1 INFO-VAX	Thu, 18 Nov 2004	Volume 2004 : Issue 640       Contents: Re: DS25 under VMS V6.2 : Re: Extending DCL [was: Re: DCL suggestion for f$verify()]: Re: Extending DCL [was: Re: DCL suggestion for f$verify()]: Re: Extending DCL [was: Re: DCL suggestion for f$verify()]: Re: Extending DCL [was: Re: DCL suggestion for f$verify()]: Re: Extending DCL [was: Re: DCL suggestion for f$verify()]: Re: Extending DCL [was: Re: DCL suggestion for f$verify()]4 Re: Looking for feedback on using EDM to back up VMS Re: Newsgroup moderation on VMS  Re: Newsgroup moderation on VMS , Re: Online forums for former digits/deccies?, Re: Online forums for former digits/deccies? Re: Radeon 7500 problem  Re: raid hardware % Re: RMS Files and directories Indexes % Re: RMS Files and directories Indexes % Re: RMS Files and directories Indexes % Re: RMS Files and directories Indexes % Re: RMS Files and directories Indexes % Re: RMS Files and directories Indexes " Re: Solaris 10 - no charge for o/s" Re: Solaris 10 - no charge for o/s" Re: Solaris 10 - no charge for o/s" Re: Solaris 10 - no charge for o/s" Re: Solaris 10 - no charge for o/s" Re: Solaris 10 - no charge for o/s" Re: Solaris 10 - no charge for o/s Re: t1lib, %SYSTEM-F-HPARITH' Re:  Upgrade base/path for V8.2 release & Re: Upgrade base/path for V8.2 release& Re: Upgrade base/path for V8.2 release& Re: Upgrade base/path for V8.2 release& Re: Upgrade base/path for V8.2 release  F ----------------------------------------------------------------------  % Date: Wed, 17 Nov 2004 21:17:26 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>  Subject: Re: DS25 under VMS V6.2+ Message-ID: <419C1445.5A365EAD@comcast.net>    Ferry Bolhar wrote:  >  > Hello people,  > N > does someone know whether an AlphaServer DS25 is running under OpenVMS V6.2? > & > MTIA and kind greetings from Vienna,  G DS25 is a fairly new development, in relative terms. I would expect its @ minimum version to be either V7.2-2 plus ECOs or V7.3 plus ECOs.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    ------------------------------  # Date: Wed, 17 Nov 2004 19:47:22 GMT % From: "John Vottero" <John@mvpsi.com> C Subject: Re: Extending DCL [was: Re: DCL suggestion for f$verify()] : Message-ID: <eVNmd.167$iU3.154@newssvr31.news.prodigy.com>  6 "David Froble" <davef@tsoft-inc.com> wrote in message & news:419A96B1.2010804@tsoft-inc.com... > John Vottero wrote:  > : >> "Anthony Borla" <ajborla@bigpond.com> wrote in message 9 >> news:NFJld.36254$K7.7440@news-server.bigpond.net.au...  >> >> [Many good ideas snipped] >> >>I >>>Obviously this is just a wish-list, one presented entirely from a user I >>>perspective. However, it is one that seeks to see DCL made into a more D >>>useful tool, one perhaps better reflecting the needs of today. A  >>>deficientI >>>command language will only see users resort more to alternatives like   >>>Perl > >>>and Python, even perhaps, to alternative operating systems. >>>  >>>  >>K >> My wish is that HP enhances DCL by starting over, with a clean sheet of  F >> paper.  Provide backwards compatibility with the /CLI qualifier in A >> AUTHORIZE.  DCL was fantastic 25 years ago but, it's outdated.  >> >> You can go to:  >>; >> http://msdn.microsoft.com/theshow/episode043/default.asp  >>F >> and watch a video that demos MSH, the command line environment for K >> Longhorn. Just watch part 5, "Enter the Programmer".  They even mention  " >> OpenVMS and DCL at least twice. >  > 2 > I didn't look at the reference.  Dial-up access. > L > I'm not one to ever say that things can't be improved.  However, for what G > I use DCL for, there aren't many deficiencies.  I'd be interested is  K > seeing what could be better.  Then again, I didn't really understand the   > GUI when it first appeared.   J There are two key ideas that got my attention.  First is that results are H objects rather than text.  When you pipe stuff between commands, you're L piping objects not text.  The command processor knows that the objects have 2 to be turned into text to be displayed so (in DCL)   $ DIR/MODI/SINCE=TODAY  G will display the files that have been modified today but, trying to do  D something with that is a pain.  Wouldn't it be cool if you could do:  ! $ filelist = DIR/MODI/SINCE=TODAY   D and 'filelist' was a collection of File objects that you could then  manipulate?  Something like:  ! $ filelist = DIR/MODI/SINCE=TODAY  $ foreach(file in filelist) 9 $     WRITE SYS$OUTPUT file.Name, "  size is ", file.Size  $ endfor  K The second important item is that command definitions are object oriented.  J If you're writing the code for a new command, you base your command on an M existing base class that already implements a lot of the common stuff.  I'll  K use the SUBMIT command as an example, you could implement SUBMIT something  
 like this:   class Submit : FileCommand { J     [Qualifier("QUEUE", nonnegatable, ValueRequired, Default="SYS$BATCH")]     string batchQueue;  '     [Qualifier("AFTER", ValueRequired)]      DateTime afterTime;   3     public override object ProcessObject(File file)      {           if (ShouldProcess(file))
             { M                //  the batchQueue and afterTime are already filled in by the   command processor "                SubmittedObject so;A                so = dotherealworkandreturnasubmittedobject(file);                 return so; 
             }      }  }   K (this is basically C# because I know that better than Java but, Java would  & be a more likely language on OpenVMS).  L The FileCommand base class has already dealt with all of the file selection G qualifiers and wildcards and it just calls ProcessObject once for each  J matching file and it passes a File object.   The Submit implementation of L ProcessObject calls ShouldProcess to handle /CONFIRM etc. and then does the G work.  Note that it returns a "Submitted" object, it doesn't display a  I message.  The command processor will call the Submitted objects ToString  K method if it needs to display a message but, if your command procedure did:   * $ submitInfo = SUBMIT/QUEUE=MYQUE SOMEFILE  E then submitInfo would be a Submitted object that you could deal with.        > L > As for deficiencies, the command line interface on windows.  Pitiful.  If J > I go on, it would be about no stand-alone backup, no batch, etc.  Right L > now, DCL is part of those things that I've come to realize make using VMS 
