   
 -- snip --   >  > Here is you post.iK > > In addition, with the exception of MS, each OS provider has a backup orsM > > alternate 64bit HW plan (Power4, Alpha EV7, PA, MIPS) that will allow thetO > > Customer to migrate when they feel the timing is right - whether it is in 3i > > years or 5+ years. > >  >   > Is you memory so short ??????? > 	 > Regardsa > Andrew Harrisonv > Enterprise IT Architectr  B I believe that memory problems can only be revealed after you have executed a NDA. ;^)h   ------------------------------  % Date: Mon, 23 Jul 2001 10:33:09 -0400t+ From: "Main, Kerry" <Kerry.Main@Compaq.com>s: Subject: RE: The death of VMS has been greatly exaggeratedR Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4AD7F27@kaoexc01.americas.cpqcorp.net>   Andrew,n  = >>>Bullshit if you didn't mean backup why did you say it ?<<<   K "Backup" referred to an IA64 failback strategy (HP-PA, Alpha-EV7, NSK-MIPS,eI IBM-Power) until IA64 servers met expectations while "alternate" strategye= referred to IBM having two future strategies - IA64 or Power.n  $ >>>Is you memory so short ???????<<<   Hey, maybe I need some ECC eh ?s   :-)i  
 Kerry Main Senior Consultantb Compaq Canada Inc. Professional Services  Voice: 613-592-4660t Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----7 From: andrew harrison [mailto:andrew.nospam@uk.sun.com]u Sent: July 23, 2001 6:17 AMz To: Info-VAX@Mvb.Saic.Com0: Subject: Re: The death of VMS has been greatly exaggerated     "Main, Kerry" wrote: > 	 > Andrew,  > I > >> Kerry your choice of words makes it sound as if all the vendors that  haveI > their own architecture and IA-64 have the IA-64 platform to ensure thatn theyA > will have a backup platform if their own architecture fails.<<<- > I > Not at all. I stated "or alternate 64 bit plan" very clearly. I did not. sayoJ > IBM's Power strategy was a backup for IA64. You chose to misintepret the" > words to suit your own purposes. >   7 Bullshit if you didn't mean backup why did you say it ?r     Here is you post.5I > In addition, with the exception of MS, each OS provider has a backup orcK > alternate 64bit HW plan (Power4, Alpha EV7, PA, MIPS) that will allow the K > Customer to migrate when they feel the timing is right - whether it is ino 3n > years or 5+ years. >    Is you memory so short ???????   Regardsy Andrew Harrisonf Enterprise IT Architecth   ------------------------------    Date: 23 Jul 2001 09:15:58 -05003 From: malmberg@encompasserve.org (John E. Malmberg)nO Subject: Triggering tasks from network file arrival - Was Re: Basic VMS Questio:3 Message-ID: <ayQ23Ti4vlqm@eisner.encompasserve.org>-  M In article <CDA4BAD1E10ED41181AC00508B6051D3C3E2E7@grumpy.internal.hspg.com>,r0 Andrew Robinson <arobinson@hspg.company> writes:. > Please could anyone help with the following: >DG > I have a procedure running on one of my VMS boxes which is constantlySI > checking a directory for a file which can be FTPed into it. If the file-K > appears it then runs another routine to process and remove the file readyaL > for the next one. I have a simple routine at the moment written in Diabold# > which runs as a detached process.>N > The problem I have is that to make this process work, it is a simple looping3 > check, which eats CPU and slaughters the disk IO.iM > Is there anything in VMS which you can utilize which you can use to triggers4 > an event if a file appears in a specific location.4 > I using an Alpha 1200 running OVMS 7.2-1 + patches  F I would presume that the FTP uses an account that is dedicated to this specific application.,  L When you FTP or DECNET or SNADTF (and probably a few other protocols) a file0 into an OpenVMS system, it does a NETWORK login.  E This NETWORK login executes the SYLOGIN.COM command procedure and then6 LOGIN.COM procedure for the account that is logged in.  J After the Network process does it's work, it will change it's process name/ and wait for another transfer by the same user.Y  0 This means that you can have the LOGIN.COM test:        $if f$mode() .eqs. "NETWORK"	     $THENY  J     ! Code to see if the helper batch job is running or pending in a queueE     ! if not, submit the job with a delay estimated to allow the filei     ! transfer to complete.f  
     $ENDIF     And in the HELPER batch job.         $PRC_WAIT:@     ! The start delay may not have been long enough, so look for4     ! a network transfer in progress from this user.     !r:     ! Use f$context() to filter network jobs for this user.     ! Use f$getjpi() to check the process name;     ! if it is the transfer process name then WAIT 00:02:00t     ! before looking again.      !l<     ! Depending on the protocol and the number of files that     ! are transfered     !      $XFER_DONE:      ! Now signal your program      ! 
     $CLEANUP:eH     ! The network process, if left around will not execute the LOGIN.COM"     ! when a new request comes in./     ! Therefore, it must be located and killed.o=     ! The name will depend on what transfer protocol is used.m     $stop/id='xxxx'   G The downside of this technique is that some of the specific information*C needed to implement it may not be publically documented, but can bedD determined from observed behavior, and also from reading the commandC files that are supplied with the network protocol that is doing the:	 transfer.t  G Also, when transfering files in a drop off method by a protocol such as5J FTP which does not guarantee that the file deposited is complete, considerF transfering the file with a temporary name, and then renaming the fileF after the transfer has completed, but in the same ftp session from the sender.s  L This will insure that if a file with the target name shows up, it is intact.   -John- wb8tyw@qsl.network Personal Opinion Only    ------------------------------  % Date: Mon, 23 Jul 2001 11:23:02 +0200w0 From: Dietmar Hermanns <dietmar.hermanns@rtl.de>4 Subject: Re: Trying to load VMS on MicroVAX 3100....> Message-ID: <3b5bd0ca$0$23236$4dbef881@businessnews.de.uu.net>   David J. Dachtera wrote:   > microvax@felinet.org wrote:r > >  > > Hi,. > > H > > I have an old MicroVAX 3100 that I got free from someone.  He didn'tB > > have an OS on it, and he didn't know whether it worked or not. > > I > > I've never had the chance to do any VMS administration, so I took the.; > > box, got a VAX-VMS hobbyist CD and tried to boot to it.t > > I > > I had to add a CD-drive to the machine...a SONY CDU55s.  I repeatedlyhE > > receive the error shown below when I attempt to boot the CD.  :-(e > > H > > The CDU55s came with a few jumpers (presumably to set it to 512-byteI > > block size?) but there's absolutely no documentation for it that I'vew
 > > been ablebK > > to find.  I've tried the jumpers in several configurations and tried ith7 > > without the jumpers...more or less the same result.  > > J > > Anybody have any ideas what I need to do to get this poor VAX running? > >  > > Thanks in advance... > > 	 > > Peter  > > $ > > Here's the device configuration: > > 
 > > >>>SH DEVe > > F > >  VMS/VMB  ULTRIX     ADDR       DEVTYP    NUMBYTES    RM/FX   WP   > >  DEVNAM   REVpE > >  -------  ------    --------  ---------  ---------   ------  ---   > >  -------  ----) > >  ESA0     SE0       08-00-2B-24-46-A4W > > L > >  DKA300   RZ3       A/3/0/00    DISK       209 MB     FX           R224  > >    211BW< > >  MKA500   TZ5       A/5/0/00    TAPE    .........     RM) > >  ...HostID....      A6          INITRr > > L > >  DKB200   RZ10      B/2/0/00    DISK       209 MB     FX           RZ24  > >    211BsF > >  DKB300   RZ11      B/3/0/00    RODISK     681 MB     RM           > >  CD-ROM  1.0ts! > >  ...HostID....   ?      INITRa > >  > > >>>u > > : > > And this is what happens when I try to boot to the CD: > >  > > >>> BOOT DKB300h > >  > >  > > -DKB300rA > > %SYSBOOT-I-SYSBOOT Mapping the SYSDUMP.DMP on the System DiskhA > > %SYSBOOT-I-SYSBOOT Can not map SYSDUMP.DMP on the System Disk K > > %SYSBOOT-I-SYSBOOT Can not map SYSDUMP.DMP on the System DiskDKBTDRIVER  > > haltin4iH > > REGISTERS  R0: 00000054   R1: 00000003   R2: 017B2800   R3: 00000400H > >            R4: 00000400   R5: 00021293   R6: 00000002   R7: 200CD180H > >            R8: 00000400   R9: 00000000  R10: 812A3600  R11: 00021293; > > CMDBUF: 00000093 12020806  SENSE_BUF: 00000000 00000000 / > > STATUS_BUF: 00             SENSE_STATUS: 02=L > > MSG_IN_BUF: 00             SENSE_MSG_IN: 00        SENSE_INDEX: 00000000G > > USER_BUF_ADDR: 812A3A00    TOTAL_XFER_COUNT: 00000400   BYTES_LEFT:e > > 00000000K > > PG_TABLE_PTR: 017B2800     BASE_VPN: 0000951B      BUFFER_PTR: 812A3600- > > MAP_MODE: 00000001 > > STACK:  00001B68 > >         000025FF > >         00000000 > >         00001E9F > >         ?       000016DF > >  > > ?06 HLT INST > >     PC = 00002C8Be > C > The Sony CD is probably not supported by the VMS SCSI disk driverrI > (DKDRIVER). The "primitive" drivers in the bootstrap and the loader are G > more forgiving (less picky), but DKDRIVER has certain expectations of-/ > those devices with which it will "play nice".: > G > Try looking up a Toshiba 4x or 8x. Some of the newer ones (20x or so)lC > will work also. Remember to set 512-byte block jumper to the "on"r > position.n > H > ...of course, good ol' DEC RRD4x's will work, at least the older ones, > anyway...  > L The know CD Type  Liste is at > http://sites.inka.de/pcde/dec-cdrom-list.txt  F Or did what I did, make a disk dump of the cd on a harddisk,  like dd K if=/dev/src0 of=/dev/sda. You can do it with an other OS (Unix, Linux) but o5 it must have 512Bytes/ Block like normal disks have. s) After that you can use this disk as a cd.s3 I dont know if there is a dd ,diskdumptool for VMSa   ------------------------------    Date: 23 Jul 2001 06:21:21 -0700, From: andrew.kirk@Tarmac.co.uk (Andrew Kirk)& Subject: Re: V7.3 XFCACHE-W-DATALOSS ?= Message-ID: <1dcc9d5c.0107230521.724bc1b5@posting.google.com>(   Mark  S I would run autogen.   Reading the manual pages it says that the cache is tuneable.i   Here is the url   L http://www.openvms.compaq.com/doc/73final/6017/6017pro_078.html#vioc_size_of   Hope this is of some use!y  
