0 INFO-VAX	Wed, 30 Jan 2002	Volume 2002 : Issue 57      Contents:/ Re: Advanced Server 7.3 issues when PDC reboots  Alpha VMS file to Oracle table% Re: Alphaserver 4100: physical memory 7 RE: Apache Virtual directory within a virtual directory @ Be prepared, Encompass is using SBA to SPAM folks for MembershipD Re: Be prepared, Encompass is using SBA to SPAM folks for MembershipD Re: Be prepared, Encompass is using SBA to SPAM folks for MembershipD Re: Be prepared, Encompass is using SBA to SPAM folks for MembershipD Re: Be prepared, Encompass is using SBA to SPAM folks for MembershipD Re: Be prepared, Encompass is using SBA to SPAM folks for Membership Re: Blade announcement Re: Blade announcement Re: Blade announcement Re: Blade announcement Re: Blade announcement3 Re: comp.sys.dec, FINAL NOTICE of Newsgroup Removal ) Compaq Analyze - security for web server? $ Re: Compaq editing webcasts again!!!$ Re: Compaq editing webcasts again!!!$ Re: Compaq editing webcasts again!!!5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq 5 Re: Comparitive reliability figures as used by Compaq " Re: Cross Compiler to SUN/Solaris?! DEC ALPHA ES40 Equipment For Sale P Re: DECW$SERVER_0 dies periodically; X11 colour wheel program blackouts the  scrP Re: DECW$SERVER_0 dies periodically; X11 colour wheel program blackouts the  scrP Re: DECW$SERVER_0 dies periodically; X11 colour wheel program blackouts the scre$ Re: DHCP as a client using Multinet?$ Re: DHCP as a client using Multinet?  Re: Disaster recovery - success!  Re: Disaster recovery - success!  Re: Disaster recovery - success!+ Re: Export Monitor Data to CSV file, Excel?  FMS/CFMS Re: fopen crashes while in AST Re: fopen crashes while in AST Re: fopen crashes while in AST/ fopen() behavior difference...bug or undefined? 3 Re: fopen() behavior difference...bug or undefined? / How to change vms system ip address and domain? 3 Re: How to change vms system ip address and domain? 3 Re: How to change vms system ip address and domain? ( Re: How to get/put the symbol list in C?( Re: How to get/put the symbol list in C?( Re: How to get/put the symbol list in C?( Re: How to get/put the symbol list in C?( Re: How to get/put the symbol list in C?= How to tell if File is locked in C ?fopen extension pehraps ? A Re: How to tell if File is locked in C ?fopen extension pehraps ? A Re: How to tell if File is locked in C ?fopen extension pehraps ? ! Re: I/O Performance on Alpha 4100 ! Re: I/O Performance on Alpha 4100 ! RE: I/O Performance on Alpha 4100  Re: Installing GNV( Re: Intel Special issue of OpenVMS Times( RE: Intel Special issue of OpenVMS Times( Re: Intel Special issue of OpenVMS Times Intrusions deletion 0 Re: Logicals defined for all process of one user Re: mount problems on cluster  Re: mount problems on cluster  Re: mount problems on cluster  Re: mount problems on cluster  Re: mount problems on cluster $ Moving from CMU-IP to TCPIP services( Re: Moving from CMU-IP to TCPIP services: Re: Partitioned System Disk (was: Re: There must be a way) pathwork Re: pathwork  Re: pathwork (French -> English)@ Re: Questions about setting up an Advanced Server  Member Server Re: The demise of compaq Re: The demise of compaq Re: The value of a leader  RE: The value of a leader  Re: The value of a leader  RE: The value of a leader ( Re: Using PortMapper to run an IP Server7 Using the 'latest and greatest' (was:VMS 7.3 stability) $ Using vfork/execvp with DCL commands( Re: Using vfork/execvp with DCL commands( Re: Using vfork/execvp with DCL commands( Re: Using vfork/execvp with DCL commands" Re: VMS 7.1-1H2 and TCPIP 5.1 ECO3 VMS 7.3 stability  Re: VMS 7.3 stability  Re: VMS 7.3 stability ? Re: VMS721_SYS-V1100: Caution do not apply with Pathworks V6.0D 6 Re: Where does VMS fit in here? (Was Re: HP admits...)6 Re: Where does VMS fit in here? (Was Re: HP admits...)6 Re: Where does VMS fit in here? (Was Re: HP admits...)6 Re: Where does VMS fit in here? (Was Re: HP admits...)6 Re: Where does VMS fit in here? (Was Re: HP admits...)6 Re: Where does VMS fit in here? (Was Re: HP admits...)6 Re: Where does VMS fit in here? (Was Re: HP admits...)E www.openvms.compaq.com and the link "OpenVMS Solutions Status Matrix"   F ----------------------------------------------------------------------    Date: 29 Jan 2002 14:00:35 -0800" From: ewilts@ewilts.org (Ed Wilts)8 Subject: Re: Advanced Server 7.3 issues when PDC reboots= Message-ID: <995e39b6.0201291400.114c9f5a@posting.google.com>   Y "Tim Jackson" <tim.jackson@amsjv.com> wrote in message news:<3c556933$1@pull.gecm.com>...  > Brad,  > I > Would these patches be part of the hopefully soon forthcoming ECO 1 for 	 > AS 7.3?   ? According to the CSC person I spoke to, yes, they should be.  I E received a copy of just PWRK$LMSRV.EXE today that corrects my problem ? and a few (unspecified) others with no release notes, but we're D getting there.  This weekend I'll be putting the new image in when IC can do the AS shutdown.  ECO 1 does not appear to be ready for even : early release yet - at least I was told I couldn't get it.  	    .../Ed    > B > "Brad McCusker" <Brad.McCuskerNOSP@MCompaq.com> wrote in message+ > news:RPj48.58$am1.650@news.cpqcorp.net...  > > Ed,  > > G > > We recently had an escalation for a problem with the failover code. 	 >  Please  > > contact D > > your CSC, and ahve them escalate this to engineering.  Have them >  reference > > the  > > "-89520" patches.  > >  > > Regards, > >  > > Brad > > 3 > > "Ed Wilts" <ewilts@ewilts.org> wrote in message : > > news:P9348.26444$yC.3433370@typhoon.mn.mediaone.net...J > > > Background:  we're an AS7.3 member server .  We had a case where our	 >  domain F > > > PDC (Win2K) took a nosedive and had to be rebooted.  Immediately >  followingJ > > > that, we started to log errors from only 1 of our cluster hosts (out >  of 3 J > > > that run AS) that it could not reach a domain controller, even after >  the >  PDCC > > > was rebooted.  The other 2 cluster hosts continued to run and  >  authenticate 
 > > > fine.    [snip]   ------------------------------    Date: 29 Jan 2002 20:11:59 -0800( From: rnagaraju@yahoo.com (Raj Nagaraju)' Subject: Alpha VMS file to Oracle table = Message-ID: <5f681ccc.0201292011.3fbb4521@posting.google.com>   @ We have files with data that were produced by Alpha VMS system.   : We have NO ACCESS to the system that produced these files.  F There are about 35 fields in each record. When I open the file throughA notepad, I can read some of the fields. Some of the fields are in B really funky, weird and unreadable format which I think is in some native or proprietary format.    How do I..either..  A 1. Convert the file so that ALL the fields are readable for me to % finally load them to an Oracle table?   > 2. Is there any shareware/freeware that would help me do this?   Thanks & Regards,  Raj    ------------------------------  % Date: Tue, 29 Jan 2002 12:34:54 -0800 C From: "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam> . Subject: Re: Alphaserver 4100: physical memory0 Message-ID: <3C57076D.CEEC34E6@intel.com.nospam>  & --------------A79B0D0CBD05ED6BAD622498* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit    Fabio Cardoso wrote:  	 > Kenneth  >  > In one machine I have  >  > System Configuration:  > ---------------------  > System Information: + > System Type    AlphaServer 4100 5/600 8MB  > Primary CPU ID 00 # > Cycle Time     1.6 nsec (598 MHz)  > Pagesize       8192 Byte >  > Memory Configuration: 1 > Cluster    PFN Start    PFN Count         Range  > (MByte)        Usage2 >  #00             0          256         0.0 MB - > 2.0 MB    Console 2 >  #01           256       524023         2.0 MB - > 4095.9 MB    System 2 >  #02        524279            9      4095.9 MB - > 4096.0 MB    Console >  > And in the other >  > System Configuration:  > ---------------------  > System Information: + > System Type    AlphaServer 4X00 5/600 8MB  > Primary CPU ID 00 # > Cycle Time     1.6 nsec (598 MHz)  > Pagesize       8192 Byte >  > Memory Configuration: 1 > Cluster    PFN Start    PFN Count         Range  > (MByte)        Usage2 >  #03             0          256         0.0 MB - > 2.0 MB    Console 2 >  #04           256       786163         2.0 MB - > 6143.8 MB    System 2 >  #05        786419           13      6143.8 MB - > 6144.0 MB    Console > " > How can I check the slots here ?  >     Thanks to David Webb and Robert Deininger for pointing out: first that there is more information on subsequent screens> and second that those additional screens are not sufficient to< tell what you really have in the system. :-(  In my defense,< and it seems I need a lot of that lately(!), the only 4100's: I have access to at the moment have 4GB of memory but CLUE> CONFIG on these systesm (they're at VMS 7.1-1H2 of all things)9 doesn't list out the separate memory options or CPUs like = David showed.  OTOH, as both David and Robert point out, only ; the console knows for sure...VMS doesn't know and shouldn't 5 care.  :-}  Sorry for the miss on the details here...   -     So again, Fabio, you need console output.            -Ken --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfield ! F20 Automation VMS System Support  kenneth.h.fairfield#intel.com     & --------------A79B0D0CBD05ED6BAD622498) Content-Type: text/html; charset=us-ascii  Content-Transfer-Encoding: 7bit   > <!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html>< <font face="Courier New,Courier">Fabio Cardoso wrote:</font>m <blockquote TYPE=CITE><font face="Courier New,Courier">Kenneth</font><font face="Courier New,Courier"></font> h <p><font face="Courier New,Courier">In one machine I have</font><font face="Courier New,Courier"></font>@ <p><font face="Courier New,Courier">System Configuration:</font>A <br><font face="Courier New,Courier">---------------------</font> ? <br><font face="Courier New,Courier">System Information:</font> N <br><font face="Courier New,Courier">System Type&nbsp;&nbsp;&nbsp; AlphaServer 4100 5/600 8MB</font> = <br><font face="Courier New,Courier">Primary CPU ID 00</font> G <br><font face="Courier New,Courier">Cycle Time&nbsp;&nbsp;&nbsp;&nbsp;  1.6 nsec (598 MHz)</font> Q <br><font face="Courier New,Courier">Pagesize&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 8 8192 Byte</font><font face="Courier New,Courier"></font>@ <p><font face="Courier New,Courier">Memory Configuration:</font>Z <br><font face="Courier New,Courier">Cluster&nbsp;&nbsp;&nbsp; PFN Start&nbsp;&nbsp;&nbsp;F PFN Count&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Range</font>V <br><font face="Courier New,Courier">(MByte)&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Usage</font>v <br><font face="Courier New,Courier">&nbsp;#00&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;k 0&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 256&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  0.0 MB -</font> L <br><font face="Courier New,Courier">2.0 MB&nbsp;&nbsp;&nbsp; Console</font>j <br><font face="Courier New,Courier">&nbsp;#01&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;^ 256&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 524023&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 2.0 MB -</font> N <br><font face="Courier New,Courier">4095.9 MB&nbsp;&nbsp;&nbsp; System</font>X <br><font face="Courier New,Courier">&nbsp;#02&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;H 524279&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;2 9&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 4095.9 MB -</font>w <br><font face="Courier New,Courier">4096.0 MB&nbsp;&nbsp;&nbsp; Console</font><font face="Courier New,Courier"></font> c <p><font face="Courier New,Courier">And in the other</font><font face="Courier New,Courier"></font> @ <p><font face="Courier New,Courier">System Configuration:</font>A <br><font face="Courier New,Courier">---------------------</font> ? <br><font face="Courier New,Courier">System Information:</font> N <br><font face="Courier New,Courier">System Type&nbsp;&nbsp;&nbsp; AlphaServer 4X00 5/600 8MB</font> = <br><font face="Courier New,Courier">Primary CPU ID 00</font> G <br><font face="Courier New,Courier">Cycle Time&nbsp;&nbsp;&nbsp;&nbsp;  1.6 nsec (598 MHz)</font> Q <br><font face="Courier New,Courier">Pagesize&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 8 8192 Byte</font><font face="Courier New,Courier"></font>@ <p><font face="Courier New,Courier">Memory Configuration:</font>Z <br><font face="Courier New,Courier">Cluster&nbsp;&nbsp;&nbsp; PFN Start&nbsp;&nbsp;&nbsp;F PFN Count&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Range</font>V <br><font face="Courier New,Courier">(MByte)&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Usage</font>v <br><font face="Courier New,Courier">&nbsp;#03&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;k 0&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 256&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  0.0 MB -</font> L <br><font face="Courier New,Courier">2.0 MB&nbsp;&nbsp;&nbsp; Console</font>j <br><font face="Courier New,Courier">&nbsp;#04&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;^ 256&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 786163&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 2.0 MB -</font> N <br><font face="Courier New,Courier">6143.8 MB&nbsp;&nbsp;&nbsp; System</font>X <br><font face="Courier New,Courier">&nbsp;#05&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;c 786419&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 13&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;  6143.8 MB -</font>w <br><font face="Courier New,Courier">6144.0 MB&nbsp;&nbsp;&nbsp; Console</font><font face="Courier New,Courier"></font> X <p><font face="Courier New,Courier">How can I check the slots here ?</font></blockquote>( <font face="Courier New,Courier"></font>J <p><br><font face="Courier New,Courier">&nbsp;&nbsp;&nbsp; Thanks to David1 Webb and Robert Deininger for pointing out</font> I <br><font face="Courier New,Courier">first that there is more information  on subsequent screens</font>M <br><font face="Courier New,Courier">and second that those additional screens  are not sufficient to</font>M <br><font face="Courier New,Courier">tell what you really have in the system.  :-(&nbsp; In my defense,</font> F <br><font face="Courier New,Courier">and it seems I need a lot of that! lately(!), the only 4100's</font> H <br><font face="Courier New,Courier">I have access to at the moment have 4GB of memory but CLUE</font> H <br><font face="Courier New,Courier">CONFIG on these systesm (they're at! VMS 7.1-1H2 of all things)</font> I <br><font face="Courier New,Courier">doesn't list out the separate memory  options or CPUs like</font> F <br><font face="Courier New,Courier">David showed.&nbsp; OTOH, as both' David and Robert point out, only</font> M <br><font face="Courier New,Courier">the console knows for sure...VMS doesn't  know and shouldn't</font> H <br><font face="Courier New,Courier">care.&nbsp; :-}&nbsp; Sorry for theJ miss on the details here...</font><font face="Courier New,Courier"></font>G <p><font face="Courier New,Courier">&nbsp;&nbsp;&nbsp; So again, Fabio, G you need console output.</font><font face="Courier New,Courier"></font> N <p><font face="Courier New,Courier">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -Ken</font> . <br><font face="Courier New,Courier">--</font>K <br><font face="Courier New,Courier">I don't speak for Intel, Intel doesn't > speak for me...</font><font face="Courier New,Courier"></font>8 <p><font face="Courier New,Courier">Ken Fairfield</font>M <br><font face="Courier New,Courier">F20 Automation VMS System Support</font> ! <br>kenneth.h.fairfield#intel.com  <br>&nbsp;</html>2  ( --------------A79B0D0CBD05ED6BAD622498--   ------------------------------  % Date: Tue, 29 Jan 2002 12:23:54 -0700l- From: "Rowell, Bradley" <browell@state.mt.us>r@ Subject: RE: Apache Virtual directory within a virtual directory@ Message-ID: <1245D1C0C039D411933708002BB29C644B2E27@DOAISD02003>  + Yes, there was a typo in the example below. 2 putting /brad/stuff before /brad cured the problem many thanks...   -----Original Message-----$ From: winston@SSRL.SLAC.STANFORD.EDU' [mailto:winston@SSRL.SLAC.STANFORD.EDU]B( Sent: Tuesday, January 29, 2002 11:54 AM To: Info-VAX@Mvb.Saic.ComB@ Subject: Re: Apache Virtual directory within a virtual directory    I In article <1245D1C0C039D411933708002BB29C644B2E22@DOAISD02003>, "Rowell,e& Bradley" <browell@state.mt.us> writes:H >How do I create an Apache virtual directory within a virtual directory: >ie. >wK >I want to access HTTP://server/brad and HTTP://server/brad/stuff where theb0 >stuff directory is located different than brad.J >Below is how I thought it might be done but doesn't work.  Any help would be
 >appreciated.e >b >Alias /brad /fc101/brad ><Directory /fc101/stuff>i >  Options Indexes >  Order allow,denyo >  AllowOverride All >  Allow from allC
 ></Directory>p >t >Alias /brad/stuff /fc101/stuffy ><Directory /fc101/stuff>e >  Options Indexes >  Order allow,deny  >  AllowOverride All >  Allow from ally
 ></Directory>o    H I'll assume that the fact that you don't have anything giving access to D /fc101/brad and have /fc101/stuff twice is just a typo in this post.  L Did you try reversing the order of the alias statements, putting the one forJ /brad/stuff before the one for /brad ?  I believe that getting a match on E /brad as the URL prefix will make Apache bail out before looking for iJ an alias for /brad/stuff, and then it'll, of course, try to find files in ' /fc101/brad/stuff, which doesn't exist.c   -- Alans    L ============================================================================ ===e0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDU?  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:a 650/926-3056A  Physical mail to: SSRL -- SLAC BIN 69, PO BOX 4349, STANFORD, CAt
 94309-0210L ============================================================================ ===T   ------------------------------  % Date: Tue, 29 Jan 2002 15:56:49 -0700f+ From: "Barry Treahy, Jr." <Treahy@mmaz.com> I Subject: Be prepared, Encompass is using SBA to SPAM folks for Membershipo' Message-ID: <3C5728B1.2020000@mmaz.com>   @ As the subject line indicates, be prepared to receive SPAM from H Encompass.  Encompass or not, I find these tactics unacceptable and how H Encompass could imagine that they would be above everyone else and that 4 no one would object to being SPAMMED is beyond me...  G I think what I find truly insulting about this, beyond the SPAM issue,  G is that I'm already a paying member and that the context of the e-mail lH which is attempting to attract new folks is Storage and IPF, two issues H that I have the least interest in hearing what Compaq has to say unless  it is VMS-centric.   Barry    -- p  @ Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIO   A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028e   ------------------------------  % Date: Tue, 29 Jan 2002 15:38:18 -0800eC From: "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam>eM Subject: Re: Be prepared, Encompass is using SBA to SPAM folks for MembershipP0 Message-ID: <3C57326A.D352ADED@intel.com.nospam>   "Barry Treahy, Jr." wrote:  A > As the subject line indicates, be prepared to receive SPAM from I > Encompass.  Encompass or not, I find these tactics unacceptable and howrI > Encompass could imagine that they would be above everyone else and that 6 > no one would object to being SPAMMED is beyond me...  @     Could you be just a bit more specific as to what constitutes> the SPAM from Encompass?  I've received three solicitations toB renew my membership (and probably will but haven't yet), but since> that is explicitly member-related, I don't consider it spam...    H > I think what I find truly insulting about this, beyond the SPAM issue,H > is that I'm already a paying member and that the context of the e-mailI > which is attempting to attract new folks is Storage and IPF, two issueshI > that I have the least interest in hearing what Compaq has to say unless  > it is VMS-centric.  F     I hate spam as much as the next person.  That said, could you postF an example of what you consider to be their spam?  Is it just the factC that you (and me!) aren't interested in things non-VMS, or are they:+ selling your name to a credit card company?e       -Ken --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfielda! F20 Automation VMS System Support- kenneth.h.fairfield#intel.com-   ------------------------------  % Date: Tue, 29 Jan 2002 17:08:31 -0700E+ From: "Barry Treahy, Jr." <Treahy@mmaz.com>sM Subject: Re: Be prepared, Encompass is using SBA to SPAM folks for Membershipc' Message-ID: <3C57397F.8020404@mmaz.com>c   Kenneth H. Fairfield wrote:6   >"Barry Treahy, Jr." wrote:d > A >>As the subject line indicates, be prepared to receive SPAM from I >>Encompass.  Encompass or not, I find these tactics unacceptable and how I >>Encompass could imagine that they would be above everyone else and thatm6 >>no one would object to being SPAMMED is beyond me... >>= >Could you be just a bit more specific as to what constitutest? >the SPAM from Encompass?  I've received three solicitations topC >renew my membership (and probably will but haven't yet), but since ? >that is explicitly member-related, I don't consider it spam...u >rH >>I think what I find truly insulting about this, beyond the SPAM issue,H >>is that I'm already a paying member and that the context of the e-mailI >>which is attempting to attract new folks is Storage and IPF, two issueseI >>that I have the least interest in hearing what Compaq has to say unless9 >>it is VMS-centric. >>C >I hate spam as much as the next person.  That said, could you postoG >an example of what you consider to be their spam?  Is it just the fact.D >that you (and me!) aren't interested in things non-VMS, or are they, >selling your name to a credit card company? >dD Hi Ken, sure, I've pasted it below.  A couple points to note is the G 'sales pitch' and that they take the time to define what Encompass is,  D which would only be necessary if they were blanketing an uninformed  audience.     ? As for whether Encompass is purchasing and selling my personal uF information, it wouldn't matter, I would still object to it.  In this G case, they are using a third party for the blitz, without any apparent  I knowledge of who they are contacting, probably just a filter of addreses 1 for anyone in the IT arena....   Barrym  # From: Encompass <Encompass@SBA.COM> ) To: "'treahy@mmaz.com'" <treahy@mmaz.com>OC Subject: Special Offer for IT Professionals - Join Encompass Today!5% Date: Tue, 29 Jan 2002 15:38:27 -0600  MIME-Version: 1.0a  