 > so nice. >   H Yep, it's pretty bad and, a new command processor won't fix everything. F NTBACKUP can be used in a command/batch environment but, if it hits a J problem it pops up a dialog box in never-never land to tell you about the E problem.  You know you have a problem when your batch job hangs but,  5 figuring out what that dialog box says is impossible.   D It looks like Microsoft realizes that a command line environment is K important.  In another year or two they'll have a good base.  3 to 5 years  G after that, it will be widely deployed and supported by most 3rd party   developers.   B How long should HP wait before they start working on "Son of DCL"?   ------------------------------    Date: 17 Nov 2004 15:25:02 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) C Subject: Re: Extending DCL [was: Re: DCL suggestion for f$verify()] 3 Message-ID: <CvjGXRxcMr9x@eisner.encompasserve.org>   j In article <NFJld.36254$K7.7440@news-server.bigpond.net.au>, "Anthony Borla" <ajborla@bigpond.com> writes: > F > Seeing this request for an alteration of DCL functionality set me toM > thinking about how useful it might be to implement a facility that allows a E > user to 'extend' DCL on an ad hoc basis. More specifically, why not 7 > implement something like a new lexical function, say:   E    A yes, user written lexical functions.  Has been on the whish list     for decades.   E    And it showed up in a POSSIBLE FUTURES slide presented last month. ?    Sorry, I don't have the URL, maybe soeone else will post it.    ------------------------------    Date: 17 Nov 2004 15:28:16 -0600; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) C Subject: Re: Extending DCL [was: Re: DCL suggestion for f$verify()] 3 Message-ID: <5COlen3$rfoq@eisner.encompasserve.org>   Q In article <419B7B72.78D3F571@adldata.com>, sol gongola <sol@adldata.com> writes:  >>   >> >; >> > My wish is that HP enhances DCL by starting over, with  >> > a clean sheet of paper. >> >? >> > Provide backwards compatibility with the /CLI qualifier in  >> > AUTHORIZE.  >> > >>   >>   > ; > Is there a way to specify something other than /CLI=DCL ?   A    Yes.  MCR, DECshell (a Bourne shell), and the POSIX shell have     all been useable as a CLI.   %    None of which ship with VMS today.   @    Sadly, writing a CLI is not documented.  But if you study the@    internals manual and source listings a while, you can reverse8    engineer it.  Then you get to hope it doesn't change.  C    It would be nice if there was support both user written CLIs and C    user written ACPs.  Support could be limitted to the level given F    to user written device drivers, which haven't made any real changes    since SMP was introduced.          ------------------------------    Date: 17 Nov 2004 15:37:53 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen) C Subject: Re: Extending DCL [was: Re: DCL suggestion for f$verify()] 3 Message-ID: <$jVqkv1ScujA@eisner.encompasserve.org>   q In article <5COlen3$rfoq@eisner.encompasserve.org>, koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:   E >    It would be nice if there was support both user written CLIs and E >    user written ACPs.  Support could be limitted to the level given H >    to user written device drivers, which haven't made any real changes >    since SMP was introduced.  E I would say that user-written ACPs _are_ as supported as user-written G device drivers.  So long as the ACP and the corresponding device driver G are written by the same entity they can cooperate in any manner that is % otherwise legal within VMS internals.   ; An  ACP is just a process-based adjunct to a device driver.    ------------------------------  % Date: Wed, 17 Nov 2004 20:29:44 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>C Subject: Re: Extending DCL [was: Re: DCL suggestion for f$verify()] + Message-ID: <419C0918.72FC48C7@comcast.net>    John Vottero wrote:  > A > "David J Dachtera" <djesys.nospam@comcast.net> wrote in message ' > news:41997373.FEA5FC27@comcast.net...  > > John Vottero wrote:  > >> [snip] / > >> Just watch part 5, "Enter the Programmer".  > > F > > Circa. the 00:50:00 minute mark. The whole thing is 1:55:00 or so. >  > I skipped straight to part 5  @ How did you do that? What version of Media Player are you using?  & > so, I don't know if there's anything  > interesting in the first part.   Not really - just boring.    > > " > >> They even mention OpenVMS and > >> DCL at least twice. > > I > > What they seem to call "MO-NET" looks absolutely HORRIBLE compared to H > > DCL. Then again, if you compare actual DCL to the Command DefinitionG > > Language (.CLD), that would look incomprehensible, also. (actually, J > > after I've been "in the groove" for a day or two, most of my DCL looks) > > incomprehensible a day or so later!).  > M > What do you think looks so horrible?  They stole (ok, borrowed) many of the $ > good things from DCL, basics like:N > - Qualifiers can have long names but, you only have to use enough to make it > unique@ > - If a parameter is required, it will prompt you for the value > : > Command definition looks to be even better that DCL/CLD.  C The part I saw looked like you needed and extensive background in C C programming and full knowledge of the entire RTL (sort of like TPU, 	 y'know?).    > > J > > Still, the result would be very UN*X-like, it seems. I haven't watchedK > > the whole thing yet. Dunno if they actually show how to use a "command" L > > built from "command-lets". So, dunno if if you're stuck re-inventing theC > > command line parsing wheel in every program, like in UN*X-land.  > N > The do a lot of demos.  I don't think you ever have to parse a command line.H > You also don't have to do anything to implement common qualifiers like > CONFIRM, WHATIF, LOG etc.  >  > > K > > Maybe we could let DCL be sensitive to the presence of a pipe cymbol in I > > a command line rather than using the PIPE verb. That'd help, as would ? > > looping constructs, conditional blocks (CASE, UNTIL, etc.).  > >  > N > There are plenty of things that could be done but I think the most importantN > is, extensibility.  I want to be able to add lexicals.  I want to be able toN > add qualifiers that can be applied to commands that I didn't write.  I don'tM > want to have to reimplement /CONFIRM, /LOG, /PAGE, /BEFORE etc every time I = > write a command (yes, I know about the UTIL$CQUAL routines)   F Then write your own "library" code to call the common qualifier stuff.