 Kind regards,  Andrew Kirk.      e Mark Daniel <Mark.Daniel@wasd.vsm.com.au> wrote in message news:<3B592C1D.563E021@wasd.vsm.com.au>... F > Recently (just) installed V7.3 on DS20 (1024MB) has started spitting > OPA0 messages reportingn > H >   %XFCCACHE-W-DATALOSS, read I/O completetion with <less bytes that it > thinks it should>" > J > We have been COPYing a *lot* of small files around (documentation CDs to
 > hard-disk).1 > I > This is not in the HELP message database (yet).  Anyone else seen this?  >  > As always, TIA.M   ------------------------------  + Date: Mon, 23 Jul 2001 12:14:34 +0000 (UTC)fB From: Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) Subject: vms backupuB Message-ID: <23DC1A5C6E91D411BCD300508B6573E8012CA9A5@SEMARMN0006>   Hi& I need an information about cd rom VMS) what is the name of cdrom in VMS system ?o     Massimiliano Mauceri Delivery and Support EngineerS SchulmbergerSema Telecomse- e-mail: massimiliano.mauceri@tlc.semagroup.it  Tel: +39 06 41536371           -- i Posted from [213.255.32.100] t1 via Mailgate.ORG Server - http://www.Mailgate.ORGp   ------------------------------    Date: 23 Jul 2001 08:34:54 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)  Subject: Re: vms backupu3 Message-ID: <gJFM78Pxifwg@eisner.encompasserve.org>o   In article <23DC1A5C6E91D411BCD300508B6573E8012CA9A5@SEMARMN0006>, Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) writes:] > Hi( > I need an information about cd rom VMS+ > what is the name of cdrom in VMS system ?   C Presuming a running VMS system (since your title indicated backup), C it can be named DKA400, DQB0 or many other things depending on yourf@ complement of IO busses and addresses thereon.  Use the command:   	$ SHOW DEVICE D  ) and look for an entry that says "wrtlck".    ------------------------------  + Date: Mon, 23 Jul 2001 12:53:02 +0000 (UTC)MB From: Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) Subject: Re: Re: vms backupeB Message-ID: <23DC1A5C6E91D411BCD300508B6573E8012CA9A6@SEMARMN0006>   Hi$ About cd rom device another question* If I have this situation on my VMS system: $1$DQA0: $1$DQA1: $1$DQB0: $1$DQB1:I which is the cd rom device if the all devices have device status online ?  tank's y Massimiliano Mauceri Delivery and Support Engineert SchulmbergerSema Telecomse- e-mail: massimiliano.mauceri@tlc.semagroup.itp Tel: +39 06 41536371           -- t Posted from [213.255.32.100] a1 via Mailgate.ORG Server - http://www.Mailgate.ORGS   ------------------------------  % Date: Mon, 23 Jul 2001 10:14:47 -0300 ) From: fabio_compaq@ep-bc.petrobras.com.bre Subject: Re: vms backup-L Message-ID: <OFBD166EAE.F0AEFA97-ON03256A92.0048813D@ep-bc.petrobras.com.br>  I It is recognized as a "normal disk", like DKA500: for example. Depends one2 the controller it is connected: DKA, DKB, etc ....  F Type $ show device DK, and if appears "Online wrtlck"  it is the CDROM device.e   Regards.   FC        J Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) em 23/07/2001 09:14:34  @ Favor responder a Massimiliano.Mauceri@tlc.semagroup.it (Mauceri       Massimiliano)e             Info-VAX@Mvb.Saic.Com        Assunto: vms backupi     Hi& I need an information about cd rom VMS) what is the name of cdrom in VMS system ?h     Massimiliano Mauceri Delivery and Support Engineer  SchulmbergerSema Telecoms!- e-mail: massimiliano.mauceri@tlc.semagroup.itd Tel: +39 06 41536371           -- Posted from [213.255.32.100]1 via Mailgate.ORG Server - http://www.Mailgate.ORGo   ------------------------------  % Date: Mon, 23 Jul 2001 11:06:48 -0400 2 From: rdeininger@mindspring.com (Robert Deininger) Subject: Re: vms backupfL Message-ID: <rdeininger-2307011106480001@user-2ive6lk.dialup.mindspring.com>  B In article <23DC1A5C6E91D411BCD300508B6573E8012CA9A6@SEMARMN0006>,C Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) wrote:    > Hi& > About cd rom device another question, > If I have this situation on my VMS system:
 > $1$DQA0:
 > $1$DQA1:
 > $1$DQB0:
 > $1$DQB1:  9 On our Digital Personal Workstations, the CD-ROM is DQB0.   G You can likely try all 4 devices with no ill effects.  Only one of themm
 will work.   -- e Robert Deininger rdeininger@mindspring.come   ------------------------------   Date: 23 JUL 2001 15:13:19 GMT+ From: Dave Greenwood <greenwoodde@ornl.gov>t Subject: Re: Re: vms backupa2 Message-ID: <23JUL01.15131902@feda01.fed.ornl.gov>  Z In a previous article, Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) wrote: > Hi& > About cd rom device another question, > If I have this situation on my VMS system:
 > $1$DQA0:
 > $1$DQA1:
 > $1$DQB0:
 > $1$DQB1:K > which is the cd rom device if the all devices have device status online ?   J Use  SHOW DEVICE/FULL DQ.  The actual cdrom (DQB0 on my systems) will have9 it's device type displayed (at least on the host system).h   Dave --------------9 Dave Greenwood                Email: Greenwoodde@ORNL.GOVcH Oak Ridge National Lab        %STD-W-DISCLAIMER, I only speak for myself   ------------------------------  # Date: Mon, 23 Jul 2001 17:35:08 GMTA2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: Re: vms backuph3 Message-ID: <gfZ67.1107$rc5.72929@news.cpqcorp.net>    In article <23DC1A5C6E91D411BCD300508B6573E8012CA9A6@SEMARMN0006>, Massimiliano.Mauceri@tlc.semagroup.it (Mauceri Massimiliano) writes:E+ :If I have this situation on my VMS system:r  A   When posting, please include your OpenVMS version and the exact B   name of the particular VAX or Alpha model in use.  Please do notB   abbreviate these.  (For further information on posting questionsB   to newsgroups, and particularly on the sort of information that E   is often necessary when answering your question(s), please see the  G   introductory section of the OpenVMS Frequently Asked Questions (FAQ) n*   document.  The FAQ URL is listed below.)  m	 :$1$DQA0:>	 :$1$DQA1:e	 :$1$DQB0:<	 :$1$DQB1:BJ :which is the cd rom device if the all devices have device status online ?  H   Based solely on the apparent use of IDE-type disks, this system would G   likely be an AlphaServer DS10, AlphaServer DS10L, AlphaStation DS10, IG   or AlphaStation XP900 series.  (Comparatively few other Alpha systemsnG   have an IDE bus, and thus few use the IDE-specific SYS$DQDRIVER.  TheyH   VAX-11/730 series also had a DQ device, but did not particularly have    a CD-ROM option available.)h  J   To find the specific identification of the devices with names beginning I   with D involved (these devices are usually disks), use the DCL command:       SHOW DEVICE/FULL Dy  G   Look for a CD-ROM among the devices you have present.  If you are nottG   sure which one is a CD-ROM, post the entry for the "device type", as .G   listed in the first line of output for each "D" device, as displayed     by the above SHOW command.  D   I will assume you have the OpenVMS Installation documentation.  IfG   not, then please see the documentation that is posted at the OpenVMS e   website.  (URL below.)  N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------    Date: 23 Jul 2001 12:17:08 -05009 From: kaplow_r@eisner.encompasserve.org.mars (Bob Kaplow) 5 Subject: Re: VMS remains secure at DEFCON hacker fest 3 Message-ID: <oy6glDuF7l$Q@eisner.encompasserve.org>a  [ In article <3B58EF53.F79C43FB@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:6 > Patrick Jankowiak wrote:	 >> [snip]l& >> #3||0!, is VMS Marketing listening? >  > Is the Pope Jewish?      Pope Rosenberg?"   ------------------------------  % Date: Mon, 23 Jul 2001 11:28:36 +0100a% From: Alan Greig <a.greig@virgin.net>o  Subject: Re: VMS Trivia Question8 Message-ID: <uhunlt877r04qo3f39rmghubaoajgf5ucu@4ax.com>  F On Sat, 21 Jul 2001 13:08:46 GMT, "David Cressey" <david@dcressey.com> wrote:     >,G >When the VAX first made its debut,  my expertise was with TOPS-10, thepI >operating system for the DECsystem-10.  TOPS-10 maintained doubly linkedl  ? Hey that's maintains not maintained :) There's been a flurry of B activity over in alt.sys.pdp10 recently now that the emulators are reliably booting TOPS-10.   L >lists to keep track of the job queues  (process queues, if you like).  