 Dear Barry Fellow IT professional:o  L Don't you often wish you could quickly learn the tricks of the trade for allE the new Compaq related technology?  Wouldn't you love a jump start onnD Compaq's next generation ENSA-2 storage technology or Intel's 64-bitJ Itanium(tm) Processor Family(IPF)? Join Encompass, A Compaq user group, toK get straightforward, unbiased information from peers who understand exactly- where you are coming from.  K You're not alone. Thousands of people just like you talk on a regular basis.F about our industry's successes and failures; about how to successfullyK deploy complex state-of-the-practice IT on all of Compaq's operating system L environments. We're called Encompass, a Compaq User Group, recognized by ourI members, Compaq and the industry as the first, best source of informationy8 about enterprise computing and we'd like you to join us!  L Encompass invites you to take advantage of a special Premier Membership rateL through January 31, 2002. Join or renew your Encompass membership by JanuaryK 31, 2002 and pay only $79 for your first year as a new Encompass IndividualyI Member(or $99 for your first year as an International Individual Member). I Why not purchase multiple memberships for your company and use them as an9I incentive for top performers? (Corporate memberships are also available.)0           --    @ Barry Treahy, Jr  *  Midwest Microwave  *  Vice President & CIO   A E-mail: Treahy@mmaz.com * Phone: 480/314-1320 * FAX: 480/661-7028    ------------------------------  # Date: Wed, 30 Jan 2002 00:58:09 GMT 4 From: "Terry C. Shannon" <terryshannon@mediaone.net>M Subject: Re: Be prepared, Encompass is using SBA to SPAM folks for Membership , Message-ID: <ByH58.7206$9U4.12847@rwcrnsc54>  F "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam> wrote in2 message news:3C57326A.D352ADED@intel.com.nospam... > "Barry Treahy, Jr." wrote: >4C > > As the subject line indicates, be prepared to receive SPAM fromsK > > Encompass.  Encompass or not, I find these tactics unacceptable and howeK > > Encompass could imagine that they would be above everyone else and thata8 > > no one would object to being SPAMMED is beyond me... >.B >     Could you be just a bit more specific as to what constitutes@ > the SPAM from Encompass?  I've received three solicitations toD > renew my membership (and probably will but haven't yet), but since@ > that is explicitly member-related, I don't consider it spam... >   I The email blast was sent by Smith Bucklin Associates, Encompass US Inc.'soK association management firm, to past and current members of DECUS/EncompasshJ US. The goal of the email is to boost membership. Nothing more and nothingK less. Neither Encompass US Inc. or SBA buy, sell, or trade mailing lists ora% other member demographic information.p   ------------------------------  % Date: Tue, 29 Jan 2002 21:34:55 -05002' From: Jim Becker <jbecker@ui.urban.org>-M Subject: Re: Be prepared, Encompass is using SBA to SPAM folks for Membershiph, Message-ID: <3C575BCF.C8C8B0E4@ui.urban.org>   "Barry Treahy, Jr." wrote: [snip]E > Hi Ken, sure, I've pasted it below.  A couple points to note is thenH > 'sales pitch' and that they take the time to define what Encompass is,E > which would only be necessary if they were blanketing an uninformedr > audience.   C It would be a mistake for a message to fail to identify itself to afF broad-based audience. Some recipients will know exactly what EncompassE is, and some might be no more than vaguely aware. Even if a recipient"@ is fully familiar with the organization, the message could quiteF conceivably get passed to another person who's not. There's absolutelyF no way the message can be highly customized for each recipient's levelC of familiarity with the organization, or degree of hostility to theeE mention of selected topics, or the level of familiarity of anyone then message might get forwarded to."  @ > As for whether Encompass is purchasing and selling my personalG > information, it wouldn't matter, I would still object to it.  In thisNH > case, they are using a third party for the blitz, without any apparentJ > knowledge of who they are contacting, probably just a filter of addreses > for anyone in the IT arena...;  B If you don't want to receive e-mail from Encompass, you can send a6 simple removal request to information@encompassus.org.  D SBA (Smith, Bucklin & Associates, Inc.) is an association management? firm hired by Encompass. SBA says it's the "largest provider of B management services to technology user groups in the world." Other9 user groups working with SBA include ITUG, SHARE, and thesE International Oracle Users Group. In any case, the SBA people workingpF on the Encompass contract constitute the Encompass staff, which is why0 SBA has been sending out the Encompass mailings.   --
 Jim Becker+ The Urban Institute (http://www.urban.org/)f' Encompass (http://www.encompassus.org/)n. ESILUG (http://encompasserve.org/lugs/esilug/)   ------------------------------  % Date: Tue, 29 Jan 2002 22:09:12 -0500&% From: JF Mezei <jfmezei@videotron.ca>nM Subject: Re: Be prepared, Encompass is using SBA to SPAM folks for Membershipo, Message-ID: <3C5763D6.ABA1DABB@videotron.ca>  L Out of curiosity, what is the point of a Compaq user group trying to recruitJ members when Compaq's may cease to exist in 2 months at which point the HP< usergroups will probably become the relevant organisations ?   ------------------------------  % Date: Tue, 29 Jan 2002 22:33:23 -0000/- From: wspencer@ap.nospam.org (Warren Spencer)s Subject: Re: Blade announcemente7 Message-ID: <91A5B5D38warrenspencer1977@209.249.90.100>,  ( jfmezei@videotron.ca (JF Mezei) wrote in" <3C567597.FEF09BB3@videotron.ca>:    >"Terry C. Shannon" wrote: >> sD >> QuickBlade + Marvel + Adaptive Infrastructure software + Galaxy = >> FIRE/ICE/WIND >&I >The way the announcement was worded, it was simply a way to stack wintel<F >8086 CPUs in a dense cabinet allowing 286 wintel servers to be in theG >same cabinet. This seems to me to have nothing to do with wilffire and; >even less galaxy. p >bH >Do these blade 8086 servers share memory or some form of interconnect ?  I Terry I think you've stumbled across the "next big thing".  It's cooling n fans.  f   ws   -- r   Warren Spencer' Senior Software Engineer (not a writer)s The Associated Press  @ ** When Windows is your hammer, everything looks like a thumb **   ------------------------------  # Date: Tue, 29 Jan 2002 22:35:37 GMTt4 From: "Terry C. Shannon" <terryshannon@mediaone.net> Subject: Re: Blade announcementf9 Message-ID: <ZsF58.580$dZ1.911@rwcrnsc52.ops.asp.att.net>n  : "Warren Spencer" <wspencer@ap.nospam.org> wrote in message1 news:91A5B5D38warrenspencer1977@209.249.90.100...n* > jfmezei@videotron.ca (JF Mezei) wrote in# > <3C567597.FEF09BB3@videotron.ca>:4 >  > >"Terry C. Shannon" wrote: > >>F > >> QuickBlade + Marvel + Adaptive Infrastructure software + Galaxy = > >> FIRE/ICE/WIND > >nK > >The way the announcement was worded, it was simply a way to stack wintelpH > >8086 CPUs in a dense cabinet allowing 286 wintel servers to be in theI > >same cabinet. This seems to me to have nothing to do with wilffire and/ > >even less galaxy. > >iJ > >Do these blade 8086 servers share memory or some form of interconnect ? >eJ > Terry I think you've stumbled across the "next big thing".  It's cooling > fans.n  C You may be right. Heck, you cram 280 CPUs into a rack, you gotta do  something about cooling!   ------------------------------  # Date: Tue, 29 Jan 2002 23:09:08 GMT<* From: "Bill Todd" <billtodd@metrocast.net> Subject: Re: Blade announcement A Message-ID: <oYF58.99302$vH6.5751904@bin6.nnrp.aus1.giganews.com>,  ? "Terry C. Shannon" <terryshannon@mediaone.net> wrote in message 3 news:ZsF58.580$dZ1.911@rwcrnsc52.ops.asp.att.net..." >d< > "Warren Spencer" <wspencer@ap.nospam.org> wrote in message3 > news:91A5B5D38warrenspencer1977@209.249.90.100...e, > > jfmezei@videotron.ca (JF Mezei) wrote in% > > <3C567597.FEF09BB3@videotron.ca>:, > >e > > >"Terry C. Shannon" wrote: > > >>H > > >> QuickBlade + Marvel + Adaptive Infrastructure software + Galaxy = > > >> FIRE/ICE/WIND > > >mF > > >The way the announcement was worded, it was simply a way to stack wintelJ > > >8086 CPUs in a dense cabinet allowing 286 wintel servers to be in theK > > >same cabinet. This seems to me to have nothing to do with wilffire andn > > >even less galaxy. > > >fL > > >Do these blade 8086 servers share memory or some form of interconnect ? > >aL > > Terry I think you've stumbled across the "next big thing".  It's cooling	 > > fans.u >rE > You may be right. Heck, you cram 280 CPUs into a rack, you gotta do< > something about cooling!  A That's one of the reasons they're P3s:  they generate *much* moren? performance per Watt (especially the mobile versions) than P4s.,  K Of course, you'd need a lot more then fans to cool the suckers if you triedl9 to replace those P3s with Itanics (think early Crays...).    - bill   ------------------------------  % Date: Tue, 29 Jan 2002 18:32:20 -0500f% From: JF Mezei <jfmezei@videotron.ca>m Subject: Re: Blade announcement , Message-ID: <3C5730C9.E5DA7AB7@videotron.ca>   "Terry C. Shannon" wrote:BE > You may be right. Heck, you cram 280 CPUs into a rack, you gotta do2 > something about cooling!  I You cram the cabinet into a refrigirator, and add the ice dispenser as aniL option to provide operators with cold crinks while they do the annual reboot$ sequence for each of the 280 CPUs...   ------------------------------  # Date: Tue, 29 Jan 2002 23:37:30 GMTt4 From: "Terry C. Shannon" <terryshannon@mediaone.net> Subject: Re: Blade announcementO, Message-ID: <_mG58.6768$9U4.11615@rwcrnsc54>  2 "JF Mezei" <jfmezei@videotron.ca> wrote in message& news:3C5730C9.E5DA7AB7@videotron.ca... > "Terry C. Shannon" wrote:rG > > You may be right. Heck, you cram 280 CPUs into a rack, you gotta doA > > something about cooling! >rK > You cram the cabinet into a refrigirator, and add the ice dispenser as aniG > option to provide operators with cold crinks while they do the annuald reboot& > sequence for each of the 280 CPUs...  K Ssshhhh... keep quiet about that! The ice crusher and refrigeration moduleseG are mid-tier Blades that won't be announced to the general public until/ 2CQ02 at the earliest!   ------------------------------  # Date: Tue, 29 Jan 2002 19:45:34 GMTe" From: "Hans Vlems" <hvlems@iae.nl>< Subject: Re: comp.sys.dec, FINAL NOTICE of Newsgroup Removal0 Message-ID: <yZC58.1960$A3.9393@typhoon.bart.nl>  3 Somebody was indeed fooling around on the Internet.y+ And yes I now know that .kr is not Kroatia.n   Hans  1 Vladimir Woelfl <gandalf@mmc.hr> wrote in message0? news:Pine.BSF.4.33.0201280106200.77397-100000@palach.okri.hr...n( > On Mon, 14 Jan 2002, Hans Vlems wrote: >t+ > > Since when does Kroatia run Usenet? ;-)r >?) > Kroatia is .hr, local name Hrvatska :-) J > concerning the question... i really don't know, but there is very littleI > activity on Croatian Usenet... I'm trying hard to change this, but whataH > could a man with DEC HiNote Ultra II laptop & Acer's i686 server do to > change this? >=I > Why is anyway comp.sys.dec be removed? Or is that "admin@ateziod.ac.kr"T' > just somebody trying too play a joke?  >e > >o > > Hans > > A > > Terry C. Shannon <terryshannon@mediaone.net> wrote in messageB; > > news:L4m08.26078$JF.202537@rwcrnsc52.ops.asp.att.net...= > > >=D > > > "Usenet Administration" <admin@ATEZTOD.AC.KR> wrote in message5 > > > news:fgghpo7q5rfa7jmlyycqq1ygyujehcb@4ax.com... : > > > > As many of you are aware, this newsgroup is slated8 > > > > for removal on January 31st, 2002.  The vote was< > > > > taken at the end of last year, and there was a large6 > > > > marjority who wants to see this group removed. > > > >nI > > > > If you are a regular reader, you need to make other arrangements.  > > > H > > > You're welcome to visit www.openvms.org and www.tru64.org for your Compaq, > > > enterprise OS and product discussions. > > >u > > >b > >h > >e > >t >S   ------------------------------   Date: 29 JAN 2002 21:48:45 GMT4 From: karcher@thuria.waisman.wisc.edu (Carl Karcher)2 Subject: Compaq Analyze - security for web server?6 Message-ID: <29JAN02.21484559@thuria.waisman.wisc.edu>  H I've listened to the numerous complaints about CA in this forum and haveG dreaded the day when new hardware will require it. Today I've installed0F it (WEBES V4.0 with SP3) on a recently acquired ES40 (Decevent doesn't) support the ES40) with low expectations. K  D While the installation was successful and everything appears to workH (although at a snail's pace due to the dependance on Java I assume), I'mF wondering how one controls access to the web server. There's nothing IE can find in the documentation about limiting access to the web server ; other than disabling it. There's a password.txt file in thep@ [desta.config] directory that looks hopeful but adding a line of? "username password" appears to have no effect. I've deleted theiF configuration.dat file and restarted DESTA as well (as this appears toF be how some configuration data is changed). While I doubt that the webF server will be of any actual value, was the thought of security really left out of this product?e  G -- Carl Karcher, Waisman Computing Services, Waisman Center, UW-Madisonn2 --                      karcher@waisman.wisc.edu     ------------------------------  % Date: Tue, 29 Jan 2002 15:12:57 -0500m( From: David Froble <davef@tsoft-inc.com>- Subject: Re: Compaq editing webcasts again!!!e, Message-ID: <3C570249.7030305@tsoft-inc.com>   Alan Greig wrote:   B > Ok, this is getting ridiculous. Last year I reported that CompaqE > appeared to have cut some segments from the replayed version of thea. > webcast.. Looks like they've done it again.     H Well, since they appear to really like shooting themselves in the foot, D maybe they're just cleaning up some loose ends around the hugh hole  already there.  :-)o    E > I watched some of the webcast this year live then replayed the rest-A > later (including the Q+A) expecting to be able to further studyBG > interesting bits later. In the Q+A section one of the first questions F > from the floor was on the subject of Alpha and customer satisfactionD > or otherwise. After giving the standard answer that customers were> > deliriously happy Elias(?) added that there was one proviso:D > customers, he said were pressing them on completion of the ItaniumG > port0 and Compaq understood there was a lot of customer concern aboute
 > the matter.d > D > Today, I returned to the webcast to replay the Q+A section in moreC > detail prior to commenting here. Lo and Behold it isn't there anyrF > more. Yes the entire Q+A section has vanished. I want to stress thatF > it *was* available for replay until yesterday. The bit at the end ofF > Capellas wrap up speech where he introduces the Q+A section has goneF > as well. Due to an editing mixup though the wrap up and introduction> > to the Q+A is tacked on to the end of the Financial section. > H > So what are Compaq playing at? Why, two years in a row, do they appear1 > to have edited the webcast? All very Orwellian.s  I It's the PC way of doing things.  No practice or planning, just freehand .7 it, and then go back and correct all the mistakes.  :-)s   Compaq IS a PC company.i   Dave   -- .4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com6 T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   ------------------------------  # Date: Tue, 29 Jan 2002 21:53:23 GMTe4 From: "Terry C. Shannon" <terryshannon@mediaone.net>- Subject: Re: Compaq editing webcasts again!!!m9 Message-ID: <nRE58.207$dZ1.467@rwcrnsc52.ops.asp.att.net>l  5 "David Froble" <davef@tsoft-inc.com> wrote in messager& news:3C570249.7030305@tsoft-inc.com... > Alan Greig wrote:e >rD > > Ok, this is getting ridiculous. Last year I reported that CompaqG > > appeared to have cut some segments from the replayed version of thea/ > > webcast.. Looks like they've done it again.l >f >fI > Well, since they appear to really like shooting themselves in the foot,2E > maybe they're just cleaning up some loose ends around the hugh hole  > already there.  :-)u  D Hmmm... I wonder if Compaq has added Rose Mary Woods to the Board of
 Directors ;-}?   ------------------------------  % Date: Tue, 29 Jan 2002 22:59:26 +0100g' From: Brass Christof <welcome@spam.net>m- Subject: Re: Compaq editing webcasts again!!!a( Message-ID: <3C571B3E.EC98242C@spam.net>   David Froble wrote:p  = > It's the PC way of doing things.  No practice or planning, eG > just freehand it, and then go back and correct all the mistakes.  :-)b >  > Compaq IS a PC company.d >  > Dave  < PC companies buying serious IT companies is like AOL buying + Time Warner or as the tail walking the dog.o   > --6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com8 > T-Soft, Inc.  170 Grimplin Road  Vanderbilt, PA  15486   -- a6 moc dot slupofni at ssarb - please revert the sequence   ------------------------------  % Date: Tue, 29 Jan 2002 20:41:52 +0000l% From: Alan Greig <a.greig@virgin.net>m> Subject: Re: Comparitive reliability figures as used by Compaq* Message-ID: <3C57090F.8DEB70C8@virgin.net>   JF Mezei wrote:h   > Alan Greig wrote:d > >t8 > > In Elias's presentation the following slide appears: > >s5 > > System                          Hrs lost per year ( > > P/Sysplex, Himalaya             2..2/ > > Single Mainframe                        8.73( > > Proprietary (ie OVMS)           17.5. > > Proprietary Unix                        350 > > NT and MS Cluster Server                69.9 >mG > Did you add the "OVMS" in there or was it just "proprietary" with youl& > speculating it was refering to VMS ? >$  H I am quoting exactly what it said on the slide. I did not add the "OVMS"     >mN > I find it odd that it would put the Sysplex in the same line item as Tandem. >. > I would have:  > Tandem	 > Sysplexn > VMS Clusters > Single Mainframe > Unix clustersp > Proprietary single nodei > Unix single node	 > Windowsn >oM > Since Ellias is supposed to be championing Tandem, VMS and Tru64, I find itaM > odd that he would not have added comments about VMS clusters in that table. N > Then again, I am not surprised since he is simply following corporate policy+ > to ignore VMS in financial presentations.  >mP > When I was close to the Tandem technical folks at a bank in the early 1990s, II > can tell you that they had greater than 2.2 hours of downtime per year. N > Upgrades of software (mandated on specific dates) required the systems to beM > unavailable during a few hours at a time. So perhaps they are talking abouto > unscheduled downtime.o >aP > Another point to ponder. In a properly configured VMS cluster, it can be quiteN > easy to shutdown a node without stopping service. having more flexibility onO > how one manages a VMS systems with regards to upgrades etc, the would be moregM > "downtime"  for each node in a VMS cluster simply because it is possible toiL > bring down a node without affecting service. (Some Tandem configs do allow > this for the applications)   --
 Alan Greig   ------------------------------  % Date: Tue, 29 Jan 2002 16:18:25 -0500m% From: JF Mezei <jfmezei@videotron.ca> > Subject: Re: Comparitive reliability figures as used by Compaq, Message-ID: <3C57116F.F2123081@videotron.ca>   Alan Greig wrote:t* > > > Proprietary (ie OVMS)           17.5  J > I am quoting exactly what it said on the slide. I did not add the "OVMS"    L Ouch. This would mean that it was Compaq that added "ie OVMS" to the GartnerG numbers and that it was Compaq who chose where to put VMS in that list.t    H At least, he was able to list Windows's numbers that still show it as an unreliable number.2 > > > NT and MS Cluster Server                69.9  0 That is almost 3 full days of downtime per year.  M What would be also interesting to know is the average downtime length and thee number of occurances.a  K I once heard of a large Alpha cluster that bugchecked and it took somethingoM like an hour to write its dump file. Can anyone confirm that a crashing alphahB with oodles of memory could take that long to write the dumpfile ?   ------------------------------  % Date: Tue, 29 Jan 2002 22:15:38 +0100w1 From: John McLean <mcleanj@swissonline.delete.ch>a> Subject: Re: Comparitive reliability figures as used by Compaq5 Message-ID: <3C5710FA.9658E9F2@swissonline.delete.ch>a   Alan Greig wrote:b > =o   > JF Mezei wrote:t > =a   > > Alan Greig wrote:e > > >r: > > > In Elias's presentation the following slide appears: > > >o7 > > > System                          Hrs lost per yeara* > > > P/Sysplex, Himalaya             2..21 > > > Single Mainframe                        8.7e* > > > Proprietary (ie OVMS)           17.50 > > > Proprietary Unix                        352 > > > NT and MS Cluster Server                69.9 > >nJ > > Did you add the "OVMS" in there or was it just "proprietary" with you=  ( > > speculating it was refering to VMS ? > >a > =u  J > I am quoting exactly what it said on the slide. I did not add the "OVMS= "R    B To me it looks like Elias (Senior Vice President and Group GeneralF Manager, Business Critical Server Group) knows which side his bread is2 buttered and is trying to work his way to the top.  @ Take this extracted quote for example, "We offer a full range ofH innovative products, all the way from our best-selling iPAQ=99 handheldsF to the pinnacle of computing scalability and availability, the NonStopF Himalaya server. By integrating these products with partner offerings,C we provide market-driven solutions for customers around the world."   / Doesn't this sound just like a Capellas clone ?.  H And the percentages of uptime were shown as 99.99% (Sysplex & Himalaya),F 99.9% (Single mainframe), 99.8% (VMS), 99.6% (proprietary unix), 99.2%" (NT and Microsoft Cluster server).  F (I assume that proprietary unix is Tru64, that OS now making its final appearance ...)n  F Yep, sourced from ye olde accurate prophets, Gartner (which gives it a7 probability of accurateness at about a zillion to one).h  - Question 1 - What on earth is Elias smoking ?o@ Question 2 - Has this redefined Business Critical so that it now
 includes NT ?tH Question 3 - Won't selling more NT systems (at lower margins) reduce the income ?2 Question 4 - when will Compaq's stupidity stop ? =       John McLeanv   ------------------------------  % Date: Tue, 29 Jan 2002 16:25:53 -0500u% From: JF Mezei <jfmezei@videotron.ca>r> Subject: Re: Comparitive reliability figures as used by Compaq, Message-ID: <3C57132F.7F172037@videotron.ca>   John McLean wrote:2 > Question 4 - when will Compaq's stupidity stop ?  K When Compaq's role model (Dell) goes belly up. As long as Dell is succesfuliK with its wintel business, Compaq sees no reason why it shouldn't regain the": position it used to have and that is now occupied by Dell.   ------------------------------    Date: 29 Jan 2002 15:32:40 -0600+ From: young_r@encompasserve.org (Rob Young)s> Subject: Re: Comparitive reliability figures as used by Compaq3 Message-ID: <cm97X5uA90K$@eisner.encompasserve.org>i  T In article <3C57116F.F2123081@videotron.ca>, JF Mezei <jfmezei@videotron.ca> writes: > Alan Greig wrote: + >> > > Proprietary (ie OVMS)           17.5e > K >> I am quoting exactly what it said on the slide. I did not add the "OVMS"s >  > N > Ouch. This would mean that it was Compaq that added "ie OVMS" to the GartnerI > numbers and that it was Compaq who chose where to put VMS in that list., >  > J > At least, he was able to list Windows's numbers that still show it as an > unreliable number.3 >> > > NT and MS Cluster Server                69.9i > 2 > That is almost 3 full days of downtime per year. > O > What would be also interesting to know is the average downtime length and the  > number of occurances.r > M > I once heard of a large Alpha cluster that bugchecked and it took something.O > like an hour to write its dump file. Can anyone confirm that a crashing alphanD > with oodles of memory could take that long to write the dumpfile ?    A 	What was the DUMPSTYLE setting , 9 or something that flips thoseCA 	bits?  In other words, most of us are doing selective compressedm 	dumps.   A 	A quick calculation shows a 256 GByte memory would require < 1/3e> 	(if compressed selective) or 76 GByte dump file... assuming aF 	reasonable 40 MByte/sec write rate (should be able to do much better F 	than that , i.e. that dump file would/should be on a nice RAID10 withF 	writeback cache, expect to saturate the FC connection at 75-80 MByte/@ 	sec.) you should be able to dump in 32 minutes at 40 MByte/sec.   				Robn   ------------------------------  % Date: Tue, 29 Jan 2002 22:41:21 +0100e1 From: John McLean <mcleanj@swissonline.delete.ch>f> Subject: Re: Comparitive reliability figures as used by Compaq5 Message-ID: <3C571701.375F20CE@swissonline.delete.ch>s   Alan Greig wrote:  > 6 > In Elias's presentation the following slide appears: > 3 > System                          Hrs lost per year2& > P/Sysplex, Himalaya             2..2- > Single Mainframe                        8.7a& > Proprietary (ie OVMS)           17.5, > Proprietary Unix                        35. > NT and MS Cluster Server                69.9