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    ------------------------------  % Date: Thu, 18 Nov 2004 14:57:32 +1100 4 From: Paddy O'Brien <paddy.o'brien@transgrid.com.au>C Subject: Re: Extending DCL [was: Re: DCL suggestion for f$verify()] / Message-ID: <419C1DAC.4070403@transgrid.com.au>    David J Dachtera wrote:  > John Vottero wrote:    [Snips]   N >>There are plenty of things that could be done but I think the most importantN >>is, extensibility.  I want to be able to add lexicals.  I want to be able toN >>add qualifiers that can be applied to commands that I didn't write.  I don'tM >>want to have to reimplement /CONFIRM, /LOG, /PAGE, /BEFORE etc every time I = >>write a command (yes, I know about the UTIL$CQUAL routines)  >  > H > Then write your own "library" code to call the common qualifier stuff. >   C I keep the following code snippets in a file to be included in any  G command file which I want to have a /confirm qualifier.  (One line has  C wrapped and I don't want to fiddle about just to unwrap none line).   A Two bits that might need changing are where the * is in C*ONFIRM  G depending on other qualifiers.  This tells the other command file what  ( is the shortest acceptable abbreviation.  6 The second bit is t6he prompt on the read/prompt line.   $       CONFIRM_QUALIFIER = 0 & $       if f$extract(0,1,P2) .eqs. "/" $       then/ $               if f$mode() .eqs. "INTERACTIVE"  $               thenG $                       @COM_UTILITY:CHECK_QUALIFIER "''P2'" "C*ONFIRM" ; $                       CONFIRM_QUALIFIER = FOUND_QUALIFIER  $               endif  $! Get any other qualifiers  $       else $               P3 = P2 
 $       endif  $ 
 $ CHECK_LOOP: B $! code to get something into the symbol FILE.  Usually a file or  $!library module $  $ CONFIRM_PROMPT:  $       if CONFIRM_QUALIFIER $       thenC $               read/prompt="Store ''FILE'? [N] " SYS$COMMAND REPLY = $               REPLY = f$edit(f$extract(0,1,REPLY),"upcase") K $               if REPLY .eqs. "" .or. REPLY .eqs. "N" then goto CHECK_LOOP ; $               if REPLY .eqs. "Q" then goto END_CHECK_LOOP " $               if REPLY .eqs. "A" $               then- $                       CONFIRM_QUALIFIER = 0  $               elseC $                       if REPLY .nes. "Y" then goto CONFIRM_PROMPT  $               endif 
 $       endif  $  $! Do what you want to do  $	goto CHECK_LOOP  $  $ END_CHECK_LOOP: 3 $! Carry on DCL - yet another in the comedy series.   E I also have a Fortran function which simulates the CONFIRM_QUALIFIER  H code.  The function return is .TRUE. for A or Y, and .FALSE. for N or Q.I It is not pure as it can alter the Confirm_Qualifier input variable as A   does in the DCL code.   G I don't see /log or /before as much of a problem.  But if I were going  F to use a /page with the full DCL functionality, I quess I would write > either a utility command file or code to be included as above.   Regards, Paddy      G ***********************************************************************   C "This electronic message and any attachments may contain privileged > and confidential information intended only for the use of the B addressees named above.  If you are not the intended recipient of C this email, please delete the message and any attachment and advise B the sender.  You are hereby notified that any use, dissemination, 7 distribution, reproduction of this email is prohibited.   A If you have received the email in error, please notify TransGrid  A immediately.  Any views expressed in this email are those of the  = individual sender except where the sender expressly and with  C authority states them to be the views of TransGrid.  TransGrid uses > virus-scanning software but excludes any liability for viruses contained in any attachment.  < Please note the email address for TransGrid personnel is now$ firstname.lastname@transgrid.com.au"  G ***********************************************************************    ------------------------------  % Date: Wed, 17 Nov 2004 20:34:44 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>= Subject: Re: Looking for feedback on using EDM to back up VMS + Message-ID: <419C0A43.196B2783@comcast.net>    David Froble wrote:  >  > Lee wrote: > @ > > Rob is almost correct.  We'll eventually put the 4-node ES45B > > cluster on a GS1280 cluster.  The ES45 cluster is still on HSJ4 > > but will be migrating to foreign SAN eventually.= > > I've set up 8 foreign SAN shadow disks on a second GS1280 2 > > cluster for running a third-party application.@ > > The plan is to drop our current DLT tape systems and use EDMB > > for backup.  Hence my quest for information on the EDM backup. > > = > > Our 3-site, 5-node cluster remains running as long as any 3 > > combination of two sites AND 3 nodes remain up.  > Q > David seems to be worried about restoring from scratch.  There is a real simple  > solution for that. > M > Assuming that you're running your system off the SAN, HSJ, or whatever, ...   ? ...and if the SAN, CI, etc. and the building within which it is  contained gets destroyed, ... ?     Keyphrase: "Business Continuity"   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    ------------------------------  + Date: Thu, 18 Nov 2004 06:00:05 +0100 (CET) % From: Nomen Nescio <nobody@dizum.com> ( Subject: Re: Newsgroup moderation on VMS8 Message-ID: <420bc2114900a4da8a40ae6e6f998577@dizum.com>  h In article <419A9F99.D9360E2F@teksavvy.com> on news:comp.os.vms JF Mezei <jfmezei@teksavvy.com> trolled: / 4 >what would it take to moderate a newsgroup on VMS ?  B Still on your rec.travel.air moderation fantasy trip, are you, JF?  d Don't forget, the moment you start with your "Bush Regime" crap, you'll be the first one kicked out!  ] And the contender for moderator, Jeff Hacker, is an ardent Bushite, so you'll have to stiffle  yourself.  Rotflmao!   http://groups.google.com/groups?as_q=bush&safe=images&as_ugroup=rec.travel.air&as_uauthors=nobody&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en    http://groups.google.com/groups?as_q=usa&safe=images&as_ugroup=rec.travel.air&as_uauthors=nobody&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en   http://groups.google.com/groups?as_q=bush&safe=images&as_uauthors=jf%20mezei&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en   http://groups.google.com/groups?as_q=usa&safe=images&as_uauthors=jf%20mezei&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en    ------------------------------  + Date: Thu, 18 Nov 2004 06:20:02 +0100 (CET) % From: Nomen Nescio <nobody@dizum.com> ( Subject: Re: Newsgroup moderation on VMS8 Message-ID: <a0be8add95d697e7d871ae86913de2a1@dizum.com>  h In article <419A9F99.D9360E2F@teksavvy.com> on news:comp.os.vms JF Mezei <jfmezei@teksavvy.com> trolled:  4 >what would it take to moderate a newsgroup on VMS ?  