EachJ >link in the chain consisted of two job numbers, stored in halfwords,  and* >used as indexes into the job queue array. > M >With the PDP-10 instructions set, it took six instructions to put a job intoE  C Well with an emulator you can always enhance the instruction set :)   C >a new queue and to close up the old queue.  Before doing those sixoL >instructions,  they had to turn off interrupts,  and turn them on after theE >six instructions.  Otherwise interrupt level code might trip over an & >inconsistent state of the job queues. >rJ >This worked, but it was highly unaesthetic.  My understanding is that theF >VAX "insert queue" instructions,  or at least some of their variants,L >enabled VMS to maintain its queues without turning off the interrupts.  But: >I never learned VMS internals,  so I don't actually know. >s0 >Anybody out there who knows the answer to this?   -- Alan   ------------------------------    Date: 23 Jul 2001 09:11:34 -0500- From: koehler@encompasserve.org (Bob Koehler)7  Subject: Re: VMS Trivia Question3 Message-ID: <5zrABRAsE1w6@eisner.encompasserve.org>s  _ In article <y9f67.24$YQ.678@petpeeve.ziplink.net>, "David Cressey" <david@dcressey.com> writes:  > K > This worked, but it was highly unaesthetic.  My understanding is that theuG > VAX "insert queue" instructions,  or at least some of their variants,hM > enabled VMS to maintain its queues without turning off the interrupts.  Butl; > I never learned VMS internals,  so I don't actually know.t > 1 > Anybody out there who knows the answer to this?N >   D The INSQUE/REMQUE instructions make queue (aka doubly linked lists) E management quite easy.  INSQTI, INSQHI, REMQTI, and REMQHI add memorynD interlocks on the queue header so that multiple simultaneous threads? can safely access the queue without additional synchronization.o  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporatione= NASA GSFC Flight Software       | Federal Sector, Civil GroupiE                                 | please remove ".aspm" when replyingt   ------------------------------  % Date: Mon, 23 Jul 2001 10:20:11 -0400 & From: "Mike Duffy" <mdduffy@erols.com>" Subject: VS3100 & (dead?) ST51080N+ Message-ID: <9jhbpm$1mj$1@bob.news.rcn.net>   , Someone gave me a Seagate ST51080N disk with< the caveat "It's been on a shelf for a couple years.  Nobody knows if it works."   7 I plugged it into a VS3100 M76/SPX, V5.5-2H4.  It spinso. up, and all looks normal with >>> SHOW DEV and >>> TEST 6 and  >>> TEST 0.s  6 However, when the system is running, any MOUNT or INIT8 command fails with "Fatal drive error" and the following? entry appears in the error log.  Note the geometry information.n  8 I thought I'd make an effort here before dispatching the9 drive to the landfill.  Any suggestions to make the thing  work will be appreciated.l    /  ******************************* ENTRY     265.p *******************************nF  ERROR SEQUENCE 322.                             LOGGED ON:        SID 0B000006F  DATE/TIME 23-JUL-2001 10:03:58.52                            SYS_TYPE 04010002  SYSTEM UPTIME: 0 DAYS 00:01:34mF  SCS NODE: WOOGIE                                              VAX/VMS V5.5-2H4  8  DEVICE ERROR  KA43  CPU FW REV# 6.  CONSOLE FW REV# 0.1  ,  GENERIC DK SUB-SYSTEM, UNIT _WOOGIE$DKA200:          HW REVISION     33313930 9                                        HW REVISION = 0913R        ERROR TYPE            06 G                                        INVALID MODE SENSE DATA RETURNED         SCSI ID               02i3                                        SCSI ID = 2.         SCSI LUN              00I4                                        SCSI LUN = 0.        SCSI SUBLUN           00 7                                        SCSI SUBLUN = 0.         PORT STATUS     00000001 J                                        %SYSTEM-S-NORMAL, NORMAL SUCCESSFUL2                                         COMPLETION        SCSI CMD        007F001A                             0096m1                                        MODE SENSE         SCSI STATUS           00 +                                        GOODe    MODE SENSE DATA          MEDIUM TYPE           00e:                                        DEFAULT MEDIUM TYPE        FLAGS                 00i4                                        WRITE ENABLED        BLOCK DESC 1.   90312000a                        00020000AA                                        _DEFAULT DENSITY OF MEDIUM C                                        _NUMBER OF BLOCKS = 2109840. ;                                        _BLOCK LENGTH = 512.           PAGE DESC 1.    FFFF0A81-                        00000000                         000000FF-8                                        PAGE CODE = 01(X)@                                        ERROR RECOVERY PARAMETERS@                                        _DISABLE ERROR CORRECTIONA                                        _DISABLE TRANSFER ON ERRORT2                                        _POST ERROR?                                        _ENABLE EARLY CORRECTIONZ7                                        _READ CONTINUOUS A                                        _TRANSFER FAILING DATA BLKuE                                        _ENABLE AUTOMATIC READ REALLOC F                                        _ENABLE AUTOMATIC WRITE REALLOC8                                        _RETRY CNT = 255.<                                        _CORRECTION SPAN = 0. OF  V A X / V M S        SYSTEM ERROR REPORT         COMPILED 23-JUL-2001 10:06:13J                                                                       PAGE 2.  <                                        _HEAD OFFSET CNT = 0.C                                        _DATA STROBE OFFSET CNT = 0.0E                                        _RECOVERY TIME LIMIT = 0. MSEC           PAGE DESC 2.    FFFF0E82S                        00000000                         00000000S                        00000000F8                                        PAGE CODE = 02(X)J                                        DISCONNECT/RECONNECT CONTROL PARAMS@                                        _BUFFER FULL RATIO = 255.A                                        _BUFFER EMPTY RATIO = 255. F                                        _BUS INACTIVITY LIMIT = 0. USECG                                        _DISCONNECT TIME LIMIT = 0. USECeD                                        _CONNECT TIME LIMIT = 0. USEC          PAGE DESC 3.    00001683c                        00000000e                        00000000h                        00000000                         000000008                        00000000 8                                        PAGE CODE = 03(X)I                                        DIRECT ACCESS DEVICE FORMAT PARAMSl6                                        _0. TRACKS/ZONEA                                        _0. ALTERNATE SECTORS/ZONEt@                                        _0. ALTERNATE TRACKS/ZONE?                                        _0. ALTERNATE TRACKS/VOLe8                                        _0. SECTORS/TRACKA                                        _0. DATA BYTES/PHYS SECTOR 5                                        _0. INTERLEAVE >                                        _TRACK SKEW FACTOR = 0.A                                        _CYLINDER SKEW FACTOR = 0.           PAGE DESC 4.    000016841                        00000000n                        00000000                         00000000                         00000000u                        00000000t8                                        PAGE CODE = 04(X)G                                        RIGID DISK DRIVE GEOMETRY PARAMSK@                                        _NUMBER OF CYLINDERS = 0.<                                        _NUMBER OF HEADS = 0.D                                        _START CYL WRITE PRECOMP = 0.H                                        _START CYL REDUCE WRITE CURR = 0.<                                        _DRIVE STEP RATE = 0.=                                        _LANDING ZONE CYL = 0.           PAGE DESC 5.    FF0F0A87t                        00000000A                        00000000a8                                        PAGE CODE = 07(X)/                                        RESERVEDm          PAGE DESC 6.    00071288u iF  V A X / V M S        SYSTEM ERROR REPORT         COMPILED 23-JUL-2001 10:06:13J                                                                       PAGE 3.                          FFFF0000                         FFFFFFFFa                        0000FF20r                        00000000a8                                        PAGE CODE = 08(X)/                                        RESERVEDl          PAGE DESC 7.    F1000A8At                        00000000-                        00000000:8                                        PAGE CODE = 0A(X)/                                        RESERVED           PAGE DESC 8.    