 > (c) Gartneru > D > Now in 15 years of managing VMS systems and over 20 years of usingH > them I can't think of a single system which had a downtime of 17.5 hrsC > in one year including hardware and software fails - except for antC > early Pathworks V5 server which I won't count. If a cluster these G > figures are even further out. Where do they come from. Well the slide ! > shown by Elias credits Gartner.e > @ > In the two years since the Alphacluster at my current employerD > replaced the VAX systems we have had *zero* cluster downtime otherE > than in total power shutdown (planned). There has been one softwareUE > crash of one node and no hardware crashes. The node rebooting afterrF > the software crash (LOCKMGRERR) was back online within 5 minutes andH > problem was resolved via DSN and fix implemented. Cluster availability > was maintained throughout. > F > At my previous employer our Admin cluster was up for five years with2 > no downtime and that included a rolling upgrade. > F > Does Elias believe these figures? If so no wonder he thinks Himalaya > is far better than VMS.a > H > Slide 9 at ftp://ftp.compaq.com/pub/corporate/ir/Analyst2002_Elias.ppt# > if you want to look for yourself.N > -- > Alan    E There are three very disturbing things about this, and plenty of only ! slightly-less disturbing aspects.n  E For my money the first big issue is that Gartner have been predictingnF the demise of VMS for many years and this was a perfect opportunity toG hasten that demise.  Hmmm, no wonder Compaq were very quiet to refute a F Gartner statement very recently, Compaq were about to quote them as anA authoritive source.  I'm very glad that The Inquirer exposed thatw/ refutation.  Playing both ends were we Compaq ?n  D Next problem is slide No. 4 of this series.  It shows a pie chart ofE revenue with "Access" taking 45%, Enterprise with 32% and Services at F 23%.  What a pity it wasn't of INCOME.  It would have been interestingB to see how they showed the NEGATIVE INCOME (ie. LOSS) by Access ofA -92%,  because Enterprise returned 25% of total income and GlobaleE services 166.5% of income.  Or to put it more sensibly, if Access had E disappeared off the planet in a puff of smoke, Compaq would have madeaC $1.22 billion from the two remaining segments, rather than the $638sA million they made from 3 segments.  Access didn't quite halve thea) possible profit but they sure went close.   F Final point is the total absence of anything positive about VMS any ofH the slides.  In fact it only appeared in the background when it was justE too difficult to omit it.  VMS appears on slide 9 (that Alan referredeC to) and on slides 12 & 13 (unlucky for some !).  Slide 12 shows the D strategic path (aka the path they lead customers along) and 13 talks4 about Shitanium and has VMS as a label on the slide.  B Today Terry S mentioned Fire, Ice and Wind.  Put the Fire and Wind@ together and we have Hot Air, of which there is no shortage fromH Compaq.  And on Ice is where it looks like Compaq would like to put VMS.     John McLeana   ------------------------------  % Date: Tue, 29 Jan 2002 16:46:47 -0500l% From: JF Mezei <jfmezei@videotron.ca>(> Subject: Re: Comparitive reliability figures as used by Compaq, Message-ID: <3C571814.1AC39744@videotron.ca>   Rob Young wrote:O >         than that , i.e. that dump file would/should be on a nice RAID10 with O >         writeback cache, expect to saturate the FC connection at 75-80 MByte/oI >         sec.) you should be able to dump in 32 minutes at 40 MByte/sec.i  I Thanks for the reality check. I guess it would be a hour of down time (35xL minutes to crash, and then about 20 minutes to boot and start all apps. Now,K in that incident I had heard about, the guy did state that all nodes in therH cluster crashed (he stated "yes, the WHOLE cluster went down".) So is itN possible then that the disk drives would take much longer if you have multiple( hosts writing to the various dumpfiles ?  J In the case of a mission critical cluster where downtime costs real money,K isn't it a mistake to request to have a dumpfile be written and stop you an. extra 30 minutes ?   ------------------------------  # Date: Tue, 29 Jan 2002 22:06:26 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>> Subject: Re: Comparitive reliability figures as used by Compaq0 Message-ID: <3C571BF2.5AF18891@blueyonder.co.uk>   John McLean wrote:  yB > Question 2 - Has this redefined Business Critical so that it now > includes NT ?r  Q TCO is higher for NT solutions. Compaq provide systems as part of site managementp all in deal. Seems to me  J > Question 3 - Won't selling more NT systems (at lower margins) reduce the
 > income ?  Q TCO is higher for NT solutions. Compaq provide systems as part of site management L all in deal. Seems to me they might make more on services than VMS this way.; At my last place of work if the desktop needed reinstalling R due to random windoze weirdness this was chargeable as a cost against the businessO unit, not the IT support group. Maybe Compaq have worked out a way to introduce  a support charge per reboot?   Just thinking aloud. v  :2 > Question 4 - when will Compaq's stupidity stop ?  G When the chicoms walk all over the US once windows is fully bedded intoeB all the strategic US defence systems (sorry Terry, thats your line
 isn't it)?   -- Q Tim.Llewellyn@cableinet.co.uk  r  C Standard disclaimer applies. My views in no way represent those of  ! my employers or service provider.w   ------------------------------  % Date: Tue, 29 Jan 2002 23:27:57 +0100t1 From: John McLean <mcleanj@swissonline.delete.ch>i> Subject: Re: Comparitive reliability figures as used by Compaq5 Message-ID: <3C5721ED.A5A85A07@swissonline.delete.ch>A   JF Mezei wrote:e >  > John McLean wrote:4 > > Question 4 - when will Compaq's stupidity stop ? > M > When Compaq's role model (Dell) goes belly up. As long as Dell is succesful M > with its wintel business, Compaq sees no reason why it shouldn't regain the < > position it used to have and that is now occupied by Dell.    F In the meanwhile, Compaq continues to look like the knight holding the0 bridge in "Monty Python and the Holy Grail" ....     John   ------------------------------  % Date: Tue, 29 Jan 2002 23:05:41 -0000o= From: "David McKenzie" <david.mckenzie@spitfire0.demon.co.uk>-> Subject: Re: Comparitive reliability figures as used by CompaqB Message-ID: <1012345571.13563.0.nnrp-12.c1edba74@news.demon.co.uk>  L > In the case of a mission critical cluster where downtime costs real money,J > isn't it a mistake to request to have a dumpfile be written and stop you an > extra 30 minutes ?  L Well, I guess the idea is that not all nodes should go down at once. Also inK a well designed DR cluster one should be able to meet the load requirementsy with one or more nodes down.  2 "JF Mezei" <jfmezei@videotron.ca> wrote in message& news:3C571814.1AC39744@videotron.ca... > Rob Young wrote:L > >         than that , i.e. that dump file would/should be on a nice RAID10 withJ > >         writeback cache, expect to saturate the FC connection at 75-80 MByte/K > >         sec.) you should be able to dump in 32 minutes at 40 MByte/sec.a >0K > Thanks for the reality check. I guess it would be a hour of down time (35mI > minutes to crash, and then about 20 minutes to boot and start all apps.  Now,I > in that incident I had heard about, the guy did state that all nodes inn the J > cluster crashed (he stated "yes, the WHOLE cluster went down".) So is itG > possible then that the disk drives would take much longer if you haveT multiple* > hosts writing to the various dumpfiles ? >l   ------------------------------   Date: 29 Jan 2002 23:50:44 GMT) From: leslie@clio.rice.edu (Jerry Leslie)I> Subject: Re: Comparitive reliability figures as used by Compaq' Message-ID: <a37cgk$d77$1@joe.rice.edu>n  2 John McLean (mcleanj@swissonline.delete.ch) wrote:2 : Question 4 - when will Compaq's stupidity stop ?  / "Have we arrived at the bottom of stupid yet ?"2  8 "Every time I think we've reached the bottom of stupid,   someone throws down a shovel."r  4 --Jerry Leslie     (my opinions are strictly my own)   ------------------------------  # Date: Wed, 30 Jan 2002 00:17:37 GMTt( From: "Ken Farmer" <kfarmer@openvms.org>> Subject: Re: Comparitive reliability figures as used by Compaq> Message-ID: <BYG58.23566$A51.9207058@typhoon.southeast.rr.com>  L Our W2K servers each have been down 5 days minimum installing patches.  That1 doesn't include them crashing on a regular basis.5   KenT   -- Kenneth Farmer http://www.OpenVMS.org http://www.Tru64.org      2 "JF Mezei" <jfmezei@videotron.ca> wrote in message& news:3C57116F.F2123081@videotron.ca... > Alan Greig wrote:y, > > > > Proprietary (ie OVMS)           17.5 >ML > > I am quoting exactly what it said on the slide. I did not add the "OVMS" >s >nF > Ouch. This would mean that it was Compaq that added "ie OVMS" to the GartneraI > numbers and that it was Compaq who chose where to put VMS in that list.d >n >oJ > At least, he was able to list Windows's numbers that still show it as an > unreliable number.4 > > > > NT and MS Cluster Server                69.9 >b2 > That is almost 3 full days of downtime per year. >lK > What would be also interesting to know is the average downtime length andm theh > number of occurances.d >tC > I once heard of a large Alpha cluster that bugchecked and it tookh	 somethingtI > like an hour to write its dump file. Can anyone confirm that a crashingo alphafD > with oodles of memory could take that long to write the dumpfile ?   ------------------------------    Date: 29 Jan 2002 21:42:01 -0600+ From: young_r@encompasserve.org (Rob Young)r> Subject: Re: Comparitive reliability figures as used by Compaq3 Message-ID: <efdMF7nfS8X3@eisner.encompasserve.org>t  T In article <3C571814.1AC39744@videotron.ca>, JF Mezei <jfmezei@videotron.ca> writes: > Rob Young wrote:P >>         than that , i.e. that dump file would/should be on a nice RAID10 withP >>         writeback cache, expect to saturate the FC connection at 75-80 MByte/J >>         sec.) you should be able to dump in 32 minutes at 40 MByte/sec. > K > Thanks for the reality check. I guess it would be a hour of down time (35iN > minutes to crash, and then about 20 minutes to boot and start all apps. Now,M > in that incident I had heard about, the guy did state that all nodes in the J > cluster crashed (he stated "yes, the WHOLE cluster went down".) So is itP > possible then that the disk drives would take much longer if you have multiple* > hosts writing to the various dumpfiles ? >   A 	Sure... especially if all the dumpfiles were on one volume.  OneuF 	wouldn't necessarily expect to crash the whole cluster and I wouldn'tA 	be too upset if someone designed without that in mind.  The onlyn< 	time I saw the "whole" cluster crash was with "InconsistentH 	I/O database" and that has been over 4 years now at another site... and5 	the issue was eventually tracked down and corrected.o  D 	But to throw this wrinkle... was it really 2+ nodes with 256 Gbytes 	of RAM?  L > In the case of a mission critical cluster where downtime costs real money,M > isn't it a mistake to request to have a dumpfile be written and stop you ano > extra 30 minutes ?  > 	Absolutely... so the Galaxy designers are ahead of the curve!   	There is "big" iron coming:  s http://groups.google.com/groups?q=kleinsorge+terabyte&hl=en&selm=3674FCA8.E663596%40star.enet.dec_nospam.com&rnum=3o  H I picked 8 as the top end of the sweet spot for SMP, because essentially@ most SMP systems under a wide variety of workloads will scale upG reasonably well to 8.  A well behaved application may scale up into the E 20s, and you need something with real MPP characteristics to go a lotrG higher - unless you do a wholesale, hive-like, roto-tilling of the OS.  F So with systems having 32-256 CPUs, a terabyte of memory, and hundredsE of IO busses comming along, how do we best handle them?  You can make?D the argument that it is better to simply build 2-8 CPU SMP systems -G which is the PC model of the world, instead of these huge systems.  ButiF they suffer from having the communications interface as the bottleneck5 when each system is working on a common task or data.h  ( 	How do you work with a Terabyte of RAM?  http://groups.google.com/groups?hl=en&as_drrb=b&as_mind=29&as_minm=3&as_miny=1995&as_maxd=10&as_maxm=12&as_maxy=1998&frame=right&rnum=71&thl=1562474696,1562474697,1562442327,1562055457,1562696075,1562818913,1562053085,0&seekm=u4sr5dsnx.fsf%40garlic.com#link75e  M But future hardware designs can mitigate many of the problems by implementingsF programmable "firewalls" to seperate "globally" accessable memory fromM "Instance Private" memory.  And other mechanisms can be implemented to assist,O in fault isolation.  Nonetheless, single points of failure, although rare, will K continue to exist - for a while - such as an uncorrectable error on addressn1 data.  Eventually, the HW will catch up to us ;-)r    A 	What I suspect is the answer to your "large dumpfile problem" ist> 	to separate the data out from the code.  The data resides in @ 	Galactic shared memory, instance private memory is much smallerA 	of course.  Box crashes, you dump private memory.  Terabyte box,r> 	75% ? in Galactic common, 25% divided among instances... 8 ofF 	those instances each have to dump 32 Gbyte when they crash (actually,@ 	selective compressed dump.... 9 GBytes :-).  I'm assuming a bitD 	as the Galactic goal as noted above is to prevent a single instance> 	crash from bringing down the Galaxy.  The comp.arch sharpies ? 	noted in 1998 that a "wild write" could bring down the Galaxy. D 	Future hardware will help prevent that as the discussion went then. 	(Fred mentioned then).u  ? 	The flip-side or interesting side of this debate is how do thenA 	Tru64 (AIX, HP/UX, Sun, SGI, etc.) folks work with a Terabyte of-D 	memory?  We know Tru64 scales much better than VMS SMP-wise.  TheirF 	model is SSI... so when that SSI Terabyte box crashes, that must make* 	for one nasty dump and one nasty startup!  > 	One hidden Galaxy advantage appears to be de-coupling data so? 	those large beasts don't have to come down and if they do comeg5 	crashing down , they come down a tiny bit at a time.    				Rob    ------------------------------  # Date: Tue, 29 Jan 2002 22:09:17 GMT 4 From: Tim Llewellyn <tim.llewellyn@blueyonder.co.uk>+ Subject: Re: Cross Compiler to SUN/Solaris?h0 Message-ID: <3C571C9C.4FA10811@blueyonder.co.uk>   JF Mezei wrote:  >  > Larry Kilgallen wrote:V > > > So, if the Customer decided to move from VMS to SUN, or to Winsh*t, my job is to > > > help him to move.6 > > >  > > > Sad, but reality.  > >o5 > > Not sad for you.  I predict many billable hours !i > M > Not necessarily. I was involved in a similar project (move from VMS to NT). N > The company had hired some "NT specialists" and my involvement was simply toM > explain the VMS specific stuff they didn't understand. They rewrote the app " > from scratch. Took them 2 years.  L yup, unfortunately this is the way I have observed things to happen. The VMSM people are kept hanging around until the replacement system is needed but noteE involved in the replacement project except when absolutely necessary.    regardsr -- / Tim.Llewellyn@cableinet.co.uk     C Standard disclaimer applies. My views in no way represent those of s! my employers or service provider.t   ------------------------------    Date: 29 Jan 2002 22:36:03 -0800$ From: jmon@retechwest.com (Joseph M)* Subject: DEC ALPHA ES40 Equipment For Sale= Message-ID: <c8732ca9.0201292236.670d3a37@posting.google.com>o  D Retechwest, http://www.retechwest.com has the following Compaq / DEC parts available for sale.   ; All parts are for the ES40 series Alpha Server and are NEW.i   Guaranteed not DOA.   A Item 1001: Server Parts for Model ES40 Alpha 6/500 Model 1 0MB : n Server Part Number DH-62AAA-AA p, Processor Card Module w/ Processor - 12 each, Memory Board Module (0) MB Memory - 48 each ) Motherboard - 12 each I/O Card - 12 each . AC Power Supply - 12 each  System Fan Assembly - 12 eachi    G Item 1002: Server Parts for Model ES40 6/500 Model 2 (0)MB OVMS: Server  Part Number DH-64AAA-AA + Processor Card Module w/ Processor - 5 eacho, Memory Board Module (0) MB Memory - 20 each  Motherboard - 5 each h I/O Card - 5 each  AC Power Supply - 5 each   System Fan Assembly - 5 each    E Item 1003: Server Parts for Model ES40 6/667 Model 1 0MB: Server Parto Number DH-62BAA-AA  - Processor Card Module w/ Processor - 11 each  , Memory Board Module (0) MB Memory - 44 each  Motherboard - 11 each - I/O Card - 11 each AC Power Supply - 11 each s System Fan Assembly - 11 eachu    E Item 1004: Server Parts for Model ES40 6/500 Model 2 0MB: Server Parti Number DH-64AAA-AA o- Processor Card Module w/ Processor - 10 each e+ Memory Board Module (0) MB Memory - 40 eachr Motherboard - 10 each  I/O Card - 10 each o AC Power Supply - 10 each  System Fan Assembly - 10 eacha    5 Entire part inventory listed above, price, $60,000.00s  @ Please contact reply to RETECHWEST, email info@retechwest.com if interested.   
 Thank You.
 RETECHWEST   ------------------------------  % Date: Tue, 29 Jan 2002 17:36:45 -0500o! From: Sam Hoblit <hoblit@bnl.gov>sY Subject: Re: DECW$SERVER_0 dies periodically; X11 colour wheel program blackouts the  scra$ Message-ID: <3C5723FD.90209@bnl.gov>   Brass Christof wrote:l  
 > Hi folks > A > After investigating my problems, searching in FAQ, comp.os.vms i6 > and ask the wizard archives I have to resort to you! > A > I expect the 2.5 problems to be related, i.e. to have the same t@ > source which could be a configuration mistake on my side or a 1 > problem with the DECwindowx X11 display server.m > ' > The two and a half problems in short:t > B > 1.The display server dies periodically every three or four days.> > Normally it dies while I do something but it died also once < > when I wasn't there. VMS continues running also I have to ; > reboot because restarting the window server doesn't work.i > < > 2.The example program from section 5.4.2 (Allocating Color	 > Cells) tC > http://www.openvms.compaq.com:8000/73final/5642/5642pro_006.html e? > blackouts the screen after pressing MB1 to try to rotate the uA > colour wheel if executed on a TrueColor default visual screen. 6@ > The screen is black and remains black afterwards although the ? > display server continues running. I didn't find any means to  @ > get anything displayed after that happens therefore I have to > > reboot. Pressing MB2 to stop the program always works, i.e. 8 > before trying to press MB2. The program also works if ? > executed on a screen with a PseudoColor default visual or on f= > an older machine with a different graphics board, an older i9 > VMS version and an older DECwindows X11 display server. > > Also, resizing the colour wheel to a bigger window leads to 7 > a somewhat distorted image. Some background stripes, h< > horizontally and vertically, adjacent to the former right < > and bottom borders cut the wheel into pieces. The program ? > works perfectly wrt that phenomenon on the older environment.t > 2 > Some data for the machine that shows the problem2 > ------------------------------------------------ > , > This is a machine with a hobbyist license. > $ > - DS20E, single cpu 500 MHz, 512MBA > - Two PowerStorm 300, devices GBB0 (head 1) and GBA0 (head 2). a > > > The default visual for head 2 is changed to PseudoColor and ? > pixel depth 8. Head 2 is still TrueColor with pixel depth 24.  >  > $ product show history * > ; > PRODUCT                             KIT TYPE    OPERATIONs= > ----------------------------------- ----------- ----------- 9 > DEC AXPVMS VMS721_UPDATE V3.0       Patch       Installh9 > DEC AXPVMS VMS721_PCSI V1.0         Patch       Installo9 > DEC AXPVMS DECNET_OSI V7.2-1        Full LP     Installo9 > DEC AXPVMS DWMOTIF V1.2-5           Full LP     Installd9 > DEC AXPVMS OPENVMS V7.2-1           Platform    Installr9 > DEC AXPVMS TCPIP V5.0-10            Full LP     Install 9 > DEC AXPVMS VMS V7.2-1               Oper System Installe >   > $ type decw$server_0_error.log > 8 > This is the DECwindows X11 display server for OpenVMS 5 > AXP V7.1-001212 compiled on Dec 12 2000 at 16:34:31m > ...t >  > ? > Some data from the machine that works okay wrt "colour wheel"s? > -------------------------------------------------------------: > < > This is a machine with a commercial license but also only : > used for private purposes. At the time when I baught it + > the hobbyist programme didnd't exist yet.. >  > - DEC 3000 300LX< > - on board graphics supports only PseudoColor with depth 8 > 	 > VMS 6.2  >   > $ type decw$server_0_error.log > 8 > This is the DECwindows X11 display server for OpenVMS 5 > AXP T6.2-950505 compiled on May  5 1995 at 04:29:36l > ? > -------------------------------------------------------------  > A > I understand that DECwindows X11 display server on the problem g> > machine has been compiled for VMS 7.1 but is now running on . > 7.2-1. But this is the factory installation. > A > I really appreciate any hints or directions how to investigate s; > the problem any further. Please let me know whether some  = > necessary information is missing I'm eager to supply that. c: > If the topic is not appropriate for the NG or for other 8 > reasons please email me directly (brass@infoplus.com).    B I don't know how to help you, just wanted to say we have a similarC problem - DECW$SERVER_0 dies on a couple of machines in our cluster F every few days or so. It doesn't seem to be related to any particular.G application running, and sometime dies when left all alone. One machinevC that has the most frequent problems is an Alphastation 5/500 with 2 E graphics cards. The other is a DPS 600au. Both running 7.2-2 with alliF but ECO's applied. Motif version 1.2-6. The contents of the last crash5 on the DPS from the end of DECW$SERVER_0_ERROR.LOG ::M    =   27-JAN-2002 11:40:24.5 Invalid access from transport: TCPIPCC                                                  node: ns1.phy.bnl.b8                                                  user: ?=   27-JAN-2002 11:40:25.1 Invalid access from transport: TCPIP C                                                  node: ns1.phy.bnl.e8                                                  user: ?H   27-JAN-2002 11:40:25.2 Connection d002c40 is closed by Txport (status  =-   20dc)-C   27-JAN-2002 11:40:25.3 %SYSTEM-F-ACCVIO, access violation, reasonMA   mask=00, virtual address=FFFFFFFFCCD56C4C, PC=0000000000000000,a  >   Unrecoverable server internal error (error code = 12) found,   terminating all connections.   Mapped Images...  .    START       END        LENGTH    IMAGE NAME.    -----       ---        ------    ----------3         0      201ff      201ff    DECW$SERVER_MAIN 2     22000     1c3bff     1a1bff    DECW$SERVER_DIX3 7baca000   7bafbfff      31fff    DECW$SECURITY_VMS + 7b81c000   7b88dfff      71fff    SECURESHR , 7b35e000   7b3cffff      71fff    SECURESHRP ...d  : I'm not sure what to try. any/all suggestions appreciated.   ------------------------------  % Date: Wed, 30 Jan 2002 03:39:12 +0100h' From: Brass Christof <welcome@spam.net>dY Subject: Re: DECW$SERVER_0 dies periodically; X11 colour wheel program blackouts the  scro( Message-ID: <3C575CD0.7FA0957A@spam.net>   Sam Hoblit wrote:p  D > I don't know how to help you, just wanted to say we have a similarE > problem - DECW$SERVER_0 dies on a couple of machines in our clusternH > every few days or so. It doesn't seem to be related to any particular.I > application running, and sometime dies when left all alone. One machinesE > that has the most frequent problems is an Alphastation 5/500 with 2eG > graphics cards. The other is a DPS 600au. Both running 7.2-2 with all H > but ECO's applied. Motif version 1.2-6. The contents of the last crash7 > on the DPS from the end of DECW$SERVER_0_ERROR.LOG ::i > ? >   27-JAN-2002 11:40:24.5 Invalid access from transport: TCPIPrE >                                                  node: ns1.phy.bnl.M: >                                                  user: ?? >   27-JAN-2002 11:40:25.1 Invalid access from transport: TCPIPoE >                                                  node: ns1.phy.bnl.o: >                                                  user: ?Q >   27-JAN-2002 11:40:25.2 Connection d002c40 is closed by Txport (status = 20dc) E >   27-JAN-2002 11:40:25.3 %SYSTEM-F-ACCVIO, access violation, reason>C >   mask=00, virtual address=FFFFFFFFCCD56C4C, PC=0000000000000000,s > @ >   Unrecoverable server internal error (error code = 12) found,  >   terminating all connections. >   Mapped Images... > 0 >    START       END        LENGTH    IMAGE NAME0 >    -----       ---        ------    ----------5 >         0      201ff      201ff    DECW$SERVER_MAIN<4 >     22000     1c3bff     1a1bff    DECW$SERVER_DIX5 > 7baca000   7bafbfff      31fff    DECW$SECURITY_VMS - > 7b81c000   7b88dfff      71fff    SECURESHR . > 7b35e000   7b3cffff      71fff    SECURESHRP > ...  > < > I'm not sure what to try. any/all suggestions appreciated.   Thanks for joining!o  > Kleinsorge Fred from DEC who is very well informed and helped ? a lot already responded. You'll see his message probably soon.  = He states that my two problems are probably unrelated and he .; asked for the Log file which I couldn't supply. Thanks for  ; posting the Log file. I'm sure Fred will read your message w  and have a look at the Log file.  < He also mentioned that he will tell the developer about the 9 strange phenomenon when running the colour wheel example c program.  : Would you like to mention the graphics cards you're using?6 Could you eventually compile and run the colour wheel ; example program? This could help in isolating the problem: -6 Is it related to the graphics card or only to the X11  display server?n8 If you don't want to download and edit the colour wheel : example program just let me know and I'll send it to you. # Do you have a C compiler installed?    -- c6 moc dot slupofni at ssarb - please revert the sequence   ------------------------------  % Date: Tue, 29 Jan 2002 23:26:28 +0100d' From: Brass Christof <welcome@spam.net> Y Subject: Re: DECW$SERVER_0 dies periodically; X11 colour wheel program blackouts the scre ( Message-ID: <3C572194.2BD9BC42@spam.net>  ' Thanks, Fred, that was really fast! :-)f   Fred Kleinsorge wrote: > $ > Looks like two unrelated problems.  