B Still on your rec.travel.air moderation fantasy trip, are you, JF?  d Don't forget, the moment you start with your "Bush Regime" crap, you'll be the first one kicked out!  ] And the contender for moderator, Jeff Hacker, is an ardent Bushite, so you'll have to stiffle  yourself.  Rotflmao!   http://groups.google.com/groups?as_q=bush&safe=images&as_ugroup=rec.travel.air&as_uauthors=nobody&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en    http://groups.google.com/groups?as_q=usa&safe=images&as_ugroup=rec.travel.air&as_uauthors=nobody&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en   http://groups.google.com/groups?as_q=bush&safe=images&as_uauthors=jf%20mezei&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en   http://groups.google.com/groups?as_q=usa&safe=images&as_uauthors=jf%20mezei&as_drrb=b&as_mind=12&as_minm=5&as_miny=1981&as_maxd=31&as_maxm=12&as_maxy=2004&as_scoring=d&lr=&num=100&hl=en    ------------------------------    Date: 17 Nov 2004 16:56:00 -0800$ From: leeroth@my-deja.com (Lee Roth)5 Subject: Re: Online forums for former digits/deccies? = Message-ID: <d0e744c9.0411171656.1c437b5a@posting.google.com>   z "Kenneth Farmer" <kfarmer@NOSPAM.spyderbyte.com> wrote in message news:<qjumd.4170$iI2.415588@twister.southeast.rr.com>...K > You want me to set one up for you on OpenVMS.org or Tru64.org?  For that  I > matter I could start a mailing list.  Either way, no sweat off my back.   < Actually, I was checking what was available before I started< one of my own. A mailing list is definitely not a desireable
 option to me.    Thanks!    ------------------------------  % Date: Wed, 17 Nov 2004 20:08:33 -0800  From: Z <z@no.spam> 5 Subject: Re: Online forums for former digits/deccies? 0 Message-ID: <10po80ofee0lv3d@corp.supernews.com>   Lee Roth wrote: 8 > Anyone have any pointers to active forums specifically  > aimed at former DEC employees? > B > One interesting site is http://www.decalumni.com/ but they don't, > appear to have an online discussion forum.  H Lee Roth ... geez, your name sounds familiar ... did you use Notes much?   ------------------------------  % Date: Wed, 17 Nov 2004 22:14:14 +0100 3 From: Michael Unger <spam.to.unger@spamgourmet.com>   Subject: Re: Radeon 7500 problem+ Message-ID: <301tm1F2rdqpvU1@uni-berlin.de>   - On 2004-11-17 17:58, "Fred Kleinsorge" wrote:   N > Because the DVI port was being configured by "BIOS magic".  The BIOS is onlyN > run on the first card, and not on the others.  So the driver needs to do theI > magic.  The next graphics patch kit should allow the DVI port (with the ?                                       ^^^^^^^^^^^^^^^^^^^^^^^^^ " > analog dongle attached) to work.   ^^^^^^^^^^^^^^^^^^^^^^ > E > We are looking at dual head and direct DVI-DVI for a future update. 0                                   ^^^^^^^^^^^^^^  E So using the digital output (DVI-D, single head) isn't supported yet?    Michael    --  ; Real names enhance the probability of getting real answers. 5 My e-mail account at DECUS Munich is no longer valid.    ------------------------------    Date: 17 Nov 2004 17:41:34 -0800% From: sean@obanion.us (Sean O'Banion)  Subject: Re: raid hardware= Message-ID: <dffe4465.0411171741.453e5156@posting.google.com>    Jim,  A What I'm using now looks like it's a lot bigger than what you are  after.D I'm using some EVA5000 (aka HSV110) with over 70TB of usable storageA in a Fibre Channel SAN, mostly presented to HP-UX servers running 3 Oracle and some COBOL business logic as a back end.   @ The EVA5000 and the smaller EVA3000 are managed with what CompaqF called a SAN Management Appliance: a custom tailored Win2K system withF a number of custom applications, most importantly the Command View EVA: application with a browser GUI interface to manage the EVAF configuration.  There is also a command line utility (SSSU) for hosts,B including VMS, that replicates Command View's functionality.  SSSUC sends commands to Command View which then implements the commands.  F Note that this means without the SAN Appliance, no changes can be madeF to the EVA.  EVA configuration metadata is stored on disks on the EVA,B and SSSU, but not the Appliance, can also store configuration on aC host.  Because the EVA sets the RAID level when physical storage is 7 presented to a host, a VMS host can manage the storage.   E The EVA also has snap replication, which we use for online backup, as D well as disk-disk-tape backup, but if the Appliance goes down, there$ is no way to send the snap commands.  D Command View and the EVA can be configured to send error messages toA ISEE and then on to HP, which is supposed to automatically open a ? service ticket, but ISEE has yet to work well, at least for us.   C The EVA controllers are expensive, so the cost only comes down when E they are nearly fully populated, though they are cheaper than HP's XP  series, HDS or EMC.   < The only time we've really seen a performance issue that wasB definitely the EVA was early in the project when we were running aA test load into Oracle without archiving enabled.  On a twenty CPU D PA8700 partition of a Superdome, we saturated a 24 member disk groupD configured as six or seven RAID5 virtual disks.  We since went to 48C member disk group as a minimum, and in production we have archiving C enabled, limiting the maximum write rate: we haven't seen a problem F that was really the EVA's issue.  A poorly designed Oracle application< did create far more IO than we planedbut it's been "fixed".     Sean  q JimStrehlow@data911.com (Jim Strehlow) wrote in message news:<4b6ec350.0411160906.50ee84d1@posting.google.com>... ' > sean@obanion.us (Sean O'Banion) wrote  > > Jim,J > > That depends on what you want: local attached (PCI card), or contoller > > managed. > F > What is a useful comparison chart that shows throughput versus cost? > H > I am mostly curious about what people are using and how satisfied they > are. > G > We once configured a StorageWorks RA3000 Pedestal Shelf Storage using - > a serial port cable and Microsoft software. ? > Is that the "industry standard" way to configure RAID arrays? G > We had to configure the RAID before loading OpenVMS onto that server.  > H > Is there OpenVMS software to examine, configure, backup, restore, etc. > the RAID controller settings?  >   > Jim Strehlow, Alameda, CA, USA   ------------------------------    Date: 17 Nov 2004 17:41:08 -0800. From: spamsink2001@yahoo.com (Alan E. Feldman). Subject: Re: RMS Files and directories Indexes= Message-ID: <b096a4ee.0411171741.44264787@posting.google.com>   ` Roy Omond <Roy.Omond@BlueBubble.UK.Com> wrote in message news:<2vuh4oF2pn1abU1@uni-berlin.de>... > Bob Koehler wrote: > 7 > > In article <HTieN5m4Xw4L@eisner.encompasserve.org>, ; >  > kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) writes:  > > M > >>Which makes it interesting that a VMS DELETE *.*.* deletes files first to G > >>last instead of last to first. I wonder if this will ever be fixed?  > >  > > K > >    IIRC both DFU and backup/delete will do last file first.  Get DFU if H > >    you need to delete a lot of files and let VMS Engineering work on > >    select() and fork().  > @ > No, that's not how DFU does it.  It's much cleverer than that,: > and the speedup is orders of magnitude better than doing > last-file-first. > @ > Basically DFU marks the directory file as no-directory, having@ > noted the FIDs of all files in the directory.  Deletion of theC > files by FID no longer has any of the directory updating overhead 
 > after that.  > A > Remember: DFU is your friend, and *every* system manager should ! > make use of such a good friend.   C By using DFU to delete a directory tree and aborting it midway with @ ^Y, you can see many "no such file" entries. This means the fileA headers of these files have been recycled for new files while the C corresponding .DIR;1 files are still functional. When all the FID's E listed in a directory have been marked for reuse, the .DIR;1  file is F deleted. This avoids the need for continually updating the .DIR;1 file1 with the cost of not being able to abort cleanly.    ------------------------------  % Date: Wed, 17 Nov 2004 20:47:34 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>. Subject: Re: RMS Files and directories Indexes+ Message-ID: <419C0D46.92697410@comcast.net>    Larry Kilgallen wrote: > b > In article <41997501.943ED01C@comcast.net>, David J Dachtera <djesys.nospam@comcast.net> writes: > > Bob Kaplow wrote:  > >> [snip] / > >> "VMS File SYstem Internals" by Kirby McCoy  > >> ISBN 1-55558-056-4  > >> Digital Press 1990  > >>M > >> As much as this book needs updating, we REALLY need updates to Ken Bates O > >> "VAX I/O Subsystem: Optimizing Performance" [set in the days of the HSC50] 7 > >> and Roy Davis "VAXCluster Principles" [pre Alpha].  > > ! > > Note to Sue, Hoff, whoever...  > H > Well, I certainly qualify as "whoever" and I have heard the publishers > speak to this issue. > B > > How receptive are the publishers to new/updated book concepts? > A > Regarding suggestions for new publications I would classify the 5 > Digital Press (a division of ...) folk as "polite".  > C > For suggestions from a would-be _author_ I would classify them as D > "enthusiastic".   It is not the case that Digital Press has a vast? > staff of authors waiting in the wings for something to write.   ( I should think that goes without saying.  F The idea was to look for opportunities to update the existing library,> then perhaps extend it as VMS has advanced beyond the existing
 publications.   D The "Who" would be me and others like me (actually, I would hope for; folks better than me, but we're all so busy these days...).   G I'd actually like to collaborate with Steve Schweda on a "ZIP and UNZIP D for VMS" book; but, given his opinion of me, that's not very likely.  1 I'd like to collaborate with Hein on an RMS book.   B If there were a TPU guru around, I'd like to collaborate with that@ person on an "Applied TPU" book, complete with a CD containing aE complete emulation of EDT (when you press CTRL+K, it responds exactly & the way EDT does - no EVE/TPU syntax).  / I'd like to collaborate with Guy on a DCL book.   G I'd like to collaborte with Jeff and Saul on a DCSC / SLS book focusing D on implementing the STK L700e and other tape libraries (that SLS was never intended to support).   D Then, of course, there's all the new features like PIPE, support forG extending volume sizes, etc. that needs a straight-forward explanation,  with examples.  H So much to do, so little time before Itanic heads to Davy Jones's locker and VMS's time runs out.   --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    ------------------------------  % Date: Wed, 17 Nov 2004 20:48:02 -0600 2 From: David J Dachtera <djesys.nospam@comcast.net>. Subject: Re: RMS Files and directories Indexes+ Message-ID: <419C0D62.A912ED7C@comcast.net>    Bob Koehler wrote: > l > In article <8UC9Ea+mubre@eisner.encompasserve.org>, kaplow_r@encompasserve.org.TRABoD (Bob Kaplow) writes: > >>4 > >> 1. How many files should you keep in directory? > >  > > One screenful! > I >    Exactly my feeling.  I've been using that rule of thumb for decades.   % D'ya ever have to deal with All-in-1?    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/C  " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    ------------------------------    Date: 17 Nov 2004 19:39:37 -0800. From: spamsink2001@yahoo.com (Alan E. Feldman). Subject: Re: RMS Files and directories Indexes= Message-ID: <b096a4ee.0411171939.5cd837da@posting.google.com>   a JF Mezei <jfmezei.spamnot@teksavvy.com> wrote in message news:<419A71F0.B135A6CB@teksavvy.com>...e
 > Hein wrote:iN > > Yeah. If one were to seriously consider that though, then I would probablyM > > design towards avoiding/minimizing splitting those records over directoryeB > > blocks. That would mean roughly 60 files per name = per block. > O > Ok, so 32k files would still represent 500 records in the directory file. Noti > quite fully cached.e >  > P > > If I was really realy serious about directory performance I might choose NOTP > > to use a directory. Just create the files with 'tmd' and use an indexed fileP > > to maintain 'names' to fid info, or use such indexed file in addition to the > > regular directory. > M > Problem with this is that there are issues with system management since youAM > have so many files that are "unmanaged" by DCL utilities sur as DIR, and ifrO > someone does a ANA/DISK/REPAIR, it will add all those files to SYSLOST (or isl > it SYS$LOST ?).a  	 [SYSLOST]8  K > ALLIN1 used 9 character file names (+4 for extension), as well as indexed-O > files to map the document names to the obscrure VMS file names. It also splits; > files amongst a number of directories to spread the load.D   ------------------------------  % Date: Wed, 17 Nov 2004 23:19:31 -0500p- From: JF Mezei <jfmezei.spamnot@teksavvy.com>e. Subject: Re: RMS Files and directories Indexes, Message-ID: <419C22C0.C382D62E@teksavvy.com>  K If one were to SET FILE MYDIR.DIR/GLOBAL_BUFFERS=xx where XX is the size ofsE the directory file, would this make a big difference in performance ?a   ------------------------------  % Date: Thu, 18 Nov 2004 04:40:34 +0000y7 From: David B Sneddon - bigpond <dbsneddon@bigpond.com> . Subject: Re: RMS Files and directories Indexes* Message-ID: <419C27C2.7000704@bigpond.com>   JF Mezei mentioned in passing:M > If one were to SET FILE MYDIR.DIR/GLOBAL_BUFFERS=xx where XX is the size of4G > the directory file, would this make a big difference in performance ?n >   ? I though directory access was via XQP, not RMS.  