0000168Co                        FFFF00004                        00000000N                        00000000e                        00000000                         00000000m8                                        PAGE CODE = 0C(X)/                                        RESERVEDc        UCB$B_ERTCNT          C9 =                                        201. RETRIES REMAININGy        UCB$B_ERTMAX          DA =                                        218. RETRIES ALLOWABLE         ORB$L_OWNER     00000000c:                                        OWNER UIC [000,000]        UCB$L_CHAR      1CC54008 ;                                        DIRECTORY STRUCTURED 4                                        FILE ORIENTED/                                        SHARABLE 0                                        AVAILABLE4                                        ERROR LOGGING0                                        ALLOCATED7                                        CAPABLE OF INPUTs8                                        CAPABLE OF OUTPUT4                                        RANDOM ACCESS        UCB$W_STS           00001        UCB$L_OPCNT     0000000039                                        0. QIO'S THIS UNITt        UCB$W_ERRCNT        0001-:                                        1. ERRORS THIS UNIT        IRP$W_BCNT          0000 ?                                        TRANSFER SIZE 0. BYTE(S)6        IRP$W_BOFF          0000b<                                        TRANSFER PAGE ALIGNED        IRP$L_PID       00010012r6                                        REQUESTOR "PID"        IRP$Q_IOSB      008A78C1lE                        00000000        IOSB, 138. BYTE(S) TRANSFERREDM< ANAL/ERR/SINCE=23-JUL-2001 00:00:00.00/OUT=SEAGATE_ERROR.TXT   ------------------------------  % Date: Mon, 23 Jul 2001 11:04:51 -0400i2 From: rdeininger@mindspring.com (Robert Deininger)& Subject: Re: VS3100 & (dead?) ST51080NL Message-ID: <rdeininger-2307011104510001@user-2ive6lk.dialup.mindspring.com>  8 In article <9jhbpm$1mj$1@bob.news.rcn.net>, "Mike Duffy" <mdduffy@erols.com> wrote:  . > Someone gave me a Seagate ST51080N disk with> > the caveat "It's been on a shelf for a couple years.  Nobody > knows if it works."G > 9 > I plugged it into a VS3100 M76/SPX, V5.5-2H4.  It spins-0 > up, and all looks normal with >>> SHOW DEV and > >>> TEST 6 and  >>> TEST 0.R > 8 > However, when the system is running, any MOUNT or INIT: > command fails with "Fatal drive error" and the followingA > entry appears in the error log.  Note the geometry information., > : > I thought I'd make an effort here before dispatching the; > drive to the landfill.  Any suggestions to make the thingI > work will be appreciated.S >  > 1 >  ******************************* ENTRY     265. ! > *******************************sH >  ERROR SEQUENCE 322.                             LOGGED ON:        SID
 > 0B000006H >  DATE/TIME 23-JUL-2001 10:03:58.52                            SYS_TYPE
 > 04010002! >  SYSTEM UPTIME: 0 DAYS 00:01:34 H >  SCS NODE: WOOGIE                                              VAX/VMS
 > V5.5-2H4 > : >  DEVICE ERROR  KA43  CPU FW REV# 6.  CONSOLE FW REV# 0.1 > . >  GENERIC DK SUB-SYSTEM, UNIT _WOOGIE$DKA200: > ! >        HW REVISION     33313930r; >                                        HW REVISION = 0913v! >        ERROR TYPE            06 I >                                        INVALID MODE SENSE DATA RETURNEDe  J This is often a sign that the device driver sees something about the driveI that is doesn't like or doesn't understand.  It doesn't mean the drive isd definitely bad.   A In at least 3 cases, I've seen VMS use these drives happily after H upgrading to a more modern version.  5.5-2H4 is pretty old.  Can you try( the drive on a system with 7.1 or later?   -- - Robert Deininger rdeininger@mindspring.com0   ------------------------------   Date: 23 Jul 2001 16:24:15 GMT& From: Wolfgang Rupp <rupp@coredump.at>& Subject: Re: VS3100 & (dead?) ST51080N3 Message-ID: <9jhj3c$o212k$1@ID-3131.news.dfncis.de>M  % Mike Duffy <mdduffy@erols.com> wrote:r1 > I plugged it into a VS3100 M76/SPX, V5.5-2H4.  t  B I am not sure about using it, but the M76 cannot boot from a driveB larger than 1GB because of some ROM limitation. Some overflow that? wraps around addressing. This also applies to 3100M30/40/38/48.e  
 Wolfgang Ruppp   ------------------------------  # Date: Mon, 23 Jul 2001 16:45:15 GMTa0 From: "P. Thompson" <thompson-nospam@new.rr.com>& Subject: Re: VS3100 & (dead?) ST51080NJ Message-ID: <Pine.LNX.4.33.0107231143470.26994-100000@malacandra.localnet>  H You might consider installing RZDISK from the freeware CD.  It will tell/ you possibly useful information about the disk.l  J There are also options to restore the drive's code to the factory defaultsD or modify some of the parameters which might help in this situation.  & On Mon, 23 Jul 2001, Mike Duffy wrote:  . > Someone gave me a Seagate ST51080N disk with> > the caveat "It's been on a shelf for a couple years.  Nobody > knows if it works.". >h9 > I plugged it into a VS3100 M76/SPX, V5.5-2H4.  It spinsa0 > up, and all looks normal with >>> SHOW DEV and > >>> TEST 6 and  >>> TEST 0.f >t8 > However, when the system is running, any MOUNT or INIT: > command fails with "Fatal drive error" and the followingA > entry appears in the error log.  Note the geometry information.o >a: > I thought I'd make an effort here before dispatching the; > drive to the landfill.  Any suggestions to make the thingi > work will be appreciated.D >e >e1 >  ******************************* ENTRY     265. ! > *******************************IH >  ERROR SEQUENCE 322.                             LOGGED ON:        SID
 > 0B000006H >  DATE/TIME 23-JUL-2001 10:03:58.52                            SYS_TYPE
 > 04010002! >  SYSTEM UPTIME: 0 DAYS 00:01:34-H >  SCS NODE: WOOGIE                                              VAX/VMS
 > V5.5-2H4 >-: >  DEVICE ERROR  KA43  CPU FW REV# 6.  CONSOLE FW REV# 0.1 >-. >  GENERIC DK SUB-SYSTEM, UNIT _WOOGIE$DKA200: >i! >        HW REVISION     33313930m; >                                        HW REVISION = 0913:! >        ERROR TYPE            06oI >                                        INVALID MODE SENSE DATA RETURNED ! >        SCSI ID               0275 >                                        SCSI ID = 2.F! >        SCSI LUN              00a6 >                                        SCSI LUN = 0.! >        SCSI SUBLUN           00M9 >                                        SCSI SUBLUN = 0.e! >        PORT STATUS     00000001-L >                                        %SYSTEM-S-NORMAL, NORMAL SUCCESSFUL4 >                                         COMPLETION! >        SCSI CMD        007F001Aa! >                            009623 >                                        MODE SENSE>! >        SCSI STATUS           00s- >                                        GOOD0 >h >  MODE SENSE DATA >o! >        MEDIUM TYPE           00i< >                                        DEFAULT MEDIUM TYPE! >        FLAGS                 00o6 >                                        WRITE ENABLED! >        BLOCK DESC 1.   90312000s! >                        00020000 C >                                        _DEFAULT DENSITY OF MEDIUM E >                                        _NUMBER OF BLOCKS = 2109840.o= >                                        _BLOCK LENGTH = 512.s >i! >        PAGE DESC 1.    FFFF0A81e! >                        00000000e! >                        000000FFa: >                                        PAGE CODE = 01(X)B >                                        ERROR RECOVERY PARAMETERSB >                                        _DISABLE ERROR CORRECTIONC >                                        _DISABLE TRANSFER ON ERRORi4 >                                        _POST ERRORA >                                        _ENABLE EARLY CORRECTIONm9 >                                        _READ CONTINUOUSiC >                                        _TRANSFER FAILING DATA BLKsG >                                        _ENABLE AUTOMATIC READ REALLOC H >                                        _ENABLE AUTOMATIC WRITE REALLOC: >                                        _RETRY CNT = 255.> >                                        _CORRECTION SPAN = 0. >iH >  V A X / V M S        SYSTEM ERROR REPORT         COMPILED 23-JUL-2001