 Okay then.  4 > What is in DECW$SERVER_0_ERROR.LOG when it "dies"?  8 I had a look the first time it happened and didn't find ; anything different from normal operation, i.e. no messages e< related to any problems. Unfortunately I just detected that < ony two generations of this file are kept and the last file = after it died had already gone. Since I rebooted the machine p; this afternoon I had to wait until it will happen again. I d< will then send it to you - it's not that big, only a few k. ; It contains only "Connection <nr>" message lines after the T initialisation section.   = I removed some extensions especially DEC-XTRAP. The complete t list of loaded exts is  * Activating extension image DECW$SVEXT_Xie,* Activating extension image DECW$SVEXT_glx,3 Activating extension image DECW$SVEXT_DEC_PEER_GLX,t  $ If necessary I'll load some of them.  M > I'll ask the developer to run the colorwheel on a P300 and try to reproducea > the colormap problem.u  * This *are* really great news. Many thanks!  : I detected a typo in my message: pressing MB2 to stop the 8 program always works, i.e. before trying to press *MB1*.   Christof   -- o6 moc dot slupofni at ssarb - please revert the sequence   ------------------------------  % Date: Tue, 29 Jan 2002 15:49:03 -0500s1 From: Michael Austin <maustin@firstdbasource.com>u- Subject: Re: DHCP as a client using Multinet? 2 Message-ID: <3C570ABF.B4CA1AA9@firstdbasource.com>   William Barnett-Lewis wrote: > F > I'm probably missing something obvious, if so please point me at the# > right place in the documentation.- > I > I have a router that provides a DHCP server and I would like to have my F > Vaxstation get it's IP from that. I have installed Multinet and it'sG > PAK. I am not able to find in the documentation how to set it up as asI > DHCP client. Again, if this is an FAQ, a pointer in the right direction. > is appreciated.  > 	 > Thanks,  > 	 > Williami > --, > You better watch out    What you wish for;- > It better be worth it   So much to die for.o/ >                                 Courtney Lovee  @ I would prefer to hard-code my IP (and I do) such that I can useH port-fowarding and I like to be able to connect with reflections without having to guess the IP address.,  A With DHCP, you don't always know what your IP address is for thatuF purpose. I have 2 vms systems (tpcpip V5.0a eco-3) connected to my DSL7 service. I also have a couple of unix boxes  as well.  e -- e   Regards,   Michael Austin7 First DBA Source, Inc. -- http://www.firstdbasource.coms Sr. Consultant   ------------------------------  % Date: Tue, 29 Jan 2002 16:21:49 -0500t% From: JF Mezei <jfmezei@videotron.ca>n- Subject: Re: DHCP as a client using Multinet?a, Message-ID: <3C57123C.EB80EC8A@videotron.ca>   Michael Austin wrote: K > > I have a router that provides a DHCP server and I would like to have my.% > > Vaxstation get it's IP from that.d  . > I would prefer to hard-code my IP (and I do)  M I agree about the static IP address for VMS and let the router do the addressa translation for you.   ------------------------------  % Date: Tue, 29 Jan 2002 16:36:29 -0500n0 From: "Syltrem" <syltrem@videotron.spammenot.ca>) Subject: Re: Disaster recovery - success! 3 Message-ID: <NCE58.2462$EI.13291@tor-nn1.netcom.ca>l   Hi!n  : QMAN$MASTER. I have the logical defined in SYLOGICALS.COM.K The QMAN$MASTER logical for the node they booted (actually boot partition 0rH for which the VAX is long gone now) may not have everything setup right.K But then why would this affect my file? I also tried to stop queue manager, K restore (again) the file from tape then restart the queue manager and, samem problem.  L Can't be sure if the file was any good on tape. Is there any way I can check6 this here (without screwing up my production system!?)   Thanks!    --   SyltremrI http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)o> To reply to myself directly, remove .spammenot from my address  ? "JF Mezei" <jfmezei@videotron.ca> a crit dans le message news: ! 3C56CD2F.92770A0D@videotron.ca...a > Syltrem wrote:F > > - After booting there were no jobs in the batch queues (???). It's possibleL > > that this was caused by the people at the recovery site who did boot ourJ > > system without asking permission. If anyone experienced this and has a clue > > as to why, pls let me know!, > K > On yor main production site, you need to look at the QMAN$MASTER logical.a >yI > If you have heterogenous cluster (multiple system disks) you have to be J > careful about the order they are booted since you would want the one whoI > actually has the main queue manager database to boot first and then theo otherlI > ones defining the QMAN$MASTER before doing the START/QUEUE so that they  know1 > where the real queue manager files are located.i   ------------------------------  % Date: Tue, 29 Jan 2002 16:38:14 -0500g0 From: "Syltrem" <syltrem@videotron.spammenot.ca>) Subject: Re: Disaster recovery - success!c3 Message-ID: <qEE58.2463$EI.13629@tor-nn1.netcom.ca>/  G RAID. That's possible. What I know for sure is that they had stripesets  (2x18GB making 1x36GB). 3 But those were "hardware", VMS would not know this.-   I can ask them.-   --   Syltrem2I http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)A> To reply to myself directly, remove .spammenot from my address  H "Koska, John C. (LNG-MBC)" <John.C.Koska@lexisnexis.com> a crit dans le
 message news:/B 3D35AD137AAAD411A6BA0008C7B1B12D016026FE@MBCALBEXC03.BENDER.COM... > > -----Original Message-----9 > > From: Syltrem [mailto:syltrem@videotron.spammenot.ca]S, > > Sent: Tuesday, January 29, 2002 11:08 AM > > To: Info-VAX@Mvb.Saic.Coma) > > Subject: Disaster recovery - success!a > >d >  > Congrats!  >8 > > @ > > Just a note to say that we did the 1st test for our disaster > > recovery planAH > > last week end, and it was a success thanks to good old VMS (and me)! > .) > <snip> >e= > > All went very smooth, but due to the hardware config overr > > there, the disk IO> > > was terribly bad and it took us 3 times as long to restore > > than it took tor > > create the backup tapes. >s< > If you were using RAID at the disaster recovery site (likeF > RAID-1/mirroring or RAID-5), then you should expect some degradationD > in restores versus backups. (ie: Writing out to two physical disks% > and/or writing out the parity bit.)1 >aE > If your restores were considered really bad, then perhaps next timemF > you can test doing a backup of the disaster recovered system, so youG > know if you can safeguard your business while in a disaster recoveredBI > situation.  I would posit that if you can not also backup the recovered H > system within your normal backup window, that you do not have adequateI > hardware/configuration at the disaster recovery site, and you will needyA > to contract or otherwise make sure you have such in the future.  >.G > I usually like to be able to restore in my normal backup window, that1H > way I know I have adequate resource on hand, and the business might beL > up and running with less financial loss.  Besides, your time will be eatenC > up by all sorts of unplanned and uncontrolable things in disaster 	 recovery. H > Things like having other OS system admins asking if you could get themH > something to eat and drink, since your system is restored and running.	 > <grin!>e >i > :) jck > John Koska > Matthew Bender & Co., Inc. -$ >   A Member of the LexisNexis Group > 1275 BroadwayC > Albany, NY  12204e > USA  > 518-487-3255 > John.C.Koska@lexisnexis.com  > + > I post personal opinion only, and all thet, > disclaimers one could imagine apply.  That* > includes, I speak for myself only and my+ > views in no way represent my employer(s). - > One should also take note of the Electronicw+ > Communications Privacy Act of 1986, whichr- > imposes civil and criminal liability on anye* > person who intentionally intercepts "any* > wire, oral or electronic communication." >h   ------------------------------  # Date: Tue, 29 Jan 2002 23:57:42 GMTc) From: rob.buxton@wcc.govt.nz (Rob Buxton)6) Subject: Re: Disaster recovery - success!f2 Message-ID: <3c573661.2595218346@news.wcc.govt.nz>  - On Tue, 29 Jan 2002 16:36:29 -0500, "Syltrem"i' <syltrem@videotron.spammenot.ca> wrote:    >Hi! > ; >QMAN$MASTER. I have the logical defined in SYLOGICALS.COM. L >The QMAN$MASTER logical for the node they booted (actually boot partition 0I >for which the VAX is long gone now) may not have everything setup right.iL >But then why would this affect my file? I also tried to stop queue manager,L >restore (again) the file from tape then restart the queue manager and, same	 >problem.e >tM >Can't be sure if the file was any good on tape. Is there any way I can checkj7 >this here (without screwing up my production system!?)w >l >Thanks! >   @ As well as the Qman$master logical I've also found the following? should be done if the device the Que Files were on has changed.m  :  start/que/manager/on=(node1,node2.,...) device:[que.dire]   Rob.   >--  >h >SyltremJ >http://pages.infinit.net/syltrem (OpenVMS related web site - en franais)? >To reply to myself directly, remove .spammenot from my addressi >l@ >"JF Mezei" <jfmezei@videotron.ca> a crit dans le message news:" >3C56CD2F.92770A0D@videotron.ca... >> Syltrem wrote:dG >> > - After booting there were no jobs in the batch queues (???). It's 	 >possible M >> > that this was caused by the people at the recovery site who did boot ourSK >> > system without asking permission. If anyone experienced this and has a  >cluec  >> > as to why, pls let me know! >>L >> On yor main production site, you need to look at the QMAN$MASTER logical. >>J >> If you have heterogenous cluster (multiple system disks) you have to beK >> careful about the order they are booted since you would want the one whoSJ >> actually has the main queue manager database to boot first and then the >otherJ >> ones defining the QMAN$MASTER before doing the START/QUEUE so that they >knowl2 >> where the real queue manager files are located. >d >    ------------------------------  % Date: Tue, 29 Jan 2002 21:33:47 +0100-2 From: martin@radiogaga.harz.de (Martin Vorlaender)4 Subject: Re: Export Monitor Data to CSV file, Excel?; Message-ID: <3c57072b.524144494f47414741@radiogaga.harz.de>   G Bart Z. Lederman (lederman@star.enet.dec.DISABLE-JUNK-EMAIL.com) wrote: > > I wrote a program a few years ago called MONITOR-TO-CSV.MAR, ....+ > I'm reasonably certain it went out on one - > of the Freeware and / or DECUS collections.    It's on Freeware CD v4 #1.   cu,o   Martin -- tD                         | Martin Vorlaender  |  VMS & WNT programmer1  VMS is today what      | work: mv@pdv-systeme.dedE  Microsoft wants        |    http://www.pdv-systeme.de/users/martinv/ 8  Windows NT 8.0 to be!  | home: martin@radiogaga.harz.de   ------------------------------    Date: 29 Jan 2002 22:35:25 -0800- From: kanpeter@graduate.hku.hk (Peter CK Kan)s Subject: FMS/CFMSe= Message-ID: <50c45596.0201292235.120ebe7e@posting.google.com>y   Help...   D How can I toggle insert mode and overwrite mode during data input onB screens written by FMS/CFMS  ... (I know I can do that by pressing ctrl-A in command line).   -- peter   ------------------------------  % Date: Tue, 29 Jan 2002 21:47:46 +0000e) From: Greg Thomas <thomasgd@omc.bt.co.uk>o' Subject: Re: fopen crashes while in ASTo8 Message-ID: <d16e5u4l9i8nt6m0571v48dkr2to10bl36@4ax.com>  C On Fri, 25 Jan 2002 13:22:46 -0500, JF Mezei <jfmezei@videotron.ca> 2 wrote in article <3C51A272.F3B20140@videotron.ca>:  N >The AST works fine the first time it is called (after about 500 transactions)F >but upon its second invocation, it will generally fail at the fopen() >statement.0  F FWIW, I had the exact same problem a few years ago on OpenVMS VAX 5.5.C In the end a minor re-design to ensure the fopen() wasn't called at B AST level and the problem went away. Not much help, I know, but at least you're not alone.C   Greg --  5 This post represents the views of the author and doesb5 not necessarily accurately represent the views of BT.u   ------------------------------  % Date: Tue, 29 Jan 2002 21:51:09 -0000a- From: wspencer@ap.nospam.org (Warren Spencer)Y' Subject: Re: fopen crashes while in AST 7 Message-ID: <91A5AB863warrenspencer1977@209.249.90.100>0  ( jfmezei@videotron.ca (JF Mezei) wrote in" <3C56F08A.1315EC2F@videotron.ca>:    >Warren Spencer wrote:G >> There are better experts in this ng than me, but it seems to me that G >> doing I/O in an AST handler is a bad idea, and I've heard rumor this0/ >> can lead to the dreaded RWAST process state.  >1F >No, it wasn't RWAST, it was simply crashing on the fopen (and only on% >that statement, but not everytime).    L The point I'm trying to make is that (I believe) it *can* lead to the RWAST J state and is therefore to be avoided.  Whether you experience the problem B or not will depend on a bazillion factors local to your situation.  # > In any event, I ran it during theEI >weekend and restarted the instances whenever they failed and reduced thesG >checkpoint to every half hour instead of 10 minutes and managed to getn >the work done on that one.  >oF >However, I would like to know more since I have used C IO in AST mode >many times without problems.   I If you did all your I/O exclusively within an AST, you may be fine since oI they're effectively single-threaded (within a given access mode).  But I t% believe you still run the RWAST risk.a  F >In the case of last weekend, the main program used *no* C file IO. ItE >only used RMS $GET and $UPDATE to file and $QIO to internet. And thenH >fopen was always triggered in AST while the main program was in a $QIOWI >(waiting for internet data, which is where the program spent most of its  >time).   C As I indicated in my first response, I believe *any* I/O is risky,  B regardless of which interface (c, rms, ...) you use to perform it.  F >> Move the "work" part of the code into the normal processing thread,H >> restrict your AST activities to setting event flags or calling $WAKE,2 >> and I'm sure you program will run like a champ. >5G >Nop. At the end of the program, the routine that updates the config is0A >called in non AST mode and it also crashed there once. There was0H >something I was doing in the program that corrupted "fopen" after a few< >calls to it. (it wouldn't fail on the first call to fopen).  H As I said above, restrict your AST activities to setting event flags or K calling $WAKE. If you mix background I/O with AST I/O, I would be suprised  ' if you *didn't* get problems like this.$  K Of course, this problem could be something as simple (yet hard-to-find) as   a wild pointer.f   ws   -- C   Warren Spencer' Senior Software Engineer (not a writer)  The Associated Press  @ ** When Windows is your hammer, everything looks like a thumb **   ------------------------------  # Date: Wed, 30 Jan 2002 03:36:48 GMT + From: Ryan Moore <rmoore@rmoore.dyndns.org> ' Subject: Re: fopen crashes while in AST = Message-ID: <Pine.LNX.4.31.0201291931540.20015-100000@jaipur>o  E Did you use the decc$set_reentrancy() function to specify you will be G using ASTs?  You should really do this if you want to use fopen() in anw
 AST.  See:  S http://www.openvms.compaq.com/commercial/c/5763p023.htm#decc_set_reentrancy_routine   8 You should call decc$set_reentrancy(C$C_AST);  Try that.    $ On Tue, 29 Jan 2002, JF Mezei wrote:   > Warren Spencer wrote:?N > > There are better experts in this ng than me, but it seems to me that doingN > > I/O in an AST handler is a bad idea, and I've heard rumor this can lead to$ > > the dreaded RWAST process state. >dL > No, it wasn't RWAST, it was simply crashing on the fopen (and only on thatN > statement, but not everytime). In any event, I ran it during the weekend andL > restarted the instances whenever they failed and reduced the checkpoint toU > every half hour instead of 10 minutes and managed to get the work done on that one.n > L > However, I would like to know more since I have used C IO in AST mode many > times without problems.  >5L > In the case of last weekend, the main program used *no* C file IO. It onlyK > used RMS $GET and $UPDATE to file and $QIO to internet. And the fopen was L > always triggered in AST while the main program was in a $QIOW (waiting forD > internet data, which is where the program spent most of its time). >nG > > Move the "work" part of the code into the normal processing thread,tM > > restrict your AST activities to setting event flags or calling $WAKE, andt/ > > I'm sure you program will run like a champ.i >pO > Nop. At the end of the program, the routine that updates the config is called K > in non AST mode and it also crashed there once. There was something I was"J > doing in the program that corrupted "fopen" after a few calls to it. (it, > wouldn't fail on the first call to fopen). >o   ------------------------------  # Date: Tue, 29 Jan 2002 23:01:34 GMTg! From: Ernest <wmozart5@yahoo.com>c8 Subject: fopen() behavior difference...bug or undefined?A Message-ID: <iRF58.99281$vH6.5746667@bin6.nnrp.aus1.giganews.com>   $ Consider this C source code snippet:  	 FILE *fd; 
 int writable;m4 writable = ((fd = fopen( "foo.txt", "a" )) != NULL);
 fclose( fd );   H On MS-Windows and several Unix platforms I've tested this on, this code I does NOT update foo.txt's "last modified" time (nothing is written).  On aK OpenVMS/Alpha 7.2 with Compaq C/C++ 6.4, it does.  Is the behavior of such  J code considered C library-implementation-specific and thus no assumptions J can be made?  I've found no specific mention of the behavior in man pages J or similar docs on various platforms, but it certainly seems to be the de  facto standard.v  H I'll have to use access() instead, which is fine, but it's not quite as F warm-and-fuzzy as getting a non-NULL from trying to open the file for C writing.  I'm more curious about the above for pedantic reasons, I n suppose.   Cheers,o Ernest   ------------------------------  # Date: Wed, 30 Jan 2002 03:19:02 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network>t< Subject: Re: fopen() behavior difference...bug or undefined?* Message-ID: <3C576A43.3060700@qsl.network>  
 Ernest wrote:e  & > Consider this C source code snippet: >  > FILE *fd;  > int writable;L6 > writable = ((fd = fopen( "foo.txt", "a" )) != NULL); > fclose( fd );  > J > On MS-Windows and several Unix platforms I've tested this on, this code K > does NOT update foo.txt's "last modified" time (nothing is written).  On dM > OpenVMS/Alpha 7.2 with Compaq C/C++ 6.4, it does.  Is the behavior of such dL > code considered C library-implementation-specific and thus no assumptions L > can be made?  I've found no specific mention of the behavior in man pages L > or similar docs on various platforms, but it certainly seems to be the de  > facto standard.t    H The behavior is documented in the Compaq C RTL manual.  OpenVMS updates . the modification time when the file is opened.  G In UNIX, that same time field is used to indicate that a "significant" PB event occured to the file.  The meaning of what is significant is + apparently left up to the operating system.I  I This is one of the fun things of porting things from UNIX to OpenVMS for rD someone that does not really know the UNIX api.  There appear to be E behaviors of UNIX routines that are "well known", but not explicitly v documented anywhere.  C Sometimes I am forced to actually try to figure out what the LINUX I source is doing.   -John  wb8tyw@qsl.network Personal Opinion Onlye   ------------------------------  % Date: Tue, 29 Jan 2002 18:01:43 -0500(# From: sol gongola <sol@adldata.com> 8 Subject: How to change vms system ip address and domain?+ Message-ID: <3C5729D7.243755DA@adldata.com>   9 I am looking for specific directions for changing the ip e/ address and the domain name (a.b.com-->a.c.com)m  < Last time I tried it, I wound up with a local host mismatch > error in 'ucx sho name', couldn't fix it, and finally deleted ? the tcpip$configuration.dat file and started from scratch with i