Global buffers1 only affect RMS access..   Regards, Dave -- FB David B Sneddon (dbs) VMS Systems Programmer dbsneddon@bigpond.comB Sneddo's quick guide ...   http://www.users.bigpond.com/dbsneddon/B DBS freeware   http://www.users.bigpond.com/dbsneddon/software.htm   ------------------------------    Date: 17 Nov 2004 14:39:22 -06004 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)+ Subject: Re: Solaris 10 - no charge for o/se3 Message-ID: <fS8xqAvXTkGI@eisner.encompasserve.org>   ` In article <41996326.6E70976C@comcast.net>, David J Dachtera <djesys.nospam@comcast.net> writes: > Bob Koehler wrote: >> eU >> In article <76adnTV1eMeTMAXcRVn-gg@igs.net>, "John Smith" <a@nonymous.com> writes: Q >> > http://story.news.yahoo.com/news?tmpl=story&ncid=738&e=1&u=/ap/20041115/ap_oo >> > n_hi_te/sun_solaris10 >> >. >> > Sun to Give Out Operating System for Free >> q* >>    Sun is going to get what it's worth. > C > Perhaps, but the principle isn't really that goofy: give away the:F > product, and sell the support (say, "Linux"). Support is an expense.6 > Software licenses are a capital expenditure (asset). > H > See, unlike certain folks we won't mention in this post (since we beatH > them bloody all too often in other posts), Sun understands that nobody< > gives a <censored> about TCO if the "sticker shock" of theC > "cost-to-acquire" keeps the product locked out of customer sites.e > ; > "What it's worth" rather depends on one's point of view: l > G > o Is it "worth it" to pick up a (relative) freebie and negotiate withs) > the bean counters for support dollars? t > J > o Is it "worth it" to beat the bean counters into submission for capitalH > dollars and then extort support money out of them with the promise notI > to beat them up for capital dollars again (until the next time you haved+ > to actually "buy" an upgrade, of course)?Y  C What happened at my former company was that putting in Dell servers D running M$lop was done because it was cheaper.  Then they had to addD more servers because doing multiple functions on one server was too F taxing.  Then they had to add more servers because performance scalingD wasn't going as planned.  Then they had to add licenses for various F tools they needed to backup these machines and put databases and otherG stuff on them (multiplied by the number of servers).  Then they had to eD get twice as many servers as were in place at that time to put in a F different location to have some sort of disaster tolerance.  Then theyF had to buy expensive disk farms with replication software and databaseG replication software to make the DT possible.  OOooops, have to add all D sorts of virus checking software, too.  And make sure that you do a C daily patch check, install, and reboot.  And so forth, and so on...   E It got to the point where they spent about 10x the amount of doing itlA with a VMScluster solution.  By then they'd expended too much to   reverse course.  t   ------------------------------    Date: 17 Nov 2004 14:49:03 -06004 From: kuhrt@nospammy.encompasserve.org (Marty Kuhrt)+ Subject: Re: Solaris 10 - no charge for o/sb3 Message-ID: <X3cDCh8B25Rq@eisner.encompasserve.org>   c In article <ML07lEOmxL0a@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:rO > In article <419A4CFA.3060701@prodigy.net>, CJT <abujlehc@prodigy.net> writes:t > D >> Perhaps they'll outsource the engineering to somebody like Canon. >> sJ >> Which suggests you may be right about the paint being thrown (as from a >> cannon).a > 2 > There ought to be some sort of law against that.  A If this were on DECUServe would it violate the Canons of Conduct?t   (groan)    ------------------------------  % Date: Wed, 17 Nov 2004 16:38:45 -0500t# From: "John Smith" <a@nonymous.com>n+ Subject: Re: Solaris 10 - no charge for o/sd, Message-ID: <8PGdnWBfIol7WQbcRVn-iw@igs.net>   Marty Kuhrt wrote:> > In article <41996326.6E70976C@comcast.net>, David J Dachtera% > <djesys.nospam@comcast.net> writes:t >> Bob Koehler wrote:c >>>w= >>> In article <76adnTV1eMeTMAXcRVn-gg@igs.net>, "John Smith"l >>> <a@nonymous.com> writes: >>>>L http://story.news.yahoo.com/news?tmpl=story&ncid=738&e=1&u=/ap/20041115/ap_o >>>> n_hi_te/sun_solaris10 >>>>. >>>> Sun to Give Out Operating System for Free >>>g+ >>>    Sun is going to get what it's worth.n >>D >> Perhaps, but the principle isn't really that goofy: give away theG >> product, and sell the support (say, "Linux"). Support is an expense.n7 >> Software licenses are a capital expenditure (asset).o >>D >> See, unlike certain folks we won't mention in this post (since weG >> beat them bloody all too often in other posts), Sun understands that D >> nobody gives a <censored> about TCO if the "sticker shock" of theD >> "cost-to-acquire" keeps the product locked out of customer sites. >>; >> "What it's worth" rather depends on one's point of view:- >>H >> o Is it "worth it" to pick up a (relative) freebie and negotiate with) >> the bean counters for support dollars?: >>C >> o Is it "worth it" to beat the bean counters into submission for E >> capital dollars and then extort support money out of them with the7C >> promise not to beat them up for capital dollars again (until the>? >> next time you have to actually "buy" an upgrade, of course)?t >aE > What happened at my former company was that putting in Dell serversfF > running M$lop was done because it was cheaper.  Then they had to addE > more servers because doing multiple functions on one server was toooH > taxing.  Then they had to add more servers because performance scalingE > wasn't going as planned.  