 > 10:06:13L >                                                                       PAGE > 2. >c> >                                        _HEAD OFFSET CNT = 0.E >                                        _DATA STROBE OFFSET CNT = 0.tG >                                        _RECOVERY TIME LIMIT = 0. MSEC  >e! >        PAGE DESC 2.    FFFF0E82 ! >                        00000000h! >                        00000000E! >                        00000000a: >                                        PAGE CODE = 02(X)L >                                        DISCONNECT/RECONNECT CONTROL PARAMSB >                                        _BUFFER FULL RATIO = 255.C >                                        _BUFFER EMPTY RATIO = 255.-H >                                        _BUS INACTIVITY LIMIT = 0. USECI >                                        _DISCONNECT TIME LIMIT = 0. USECiF >                                        _CONNECT TIME LIMIT = 0. USEC >y! >        PAGE DESC 3.    00001683-! >                        000000001! >                        00000000f! >                        00000000a! >                        00000000m! >                        00000000a: >                                        PAGE CODE = 03(X)K >                                        DIRECT ACCESS DEVICE FORMAT PARAMSt8 >                                        _0. TRACKS/ZONEC >                                        _0. ALTERNATE SECTORS/ZONEeB >                                        _0. ALTERNATE TRACKS/ZONEA >                                        _0. ALTERNATE TRACKS/VOLl: >                                        _0. SECTORS/TRACKC >                                        _0. DATA BYTES/PHYS SECTORg7 >                                        _0. INTERLEAVE @ >                                        _TRACK SKEW FACTOR = 0.C >                                        _CYLINDER SKEW FACTOR = 0.L >E! >        PAGE DESC 4.    00001684I! >                        00000000 ! >                        000000002! >                        00000000 ! >                        00000000 ! >                        00000000 : >                                        PAGE CODE = 04(X)I >                                        RIGID DISK DRIVE GEOMETRY PARAMSIB >                                        _NUMBER OF CYLINDERS = 0.> >                                        _NUMBER OF HEADS = 0.F >                                        _START CYL WRITE PRECOMP = 0.J >                                        _START CYL REDUCE WRITE CURR = 0.> >                                        _DRIVE STEP RATE = 0.? >                                        _LANDING ZONE CYL = 0.  > ! >        PAGE DESC 5.    FF0F0A870! >                        00000000 ! >                        00000000 : >                                        PAGE CODE = 07(X)1 >                                        RESERVED  > ! >        PAGE DESC 6.    00071288  >0H >  V A X / V M S        SYSTEM ERROR REPORT         COMPILED 23-JUL-2001
 > 10:06:13L >                                                                       PAGE > 3. > ! >                        FFFF0000 ! >                        FFFFFFFFU! >                        0000FF20 ! >                        00000000 : >                                        PAGE CODE = 08(X)1 >                                        RESERVED  > ! >        PAGE DESC 7.    F1000A8AM! >                        00000000 ! >                        000000008: >                                        PAGE CODE = 0A(X)1 >                                        RESERVED  > ! >        PAGE DESC 8.    0000168C ! >                        FFFF0000 ! >                        000000000! >                        00000000 ! >                        00000000 ! >                        00000000 : >                                        PAGE CODE = 0C(X)1 >                                        RESERVED ! >        UCB$B_ERTCNT          C9R? >                                        201. RETRIES REMAININGZ! >        UCB$B_ERTMAX          DA ? >                                        218. RETRIES ALLOWABLER! >        ORB$L_OWNER     00000000 < >                                        OWNER UIC [000,000]! >        UCB$L_CHAR      1CC54008 = >                                        DIRECTORY STRUCTURED 6 >                                        FILE ORIENTED1 >                                        SHARABLEM2 >                                        AVAILABLE6 >                                        ERROR LOGGING2 >                                        ALLOCATED9 >                                        CAPABLE OF INPUT : >                                        CAPABLE OF OUTPUT6 >                                        RANDOM ACCESS! >        UCB$W_STS           0000E! >        UCB$L_OPCNT     00000000 ; >                                        0. QIO'S THIS UNIT ! >        UCB$W_ERRCNT        0001 < >                                        1. ERRORS THIS UNIT! >        IRP$W_BCNT          0000EA >                                        TRANSFER SIZE 0. BYTE(S)F! >        IRP$W_BOFF          0000 > >                                        TRANSFER PAGE ALIGNED! >        IRP$L_PID       00010012N8 >                                        REQUESTOR "PID"! >        IRP$Q_IOSB      008A78C1UG >                        00000000        IOSB, 138. BYTE(S) TRANSFERREDC> > ANAL/ERR/SINCE=23-JUL-2001 00:00:00.00/OUT=SEAGATE_ERROR.TXT >0 >  >  >  >  >    -- 0* Leave the nospam in, correct email address   ------------------------------  % Date: Mon, 23 Jul 2001 10:39:28 +0100 0 From: andrew harrison <andrew.nospam@uk.sun.com>Y Subject: Re: Wailing and moaning.... (was: Reward for the first of the next 50...) 50...) * Message-ID: <3B5BF0D0.DEEE63E5@uk.sun.com>   Larry Kilgallen wrote: > _ > In article <3B56A150.8B23833D@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:E > >  > > Fred Kleinsorge wrote: > >>O > >> A quick check of the source listings seems to indicate that I am.  And theAH > >> hot air you blow seems to be the only indication of what you do ;-) > >> > > 8 > > Really, I would suggest that since the 25th your own1 > > contributions have had an element of warm air 1 > > polluting what is normally relatively factualA1 > > content :):):) hence my re-evaluation of your 	 > > role.  > I > Andrew, given Fred's position, we look to him for authoritative answers0H > on what VMS Development engineers want to do.  That is not necessarilyG > what the company management will end up doing, but even in the Compaq 8 > environment, it is a powerful influence on management. >   9 Really, note that I said since the 25th. Your point about ; "powerful influence" before that date might well have been    defensible, since then it isn't.  < From coments made by various Compaq representatives on this 9 newsgroup it is clear that VMS development engineers were ; either consulted very very close to the 25th or not at all.   9 The decision to move OpenVMS to IA-64 is one of the most G9 important development decisions that has been made about  : OpenVMS in the last 5 years. A powerfull influencer would = have been consulted, there is no evidence that this happened, = if it had then why do "powerfull influencers" post responses /> like we have only had 13 days to look at the porting issues ??     Regards  Andrew Harrison  Enterprise IT Architect    ------------------------------    Date: 23 Jul 2001 09:14:29 -0500- From: koehler@encompasserve.org (Bob Koehler) $ Subject: Re: Zero Quadword Time Poll3 Message-ID: <Quxthy3JlONl@eisner.encompasserve.org>P  \ In article <3B58B1F3.906BFBA1@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> writes: > 7 > In Unix, doesn't a time value of 0 mean 01-JAN-1970 ?   ? It means 01-JAN-1970 00:00:00.00 GMT.  You can get other values(E depending on you timezone, including 31-DEC-1969 12:00 through 23:59.C  F ----------------------------------------------------------------------? Bob Koehler                     | Computer Sciences Corporation = NASA GSFC Flight Software       | Federal Sector, Civil Group E                                 | please remove ".aspm" when replying    ------------------------------    Date: 23 Jul 2001 09:48:50 -0500 From: briggs@encompasserve.org$ Subject: Re: Zero Quadword Time Poll3 Message-ID: <V3NHm9HJnhgj@eisner.encompasserve.org>8  W In article <tlhkmm9qivbga9@news.supernews.com>, "John Vottero" <John@mvpsi.com> writes: < > "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message( > news:3B58B1F3.906BFBA1@videotron.ca... >> John Vottero wrote:J >> > No one is saying that a quadword equal to 0 means "the current time". >> > >  > [snip] >  >>F >> Also, consider the cases where one need to calculate the difference	 > betweenRL >> VMS base time and some other time value. (comparing to eras for instance, > the I >> GPS era of Jan 1 1990 with the VMS era of 1858). If a quadword of 0 noP > longer6 >> means the 1858 date, how would one calculate that ? >> > F > It seems to me that this is the case that should kill this proposal.  9 Following up the follow-up becasue I missed the original.B  > I don't follow the reasoning here.  Is someone planning to use> LIB$SUB_TIMES to compute the offset of the Unix base date from the VMS base date:   	Old:A  + 	SYS$BINTIM ( 01-jan-1970, unix-base-quad ) * 	SYS$BINTIM ( 17-nov-1858, vms-base-quad ) 	; Yielding zeroE 	LIB$SUB_TIMES ( unix-base-quad, vms-base-quad, quadword-difference ) ' 	; Yielding delta as negative quadword.I   	New:0  + 	SYS$BINTIM ( 01-jan-1970, unix-base-quad ) * 	SYS$BINTIM ( 17-nov-1858, vms-base-quad )@ 	; Failing -- let's assume vms-base-quad was initialized to zeroE 	LIB$SUB_TIMES ( unix-base-quad, vms-base-quad, quadword-difference )i' 	; Yielding delta as positive quadword.   D I would have thought that the more straightforward way of doing this would be to:  0 	SYS$BINTIM ( 01-jan-1970, quadword-difference ). 	; Which will yield delta as positive quadword  E Note that I would never consider using LIB$ADD_TIMES or LIB$SUB_TIMEShD for VMS date arithmetic.  LIB$ADDX and LIB$SUBX are vastly superior. In my opinion, of course.e   	John Briggs   ------------------------------  % Date: Mon, 23 Jul 2001 11:57:30 -0400D% From: "John Vottero" <John@mvpsi.com> $ Subject: Re: Zero Quadword Time Poll/ Message-ID: <tloiaaq4daeo67@news.supernews.com>a  : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3B5B96B9.2469AE0A@videotron.ca... > John Vottero wrote:eK > > I don't see the problem.  