 tcpip$config.r  ! vms 7.2-1 with tcpip version 5.1en  	 thank youp --  ' Sol Gongola           (sol@adldata.com)y ADL Data Systems Inc 20 livingstone ave Dobbs Ferry, NY 10522s' phone: 914 591 1800   fax: 914 591 1818a   ------------------------------  % Date: Tue, 29 Jan 2002 23:09:15 -0000 = From: "David McKenzie" <david.mckenzie@spitfire0.demon.co.uk>e< Subject: Re: How to change vms system ip address and domain?B Message-ID: <1012345786.13658.0.nnrp-12.c1edba74@news.demon.co.uk>  9 Either use tcpip$config or use the set interface commandst  0 "sol gongola" <sol@adldata.com> wrote in message% news:3C5729D7.243755DA@adldata.com...u: > I am looking for specific directions for changing the ip1 > address and the domain name (a.b.com-->a.c.com)o > = > Last time I tried it, I wound up with a local host mismatcht? > error in 'ucx sho name', couldn't fix it, and finally deletede@ > the tcpip$configuration.dat file and started from scratch with > tcpip$config.i > # > vms 7.2-1 with tcpip version 5.1ee >u > thank you2 > --) > Sol Gongola           (sol@adldata.com)r > ADL Data Systems Inc > 20 livingstone ave > Dobbs Ferry, NY 10522D) > phone: 914 591 1800   fax: 914 591 1818    ------------------------------  % Date: Wed, 30 Jan 2002 00:18:40 +0100  From: "B. Eckstein" <be@cli.de>s< Subject: Re: How to change vms system ip address and domain?% Message-ID: <3C572DD0.1080005@cli.de>n   sol gongola wrote:  ; > I am looking for specific directions for changing the ip ,1 > address and the domain name (a.b.com-->a.c.com)C    # > vms 7.2-1 with tcpip version 5.1e>  
 @tcpip$config    --    B.Eckstein, - mailto:be@epost.de5 FAQ zu de.comp.hardware.netzwerke: http://how.to/dchni   ------------------------------  % Date: Tue, 29 Jan 2002 15:10:23 -0800SC From: "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam>S1 Subject: Re: How to get/put the symbol list in C?a0 Message-ID: <3C572BDF.4FDB5E41@intel.com.nospam>  
 Ernest wrote:d  2 > forkosh@panix1.panix.com (John Forkosh) wrote in# > news:a35s8a$eqi$2@news.panix.com:s >hK > >: Assuming I can get the full list captured into a C data structure, and D > >: add/modify/delete entries as I need, how do I then output these( > >: symbols for use by a child process? > >rH > > You can run just about any external program from C with the system()A > > command.  I can't recall whether show has an /out= qualifier.iJ > > If not, you could use lib$spawn().  Then you could redirect the outputG > > to a file and parse that file.  Not very elegant, but I assume thiseH > > is some kind of initialization-time task, since I can't imagine some >pK > That's correct.  So I think I can handle getting the symbols, that's half H > the battle.  But how to output the changed values for use by the childL > process?  setenv() is obviously just for the 4 pseudo-environment vars, as > is the envp arg to execve().  G     Can you please tell us what you're trying to accomplish (program or E process design, overall goal)?  I'm having a lot of trouble trying toiC understand why (1) you need to translate all process symbols (don'teE you want the logical names as well?) and (2) what you're trying to do A to the subprocesses.  For example, a simple way to communicate toiB subprocesses is through the JOB logical name table which is sharedB between all of them (but not shared with detached processes).  But@ without a better grasp on what your problem is, it's hard to say* whether that's the best approach or not...       -Ken --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfield ! F20 Automation VMS System Support  kenneth.h.fairfield#intel.com    ------------------------------  # Date: Wed, 30 Jan 2002 00:10:50 GMTK! From: Ernest <wmozart5@yahoo.com>p1 Subject: Re: How to get/put the symbol list in C?nA Message-ID: <eSG58.99518$vH6.5805589@bin6.nnrp.aus1.giganews.com>n  F "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam> wrote in) news:3C572BDF.4FDB5E41@intel.com.nospam: P  I >     Can you please tell us what you're trying to accomplish (program or   % It's a port of the Unix make utility.   E > understand why (1) you need to translate all process symbols (don't-& > you want the logical names as well?)  I Logical names would be fine too, but not necessary.  Global symbols seem oJ to be the closest equivalent of environment variables on other platforms, K as far as my limited VMS experience shows.  And without fairly substantial wD redesign, one or the other has to be chosen.  Since most/all of the I variables that need to be set globally are not specific to the operating tJ system (which are user-configurable), using generic symbols/strings seems  to be the best approach.  9 > and (2) what you're trying to do to the subprocesses.  e  H Not much really, just run them [possibly many at once] with a different & set of global symbols than the parent.   Thanks,s Ernest   ------------------------------  # Date: Wed, 30 Jan 2002 00:25:11 GMT ! From: Ernest <wmozart5@yahoo.com>k1 Subject: Re: How to get/put the symbol list in C?lA Message-ID: <H3H58.99663$vH6.5818801@bin6.nnrp.aus1.giganews.com>w  $ Ernest <wmozart5@yahoo.com> wrote in: news:eSG58.99518$vH6.5805589@bin6.nnrp.aus1.giganews.com:   : >> and (2) what you're trying to do to the subprocesses.   > I > Not much really, just run them [possibly many at once] with a differentm( > set of global symbols than the parent.  G Actually, one other thing worth mentioning: we need to support command .K separation and redirection features like ";" "&&" ">" "<", etc., which the eJ PIPE command provides, hence the need to use that as the initial argument # to execvp for all spawned commands.>   Cheers,h Ernest   ------------------------------  # Date: Wed, 30 Jan 2002 02:18:57 GMTw- From: "John E. Malmberg" <wb8tyw@qsl.network>t1 Subject: Re: How to get/put the symbol list in C?a* Message-ID: <3C575C2F.9080500@qsl.network>  
 Ernest wrote:d  H > "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam> wrote in+ > news:3C572BDF.4FDB5E41@intel.com.nospam: a >  > I >>    Can you please tell us what you're trying to accomplish (program orQ >> > ' > It's a port of the Unix make utility.h     Another one? :-)    aE >>understand why (1) you need to translate all process symbols (don't-& >>you want the logical names as well?) >> > K > Logical names would be fine too, but not necessary.  Global symbols seem oL > to be the closest equivalent of environment variables on other platforms, M > as far as my limited VMS experience shows.  And without fairly substantial wF > redesign, one or the other has to be chosen.  Since most/all of the K > variables that need to be set globally are not specific to the operating rL > system (which are user-configurable), using generic symbols/strings seems  > to be the best approach.     As I posted earlier:  C Unless you explicitely exclude them on a LIB$SPAWN or SYS$CREPRC() oG command, the child inherits the logical names and the symbols from the sH parent.  No special programming tricks are needed.  It should work just  like it does on UN*X.n  F If you use setenv() and friends in the parent process, those are also G inherited by programs started by the exec() family of services, or the c system() command.     a9 >>and (2) what you're trying to do to the subprocesses.    >> cJ > Not much really, just run them [possibly many at once] with a different ( > set of global symbols than the parent.    E If you using either LIB$SPAWN or SYS$CREPRC to create the child, you oF will need to use some communication method to tell it what symbols to 4 create locally.  The use of mailboxes is one method.  C Or you can use the LIB$SET_SYMBOL command to set the symbol in the aG parent just before the child is created as a subprocess.  This way the f' child process will look just like UNIX.o     -Johns   wb8tyw@qsl.network   Personal Opinion Onlyi   ------------------------------  % Date: Tue, 29 Jan 2002 21:56:51 -0600eC From: "Craig A. Berry" <craig.berry@nospam.signaltreesolutions.com>e1 Subject: Re: How to get/put the symbol list in C?e= Message-ID: <3C576F03.9090708@nospam.signaltreesolutions.com>e  
 Ernest wrote:o  & > Ernest <wmozart5@yahoo.com> wrote in< > news:eSG58.99518$vH6.5805589@bin6.nnrp.aus1.giganews.com:  >  > : >>>and (2) what you're trying to do to the subprocesses.   >>>,I >>Not much really, just run them [possibly many at once] with a differentY( >>set of global symbols than the parent. >> > I > Actually, one other thing worth mentioning: we need to support command (M > separation and redirection features like ";" "&&" ">" "<", etc., which the :L > PIPE command provides, hence the need to use that as the initial argument % > to execvp for all spawned commands.d    F A number of the things you are doing have been done before in whole or: in part. There is a GNU make included with the GNV packageH <http://gnv.sourceforge.net> and there have been a variety of other makeF ports in the past. I don't know how good they are so perhaps your work is still needed.  > On getting and setting DCL symbols, go to the documentation atF <http://www.openvms.compaq.com/doc>, locate the manual called "OpenVMS< RTL Library (LIB$) Manual" and read about lib$get_symbol andE lib$set_symbol. IIRC, lib$get_symbol does not accept wildcards so yousG will have to use one of the other tricks people have been talking about ! to get the whole list of symbols.   H If you get any recent Perl distribution and look at the file [.vms]vms.cE and check out the routine called prime_env_iter(), you will see it ishE doing exactly what you are preparing to do, though it gets quite uglytF and complicated (too complicated!) because it is handling both symbolsH and logical names and various optional behaviors and storing the resultsG of its environment scan in a Perl hash. Nevertheless, somewhere betweenAD simplifying this complexity and reinventing the wheel you may find a
 happy medium.   E You will also see a routine called getredirection() that does commandsG redirection and most of the Unixy command goodies such as ending with a H "&" to run the command as a subprocess. IIRC, I don't think it handles aG list of commands separated by semicolons but I could be wrong. This mayiG not help you unless you are able to integrate it into all programs thatk* you will be executing as spawned commands.   ------------------------------  % Date: Wed, 30 Jan 2002 00:33:59 -0000t, From: "pocketmoon" <pocketmoon@ntlworld.com>F Subject: How to tell if File is locked in C ?fopen extension pehraps ?C Message-ID: <S2H58.51156$ka7.8395723@news6-win.server.ntlworld.com>b  	 Hi Folks,   " OpenVMS 7.2 Dec C (cant remember!)  J I'm scanning through a directory tree using opendir/readdir and would likeL to determine if a file is locked, i.e. currently opened by another user (who% would in fact be writing to the file)t  E I'm trying to replace a bit of C code which gets the lock status (andfG creation/modification dates) by calling a DCL script via the C system()sK function. This script creates a logical with either date set and the C code-# then reads this logical back in (!)   H The DCL gets the creation or modification timestamp via a file attributeF call which fails if the file is being written to by another user. ThisC failure is returned back to the C which assumes the file is locked.   J  So I'm thinking perhaps there's perhaps a DEC C extension to fopen which,H e.g. fp = fopen (myfile, "r", "???=???)  which will ask for an exclusiveA lock on the file, which will of course fail if it's already open.t   Cheers   Rob J.   ------------------------------  % Date: Tue, 29 Jan 2002 20:32:36 -0500?% From: JF Mezei <jfmezei@videotron.ca>nJ Subject: Re: How to tell if File is locked in C ?fopen extension pehraps ?, Message-ID: <3C574D33.6412894F@videotron.ca>   pocketmoon wrote: L >  So I'm thinking perhaps there's perhaps a DEC C extension to fopen which,J > e.g. fp = fopen (myfile, "r", "???=???)  which will ask for an exclusiveC > lock on the file, which will of course fail if it's already open.d   fopen(yourfile,"r","shr=nil")    ------------------------------  # Date: Wed, 30 Jan 2002 02:07:08 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network> J Subject: Re: How to tell if File is locked in C ?fopen extension pehraps ?* Message-ID: <3C57596A.9020904@qsl.network>   pocketmoon wrote:e   > Hi Folks,  > $ > OpenVMS 7.2 Dec C (cant remember!) > L > I'm scanning through a directory tree using opendir/readdir and would likeN > to determine if a file is locked, i.e. currently opened by another user (who' > would in fact be writing to the file)k    F Look at FRONTPORT on the OpenVMS Freeware 5.0 CDROMs.  A copy of them < are available from a link at http://www.openvms.compaq.com/.     G > I'm trying to replace a bit of C code which gets the lock status (andoI > creation/modification dates) by calling a DCL script via the C system() M > function. This script creates a logical with either date set and the C codet% > then reads this logical back in (!)  > J > The DCL gets the creation or modification timestamp via a file attributeH > call which fails if the file is being written to by another user. ThisE > failure is returned back to the C which assumes the file is locked.e    E Sounds a lot like the C stat() call, except for determining the lock oF status.  Determining if the file is locked with out trying to open it F requires some internals knowledge of RMS and possibly privilege to go K into EXEC mode.  But such information has been posted on this forum before.u     L >  So I'm thinking perhaps there's perhaps a DEC C extension to fopen which,J > e.g. fp = fopen (myfile, "r", "???=???)  which will ask for an exclusiveC > lock on the file, which will of course fail if it's already open.d    G That is the default behavior of Compaq C, formerly known as DEC C, and dI also the default behavior of VAX C.  So all you had to do was attempt to aG open the file to test if it was locked.  Of course that can change the i modification dates.y  A Unless otherwise specified, a C program on OpenVMS using UNIX or 'F STANDARD I/O will exclusively lock a file.  This knowledge could make  your task much simpler.-  H For programs that use RMS Services to access the files, the simple test G to see if the file is locked could be misleading.  You may have access n- to the file but not the records inside of it.0  E The documentation for the RMS services can be found off of a link to pI http://www.openvms.compaq.com, and the documentation for Compaq C can be xG found by following links from http://www.openvms.compaq.com/commercial/n   -Johnr wb8tyw@qsl.network Personal Opinion OnlyN   ------------------------------  % Date: Tue, 29 Jan 2002 15:16:44 -0500n. From: Catherine Salvendy <csalvend@purdue.edu>* Subject: Re: I/O Performance on Alpha 4100* Message-ID: <3C57032C.94129139@purdue.edu>   Dirk Munk wrote:  B > Look at your caching (show mem /cache /full). It depends on yourH > application of course, but your read hit rate should be something likeF > 90%. If it is lower, and there is no cache memory free, increase theI > size of the cache. The standard size is 6400 blocks, or 3.2 MB. This iseF > much to low for most systems. Increase it to let's say 128MB (if youI > have that much memory free), and you wil have a much faster system. The " > sysgen parameter is VCC_MAXSIZE. >   mF I have been looking at the vitual IO parameter, but the help in SYSGENB indicates that VCC_MAXSIZE is a Compaq parameter and should not beC changed without consulting them! So I can safely change this myself A (via autogen of course)? We have a DS20E with 1GIG memory and oura7 cache memory is always full and the hit rate shows 69%.    ------------------------------  % Date: Tue, 29 Jan 2002 15:24:49 -0500-- From: "John Eisenschmidt" <jeisensc@aaas.org>0* Subject: Re: I/O Performance on Alpha 4100+ Message-ID: <sc56becf.002@AAASMTA.aaas.org>a  I I couldn't find the email that originally suggested this, but thank you =sK Dirk. Our Alphas were all at the default of 6400 blocks. On our two DS10s =mK with 1GB I upped it to 256000 blocks and what a difference! Hit Rate went = C from 43% to 94% within a couple hours and the speed difference is =c
 noticable.   Thank you thank you thank you.  F >>> Catherine Salvendy <csalvend@purdue.edu> 01/29/2002 3:16:44 PM >>>     Dirk Munk wrote:  B > Look at your caching (show mem /cache /full). It depends on yourH > application of course, but your read hit rate should be something likeF > 90%. If it is lower, and there is no cache memory free, increase theI > size of the cache. The standard size is 6400 blocks, or 3.2 MB. This isRF > much to low for most systems. Increase it to let's say 128MB (if youI > have that much memory free), and you wil have a much faster system. Ther" > sysgen parameter is VCC_MAXSIZE. >=20 =20aF I have been looking at the vitual IO parameter, but the help in SYSGENB indicates that VCC_MAXSIZE is a Compaq parameter and should not beC changed without consulting them! So I can safely change this myselfrA (via autogen of course)? We have a DS20E with 1GIG memory and our 7 cache memory is always full and the hit rate shows 69%.$   ------------------------------  % Date: Tue, 29 Jan 2002 15:34:37 -0500 > From: "Koska, John C. (LNG-MBC)" <John.C.Koska@lexisnexis.com>* Subject: RE: I/O Performance on Alpha 4100M Message-ID: <3D35AD137AAAD411A6BA0008C7B1B12D01602707@MBCALBEXC03.BENDER.COM>9  / I believe you will find there is a tradeoff in -. balance slots and other things in relation to / VCC_MAXSIZE.  If you make VCC_MAXSIZE too big, c/ you may take away balance slots, and hence you m0 might limit the number of processes your system  can support.  3 See Hoff's post in comp.os.vms from 1/17/2002 with n5 Subject: Re: Questions about VIOC on VMS 7.3 Alpha.  c1 I believe it is informative as to what can/should  be done with VCC_MAXSIZE.c   :) jck
 John Koska Matthew Bender & Co., Inc. -"   A Member of the LexisNexis Group
 1275 Broadway- Albany, NY  12204a USAu 518-487-3255 John.C.Koska@lexisnexis.come  ) I post personal opinion only, and all theo* disclaimers one could imagine apply.  That( includes, I speak for myself only and my) views in no way represent my employer(s).d+ One should also take note of the Electronic ) Communications Privacy Act of 1986, whichm+ imposes civil and criminal liability on any ( person who intentionally intercepts "any( wire, oral or electronic communication."   > -----Original Message-----7 > From: Catherine Salvendy [mailto:csalvend@purdue.edu] ) > Sent: Tuesday, January 29, 2002 3:17 PM  > To: Info-VAX@Mvb.Saic.Com , > Subject: Re: I/O Performance on Alpha 4100 >  > Dirk Munk wrote: > D > > Look at your caching (show mem /cache /full). It depends on your< > > application of course, but your read hit rate should be  > something likeH > > 90%. If it is lower, and there is no cache memory free, increase the@ > > size of the cache. The standard size is 6400 blocks, or 3.2 
 > MB. This iswH > > much to low for most systems. Increase it to let's say 128MB (if you@ > > have that much memory free), and you wil have a much faster 
 > system. Thev$ > > sysgen parameter is VCC_MAXSIZE. > >  >  yH > I have been looking at the vitual IO parameter, but the help in SYSGEND > indicates that VCC_MAXSIZE is a Compaq parameter and should not beE > changed without consulting them! So I can safely change this myself(C > (via autogen of course)? We have a DS20E with 1GIG memory and ourm9 > cache memory is always full and the hit rate shows 69%.  >    ------------------------------  % Date: Tue, 29 Jan 2002 18:17:01 -0500t, From: "Kenneth Block" <krblock@computer.org> Subject: Re: Installing GNVs1 Message-ID: <N3G58.160$am1.4263@news.cpqcorp.net>o  H At the moment there is no install for GNV.  You just unpack the tar, runJ setup, and then run bash. I have some sources for a PCSI install, but they need to be cleaned up.  9 "Paul Repacholi" <prep@prep.synonet.com> wrote in message ' news:87sn8qdtie.fsf@prep.synonet.com...c' > "Tom Linden" <tom@kednos.com> writes:C > / > > go to http://gnv.sourceforge.net/readme.htms >i; > > Get the sources and install you will get the following:n > .... >p) > Did that. Is there a INSTALL somewhere?c >n > --> > Paul Repacholi                               1 Crescent Rd.,9 > +61 (08) 9257-1001                           Kalamunda. B >                                              West Australia 60760 > Raw, Cooked or Well-done, it's all half baked.H > EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 29 Jan 2002 15:20:14 -08001 From: susan_skonetski@hotmail.com (Sue Skonetski)n1 Subject: Re: Intel Special issue of OpenVMS Timese; Message-ID: <857e9e41.0201291520.35f36d@posting.google.com>h   Dave and Terry,l  4 Please forgive the spelling mistakes I am on a role.  F Thank you very much for your kind words.  This was my first time doingF a newsletter and I am glad you found it useful and I hope you like the Intel issue as well.  F Just so folks know, (as far as I know) this was the only portion of my& previous job that I will not be doing.  @ And in case you thought you would be rid of me in the newsgroup.  
 NO WAY ;')  D It never was part of my job I became part of the newsgroup because I? love VMS and the community that supports it.  Sometimes I agree.D sometimes not but I am planning on continuing doing everything I can  to get more VMS information out.  D If you think I am taking it personal, you are right it is personal. = Some of the people I respect the most work with VMS either aswB engineers or customer engineers (no such thing as and ignorant VMSF user), my daughters college education is being bet on VMS AND I do not want4 Nuclear Power plants running anything but VMS or NSK- healthcare running on anything but VMS or NSKn3 Stock exchanged running on anything but VMS and NSKoF Anything that can hurt major amounts of people running on anything but VMS and NSKe  9 I am not a complete VMS bigot NSK is a great product too.b   ok I am off my soap box now.   Sue       ` "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message news:<3C561E94.88D1BF4D@fsi.net>... > "Terry C. Shannon" wrote:a > > A > > "Sue Skonetski" <susan.skonetski@compaq.com> wrote in messages/ > > news:uki58.105$am1.1433@news.cpqcorp.net...  > > > Dear Newsgroup,l > > >oJ > > > The Intel issue of OpenVMS Times will be posted and on line tomorrow	 >  thanks L > > > to Warren Sander http://www.openvms.compaq.com/openvmstimes/index.html > > >eN > > > Please note that this will be the last issue of OpenVMS Times where I am >  theH > > > editor.  I have changed jobs in the VMS group.  Diane Gaudreau HPS >  Marketing% > > > will be the editor from now on.S > > >e > > J > > Sue, congratulations on your role in getting the OpenVMS Times off theM > > ground and up and running. I am certain that the comp.os.vms constituencyeH > > shares the (correct) opinion that you've done a superb job with this$ > > downstream marketing collateral! > . > Terry's statements seconded whole-heartedly!   ------------------------------  % Date: Tue, 29 Jan 2002 17:31:53 -0600o+ From: Christopher Smith <csmith@amdocs.com>n1 Subject: RE: Intel Special issue of OpenVMS TimesaL Message-ID: <3B55D7F383B0D31197D9009027541CBF1A1A382B@cmiexch1.cmi.itds.com>   > -----Original Message-----H > From: susan_skonetski@hotmail.com [mailto:susan_skonetski@hotmail.com]  6 > Please forgive the spelling mistakes I am on a role.   Roll? :)  B > And in case you thought you would be rid of me in the newsgroup.   > NO WAY ;')  F Well, Sue, you can add one more to the list of people who are glad to 
 hear that.  F > If you think I am taking it personal, you are right it is personal. ? > Some of the people I respect the most work with VMS either assD > engineers or customer engineers (no such thing as and ignorant VMSH > user), my daughters college education is being bet on VMS AND I do not > want  G "No such thing" is a strong phrase -- but I haven't met many.  Maybe a i# few really, really dense ones... ;)   ; > I am not a complete VMS bigot NSK is a great product too.-   .. and can I quote you on that?a   Regards,   Chrise    ! Christopher Smith, Perl Developero Amdocs - Champaign, IL   /usr/bin/perl -e '? print((~"\x95\xc4\xe3"^"Just Another Perl Hacker.")."\x08!\n");0 'o  m   ------------------------------   Date: 29 Jan 2002 21:16:53 CDT= From: wayne@tachysoft.xxx.120519.killspam.00c7 (Wayne Sewell)s1 Subject: Re: Intel Special issue of OpenVMS Timesy. Message-ID: <uktHwS8Uwfsk@tachxxsoftxxconsult>  o In article <857e9e41.0201291520.35f36d@posting.google.com>, susan_skonetski@hotmail.com (Sue Skonetski) writes:      > F > If you think I am taking it personal, you are right it is personal. ? > Some of the people I respect the most work with VMS either aswD > engineers or customer engineers (no such thing as and ignorant VMSH > user), my daughters college education is being bet on VMS AND I do not > want6 > Nuclear Power plants running anything but VMS or NSK/ > healthcare running on anything but VMS or NSKp5 > Stock exchanged running on anything but VMS and NSKnH > Anything that can hurt major amounts of people running on anything but
 > VMS and NSKu  L Haven't you heard?  There are *no* reliable computer systems anywhere in theJ world.  Billy says so, therefore it Must Be True.   Billy says he wants toN build a reliable computer some day, and he might even do it, ten millenia from	 now.  :-)      -- yO ===============================================================================pM Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachysoft.xxxt: http://www.tachysoft.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-)oO ===============================================================================a= Society Lady:  Are you familiar with the great wall of China? 5        Curly:  No, but I know a big fence in Chicago!g   ------------------------------  , Date: Wed, 30 Jan 2002 04:04:06 -0200 (BRST) From: valdemir-@uol.com.br Subject: Intrusions deletion4 Message-ID: <200201300604.EAA28190@wilde.uol.com.br>  
 Hello all:  :     Is there any lexical function to extract intrusion=B4s       record ?=20   @     I=B4d like made a command procedure to delete all intrusions  <     records in my system. (now i=B4m using VAX/VMS5.5-2, and  ?     the command del/intru * doesn=B4t work... only in OpenVMS).      Thanks in advance...   ------------------------------  % Date: Tue, 29 Jan 2002 14:31:12 -0800-C From: "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam>t9 Subject: Re: Logicals defined for all process of one userw0 Message-ID: <3C5722B0.708E1CEE@intel.com.nospam>   "John E. Malmberg" wrote:5   > Phillip Helbig wrote:s >  > Jouk Jansen wrote:i >lK > >>There are logicals defined in several classes (process,job,group systemyJ > >>etc.) But how do I define a logical for all jobs of one user(UIC)? Why > >>is there no LNM$USER_TABLE.r > >> > > , > > I have sometimes felt the need for this. > >gL > > This is easy enough to set up, i.e. create a table owned by the user andI > > define a logical to point to it etc.  One could also play around withr' > > the definition of LNM$FILE_DEV etc.t > >tE > > This has many practical uses.  For example, one can interactivelyeL > > redefine logicals (such as SYS$PRINT if the corresponding printer breaks0 > > down) which are used by a running batch job. > > C > > Another question is why such a table is not present by default.' >hD > My guess is that it goes against the grain of the idea that when aE > process exits, all of it's resources are cleaned up.  Having a USERoC > table, would require quotas, or privileges to prevent a user fromh> > accidently filling up what has to be a system wide resource. >aJ > And then there may also need to be some way of tracking and deleting theJ > entries in the table when a user is deleted.  You do not want a new user5 > to inherit some one else's user logical name table.  >SI > And finally the issue comes down to the usefullness of the feature V.S. I > the time to implement it, how much it would actually be used, etcetera.   D     In a past life (actually, _two_ past lives :-) we found the needE for a shared logical name table that users could write to in which we B stored things such as mailbox device names, "vectors" of username/C Decwindows-display-device/DCL-command-to-execute, etc.  The mailboxtD case allowed interactive users to issues commands and receive outputD from batch jobs; the "vectors" logic allowed us to submit batch jobsD on a remote cluster node which would (1) start a DECterm and (2) run@ a command _in_ that DECterm.  There were a variety of other uses# which I don't recall at the moment.n  A     In the first case, we granted everyone GRPPRV and defined theFA mailbox logical names in LNM$GROUP.  In the latter case, a sharedhA table was created which gave RW access to holders of a particular ? identifier.  I don't recall whether we bother to add that tableo> to LNM$FILE_DEV or whether it was always referenced explicitly' during translation...either would work.e  C     Perhaps what Jouk needs is to have such a special table created @ at boot time with an ACL that allow only his username RW access.F It would be a little "dicey" to make this work for a detached process.F Seems to me the table would have to be entered in LNM$SYSTEM_DIRECTORYD and translations from the table would need to be explicit (since theC detached process presumably doesn't run LOGINOUT and run the user's 9 LOGIN.COM...otherwise you'd just do this stuff there...).S       -Ken --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfieldt! F20 Automation VMS System Support  kenneth.h.fairfield#intel.coms   ------------------------------  % Date: Tue, 29 Jan 2002 15:16:27 -0500?; From: Robert DiRosario <robert.j.dirosario.1@gsfc.nasa.gov> & Subject: Re: mount problems on cluster- Message-ID: <3C57031B.654F27D2@gsfc.nasa.gov>C   "John E. Malmberg" wrote:d   > Robert DiRosario wrote:  >V5 > > I'm having problems mounting disks on my cluster.f > >o0 > > If I reboot a node and then do "show dev /m"7 > > some of my DSSI disks show up with "(remote mount)"d4 > > listed for the volume label.  If I try to access2 > > the disk I get errors.  I can access the disks > > just fine from other nodes.  > > / > > If I issue the mount commands on a terminals > > everything works fine. > > 5 > > If I power down everything and boot all the nodess( > > from scratch, everything works fine. > >t6 > > Why are the DSSI disks not being mounted correctly > > from sylogicals.com? >wD > Usually disk mounts are done at the SYSTARTUP_VMS.COM phase of theI > startup.  Use SYLOGICALS.COM for system wide logical names needed afterJ > the boot.q  I If I want to move stuff off of the system disk to other disks the disk(s) $ need to be mounted in sylogicals.com   >h >nH > When the system is being booted from a power up, it can take some timeH > for all of the MSCP devices to be discovered.  Until then they can not
 > be mounted.i  I It works fine when I power the cluster up.  I have problems when I rebootyG a node.  These are not disks that need to be MSCP served to the 4000's.eF They are DSSI disks that the 4000's have direct access to.  The 4000's5 need to server these disk to the rest of the cluster.   D Nodes A, B and C have direct DSSI access to disks D1 and D2 (and theG system disk).  Nodes A and B have direct DSSI access to disks D10, D11,nD D12 and D13.  Why should a node try to use "remote access" to a disk& when it has a direct path to the disk?  J The problem is NOT that the disks aren't ready.  The other cluster members# see the disk and have them mounted.c  I If I pick one of the nodes, A, B or C, that has all of the drives mounted D and do a "reboot", it won't have all of the drives mounted after the reboot.-   >- >-E > So basically at boot time, you will need a to have a loop that usessD > f$getdvi(devname,"exists") with a delay of about 5 seconds betweenJ > tries.  It should also only try for about 12 times, so that if something4 > is wrong with the disk, the startup will not hang. > J > Using the /CLUSTER on a mount command in the SYSTEM BOOT sequence is notJ > recommended.  You do not know if the other nodes in the cluster are able > to service the request.l >oJ > If you are serving disks back from a diskless satellite, then use of theF > /CLUSTER will usually work, because typically the other hosts in theI > network are already up.  You may still have to do some manual mounts iflG > you have multiple nodes booting at the same time, but usually it willhJ > come out right.  About the only way to have it always work would involveG > setting up a detached process or a batch job that makes sure that all, > the disks got mounted. >t! > See SYS$EXAMPLES:MSCPMOUNT.COM.  >lJ > Also see http://encompasserve.org/~malmberg/dcl/ for MOUNT_DISKS.TXT and > MOUNT_DISKS_DRAGON.TXT.y >nE > Mount_disks.txt looks up all the systems currently in a cluster andaM > executes the appropriate STARTUP_DIR:MOUNT_DISKS_*.COM to get them mounted.p >m> > MOUNT_DISKS_DRAGON.TXT mounts the disks for a specific node. >a4 > > Node m48 is on ethernet and works fine, but it's8 > > turned off for now.  It has it's page and swap files7 > > on a local SCSI disk, and when nodes A, B or C boot-8 > > they run sypagswpfiles.com and try to use m48's page6 > > and swap file.  (A quick search of google turns up8 > > satellite_page.com as a solution for old versions of8 > > VMS, but I'm using 7.1.  Is the node name defined by% > > the time sypagswpfiles.com runs?)n >hJ > I would recommend using f$edit(f$getsyi("SCSNODE"),"UPCASE,TRIM") to getF > the system name.  The SCSNODE name is available to all user editable > command files in the boot. >iJ > Satellite_page.com is generated by the CLUSTER_CONFIG.COM procedure whenH > you configure a satellite system.  It is used to mount the remote disk) > and install a page and swap file on it.  >tF > In this case you would not use the SYPAGSWPFILES.COM unless you were4 > adding additional page or swap files for the node.  L I just want to move the page and swap files off of the cluster DSSI disk andL onto local SCSI disks, to free up space on the cluster disk and speed thingsK up.  (I'm assuming my RZ25 and RZ35 drives are faster then the RF72 drive.)n   >a >eG > Note that SYPAGSWPFILES.COM is executed even for minimum startups, sooG > when I use one, I put some logic in it to prevent it from running, asiF > when I am booting minimum, I may need to do work on the disk(s) that2 > have the additional page and swap files on them. >-) > See SYPAGSWPFILES.TXT at the above URL.n >uJ > Because the disk for the secondary page and swap files may not be ready,I > The SYPAGESWPFILES procedure must also poll and wait for the disk to bet  > configured and placed on line. >wM > To prevent code duplication, I have the SYPAGSWAPFILES_<node>.COM file calldO > the appropriate STARTUP_DIR:MOUNT_DISKS_<node>.COM with the parameter "PAGE".c >eF > To help understand the command files that I posted on Encompasserve,H > please note that in general, I use a naming convention for files on my > systems as follows:1 >B@ > No suffix: The file can be placed on any system that I manage. >s; > A suffix of "_<node>": The file is specific to that node.  > E > A suffix of "_<cluster_alias>_COMMON.COM:  This file is common to a . > group of nodes that share the cluster_alias. >rF > This allows me to keep all of the files in the SYS$COMMON: directory$ > tree, and in a common CMS library. >o > -John? > wb8tyw@qsl.network >G > Personal Opinion Only    Thanks Robert   ------------------------------  % Date: Tue, 29 Jan 2002 23:17:25 -0000a= From: "David McKenzie" <david.mckenzie@spitfire0.demon.co.uk>b& Subject: Re: mount problems on clusterB Message-ID: <1012346275.13882.0.nnrp-12.c1edba74@news.demon.co.uk>  H I have seen something similar a long time ago with a DSSI cluster. ThereK used to be a problem with DSSI on some of the machines that the served path K responded quicker than the direct path, and hence go selected. I just can't  remember how that got fixed.  E For the other point there are lots of things in systartup_vms.com and/L sylogicals in terms of layout that are misleading, or in my opinion outrightD wrong. My pet peeve for instance is that the queue manager and hence* autostart queues get started near the top.  G You can lose the whole days batch jobs down the toilet as the run, fing K layered products not started and bomb out, leaving lots more slots for event more batch jobs to run.   @ I guess I am trying to say, that I don't think that the templateJ systartup_vms and sylogicals is some sort of relegious document written by gurus never to be touched.  H "Robert DiRosario" <robert.j.dirosario.1@gsfc.nasa.gov> wrote in message' news:3C57031B.654F27D2@gsfc.nasa.gov...g >D >8 > "John E. Malmberg" wrote:t >. > > Robert DiRosario wrote:y > > 7 > > > I'm having problems mounting disks on my cluster.h > > > 2 > > > If I reboot a node and then do "show dev /m"9 > > > some of my DSSI disks show up with "(remote mount)"d6 > > > listed for the volume label.  If I try to access4 > > > the disk I get errors.  I can access the disks! > > > just fine from other nodes.t > > >n1 > > > If I issue the mount commands on a terminale > > > everything works fine. > > >t7 > > > If I power down everything and boot all the nodesb* > > > from scratch, everything works fine. > > >i8 > > > Why are the DSSI disks not being mounted correctly > > > from sylogicals.com? > >nF > > Usually disk mounts are done at the SYSTARTUP_VMS.COM phase of theK > > startup.  Use SYLOGICALS.COM for system wide logical names needed afterk
 > > the boot.r >wK > If I want to move stuff off of the system disk to other disks the disk(s)m& > need to be mounted in sylogicals.com >o > >  > > J > > When the system is being booted from a power up, it can take some timeJ > > for all of the MSCP devices to be discovered.  Until then they can not > > be mounted.  >tK > It works fine when I power the cluster up.  I have problems when I rebootSI > a node.  These are not disks that need to be MSCP served to the 4000's. H > They are DSSI disks that the 4000's have direct access to.  The 4000's7 > need to server these disk to the rest of the cluster.. > F > Nodes A, B and C have direct DSSI access to disks D1 and D2 (and theI > system disk).  Nodes A and B have direct DSSI access to disks D10, D11,.F > D12 and D13.  Why should a node try to use "remote access" to a disk( > when it has a direct path to the disk? >aL > The problem is NOT that the disks aren't ready.  The other cluster members% > see the disk and have them mounted.i >iK > If I pick one of the nodes, A, B or C, that has all of the drives mountedvF > and do a "reboot", it won't have all of the drives mounted after the	 > reboot.h >d > >f > >oG > > So basically at boot time, you will need a to have a loop that useseF > > f$getdvi(devname,"exists") with a delay of about 5 seconds betweenL > > tries.  It should also only try for about 12 times, so that if something6 > > is wrong with the disk, the startup will not hang. > >aL > > Using the /CLUSTER on a mount command in the SYSTEM BOOT sequence is notL > > recommended.  You do not know if the other nodes in the cluster are able > > to service the request.: > >IL > > If you are serving disks back from a diskless satellite, then use of theH > > /CLUSTER will usually work, because typically the other hosts in theK > > network are already up.  You may still have to do some manual mounts ifaI > > you have multiple nodes booting at the same time, but usually it willsL > > come out right.  About the only way to have it always work would involveI > > setting up a detached process or a batch job that makes sure that alla > > the disks got mounted. > >i# > > See SYS$EXAMPLES:MSCPMOUNT.COM.n > > L > > Also see http://encompasserve.org/~malmberg/dcl/ for MOUNT_DISKS.TXT and > > MOUNT_DISKS_DRAGON.TXT.h > >gG > > Mount_disks.txt looks up all the systems currently in a cluster andtF > > executes the appropriate STARTUP_DIR:MOUNT_DISKS_*.COM to get them mounted. > > @ > > MOUNT_DISKS_DRAGON.TXT mounts the disks for a specific node. > >c6 > > > Node m48 is on ethernet and works fine, but it's: > > > turned off for now.  It has it's page and swap files9 > > > on a local SCSI disk, and when nodes A, B or C bootr: > > > they run sypagswpfiles.com and try to use m48's page8 > > > and swap file.  (A quick search of google turns up: > > > satellite_page.com as a solution for old versions of: > > > VMS, but I'm using 7.1.  Is the node name defined by' > > > the time sypagswpfiles.com runs?)g > > L > > I would recommend using f$edit(f$getsyi("SCSNODE"),"UPCASE,TRIM") to getH > > the system name.  The SCSNODE name is available to all user editable > > command files in the boot. > > L > > Satellite_page.com is generated by the CLUSTER_CONFIG.COM procedure whenJ > > you configure a satellite system.  It is used to mount the remote disk+ > > and install a page and swap file on it.  > >gH > > In this case you would not use the SYPAGSWPFILES.COM unless you were6 > > adding additional page or swap files for the node. > J > I just want to move the page and swap files off of the cluster DSSI disk and G > onto local SCSI disks, to free up space on the cluster disk and speed  thingsE > up.  (I'm assuming my RZ25 and RZ35 drives are faster then the RF72a drive.)n >e > >l > >iI > > Note that SYPAGSWPFILES.COM is executed even for minimum startups, soeI > > when I use one, I put some logic in it to prevent it from running, as H > > when I am booting minimum, I may need to do work on the disk(s) that4 > > have the additional page and swap files on them. > >u+ > > See SYPAGSWPFILES.TXT at the above URL.t > >wL > > Because the disk for the secondary page and swap files may not be ready,K > > The SYPAGESWPFILES procedure must also poll and wait for the disk to be=" > > configured and placed on line. > >=J > > To prevent code duplication, I have the SYPAGSWAPFILES_<node>.COM file callI > > the appropriate STARTUP_DIR:MOUNT_DISKS_<node>.COM with the parameter- "PAGE".  > >eH > > To help understand the command files that I posted on Encompasserve,J > > please note that in general, I use a naming convention for files on my > > systems as follows:s > >eB > > No suffix: The file can be placed on any system that I manage. > > = > > A suffix of "_<node>": The file is specific to that node.  > >eG > > A suffix of "_<cluster_alias>_COMMON.COM:  This file is common to a 0 > > group of nodes that share the cluster_alias. > >hH > > This allows me to keep all of the files in the SYS$COMMON: directory& > > tree, and in a common CMS library. > >i	 > > -Johnt > > wb8tyw@qsl.network > >m > > Personal Opinion Onlye >e > Thanks > Robert >e   ------------------------------  # Date: Tue, 29 Jan 2002 23:21:31 GMTb) From: rob.buxton@wcc.govt.nz (Rob Buxton) & Subject: Re: mount problems on cluster2 Message-ID: <3c572c27.2592600612@news.wcc.govt.nz>  : On Mon, 28 Jan 2002 17:15:39 -0800, "Kenneth H. Fairfield"- <Kenneth.H.Fairfield@intel.com.nospam> wrote:$   >Rob Buxton wrote: >  >> I'll expand. further... >>G >> The procedure I use is similar, but probably a bit more complex thane >> Robs.I >> The common disk containg UAF etc. is a shadow set, it's served by HSDs   >> connected to a pair of VAXes.I >> The VAXes are the voters of the cluster along with a quorum disk (also  >> HSD served)I >> When the Alphas boot the HSD disks are available - (the Alphas are tooe >> young to vote.) >>H >> Because it's a shadow set I need to check that each member Exists andH >> is Available before mounting. The com file then mounts either both or) >> the member it finds as the shadow set.c >  >[...]       >s@ >    Forgive me if I'm belaboring the obvious, but when you say,E >"[the procedure] then mounts either both or _the member it finds_ as F >the shadow set", that last is not sufficient.  To mount a shadow set,D >_all_ members _must_ be visible to _all_ hosts which mount the set.B >If you have a 2-member shadow set, and only one member is visible? >during boot, the mount will (or _should_) fail.  Have I misseds >something?n > 	 >    -Kenn  C I've tried to cater for the eventuality (pretty remote here) that I D might have actaully removed one of the disks that normally comprises0 the Shadow Set from the HSD - e.g. disk failure.7 Other Cluster Members would see the reduced Shadow Set.r  B So, if the Shadow Set only had one Member at the time of the AlphaD Booting I want it to be able to successfully mount the Single MemberF Shadow Set and not trip up trying to mount a disk that's not currently available to anything.  # So, in Summary my com file is like.nE If both devices exist mount the 2 disk shadow set - normal operation.oE If only one devices exists, mount a single disk shadow set using thato device.l       >--e7 >I don't speak for Intel, Intel doesn't speak for me...  >  >Ken Fairfield" >F20 Automation VMS System Support >kenneth.h.fairfield#intel.com >  >    ------------------------------  % Date: Tue, 29 Jan 2002 23:26:45 -0000n= From: "David McKenzie" <david.mckenzie@spitfire0.demon.co.uk> & Subject: Re: mount problems on clusterB Message-ID: <1012346835.14150.0.nnrp-12.c1edba74@news.demon.co.uk>  K Note that in a wide area cluster that may be a very bad strategy. DependingtJ on how the cluster fell apart due to link loss you could get both sides ofL the cluster believing ( and worse writing to ) what they see as the remainig valid disk in the shadow set.n  6 "Rob Buxton" <rob.buxton@wcc.govt.nz> wrote in message, news:3c572c27.2592600612@news.wcc.govt.nz...< > On Mon, 28 Jan 2002 17:15:39 -0800, "Kenneth H. Fairfield"/ > <Kenneth.H.Fairfield@intel.com.nospam> wrote:  >' > >Rob Buxton wrote: > >e > >> I'll expand. further... > >>I > >> The procedure I use is similar, but probably a bit more complex than-
 > >> Robs.K > >> The common disk containg UAF etc. is a shadow set, it's served by HSDss" > >> connected to a pair of VAXes.K > >> The VAXes are the voters of the cluster along with a quorum disk (also2 > >> HSD served)K > >> When the Alphas boot the HSD disks are available - (the Alphas are toog > >> young to vote.) > >>J > >> Because it's a shadow set I need to check that each member Exists andJ > >> is Available before mounting. The com file then mounts either both or+ > >> the member it finds as the shadow set.g > >r > >[...] >c >s >e > >sB > >    Forgive me if I'm belaboring the obvious, but when you say,G > >"[the procedure] then mounts either both or _the member it finds_ asdH > >the shadow set", that last is not sufficient.  To mount a shadow set,F > >_all_ members _must_ be visible to _all_ hosts which mount the set.D > >If you have a 2-member shadow set, and only one member is visibleA > >during boot, the mount will (or _should_) fail.  Have I missede