If you subtract zero from something, you windi up > > with the same thing. >lL > Subtract an absolute time from an absolute and you get a delta time. (now)L > Subtract a delta time from an absolute and you get an absolute. (proposed) > G Yes, you're right.  In this case, subtracting zero from a number should5J change it's sign (to make it a delta time).  And that's what lib$sub_timesF does.  However, sys$asctim refuses to format the result, it fails with %SYSTEM-F-INVTIME.  E This is all getting a little gray.  Doing things like subtracting the H absolute time zero from the current time should return a delta time but,J that violates the 10,000 day rule.  Was the 10,000 day rule eliminated for6 the Unix time fix?  Or was the 10,000 value increased?  L > The system calls will continue to return a success in the vast majority of the 0 > cases, but the result will be quite different. >tH > I guess it comes back to the engineer's original question: does anyoneH > calculate delta-time between an absolute date and the smitsonian date. > K > Such a change would require that everyone check their code to ensure thatt thehK > change won't affect their code. Changing their code might be easy, but itm isF > still a change to some code to make it work (with all the hassles of having1 > different code depending on version of the OS).: >3' > What benefit would the change bring ?3  0 I'm still wondering what the benefit is as well.   ------------------------------  % Date: Mon, 23 Jul 2001 17:09:29 +0100t* From: "Richard Brodie" <R.Brodie@rl.ac.uk>$ Subject: Re: Zero Quadword Time Poll, Message-ID: <9jhi7r$1m3k@newton.cc.rl.ac.uk>  Z "John Vottero" <John@mvpsi.com> wrote in message news:tloiaaq4daeo67@news.supernews.com...  ) > > What benefit would the change bring ?o >22 > I'm still wondering what the benefit is as well.  F    It would move the burden of special casing 0 from those using deltaE times to those computing absolute times. A delta time of zero must be @ more common than actually needing to refer to the VMS base time.L    Whether it's worth breaking backward compatibility for is another matter.   $ spawn/nolog/nosymb show statM   Status on  23-JUL-2001 17:01:14.33     Elapsed CPU :17-NOV-1858 00:00:00.00iF   Buff. I/O :       28    Cur. ws. :   16384    Open files :         0F   Dir. I/O :         0    Phys. Mem. :   976    Page Faults :       70   ------------------------------  % Date: Mon, 23 Jul 2001 08:52:05 +02001> From: "Jean-Francois Marchal" <jean-francois.marchal@x9000.fr>6 Subject: Re: [DWMOTIF V1.2-6] VAX EURO patch missing ?. Message-ID: <9jghgs$p82$1@reader1.imaginet.fr>  A As for the Euro Currency, is there an elegant way of embedding it < in text files to be printed on Postscript printers via DCPS, or on PCL printers ?   Jean-Franois Marchale X9000 - LYON (FR)e  D "Bob Knowles" <bob.knowles@compaq.com> a crit dans le message news:2 da08649c.0107190043.483a2e53@posting.google.com...7 > eplan@kapsch.net (Peter LANGSTOEGER) wrote in message*& news:<3b548851$1@news.kapsch.co.at>...F > > There is still no EURO currency symbol support in DECwindows-MOTIF V1.2-6.:B > > Why ? Is the EURO such a new invention ? (or is it the - we're
 americans,% > > we don't need EUROs - problem ;-)  > G > The 'we're Americans - we don't need Euros' problem certainly doesn'tOB > affect all OpenVMS products. The (UK-based) developers of CompaqD > Office Server V6.1 (due to Field Test later this year) are gettingB > there as quickly as possible. Far from ignoring the Euro we evenH > support or shortly will support the Euro symbol on older products suchC > as FMS and Office Server. For more information about the mail and 1 > messaging products supported by this group, seeFH > http://www.openvms.compaq.com/commercial/newsletter/newsletter.html or? > sign up for a copy of the newsletter to be e-mailed to you at D > http://www.openvms.compaq.com/commercial/offsvr/offsvr_signup.html   ------------------------------    Date: 23 Jul 2001 03:49:35 -0700* From: bob.knowles@compaq.com (Bob Knowles)6 Subject: Re: [DWMOTIF V1.2-6] VAX EURO patch missing ?< Message-ID: <da08649c.0107230249.b4ef583@posting.google.com>  t "Jean-Francois Marchal" <jean-francois.marchal@x9000.fr> wrote in message news:<9jghgs$p82$1@reader1.imaginet.fr>...C > As for the Euro Currency, is there an elegant way of embedding it > > in text files to be printed on Postscript printers via DCPS, > or on PCL printers ?  E I know next to nothing about DCPS except as the naivest of users, but E I'm told that DCPS from 1.7 onwards supports the Euro symbol; for the   ANSI translator you need to add:  D <ESC>.b<ESC>n$	- the first escape sequence means Select ISO Latin-9, the second prints Euro symbol   ? A useful document although a trifle dated is "Printing the Euro D Currency Symbol on DIGITAL Printers", it was written before the DCPS@ changes were made but fairly accurately describes the situation.3 <http://vmsres.zko.dec.com/euro/printing-euro.html>   8 On Peter's original query, the ECO you're looking for is VAX_DWEURO-V0101-1 >_F >Copyright (c) Compaq Computer Corporation 2001. All rights reserved. 9 >PRODUCT: Compaq DECwindows Motif V1.2-6 for OpenVMS VAX   >OP/SYS: OpenVMS VAX   >COMPONENT: DECwindows  % >SOURCE: Compaq Computer Corporation   >ECO INFORMATION: " >ECO Kit Name: VAX_DWEURO-V0101-1 & >DEC-VAXVMS-VAX_DWEURO-V0101-1-1.PCSI + >ECO Kits Superseded by This ECO Kit: None  + >ECO Kit Approximate Size: 8224 Blocks Kit A< >Applies To: Compaq DECwindows Motif V1.2-6 for OpenVMS VAX ( >OpenVMS VAX V5.5-2, V6.2, 7.2 and 7.3. % >System/Cluster Reboot Necessary: No .6 >Rolling Re-boot Supported: Information Not Available B >Installation Rating: INSTALL_2 2 - To be installed on all systemsA running the >listed version(s) of OpenVMS and using the following  feature(s): Euro currency >sign    bE   ------------------------------    Date: 23 Jul 2001 13:24:50 +0200* From: eplan@kapsch.net (Peter LANGSTOEGER)6 Subject: Re: [DWMOTIF V1.2-6] VAX EURO patch missing ?* Message-ID: <3b5c0982$1@news.kapsch.co.at>  i In article <da08649c.0107230249.b4ef583@posting.google.com>, bob.knowles@compaq.com (Bob Knowles) writes:0L >On Peter's original query, the ECO you're looking for is VAX_DWEURO-V0101-1  - And is where ? Yes, I guessed this name, too. ) But I so far found no traces of this ECO.T+ Is is already public (or field test only) ?    -- _< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888 < <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------    Date: 23 Jul 2001 15:42:27 +0200* From: eplan@kapsch.net (Peter LANGSTOEGER): Subject: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?* Message-ID: <3b5c29c3$1@news.kapsch.co.at>  J Last week, both cache batteries of our pair of HSJ40 died (two weeks afterK the service contract ended) and the Mirror- and RAID-Sets got "INOPERATIVE"  and my (CI) VMScluster freezed.P  I I looked into the HSJ, saw the situation and rebooted the HSJ for a quick I start and the cluster continued. And then the weekend (with a visit by my  newborn nieces ;-) came.  B A couple of hours later the HSJ repeated the hang. Appearantly theB cache batteries got "low" on reboot and this couple of hours laterH the HSJ declared the cache battery as "failed" (note: the CACHE is GOOD,K only the Cache Battery is FAILED) and therefore the disks got "INOPERATIVE"  again.  I So, I learned that I had the wrong CACHE_POLICY set (A, instead of B) and K fixed it, rebooted the HSJ and the cluster continued. But a couple of hoursI later, the hang returned.0  C So, I learned to switch off WRITEBACK_CACHE of all mirror/raid sets0@ (Is this correct ?) Rebooted the HSJs and the cluster continued.    , This morning, the cluster was freezed again.K Seems, I don't understand why a good cache battery is required to run disks # on the HSJ in READ_CACHE only mode.   H Currently, I have a HSJ which has a 'FAILED' cache battery just from theL startup, which does not serve disks regardless the PREFERRED_PATH in the HSJI and a HSJ which has a 'LOW' cache battery from boot on, serving all disks J until the HSJ declares the cache battery 'FAILED' a couple of hours later.  C What to do now ? I can live with reduced functionality/performance,N2 but I can't live with cluster hangs twice a day...  N Do I have to replace the batteries at all (which is hard without a contract) ?@ Do I have to dismount the batteries to get a working cache again 	(or is there some parameter) ?SI Do I have to disable the READ_CACHE on all device units, too (unlikely) ? 5 Do I have to reboot the whole VMScluster (unlikely) ?0  E Again, do I really need HSJ cache batteries to use MIRROR/RAID sets ?    Many TIA   -Peter  J PS: I just saw, that the cache battery of the 2nd HSJ "is now sufficientlyM charged". Strange things happen. Probably, the VMScluster freeze won't happen I for the next couple of days, but my understanding still lacks (and I want L the 1st HSJ to be usable as a hot-standby/load-balancing unit, for the 2nd).