 > >something?n > >n > >    -Ken  >hE > I've tried to cater for the eventuality (pretty remote here) that IkF > might have actaully removed one of the disks that normally comprises2 > the Shadow Set from the HSD - e.g. disk failure.9 > Other Cluster Members would see the reduced Shadow Set.d >UD > So, if the Shadow Set only had one Member at the time of the AlphaF > Booting I want it to be able to successfully mount the Single MemberH > Shadow Set and not trip up trying to mount a disk that's not currently > available to anything. >s% > So, in Summary my com file is like.dG > If both devices exist mount the 2 disk shadow set - normal operation.eG > If only one devices exists, mount a single disk shadow set using thatD	 > device.t >i >  >, > >-- 9 > >I don't speak for Intel, Intel doesn't speak for me...t > >i > >Ken Fairfield$ > >F20 Automation VMS System Support  > >kenneth.h.fairfield#intel.com > >T > >h >c   ------------------------------  # Date: Wed, 30 Jan 2002 02:51:31 GMTt- From: "John E. Malmberg" <wb8tyw@qsl.network> & Subject: Re: mount problems on cluster* Message-ID: <3C5763D0.1000104@qsl.network>   Robert DiRosario wrote:l   >  > "John E. Malmberg" wrote:   eK > If I want to move stuff off of the system disk to other disks the disk(s)n& > need to be mounted in sylogicals.com    I True, but since mounting a disk can be time consuming, I would only mountt3 the disks that are immediately needed at that time.o    D My personal opinion is that if I needed to do that, I would call my E MOUNT_DISK_<node>.COM procedure with a parameter to indicate that it e- should mount the disks needed for common use.a  " Just like I do for the page files.    U >>H >>When the system is being booted from a power up, it can take some timeH >>for all of the MSCP devices to be discovered.  Until then they can not
 >>be mounted.  >>K > It works fine when I power the cluster up.  I have problems when I reboot I > a node.  These are not disks that need to be MSCP served to the 4000's.rH > They are DSSI disks that the 4000's have direct access to.  The 4000's7 > need to server these disk to the rest of the cluster.o > F > Nodes A, B and C have direct DSSI access to disks D1 and D2 (and theI > system disk).  Nodes A and B have direct DSSI access to disks D10, D11, F > D12 and D13.  Why should a node try to use "remote access" to a disk( > when it has a direct path to the disk?    F The software layers that can serve the disks through MSCP are started H before the local paths for the disk are configured.  A local connection I requires discovering the local path, and actually communicating with the nC disk.  That takes more time than discovering that there is an MSCP q served path to the disk.  C As other posters have pointed out, I missed the CLU_MOUNT_DISK.COM.,  H I also have somewhere a procedure to mount shadowed disks in a cluster, 6 but I am not sure where it is at after my recent move.    S >>J >>Satellite_page.com is generated by the CLUSTER_CONFIG.COM procedure whenH >>you configure a satellite system.  It is used to mount the remote disk) >>and install a page and swap file on it.  >>F >>In this case you would not use the SYPAGSWPFILES.COM unless you were4 >>adding additional page or swap files for the node. >> > N > I just want to move the page and swap files off of the cluster DSSI disk andN > onto local SCSI disks, to free up space on the cluster disk and speed thingsM > up.  (I'm assuming my RZ25 and RZ35 drives are faster then the RF72 drive.)  >     E Use the CLUSTER_CONFIG.COM to generate a phantom cluster member, and  I then use the generated SATELLITE_PAGE.COM as template to create your own.E  G Or you can just roll your own.  It just needs to wait for a disk to be oB configured, mount it, and then install a pagefile.  Just like the ; SYPAGSWPFILES.COM.  It just runs a bit earlier in the boot.e  H As far as your existing page and swap files, after you get the new ones A working, just rename them, and delete them after the next reboot.p   -Johny wb8tyw@qsl.network Personal Opinion Onlyu   ------------------------------  % Date: Tue, 29 Jan 2002 14:52:12 -0500l% From: JF Mezei <jfmezei@videotron.ca>,- Subject: Moving from CMU-IP to TCPIP servicesd, Message-ID: <3C56FD42.7303EF1F@videotron.ca>  M One of my Vaxes has CMU-IP tcpip stack (mostly because I needed PPP which wast never ported to VAX-VMS/TCPIP.  J Now that need has gone away, and I would like to install TCPIP Services on: that node. (long term goal is to have homogenous cluster).  ? Has anyone made such a move, and if so, are there any gotchas ?   N I know that when I upgraded to 7.2 on that node, it zapped the CMUIP telnet toM put in the TCPIP service's CLD (even though TCPIP services wasn't installed).s  M If I remember correctly, the CMU IP manual does have a list of files that areaK installed, so I guess I would only need to delete those files individually,i right ?    ------------------------------  # Date: Wed, 30 Jan 2002 03:10:32 GMT0- From: "John E. Malmberg" <wb8tyw@qsl.network>i1 Subject: Re: Moving from CMU-IP to TCPIP servicesu* Message-ID: <3C576847.9060106@qsl.network>   JF Mezei wrote:.  O > One of my Vaxes has CMU-IP tcpip stack (mostly because I needed PPP which wasD  > never ported to VAX-VMS/TCPIP. > L > Now that need has gone away, and I would like to install TCPIP Services on< > that node. (long term goal is to have homogenous cluster).    O > If I remember correctly, the CMU IP manual does have a list of files that areuM > installed, so I guess I would only need to delete those files individually,n	 > right ?t >   F If you installed CMUIP according to the defaults that the most recent . version wanted to use, it created a directory ; SYS$SYSDEVICE:[CMUIP_ROOT.] with a logical name CMUIP_ROOT:n  F All files of CMUIP are in those roots.  Simply remove CMUIP from your 1 SYSTARTUP_*.COM file, and install TCPIP services.y  4 After a reboot, then you can delete the CMUIP files.  F If you overlayed CMUIP_ROOT: with SYS$SYSROOT:, then you will need to C hunt up the documentation and find and delete the individual files.f   -Johnv wb8tyw@qsl.network Personal Opinion Onlyg   ------------------------------  # Date: Tue, 29 Jan 2002 19:42:33 GMTR2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)C Subject: Re: Partitioned System Disk (was: Re: There must be a way) 1 Message-ID: <JWC58.146$am1.4211@news.cpqcorp.net>u  M In article <3C56EA88.1032E346@lbl.gov>, Jerry Tunis <mjtunis@lbl.gov> writes:      Welcome to comp.os.vms!   F :I have 3 old 380mb scsi drives (DUA0-2) on a small MicroVax cluster.   F   Specific details of the MicroVAX systems involved?  Some have limits1   in the console.  Some have controller limits.     6   Please see the OpenVMS FAQ for some related details.  H :I want to retire these drives and replace them with a new big drive ...  %   Details of the "big drive", please?k  J :by creating 3 partitions on it. Can one partition a VMS drive in a system! :running VAX/VMS version V5.5-2? n  G   Upgrade, please.  OpenVMS VAX V5.5-2 has definite limits on the totaleF   disk capacity supported.  (Again, the FAQ has some related details.)  H   Please also consider an upgrade to a more recent VAX or Alpha system, E   as the references to "DU" disks on MicroVAX systems tends to imply rC   old MicroVAX systems and/or old controllers and old disk storage.h   :There must be some way.  H   Well, no.  OpenVMS does not support partitioned system disks.  Now if J   the disk controller presents the appearance of a non-partitioned disk...  G   Please also take the time to review the OpenVMS FAQ introduction for eI   details on how to ask questions, how to provide necessary details, and DH   how to pick a title that is relevent to the question.  (No offense is G   intended here, but your choice of a title really looked like spam...)"     And again, welcome!   N  ---------------------------- #include <rtfaq.h> -----------------------------N       For additional, please see the OpenVMS FAQ -- www.openvms.compaq.com    N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------    Date: 29 Jan 2002 20:29:17 -0800$ From: nomachin69@hotmail.com (denis) Subject: pathworka= Message-ID: <ff3b44b9.0201292029.685c8604@posting.google.com>r   Bonjour,  ?    Nous utilisons de puis plusieurs annes Pathwork 5.0 pour larF cration de rpertoire partagable.  Nous pouvons nous connect de DOS,B Windows 95, Windows 98 , NT et ME, mais avec la version de WindowsA 2000 Pro ou Server il nous tes impossible de se connect il noustD donne toujours un message qu'il ne reconnait pas le nom.  Nous avonsF install PowerTerm et il me permet de me connecter sur mon serveur viaC LAT ou TCPIP sans aucun problme.  J'aimerais savoir si n'a pas uneo* possibilit de rgl le problme.   Merci.   ------------------------------  # Date: Wed, 30 Jan 2002 07:01:22 GMTr- From: michael.greenNOSP@Mvideotron.ca (MikeG)k Subject: Re: pathworkt1 Message-ID: <3c5797e1.19172650@news.videotron.ca>    Denis,C il existe un "patch" pour le probleme que vous dcrivent.  ContacteTD votre centre de support Compaq pour le recevoir.  Aussi, a le momentE ou  vous entrer le nom d'utilisateur et mot de passe,  attentions quek) le nom d'utilisateur soit en majascule.      Bone chance, Mike G t  ) PS Excuse mon Franais tres mal crite.  C        D On 29 Jan 2002 20:29:17 -0800, nomachin69@hotmail.com (denis) wrote:  	 >Bonjour,n >.@ >   Nous utilisons de puis plusieurs annes Pathwork 5.0 pour laG >cration de rpertoire partagable.  Nous pouvons nous connect de DOS,lC >Windows 95, Windows 98 , NT et ME, mais avec la version de WindowslB >2000 Pro ou Server il nous tes impossible de se connect il nousE >donne toujours un message qu'il ne reconnait pas le nom.  Nous avonsvG >install PowerTerm et il me permet de me connecter sur mon serveur via D >LAT ou TCPIP sans aucun problme.  J'aimerais savoir si n'a pas une+ >possibilit de rgl le problme.   Merci.    ------------------------------  # Date: Wed, 30 Jan 2002 05:12:12 GMTd& From: "C.W.Holeman II" <cwhii@ACM.org>) Subject: Re: pathwork (French -> English)y' Message-ID: <3C5780AA.6BA9EC7B@ACM.org>    denis wrote: > 
 > Bonjour, > A >    Nous utilisons de puis plusieurs annes Pathwork 5.0 pour laaH > cration de rpertoire partagable.  Nous pouvons nous connect de DOS,D > Windows 95, Windows 98 , NT et ME, mais avec la version de WindowsC > 2000 Pro ou Server il nous tes impossible de se connect il nouscF > donne toujours un message qu'il ne reconnait pas le nom.  Nous avonsH > install PowerTerm et il me permet de me connecter sur mon serveur viaE > LAT ou TCPIP sans aucun problme.  J'aimerais savoir si n'a pas une , > possibilit de rgl le problme.   Merci.  
 Altavista:  N Hello, We use of then several years Pathwork 5.0 for the creation of repertoryM partagable. We can connected DOS, Windows 95, Windows 98, NT and ME, but withtM the version of Windows 2000 Pro or Server it impossible be connected to us ittK always gives us a message which it did not reconnait the name. We installedlI PowerTerm and it enables me to be connected on my server via LAT or TCPIPaO without any problem. I would like to know if does not have a possibility of notc" regulated the problem. Thank you.    -- < C.W.Holeman II  cwhii@acm.org				http://emle.org- http://also.as/cwhii		http://JulianLocals.com    ------------------------------  # Date: Wed, 30 Jan 2002 03:05:32 GMT - From: "John E. Malmberg" <wb8tyw@qsl.network>,I Subject: Re: Questions about setting up an Advanced Server  Member Server * Message-ID: <3C576718.5020704@qsl.network>   Ken Robinson wrote:   M > I'm finally able to set up a Member Server on a test network before moving   > it to out Production network.N > G > The test network consists of a NT 4.0 PDC, a few other NT boxes, and lM > a DS10 running VMS 7.3 and Advanced Server 7.3 as a Member Server. On this cF > test network I have access to the NT PDC and the Admin username and ? > password. On our production network I won't have that access.     G You may want to instead make your OpenVMS Advanced Server systems into 3H their own domain, and then use a trust to the NT domain.  Pathworks 6.x  can do this also.e  G This will allow you to locally administer it, but allow you to use the 1. security policies and groups of the NT domain.  F Of course this means that you will have to reapply all of the securityI descriptors, but it will give you some insulation from some of the fun ofo managing a NT domain.b    B > I have a few concerns and questions as the result of a few days  > testing... > I > 1) It seems I need access to the PDC Admin username and password to be tG > able to add a local group to a local share. For the following, I was iH > logged into the member server. I created a local group named ANSUSER. G > When I tried to create a local share, also called ANSUSER, using the v@ > parameters "/noperm/perm=(ansuser=full)", I got an error back E > saying "no such user or group". I then created the share using all FE > defaults and it worked fine. When I went to the PDC, I was able to eA > modify the share to have the above permissions with no problem.a    G All Domain Controllers in a security domain share the local groups, so sD you must be a domain administrator to control them.  The way to get D around that is to establish a trust relationship between to domains.    nI > 2) I occassionally get prompted for a server password (I trying to pin  I > down when this happens). No passwords that I know work at this prompt,   > including hitting Return.t    $ I do not know about that one either.    F > I am trying to configure the member server to look like a currently H > running PW 5.0f system that is currently running on some old hardware 8 > that's on its last legs or as close as I can get it...    I Before I changed jobs, I had to move a PW 5.0 system up to Pathworks 6.0.s  2 I set it up as a resource domain of the NT Domain.  F In either case, if you have a domain adminstrative account for the NT F domain, you can use DCL to manage your NT domain through the Advanced  Server Admin program.g   -Johne wb8tyw@qsl.network Personal Opinion Onlyl   ------------------------------  # Date: Wed, 30 Jan 2002 06:09:40 GMT ) From: Charles Richmond <richmond@ev1.net>c! Subject: Re: The demise of compaq-' Message-ID: <3C57AA5D.AB2A275A@ev1.net>4   Charlie Gibbs wrote: > + >     [snip...]     [snip...]     [snip...]  > G > This sort of thinking has declined somewhat.  Nowadays many companies1G > would rather get rid of their veteran staff and replace them with new D > recruits fresh out of school.  The bean counters like this becauseD > the new hires start at the bottom of the pay scale.  The PHBs likeG > it because the newbies have bright, shiny new minds which are largelywG > blank, ready to be filled with the latest mission statements.  It's a5J > lot easier to sell an idea if you don't have to contend with spoilsports; > who can draw on experience to tell you why it won't work.a > H >From what I have read in the newspaper, the PHB's in the U.S. do *not* D need the recruits fresh out of school. The PHB's can hire H1B's for 1 *half* what they could get a U.S. graduate for...A   -- i? +-------------------------------------------------------------+ ? |     Charles and Francis Richmond     <richmond@plano.net>   | ? +-------------------------------------------------------------+h   ------------------------------   Date: 30 Jan 2002 06:32:16 GMT) From: leslie@clio.rice.edu (Jerry Leslie)n! Subject: Re: The demise of compaqs' Message-ID: <a3841g$27h$4@joe.rice.edu>T  * Charles Richmond (richmond@ev1.net) wrote:I : From what I have read in the newspaper, the PHB's in the U.S. do *not*  F : need the recruits fresh out of school. The PHB's can hire H1B's for 3 : *half* what they could get a U.S. graduate for...i :iF The H-1B laws are suppose to require that H-1Bs be paid the prevailingG wage. But the INS only has 40 staffers to investigate fraud in H-1B and  other visa applications:  4      http://www.siliconvalley.com/news/special/visa/(      SiliconValley.com - Special Reports  F     "...H-1B visa fraud is not uncommon, ranging from academic degreesA      faked overseas to phony job offers in the United States. ThehG      Immigration and Naturalization Service has only 40 staffers at itshD      service centers nationwide to investigate fraud in H-1B and all       other visa applications..."  1 The GAO has published a report on needed changes:e  ,     http://www.gao.gov/new.items/he00157.pdfB    "H-1B Foreign Workers: Better Controls Needed to Help Employers:     and Protect Workers", GAO/HEHS-00-157, September 2000.  ? But offshore outsourcing is even cheaper than the H-1B program:t  E    http://www.computerworld.com/storyba/0,4125,NAV47_STO58739,00.htmllE    Ford opens IT hub in India to save millions | Computerworld News &i    Feature Stories  4    http://news.cnet.com/news/0-1006-200-8324464.html>    India becoming world's back office -  Tech News -  CNET.com  	    [snip]n  E   "They are part of an emerging work force for India's latest export:-    IT-enabled services.0  H    These include telemarketing, helpdesk support, medical transcription,@    back-office accounting, payroll management, maintaining legalH    databases, insurance claim and credit-card processing, animation, andI    higher-end engineering design--all of which can be delivered by phone,s    computer and the Internet.o  	    [snip]   G    Call center services can even extend to emotional help, as BangaloretG    IT-firm Phoenix Global Solutions plans to do. It has hired 50 peoplep8    for a pilot project to counsel troubled Americans..."  > So out-of-work VMSers can find a sympathetic voice to talk to.  D Such offshore ventures are insured against political risk by a small agency of the US government:      http://www.opic.gov/o5    The Overseas Private Investment Corporation (OPIC)f  ' OPIC may have to pay claims from Enron:t  4    http://www.enrongate.com/news/index.asp?id=118898@    EnronGate: News Archives: Enron project was raised with India    4 --Jerry Leslie     (my opinions are strictly my own)   ------------------------------  % Date: Tue, 29 Jan 2002 22:39:16 -0000h- From: wspencer@ap.nospam.org (Warren Spencer)s" Subject: Re: The value of a leader7 Message-ID: <91A5B8212warrenspencer1977@209.249.90.100>e  ( jfmezei@videotron.ca (JF Mezei) wrote in" <3C56E9BD.DE78AD98@videotron.ca>:   F >IBM announced that Gerstner was to retire next year and that they are >looking for a new leader. >IBM stock down sigificantly.r >hH >I think this shows how important it is for a company the size of Compaq# >to have a real leader at the helm.c  : Hmmm, I wonder if Gerstner is looking for a new challenge?   ws   --     Warren Spencer' Senior Software Engineer (not a writer)a The Associated Press  @ ** When Windows is your hammer, everything looks like a thumb **   ------------------------------  % Date: Tue, 29 Jan 2002 15:03:07 -0800s# From: "Tom Linden" <tom@kednos.com>m" Subject: RE: The value of a leader9 Message-ID: <CIEJLCMNHNNDLLOOGNJIMEKHEAAA.tom@kednos.com>t   > -----Original Message-----6 > From: Warren Spencer [mailto:wspencer@ap.nospam.org]) > Sent: Tuesday, January 29, 2002 2:39 PMe > To: Info-VAX@Mvb.Saic.Com>$ > Subject: Re: The value of a leader >  > * > jfmezei@videotron.ca (JF Mezei) wrote in$ > <3C56E9BD.DE78AD98@videotron.ca>:  > H > >IBM announced that Gerstner was to retire next year and that they are > >looking for a new leader. > >IBM stock down sigificantly.d   > > J > >I think this shows how important it is for a company the size of Compaq% > >to have a real leader at the helm.E > < > Hmmm, I wonder if Gerstner is looking for a new challenge?  9 If you have been a General, Sargeant is hardly challenge.o   >  > ws >  > --   >  > Warren Spencer) > Senior Software Engineer (not a writer)r > The Associated Press > B > ** When Windows is your hammer, everything looks like a thumb ** >    ------------------------------  % Date: Wed, 30 Jan 2002 03:12:46 +0100y' From: Brass Christof <welcome@spam.net>e" Subject: Re: The value of a leader( Message-ID: <3C57569E.26297DAD@spam.net>   Tom Linden wrote:u  L > > >I think this shows how important it is for a company the size of Compaq' > > >to have a real leader at the helm.j > >M> > > Hmmm, I wonder if Gerstner is looking for a new challenge? > ; > If you have been a General, Sargeant is hardly challenge.t  = Compare the position of a General in a country with a lot of .9 rules and politicians telling him what to do, to buy, to l< develop and how the strategy should be to the position in a 8 smaller country which has money, a lot of new ideas and ) offers a position to build up a new army.h  < Couldn't that be a real challenge no matter how the title is called?t   > > ws > >e > > -- > >t > > Warren Spencer+ > > Senior Software Engineer (not a writer)l > > The Associated Press > >aD > > ** When Windows is your hammer, everything looks like a thumb ** > >    -- i6 moc dot slupofni at ssarb - please revert the sequence   ------------------------------  % Date: Tue, 29 Jan 2002 18:47:38 -0800b# From: "Tom Linden" <tom@kednos.com>n" Subject: RE: The value of a leader9 Message-ID: <CIEJLCMNHNNDLLOOGNJIAEKPEAAA.tom@kednos.com>.   > -----Original Message-----0 > From: Brass Christof [mailto:welcome@spam.net]) > Sent: Tuesday, January 29, 2002 6:13 PM  > To: Info-VAX@Mvb.Saic.Coma$ > Subject: Re: The value of a leader >  >  > Tom Linden wrote:r > @ > > > >I think this shows how important it is for a company the  > size of Compaq) > > > >to have a real leader at the helm.t > > >s@ > > > Hmmm, I wonder if Gerstner is looking for a new challenge? > > = > > If you have been a General, Sargeant is hardly challenge.a > ? > Compare the position of a General in a country with a lot of Y; > rules and politicians telling him what to do, to buy, to  > > develop and how the strategy should be to the position in a : > smaller country which has money, a lot of new ideas and + > offers a position to build up a new army.I > > > Couldn't that be a real challenge no matter how the title is	 > called?t > C  No.  Compaq would have a difficult time attracting that caliber of @ talent.  You forget that even though IBMs market cap is 8 times B what it was when Gerstner took over 8-9 years ago it was still the? leading company in its sector.  I think its best to stop there.i   > > > ws > > >a > > > -- > > >  > > > Warren Spencer- > > > Senior Software Engineer (not a writer)a > > > The Associated Press > > > F > > > ** When Windows is your hammer, everything looks like a thumb ** > > >  >  > --  8 > moc dot slupofni at ssarb - please revert the sequence >    ------------------------------  % Date: Tue, 29 Jan 2002 22:08:05 +0300n4 From: "Ruslan R. Laishev" <laishev@smtp.deltatel.ru>1 Subject: Re: Using PortMapper to run an IP Servern0 Message-ID: <3C56F315.D6B49EE8@smtp.deltatel.ru>  $ dir ucx$examples or tcpip$examples ?   "Michael D. Ober" wrote: > N > Does anyone have a simple sample program that uses the VMS TCP/IP PortmapperH > system to listen to a port and then spawn an external program with the/ > client connection.  I'm writing in VMS Basic.. >  > --	 > Thanks,  > Mike Ober.   -- ) Cheers, Ruslan. D +---------------------pure personal opinion------------------------+;       RADIUS Server for OpenVMS project - www.radiusvms.com 8         vms-isps@dls.net - Forum for ISP running OpenVMS*                  Mobile: +7 (901) 971-3222A    TKD (WTF) in Russia, St.-Petersburg - www.TaeKwonDo-WTF.SPb.RUi   ------------------------------  # Date: Tue, 29 Jan 2002 22:46:01 GMTs3 From: sy18889@rabbit.fmr.com (Bradford J. Hamilton)n@ Subject: Using the 'latest and greatest' (was:VMS 7.3 stability)/ Message-ID: <JCF58.52$M3.221@news-srv1.fmr.com>c   Hi Ed,  O Your experience is the most compelling reason why we wait for a "point-release"e* (i.e., V7.3-1), to be GA before upgrading.  M Unless you have a compelling reason for using a new version (for example, XFC-C functionality that you *must* have	:-)), you might want to considerF waiting.  6 Of course, other folks may have different experiences.  b In article <995e39b6.0201291416.75d12c26@posting.google.com>, ewilts@ewilts.org (Ed Wilts) writes:D >It may just be me, but I've been managing VMS systems for nearly 20B >years and VMS 7.3 is ranking down there with the V4.0 release forB >stability.  In the less than 1 month that I've been running it in@ >production on my biggest Alphas (naturally the test and smaller> >systems all went smoothly), I've had 2 bugchecks escalated toG >Engineering, and one bugcheck go unexplained.  Each of the 3 bugcheckscG >- all with different bugcheck codes - have been on a different node inc >the all-Alpha all-7.3 cluster.u >e@ >Has anyone else been stricken by the crashes like we have?  Our% >cluster isn't really that unusual...m >s >Thanks,
 >   .../Ed >mailto:ewilts@ewilts.orga   Bradford J. Hamilton  bradhamilton@mediaone.net	(home) sy18889@rabbit.fmr.com		(work)  ; "All opinions that I express are my own, not my employer's"m   ------------------------------  # Date: Tue, 29 Jan 2002 19:12:37 GMTg! From: Ernest <wmozart5@yahoo.com>t- Subject: Using vfork/execvp with DCL commandseA Message-ID: <FuC58.102209$%b.6348077@bin4.nnrp.aus1.giganews.com>o  K How can execvp be used to run DCL commands (such as PIPE and DIR)?  I know rG that it uses VAXC$PATH to search for commands, but how can one emulate eB system()'s ability (which I can't use, since I must spawn without J immediately waiting for the job to finish) to run DCL commands?  On Unix, I I can accomplish running shell built-in commands by using something like aH "/bin/ksh -c echo hello", but I am not sure how this would be done with  DCL.   Thanks,l Ernest   ------------------------------  % Date: Tue, 29 Jan 2002 15:00:10 -0800uC From: "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam>r1 Subject: Re: Using vfork/execvp with DCL commandsv0 Message-ID: <3C57297A.233BBA96@intel.com.nospam>  
 Ernest wrote:e  L > How can execvp be used to run DCL commands (such as PIPE and DIR)?  I knowH > that it uses VAXC$PATH to search for commands, but how can one emulateC > system()'s ability (which I can't use, since I must spawn without"K > immediately waiting for the job to finish) to run DCL commands?  On Unix,5J > I can accomplish running shell built-in commands by using something likeI > "/bin/ksh -c echo hello", but I am not sure how this would be done withi > DCL.  D     It would be easier if you said what you're trying to accomplish,G especially since a number of us VMSers really don't know from "execvp".e  F     That said, system() is only a simple-to-use wrapper for LIB$SPAWN.G If the DCL command can really run asynchronously to the parent process,d? just use LIB$SPAWN and with the "NOWAIT" bit set in the "flags"b@ mask.  See HELP RTL LIB$ LIB$SPAWN for the calling sequence, theF manuals (or online docs) for a more complete description and examples.       -Ken --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfield5! F20 Automation VMS System Support  kenneth.h.fairfield#intel.com    ------------------------------  # Date: Wed, 30 Jan 2002 00:21:16 GMT ! From: Ernest <wmozart5@yahoo.com>.1 Subject: Re: Using vfork/execvp with DCL commandsoA Message-ID: <00H58.99605$vH6.5814972@bin6.nnrp.aus1.giganews.com>d  F "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam> wrote in) news:3C57297A.233BBA96@intel.com.nospam: i  F >     It would be easier if you said what you're trying to accomplish,I > especially since a number of us VMSers really don't know from "execvp".   H The execXX functions are pretty standard POSIX C library stuff, AFAIK.  ? The main goal is to keep the code as portable as possible.  If 5I execvp/execlp/etc. won't work, I'll have to use LIB$SPAWN, but certainly  9 using more portable functions would be highly preferable.   C Is the DCL shell itself not runnable as an image?  I saw the SPAWN  ) command, but that's yet another built-in.S   Thanks,V Ernest   ------------------------------  # Date: Wed, 30 Jan 2002 03:28:32 GMTv1 From: "David J. Dachtera" <djesys.nospam@fsi.net> 1 Subject: Re: Using vfork/execvp with DCL commandst' Message-ID: <3C576974.CB8115BD@fsi.net>   
 Ernest wrote:  > H > "Kenneth H. Fairfield" <Kenneth.H.Fairfield@intel.com.nospam> wrote in* > news:3C57297A.233BBA96@intel.com.nospam: > H > >     It would be easier if you said what you're trying to accomplish,K > > especially since a number of us VMSers really don't know from "execvp".  > H > The execXX functions are pretty standard POSIX C library stuff, AFAIK.@ > The main goal is to keep the code as portable as possible.  IfJ > execvp/execlp/etc. won't work, I'll have to use LIB$SPAWN, but certainly; > using more portable functions would be highly preferable.u > 4 > Is the DCL shell itself not runnable as an image?    No.c   > I saw the SPAWNl+ > command, but that's yet another built-in.e  ( That's how you "fork a shell" - sort of.   -- I David J. Dachteral dba DJE Systemsc http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/g   ------------------------------  # Date: Tue, 29 Jan 2002 19:34:04 GMTs2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)+ Subject: Re: VMS 7.1-1H2 and TCPIP 5.1 ECO3e1 Message-ID: <MOC58.145$am1.4167@news.cpqcorp.net>r  ^ In article <3c56722d$1@news.euriware>, "Philippe Bocher" <philippe.bocher@euriware.fr> writes: :Hello,mI :I need to upgrade my version of TCPIP to 5.1. When I look in the releasen :notes they sayn7 :OpenVMS Alpha V7.1, V7.1-1H1, V7.1-1H2, V7.2 for ECO 1 A :OpenVMS Alpha V7.2-1,V7.2-1H1,V7.2-2 and V7.3 for ECO2 and up...k :(Not 7.1-1H2) :o :What I'm I missing ?t  F   If you are asking the question, then it is now time to upgrade your    OpenVMS Alpha release.  I   We "punted" OpenVMS Alpha V7.1-1H2 some time ago, encouraging folks to nH   move to V7.1-2 or later.  More recently, we're trying to get folks to +   upgrade to OpenVMS Alpha V7.2-2 or later.t  F   This TCP/IP Services ECO kit may (or may not) work on the specified F   OpenVMS Alpha release, but we tend to test kits only on the OpenVMS G   releases that we are currently supporting.  That said, we did provideaG   a number of fixes and add a number of enhancements in recent OpenVMS tF   Alpha releases, and it may well be the case that V7.2-1 is a minimum   requirement. 0  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: 29 Jan 2002 14:16:03 -0800" From: ewilts@ewilts.org (Ed Wilts) Subject: VMS 7.3 stability= Message-ID: <995e39b6.0201291416.75d12c26@posting.google.com>e  C It may just be me, but I've been managing VMS systems for nearly 20yA years and VMS 7.3 is ranking down there with the V4.0 release foraA stability.  In the less than 1 month that I've been running it in ? production on my biggest Alphas (naturally the test and smaller = systems all went smoothly), I've had 2 bugchecks escalated todF Engineering, and one bugcheck go unexplained.  Each of the 3 bugchecksF - all with different bugcheck codes - have been on a different node in the all-Alpha all-7.3 cluster.  ? Has anyone else been stricken by the crashes like we have?  Our $ cluster isn't really that unusual...   Thanks,-	    .../Ed- mailto:ewilts@ewilts.org   ------------------------------  % Date: Tue, 29 Jan 2002 17:16:09 -0800r0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com> Subject: Re: VMS 7.3 stability, Message-ID: <3C56D8D9.5F1360BF@Mvb.Saic.Com>   Ed Wilts wrote:d > E > It may just be me, but I've been managing VMS systems for nearly 20eC > years and VMS 7.3 is ranking down there with the V4.0 release forcC > stability.  In the less than 1 month that I've been running it in.A > production on my biggest Alphas (naturally the test and smaller ? > systems all went smoothly), I've had 2 bugchecks escalated to H > Engineering, and one bugcheck go unexplained.  Each of the 3 bugchecksH > - all with different bugcheck codes - have been on a different node in  > the all-Alpha all-7.3 cluster. > A > Has anyone else been stricken by the crashes like we have?  Ourf& > cluster isn't really that unusual...  F When you installed it, did you also install all current patches?  Make2 sure you have installed VMS73_UPDATE, available at ftp.support.compaq.com.   G I had some bugchecks with my first V7.3 system, but after applying some D patches (all included in the one given above) all of my V7.3 systems have been rock solid.t  
 Mark Berryman    ------------------------------  # Date: Wed, 30 Jan 2002 02:16:12 GMTe$ From: "Ed Wilts" <ewilts@ewilts.org> Subject: Re: VMS 7.3 stability< Message-ID: <MHI58.30608$yC.3871581@typhoon.mn.mediaone.net>  G I'm reasonably current and have most of the pieces of the update patch.uF Some of the recent stuff put into the update patch didn't make my lastH change window and are scheduled for this weekend.  Don't forget, though,K that for a crash to be escalated to Engineering means that it doesn't matchIL a known footprint - ie, these are *new* bugchecks, not any that are fixed by the current patch kits.c   Cheers,c
     .../Ed -- Ed Wilts Mounds View, MN, USA mailto:ewilts@ewilts.org= "Mark Berryman" <Mark.Berryman@Mvb.Saic.Com> wrote in messagen& news:3C56D8D9.5F1360BF@Mvb.Saic.Com... > Ed Wilts wrote:  > >mG > > It may just be me, but I've been managing VMS systems for nearly 20nE > > years and VMS 7.3 is ranking down there with the V4.0 release foruE > > stability.  In the less than 1 month that I've been running it inoC > > production on my biggest Alphas (naturally the test and smalleraA > > systems all went smoothly), I've had 2 bugchecks escalated tonJ > > Engineering, and one bugcheck go unexplained.  Each of the 3 bugchecksJ > > - all with different bugcheck codes - have been on a different node in" > > the all-Alpha all-7.3 cluster. > >hC > > Has anyone else been stricken by the crashes like we have?  OurA( > > cluster isn't really that unusual... > H > When you installed it, did you also install all current patches?  Make4 > sure you have installed VMS73_UPDATE, available at > ftp.support.compaq.com.p >oI > I had some bugchecks with my first V7.3 system, but after applying someBF > patches (all included in the one given above) all of my V7.3 systems > have been rock solid.8 >8 > Mark Berryman    ------------------------------  % Date: Wed, 30 Jan 2002 03:28:57 +0100l' From: Brass Christof <welcome@spam.net> H Subject: Re: VMS721_SYS-V1100: Caution do not apply with Pathworks V6.0D( Message-ID: <3C575A69.91D830EA@spam.net>   Carl Karcher wrote:  > F > After applying patch VMS721_SYS-V1100 (to an Alphaserver 1200 5/533)I > and rebooting, PWRK$LMSRV.EXE (from Pathworks V6.0D) crashed the system  > with:i > B >   "INVSECURESTATE, Invalid state detected by SECURITY subsystem" > ( > shortly after starting up. The PC was: > . >         NSA$DEREFERENCE_RIGHTS_CHAIN_C+0008C > - > This was repeatable (same thing next time).E > J > After renaming the SECURITY.EXE_OLD and SECURITY_MON.EXE_OLD to *.EXE inH > SYS$LOADABLE_IMAGES and rebooting the crash no longer occurred. We hadJ > previously installed VMS721_SYS-V1000 so the problem was introduced withD > V11 of this patch. The problem has been reported to CSC (Pathworks" > handed it off to the VMS group). > F > VMS721_SYS-V1100 was one of 14 patches I applied in a single PRODUCTD > INSTALL session followed by a single reboot. I've always done thisH > (except for one patch that said not too) and never had a problem. ThisH > time it was an example of the risks of applying more than one patch atJ > the same time. In this case the bugcheck code pretty much pointed to theI > SECURITY*.EXE images (and a search on DSNlink confirmed it) so recoveryrF > was fairly simple. Next time I may not be so lucky (but I had a good
 > backup).  ? While I as a hobbyist user am surely not in a position to teach- you ? anything I would like to ask you whether you carefully read the  notes + that are associated with the ECO mentioned?o  < I spent several hours in reading ECO information to find out what e; occasionally crashes my DECwindows X11 display server and In found < the following note in the above mentioned ECO's readme file:  J ************************************************************************** *                             , Note                                      * H *                                                                        *t8 * After installing the VMS721_SYS-V1100  ECO kit you may experience the  *i * following system7 crash:                                                *tH *                                                                        *o6 *   INVSECURESTATE, Invalid state detected by SECURITY subsystem.        * H *                                                                        * 5 * In order to avoid this crash, before installing the  VMS721_SYS-V1100   *7 * ECO kit, contact your services representative for newe SECURITY.EXE and *5 * SECURITY_MON.EXE images  Then, after installing theb VMS721_SYS-V1100   *8 * ECO kit, but before rebooting the system, copy the new SECURITY images *N> * to the SYS$COMMON:[SYS$LDR]  directory.  You can then reboot your      *r * F system.                                                                * H *                                                                        *     L **************************************************************************    ? From my understanding this seems pretty much an explanation fore your  9 observation. Did you proceed as recommended in that note?    -- g6 moc dot slupofni at ssarb - please revert the sequence   ------------------------------    Date: 29 Jan 2002 14:52:26 -0800( From: bob@instantwhip.com (Bob Ceculski)? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)e= Message-ID: <d7791aa1.0201291452.434cd7b6@posting.google.com>   u "Terry C. Shannon" <terryshannon@mediaone.net> wrote in message news:<BWz58.181$Tu1.483@rwcrnsc51.ops.asp.att.net>...m4 > "JF Mezei" <jfmezei@videotron.ca> wrote in message( > news:3C56C7AF.45E85024@videotron.ca... > > Alan Greig wrote:YK > > > Perhaps Compaq intend to complete the port (in fact realise that theytL > > > must) but then immediately sell VMS. Perhaps they already have a buyerJ > > > lined up? Intel? Management buyout massively subsidized by Compaq to > > > get it off their hands?S > >rL > > I don't think that Compaq will sell VMS. As long as it generates profitsI > > without any advertising, then there is no reason to allow a potential H > > competitor to maximise the potential of VMS and steal customers from
 >  HP/Compaq.- > >-K > > Compaq will slowly reduce development budgets for VMS after the port tov >  IA64tF > > and will let customers decide, with strong incentives to stay with
 >  Compaq/HP.4 > >sJ > > And if the HP takeover of Compaq succeeds, VMS customers will have one >  less @ > > choice to go to if they want to leave those who screwed VMS. > >  > F > Research conducted by Ken Farmer and I, not to mention a boatload ofN > anecdotal evidence, renders it clear that the port of choice for VMS emigresI > is none other than IBM. HWP is specified as a venue of choice by a verya# > small number of would-be emigres.  > J > IMHO IBM is the company best equipped to meet the needs of VMS users whoN > elect to leave the fold. IBM has one hell of an opportunity here; it remainsN > to be seen what, if anything, CPQ will do to deter customers from "wandering > off the reservation."   ) are you saying it is time to abandon vms?y   ------------------------------  # Date: Tue, 29 Jan 2002 23:04:25 GMTa* From: "Bill Todd" <billtodd@metrocast.net>? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)rA Message-ID: <ZTF58.104673$%b.6543379@bin4.nnrp.aus1.giganews.com>e  5 "Bob Ceculski" <bob@instantwhip.com> wrote in messagey7 news:d7791aa1.0201291452.434cd7b6@posting.google.com...eA > "Terry C. Shannon" <terryshannon@mediaone.net> wrote in messagen5 news:<BWz58.181$Tu1.483@rwcrnsc51.ops.asp.att.net>...h6 > > "JF Mezei" <jfmezei@videotron.ca> wrote in message* > > news:3C56C7AF.45E85024@videotron.ca... > > > Alan Greig wrote:eH > > > > Perhaps Compaq intend to complete the port (in fact realise that theyH > > > > must) but then immediately sell VMS. Perhaps they already have a buyer L > > > > lined up? Intel? Management buyout massively subsidized by Compaq to > > > > get it off their hands?r > > >eF > > > I don't think that Compaq will sell VMS. As long as it generates profitsuK > > > without any advertising, then there is no reason to allow a potentialeJ > > > competitor to maximise the potential of VMS and steal customers from > >  HP/Compaq.n > > >bJ > > > Compaq will slowly reduce development budgets for VMS after the port to	 > >  IA64 H > > > and will let customers decide, with strong incentives to stay with > >  Compaq/HP.a > > >	L > > > And if the HP takeover of Compaq succeeds, VMS customers will have one	 > >  less B > > > choice to go to if they want to leave those who screwed VMS. > > >: > > H > > Research conducted by Ken Farmer and I, not to mention a boatload ofH > > anecdotal evidence, renders it clear that the port of choice for VMS emigres K > > is none other than IBM. HWP is specified as a venue of choice by a veryr% > > small number of would-be emigres.c > >dL > > IMHO IBM is the company best equipped to meet the needs of VMS users whoH > > elect to leave the fold. IBM has one hell of an opportunity here; it remains E > > to be seen what, if anything, CPQ will do to deter customers fromo
 "wandering > > off the reservation."" >u+ > are you saying it is time to abandon vms?   G I doubt you'll get Terry to go that far, though even he sounds worried. K That's also a decision that's likely very installation-specific:  some havenJ made it already, some certainly are considering it more seriously now, andA some really won't have a problem as long as they can get support.o  L I wouldn't encourage anyone who falls into the last catagory to abandon VMS:I I just think people are better off making decisions with their eyes open,mI rather than just sitting back and accepting the Compaq spin until the rug   gets yanked out from under them.  K There's also the question of whether anyone cares enough about changing thesH way Compaq handles VMS to stand up and fight to change it.  And now moreK than ever it looks as if that would require taking action to try to get rid  of current management.   - bill   ------------------------------  % Date: Tue, 29 Jan 2002 18:42:59 -0500 % From: JF Mezei <jfmezei@videotron.ca>-? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)n, Message-ID: <3C573347.423382F0@videotron.ca>   Bob Ceculski wrote:N+ > are you saying it is time to abandon vms?L  F It is time to *look* at alternatives to see what is available and whatL features are matched and what features are not available elsewhere (featuresM that you make use of).  Compaq has been giving plenty of hints that VMS wouldSL not be there forever. One such hint is that VMS *may* become mature on Alpha as soon as 2006.   ------------------------------  # Date: Wed, 30 Jan 2002 02:50:13 GMTR1 From: "David J. Dachtera" <djesys.nospam@fsi.net>D? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)e' Message-ID: <3C57606B.290846BF@fsi.net>e   Bill Todd wrote: > 7 > "Bob Ceculski" <bob@instantwhip.com> wrote in messageu9 > news:d7791aa1.0201291452.434cd7b6@posting.google.com... ) > > [snip] (in response to Terry Shannon)a- > > are you saying it is time to abandon vms?n > I > I doubt you'll get Terry to go that far, though even he sounds worried. M > That's also a decision that's likely very installation-specific:  some havelL > made it already, some certainly are considering it more seriously now, andC > some really won't have a problem as long as they can get support.i > N > I wouldn't encourage anyone who falls into the last catagory to abandon VMS:K > I just think people are better off making decisions with their eyes open,vK > rather than just sitting back and accepting the Compaq spin until the rugk" > gets yanked out from under them.  H I feel safe saying that until Cerner and Sunquest decide they've reachedD their limits with VMS, my current situation is the best it's been inA almost two years. Just speaking for myself personally, of course.a  M > There's also the question of whether anyone cares enough about changing therJ > way Compaq handles VMS to stand up and fight to change it.  And now moreM > than ever it looks as if that would require taking action to try to get rida > of current management.  H Yeah, well, we've seen recent evidence (anectdotal as it may be) of what+ happens when we try to go up against 'em...r   -- e David J. DachteraL dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------    Date: 29 Jan 2002 21:55:14 -0600+ From: young_r@encompasserve.org (Rob Young)n? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)/3 Message-ID: <TQpONzEwOxlM@eisner.encompasserve.org>,  T In article <3C573347.423382F0@videotron.ca>, JF Mezei <jfmezei@videotron.ca> writes:  I > One such hint is that VMS *may* become mature on Alpha as soon as 2006.q     	That is a distortion...  * 	Compaq's rebuttal to Gartner reads as so:  ' http://www.theinquirer.net/15010212.htm   N The best evidence we have of that [long term platform support] is today we areN still providing support for Digital PDP-11 and VAX systems purchased well overO 10 years ago.  We will be doing new functionality releases of OpenVMS for AlphaaK users at least through 2006. Beyond that timeframe, Alpha-based maintenanceo releases will continue.   . 	"new functionality ... at least through 2006"   				Robi   ------------------------------  # Date: Wed, 30 Jan 2002 04:32:32 GMT,4 From: "Terry C. Shannon" <terryshannon@mediaone.net>? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)m< Message-ID: <AHK58.2858$dZ1.10204@rwcrnsc52.ops.asp.att.net>  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:TQpONzEwOxlM@eisner.encompasserve.org...r7 > In article <3C573347.423382F0@videotron.ca>, JF Mezein <jfmezei@videotron.ca> writes: >eK > > One such hint is that VMS *may* become mature on Alpha as soon as 2006.  >- >- > That is a distortion...c  J Distortion? So what??? Remember the old trade press slogan, "never let the& facts get in the way of a good story."  J Unless CPQ decides to do another spin of EV7 beyond the planned EV79, 2006K would seem about right for the Alpha end game. EV79 is supposed to come outoK in 2004, figure it's good for about two years in the CMOS9 incarnation. And0C for government customers, it'll be good for a lot longer than that.s  B Assuming that post-Madison IPF processors offer decent performanceI (Probability Factor: 0.who knows?) and that Compaq proves Gartner (dare I J say it???) wrong by delivering VMS-on-IPF as promised, the maturity of VMSB on Alpha shouldn't be much of a factor. As always, time will tell.   ------------------------------  % Date: Wed, 30 Jan 2002 07:19:27 +0100e1 From: John McLean <mcleanj@swissonline.delete.ch>e? Subject: Re: Where does VMS fit in here? (Was Re: HP admits...)f5 Message-ID: <3C57906F.B0FA305F@swissonline.delete.ch>n   Bill Todd wrote: .... > M > There's also the question of whether anyone cares enough about changing the J > way Compaq handles VMS to stand up and fight to change it.  And now moreM > than ever it looks as if that would require taking action to try to get rid  > of current management.  H What is their obligation to answer questions in the stockholders meeting? ?  Are they able to defer an answer (eg. "We'll answer that oneeC off-line") or can they decide that a question is inaproppriate (eg.M- "Please, we are discussing the merger here").s  C In order words, can they duck and weave around the more interestingtG questions about INCOME sources compared to REVENUE sources, about their H failure to advertise anything but PCS, about their future plans for VMS 
 etc, etc ?     John McLeane   ------------------------------  % Date: Tue, 29 Jan 2002 17:44:43 -0500 > From: "Koska, John C. (LNG-MBC)" <John.C.Koska@lexisnexis.com>N Subject: www.openvms.compaq.com and the link "OpenVMS Solutions Status Matrix"M Message-ID: <3D35AD137AAAD411A6BA0008C7B1B12D0160270B@MBCALBEXC03.BENDER.COM>l  ! I only see the first page of the u% "OpenVMS Applications Status Report"    with my browser [ :( , IE 5.5 ].  % However, if I download the document, t I get all of it.  ( Is there problem with the page possibly?  ' Or is it just me, and it being late in b day where I am?    :) jck
 John Koska Matthew Bender & Co., Inc. -"   A Member of the LexisNexis Group
 1275 Broadwayi Albany, NY  12204. USA  518-487-3255 John.C.Koska@lexisnexis.com   ) I post personal opinion only, and all thee* disclaimers one could imagine apply.  That( includes, I speak for myself only and my) views in no way represent my employer(s).h+ One should also take note of the Electronicc) Communications Privacy Act of 1986, whicho+ imposes civil and criminal liability on anyi( person who intentionally intercepts "any( wire, oral or electronic communication."   ------------------------------   End of INFO-VAX 2002.057 ************************ilts Mounds View, MN, USA mailto:ewilts@ewilts.org= "Mark Berryman" <Mark.Berryman@Mvb.Saic.Com> wrote
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L    
L     L    L    L    L    L    L    L    L    L    	L    
L    L    L    
L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L     L    !L    "L    #L    $L    %L    &L    'L    (L    )L    *L    +L    ,L    -L    .L    /L    0L    1L    2L    3L    4L    5L    6L    7L    8L    9L    :L    ;L    <L    =L    >L    ?L    @L    AL    BL    CL    DL    EL    FL    GL    HL    IL    JL    KL    LL    ML    NL    OL    PL    QL    RL    SL    TL    UL    VL    WL    XL    YL    ZL    [L    \L    ]L    ^L    _L    `L    aL    bL    cL    dL    eL    fL    gL    hL    iL    jL    kL    lL    mL    nL    oL    pL    qL    rL    sL    tL    uL    vL    wL    xL    yL    zL    {L    |L    }L    ~L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    L    