 So, please...    --  < Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888 < <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------  % Date: Mon, 23 Jul 2001 10:00:36 -0400 + From: "Main, Kerry" <Kerry.Main@compaq.com> > Subject: RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?R Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4AD7F26@kaoexc01.americas.cpqcorp.net>   Well,   . Something to consider as a temporary measure -  3 - set all cache stuff to write through on all units H HSJ80> SET D100 NOWRITEBACK (I assume the command is equivalent of older HSJxx's)  H - set HSJ to UPS mode (I believe this allows HSJxx's to function without cache batteries)  H I used this for a few days during a new install (no prod data, but write. back setup) until the cache batteries arrived.   Regards   
 Kerry Main Senior Consultant  Compaq Canada Inc. Professional Services  Voice: 613-592-4660  Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----0 From: eplan@kapsch.net [mailto:eplan@kapsch.net] Sent: July 23, 2001 9:42 AM  To: Info-VAX@Mvb.Saic.ComE: Subject: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?    J Last week, both cache batteries of our pair of HSJ40 died (two weeks afterK the service contract ended) and the Mirror- and RAID-Sets got "INOPERATIVE"E and my (CI) VMScluster freezed.   I I looked into the HSJ, saw the situation and rebooted the HSJ for a quick0I start and the cluster continued. And then the weekend (with a visit by my  newborn nieces ;-) came.  B A couple of hours later the HSJ repeated the hang. Appearantly theB cache batteries got "low" on reboot and this couple of hours laterH the HSJ declared the cache battery as "failed" (note: the CACHE is GOOD,K only the Cache Battery is FAILED) and therefore the disks got "INOPERATIVE"N again.  I So, I learned that I had the wrong CACHE_POLICY set (A, instead of B) andvK fixed it, rebooted the HSJ and the cluster continued. But a couple of hoursa later, the hang returned.2  C So, I learned to switch off WRITEBACK_CACHE of all mirror/raid sets:@ (Is this correct ?) Rebooted the HSJs and the cluster continued.    , This morning, the cluster was freezed again.K Seems, I don't understand why a good cache battery is required to run disksc# on the HSJ in READ_CACHE only mode.k  H Currently, I have a HSJ which has a 'FAILED' cache battery just from theL startup, which does not serve disks regardless the PREFERRED_PATH in the HSJI and a HSJ which has a 'LOW' cache battery from boot on, serving all disks J until the HSJ declares the cache battery 'FAILED' a couple of hours later.  C What to do now ? I can live with reduced functionality/performance,l2 but I can't live with cluster hangs twice a day...  L Do I have to replace the batteries at all (which is hard without a contract) ?h@ Do I have to dismount the batteries to get a working cache again 	(or is there some parameter) ?tI Do I have to disable the READ_CACHE on all device units, too (unlikely) ?o5 Do I have to reboot the whole VMScluster (unlikely) ?   E Again, do I really need HSJ cache batteries to use MIRROR/RAID sets ?l   Many TIA   -Peter  J PS: I just saw, that the cache battery of the 2nd HSJ "is now sufficientlyF charged". Strange things happen. Probably, the VMScluster freeze won't happenI for the next couple of days, but my understanding still lacks (and I wantoL the 1st HSJ to be usable as a hot-standby/load-balancing unit, for the 2nd).
 So, please...n   -- i< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888o< <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------    Date: 23 Jul 2001 11:46:30 -0500+ From: young_r@encompasserve.org (Rob Young) > Subject: Re: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?3 Message-ID: <Iul+JBJX8CjZ@eisner.encompasserve.org>o  W In article <3b5c29c3$1@news.kapsch.co.at>, eplan@kapsch.net (Peter LANGSTOEGER) writes:iL > Last week, both cache batteries of our pair of HSJ40 died (two weeks afterM > the service contract ended) and the Mirror- and RAID-Sets got "INOPERATIVE"0! > and my (CI) VMScluster freezed.e > K > I looked into the HSJ, saw the situation and rebooted the HSJ for a quick-K > start and the cluster continued. And then the weekend (with a visit by my  > newborn nieces ;-) came. > D > A couple of hours later the HSJ repeated the hang. Appearantly theD > cache batteries got "low" on reboot and this couple of hours laterJ > the HSJ declared the cache battery as "failed" (note: the CACHE is GOOD,M > only the Cache Battery is FAILED) and therefore the disks got "INOPERATIVE"e > again. > K > So, I learned that I had the wrong CACHE_POLICY set (A, instead of B) andnM > fixed it, rebooted the HSJ and the cluster continued. But a couple of hourso > later, the hang returned.a > E > So, I learned to switch off WRITEBACK_CACHE of all mirror/raid setsNB > (Is this correct ?) Rebooted the HSJs and the cluster continued. >  > . > This morning, the cluster was freezed again.M > Seems, I don't understand why a good cache battery is required to run diskss% > on the HSJ in READ_CACHE only mode.s >   > 	Because, as I pointed out here before, unbeknownst to most ofG 	us, a mirror-set uses writeback cache whether or not set to READ_CACHE  	*only*l  G 	As I pointed out earlier, try creating a RAID-1 with failed batteries. ? 	It can't be done.  Maybe you can live dangerously and tell the A 	controller you have a UPS.  Note:  I haven't tried this.  And ifnB 	the worst possible scenario strikes you (i.e. you lose datacenterB 	power, or someone *accidently* unplugs your HSJ40s, etc. you have# 	lost writes.)  And no I'm not sure1B 	it even works, i.e. with a UPS , battery status doesn't matter or 	even come into play.d    J > Currently, I have a HSJ which has a 'FAILED' cache battery just from theN > startup, which does not serve disks regardless the PREFERRED_PATH in the HSJK > and a HSJ which has a 'LOW' cache battery from boot on, serving all disksBL > until the HSJ declares the cache battery 'FAILED' a couple of hours later. > E > What to do now ? I can live with reduced functionality/performance,r4 > but I can't live with cluster hangs twice a day... > P > Do I have to replace the batteries at all (which is hard without a contract) ?  ; 	You can buy batteries in the aftermarket.  Since you state_B 	they are HSJ-40, that is a bit more painful as they are internal.. 	But still very doable (simple if you ask me).   > G > Again, do I really need HSJ cache batteries to use MIRROR/RAID sets ?D >   > 	Maybe.  Maybe if you tell them they are plugged into UPS, youH 	can live dangerously and it works.  But yes, if you use batteries :^O .   				Robm   ------------------------------    Date: 23 Jul 2001 17:46:00 +0200* From: eplan@kapsch.net (Peter LANGSTOEGER)> Subject: RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?* Message-ID: <3b5c46b8$1@news.kapsch.co.at>   In article <BE56C50EA024184DAF48F0B9A47F5CF4AD7F26@kaoexc01.americas.cpqcorp.net>, "Main, Kerry" <Kerry.Main@compaq.com> writes:/ >Something to consider as a temporary measure -e  M and what's the final measure (temporary means: until VMS is discarded here) ?m  4 >- set all cache stuff to write through on all unitsI >HSJ80> SET D100 NOWRITEBACK (I assume the command is equivalent of olderh	 >HSJxx's)e   That's what I did (as I wrote)  I >- set HSJ to UPS mode (I believe this allows HSJxx's to function withoutn >cache batteries)   ? Thats a curious idea, but I don't find commands for this so fart> (maybe HSJ80 only ?). Or did you really mean: Connect an UPS ?  6 Thanks for the hint, but the biggest question remains:C 	Is a HSJ cache battery really neccessary to run MIRROR/RAID sets ?fD 	If yes, a HSJ without a battery is as useful as a bunch of disks...  C And the other question remains also: Is cache_policy=B a big risk ?    -- i< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888i< <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------  % Date: Mon, 23 Jul 2001 12:36:05 -0400d+ From: "Main, Kerry" <Kerry.Main@compaq.com>h> Subject: RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?R Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF4AD7F2A@kaoexc01.americas.cpqcorp.net>  I >>> and what's the final measure (temporary means: until VMS is discardedc here) ? <<<   J I would presume that this is a temporary measure until such time as properG batteries can be obtained. Imho, I would not run like this for extendedd periods.  J >>> Thats a curious idea, but I don't find commands for this so far (maybe: HSJ80 only ?). Or did you really mean: Connect an UPS ?<<<  I There is a way to tell the HSJxx (and I believe this exists on the HSJ40,tK but not sure) that you have a UPS in the datacenter (even if you don't) andoL don't need cache batteries. Ok, you do, but this is a way to tell the HSJ toD ignore the fact that the batteries are not there and/or not working.  L Not sure if this is also on the HSJ40 or not, but from the CLI guide - under" "SET THIS [OTHER]option" command:    option:   
 UPS=NODE_ONLY- UPS=DATACENTER_WIDE  NOUPSo  I Specifies whether the controller should perform regular battery conditionmJ checks. Setting the UPS/NOUPS switch for either controller sets the switchI for both controllers. Both controllers must be restarted in order for thes new setting to take effect.d  H - Specify UPS=NODE_ONLY if your storage subsystem power is supplied by aJ separate uninterruptible power supply (UPS) directly connected to the PVA.L If NODE_ONLY is specified the controller continues to check the condition of the cache batteries.  K CAUTION: Setting UPS =NODE_ONLY or UPS=DATACENTER_WIDE without having a UPS4E or similar backup system in place may result in data loss if power isl interrupted.  I - Specify UPS=DATACENTER_WIDE if the whole room or building (not just the F subsystem) is supported by a UPS. If DATACENTER_WIDE is specified, theJ controller does not check the condition of the cache batteries and ignoresE the battery's state. This causes RAIDsets and mirrorsets to always bei> available, regardless of the condition of the cache batteries.  K - Specify NOUPS to instruct the controller to perform regular cache batteryr9 checks and evaluate the condition of the cache batteries.h   Regards,  
 Kerry Main Senior Consultanta Compaq Canada Inc. Professional Servicesc Voice: 613-592-4660g Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----0 From: eplan@kapsch.net [mailto:eplan@kapsch.net] Sent: July 23, 2001 11:46 AM To: Info-VAX@Mvb.Saic.Comi> Subject: RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?    
 In articleG <BE56C50EA024184DAF48F0B9A47F5CF4AD7F26@kaoexc01.americas.cpqcorp.net>,R- "Main, Kerry" <Kerry.Main@compaq.com> writes:e/ >Something to consider as a temporary measure -o  K and what's the final measure (temporary means: until VMS is discarded here)e ?g  4 >- set all cache stuff to write through on all unitsI >HSJ80> SET D100 NOWRITEBACK (I assume the command is equivalent of older 	 >HSJxx's)e   That's what I did (as I wrote)  I >- set HSJ to UPS mode (I believe this allows HSJxx's to function withouta >cache batteries)   ? Thats a curious idea, but I don't find commands for this so far > (maybe HSJ80 only ?). Or did you really mean: Connect an UPS ?  6 Thanks for the hint, but the biggest question remains:C 	Is a HSJ cache battery really neccessary to run MIRROR/RAID sets ?"D 	If yes, a HSJ without a battery is as useful as a bunch of disks...  C And the other question remains also: Is cache_policy=B a big risk ?h   -- t< Peter "EPLAN" LANGSTOEGER           Tel.    +43 1 81111-2651; Network and OpenVMS system manager  Fax.    +43 1 81111-888_< <<< KAPSCH AG  Wagenseilgasse 1     E-mail  eplan@kapsch.netH A-1121 VIENNA  AUSTRIA              "I'm not a pessimist, I'm a realist"   ------------------------------    Date: 23 Jul 2001 12:54:50 -0500+ From: young_r@encompasserve.org (Rob Young)R> Subject: RE: [HSJ40] Bad Cache Battery hangs OpenVMS Cluster ?3 Message-ID: <0QXX5DgnqOa2@eisner.encompasserve.org>O   In article <BE56C50EA024184DAF48F0B9A47F5CF4AD7F2A@kaoexc01.americas.cpqcorp.net>, "Main, Kerry" <Kerry.Main@compaq.com> writes: > J >>>> and what's the final measure (temporary means: until VMS is discarded
 > here) ? <<<t > L > I would presume that this is a temporary measure until such time as properI > batteries can be obtained. Imho, I would not run like this for extendedo
 > periods. > K >>>> Thats a curious idea, but I don't find commands for this so far (maybe3< > HSJ80 only ?). Or did you really mean: Connect an UPS ?<<< > K > There is a way to tell the HSJxx (and I believe this exists on the HSJ40,0M > but not sure) that you have a UPS in the datacenter (even if you don't) andgN > don't need cache batteries. Ok, you do, but this is a way to tell the HSJ toF > ignore the fact that the batteries are not there and/or not working. > N > Not sure if this is also on the HSJ40 or not, but from the CLI guide - under$ > "SET THIS [OTHER]option" command:  > 	 > option:E >  > UPS=NODE_ONLY  > UPS=DATACENTER_WIDE1 > NOUPSe > K > Specifies whether the controller should perform regular battery conditioneL > checks. Setting the UPS/NOUPS switch for either controller sets the switchK > for both controllers. Both controllers must be restarted in order for the  > new setting to take effect.m > J > - Specify UPS=NODE_ONLY if your storage subsystem power is supplied by aL > separate uninterruptible power supply (UPS) directly connected to the PVA.N > If NODE_ONLY is specified the controller continues to check the condition of > the cache batteries. > M > CAUTION: Setting UPS =NODE_ONLY or UPS=DATACENTER_WIDE without having a UPSiG > or similar backup system in place may result in data loss if power isn > interrupted. > K > - Specify UPS=DATACENTER_WIDE if the whole room or building (not just the H > subsystem) is supported by a UPS. If DATACENTER_WIDE is specified, theL > controller does not check the condition of the cache batteries and ignoresG > the battery's state. This causes RAIDsets and mirrorsets to always bet@ > available, regardless of the condition of the cache batteries. > M > - Specify NOUPS to instruct the controller to perform regular cache batterye; > checks and evaluate the condition of the cache batteries.e >   E 	Kerry, I think you are looking at newer controllers there.  HSOF CliC? 	Reference states on page 2-71 (note: HSJ50 manual, but a 50 ise 	a 40 so to speak):   
 	CACHE_UPS 	NOCACHE_UPS (Default)  @ 	Specifies whether the controller should perform regular battery@ 	condition checks, and whether it should ignore the CACHE_POLICY	 	setting.r  @ 	Specify CACHE_UPS if your storage subsystem receives power from@ 	an uninterruptable power supply.  The controller will not check= 	the condition of the cache batteries, and it will ignore theo? 	CACHE_POLICY setting, which means that RAIDsets and mirrorsetsy= 	will always be available, regardless of the condition of thev 	cache batteries.t  ; 	Specify NOCACHE_UPS to have the controller perform regularc> 	battery checks and follow the CACHE_POLICY setting for >> low 	batteries <<.   ---c  = 	So answering my own earlier post.  Setting UPS means it willa9 	ignore.  In your case, you got by on a low condition foreB 	a while but of course when status flipped to failed, you were/are 	sunk.   				RobM   ------------------------------   End of INFO-VAX 2001.406 ************************