Then they had to add licenses for various H > tools they needed to backup these machines and put databases and otherH > stuff on them (multiplied by the number of servers).  Then they had toE > get twice as many servers as were in place at that time to put in aeH > different location to have some sort of disaster tolerance.  Then theyH > had to buy expensive disk farms with replication software and databaseE > replication software to make the DT possible.  OOooops, have to add> > all>E > sorts of virus checking software, too.  And make sure that you do acE > daily patch check, install, and reboot.  And so forth, and so on...h >CG > It got to the point where they spent about 10x the amount of doing ityB > with a VMScluster solution.  By then they'd expended too much to > reverse course.,    G There is no advertising by HP which to explain this common situation tohJ CEO's. HP would make more money selling an Alpha/VMS solution to a companyK than they would on a shipload of PC's from Taiwan, yet I don't have to tellr you what HP does.   L CEO's watch the World Series, the Super Bowl, the PGA tour, etc... some even= watch business programs like Bloomberg TV or political shows.t  0 HP does nothing. No wonder Alpha sales are down.   ------------------------------    Date: 17 Nov 2004 15:38:45 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)-+ Subject: Re: Solaris 10 - no charge for o/s63 Message-ID: <5Qi+0$S9BETs@eisner.encompasserve.org>u  j In article <X3cDCh8B25Rq@eisner.encompasserve.org>, kuhrt@nospammy.encompasserve.org (Marty Kuhrt) writes:e > In article <ML07lEOmxL0a@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:rP >> In article <419A4CFA.3060701@prodigy.net>, CJT <abujlehc@prodigy.net> writes: >> nE >>> Perhaps they'll outsource the engineering to somebody like Canon.e >>> K >>> Which suggests you may be right about the paint being thrown (as from as >>> cannon). >> s3 >> There ought to be some sort of law against that.r > C > If this were on DECUServe would it violate the Canons of Conduct?  > 	 > (groan)n   Yes, that was my point.e   ------------------------------  % Date: Wed, 17 Nov 2004 20:54:20 -0600e2 From: David J Dachtera <djesys.nospam@comcast.net>+ Subject: Re: Solaris 10 - no charge for o/s-+ Message-ID: <419C0EDC.F84E9A7F@comcast.net>o   John Smith wrote:  >  > Marty Kuhrt wrote:@ > > In article <41996326.6E70976C@comcast.net>, David J Dachtera' > > <djesys.nospam@comcast.net> writes:D > >> Bob Koehler wrote:a > >>> ? > >>> In article <76adnTV1eMeTMAXcRVn-gg@igs.net>, "John Smith"@ > >>> <a@nonymous.com> writes: > >>>>N > http://story.news.yahoo.com/news?tmpl=story&ncid=738&e=1&u=/ap/20041115/ap_o > >>>> n_hi_te/sun_solaris10 > >>>>0 > >>>> Sun to Give Out Operating System for Free > >>>e- > >>>    Sun is going to get what it's worth.g > >>F > >> Perhaps, but the principle isn't really that goofy: give away theI > >> product, and sell the support (say, "Linux"). Support is an expense.r9 > >> Software licenses are a capital expenditure (asset).n > >>F > >> See, unlike certain folks we won't mention in this post (since weI > >> beat them bloody all too often in other posts), Sun understands that F > >> nobody gives a <censored> about TCO if the "sticker shock" of theF > >> "cost-to-acquire" keeps the product locked out of customer sites. > >>= > >> "What it's worth" rather depends on one's point of view:b > >>J > >> o Is it "worth it" to pick up a (relative) freebie and negotiate with+ > >> the bean counters for support dollars?m > >>E > >> o Is it "worth it" to beat the bean counters into submission foruG > >> capital dollars and then extort support money out of them with therE > >> promise not to beat them up for capital dollars again (until therA > >> next time you have to actually "buy" an upgrade, of course)?v > >kG > > What happened at my former company was that putting in Dell serversiH > > running M$lop was done because it was cheaper.  Then they had to addG > > more servers because doing multiple functions on one server was tootJ > > taxing.  Then they had to add more servers because performance scalingG > > wasn't going as planned.  Then they had to add licenses for variousvJ > > tools they needed to backup these machines and put databases and otherJ > > stuff on them (multiplied by the number of servers).  Then they had toG > > get twice as many servers as were in place at that time to put in a J > > different location to have some sort of disaster tolerance.  Then theyJ > > had to buy expensive disk farms with replication software and databaseG > > replication software to make the DT possible.  OOooops, have to add  > > all G > > sorts of virus checking software, too.  And make sure that you do arG > > daily patch check, install, and reboot.  And so forth, and so on...  > >lI > > It got to the point where they spent about 10x the amount of doing it D > > with a VMScluster solution.  By then they'd expended too much to > > reverse course.  > I > There is no advertising by HP which to explain this common situation to L > CEO's. HP would make more money selling an Alpha/VMS solution to a companyM > than they would on a shipload of PC's from Taiwan, yet I don't have to tellp > you what HP does.w > N > CEO's watch the World Series, the Super Bowl, the PGA tour, etc... some even? > watch business programs like Bloomberg TV or political shows.w > 2 > HP does nothing. No wonder Alpha sales are down.  @ Well, *ALPHA* sales are down because it was killed in its prime.  = VMS sales are down due to (need I say it?) stealth marketing.R  E ...and I'm kinda disappointed you didn't offer a 30-second commercialiG script siezing upon the opportunity to illustrate *WHY* TCO matters and : how to get the bean counters past the sticker shock of the cost-to-acquire.   -- > David J Dachtera dba DJE Systemsl http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:e" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/e  " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/o   ------------------------------  % Date: Wed, 17 Nov 2004 23:07:01 -0500u# From: "John Smith" <a@nonymous.com>o+ Subject: Re: Solaris 10 - no charge for o/sn, Message-ID: <A4OdnQwEQqt7ggHcRVn-qw@igs.net>   David J Dachtera wrote:e > John Smith wrote:  >> >> Marty Kuhrt wrote:j@ >>> In article <41996326.6E70976C@comcast.net>, David J Dachtera' >>> <djesys.nospam@comcast.net> writes:  >>>> Bob Koehler wrote:s >>>>>t? >>>>> In article <76adnTV1eMeTMAXcRVn-gg@igs.net>, "John Smith"t >>>>> <a@nonymous.com> writes: >>>>>> >>L http://story.news.yahoo.com/news?tmpl=story&ncid=738&e=1&u=/ap/20041115/ap_o >>>>>> n_hi_te/sun_solaris10 >>>>>>0 >>>>>> Sun to Give Out Operating System for Free >>>>> - >>>>>    Sun is going to get what it's worth.K >>>>F >>>> Perhaps, but the principle isn't really that goofy: give away the@ >>>> product, and sell the support (say, "Linux"). Support is anB >>>> expense. Software licenses are a capital expenditure (asset). >>>>F >>>> See, unlike certain folks we won't mention in this post (since weD >>>> beat them bloody all too often in other posts), Sun understandsG >>>> that nobody gives a <censored> about TCO if the "sticker shock" ofOC >>>> the "cost-to-acquire" keeps the product locked out of customerc >>>> sites.  >>>>= >>>> "What it's worth" rather depends on one's point of view:  >>>>E >>>> o Is it "worth it" to pick up a (relative) freebie and negotiate 0 >>>> with the bean counters for support dollars? >>>>E >>>> o Is it "worth it" to beat the bean counters into submission for G >>>> capital dollars and then extort support money out of them with thenE >>>> promise not to beat them up for capital dollars again (until the A >>>> next time you have to actually "buy" an upgrade, of course)?d >>>iG >>> What happened at my former company was that putting in Dell serversIH >>> running M$lop was done because it was cheaper.  Then they had to addG >>> more servers because doing multiple functions on one server was too B >>> taxing.  Then they had to add more servers because performanceG >>> scaling wasn't going as planned.  Then they had to add licenses forf> >>> various tools they needed to backup these machines and putB >>> databases and other stuff on them (multiplied by the number ofD >>> servers).  Then they had to get twice as many servers as were inG >>> place at that time to put in a different location to have some sort E >>> of disaster tolerance.  Then they had to buy expensive disk farmsnG >>> with replication software and database replication software to makee. >>> the DT possible.  OOooops, have to add allG >>> sorts of virus checking software, too.  And make sure that you do aiG >>> daily patch check, install, and reboot.  And so forth, and so on...e >>>tF >>> It got to the point where they spent about 10x the amount of doingG >>> it with a VMScluster solution.  By then they'd expended too much tot >>> reverse course.n >>G >> There is no advertising by HP which to explain this common situation F >> to CEO's. HP would make more money selling an Alpha/VMS solution toE >> a company than they would on a shipload of PC's from Taiwan, yet Ie' >> don't have to tell you what HP does.r >>E >> CEO's watch the World Series, the Super Bowl, the PGA tour, etc...*C >> some even watch business programs like Bloomberg TV or political4	 >> shows.0 >>3 >> HP does nothing. No wonder Alpha sales are down.  >jB > Well, *ALPHA* sales are down because it was killed in its prime. > ? > VMS sales are down due to (need I say it?) stealth marketing.e >rG > ...and I'm kinda disappointed you didn't offer a 30-second commercial5E > script siezing upon the opportunity to illustrate *WHY* TCO matters @ > and how to get the bean counters past the sticker shock of the > cost-to-acquire.    H Actually it's so bloody discouraging that we are giving consideration toI abandoning VMS-based solutions altogether if only to have more consistento@ cash flow in the business by focusing selling non-VMS solutions.    J I for one am getting extremely tired of C-level execs and those just below/ them in positions of influence telling us that:i  ! a) they've never heard of VMS, or   K b) people at HP tell them that they should be using Windows or Linux, maybea	 HP-UX, orn  ( c) They don't use VMS in their shops, or  L d) VMS is as good as dead so why don't we re-write our app to run on Solaris or AIX or Linux, ord  F e) they know VMS is good or better than good or even the best but that/ they'll never be able to sell it internally, ori  L f) Alpha is dead and IA-64 is in trouble and they aren't into buying trouble" (career limiting and all that), or  / g) they have not seen or heard of VMS in years.V    J The only 'spots' I feel like writing these days are the verbatim quotes ofJ these execs shooting down VMS-based solutions out of ignorance, stupidity,J timidity, or the fact that their offices are lined with trinkets from Sun, Dell, IBM, and Microsoft.8  K Maybe some 1/2 hour late-nite 'infomercial' spots that can be had for a fewmH thousand bucks a week to run each night discussing HP management in thisK regard will make me feel better. Perhaps some news agency will pick them up I as a news story and flash it around the world - have to ask my friends atgK Reuters & Bloomberg what they think. Or a streaming video presentation withaK a link sent out in about 100,000 e-mails to the computing press, investment D banking firms technology analysts, etc....   I'm feeling better justL thinking about this.  Maybe I'll call it "Farenheit 6/25" - How to Hand Your Customers to Your Competitors".m    B As we all know - it isn't the product that's the problem, it's theK management.   HP seems to prefer that the ISV's push on a rope or piss intob7 the wind before they will lift a finger to promote VMS.a   ------------------------------  % Date: Wed, 17 Nov 2004 23:48:35 -0500p- From: JF Mezei <jfmezei.spamnot@teksavvy.com>a+ Subject: Re: Solaris 10 - no charge for o/s_, Message-ID: <419C298F.DED4B23E@teksavvy.com>   John Smith wrote:tM > Maybe some 1/2 hour late-nite 'infomercial' spots that can be had for a fewrN > thinking about this.  Maybe I'll call it "Farenheit 6/25" - How to Hand Your! > Customers to Your Competitors".-   Celcius 549  (Farenheit 911 :-),  L I hear you mate.  One comes to realise that there is nothing that we can do,L all alone. We can only continue to point out HP's mistakes in the hopes thatN someone who can so something will eventually realise the problems and mount an8 internal campaign. However, it may be too late already.   D > As we all know - it isn't the product that's the problem, it's theM > management.   HP seems to prefer that the ISV's push on a rope or piss intow9 > the wind before they will lift a finger to promote VMS.   L Those who do want to promote VMS still need a minimal amount of support fromN the owner of VMS. And that requires the owner give some exposure of VMS in its! press releases at the very least..   ------------------------------  % Date: Wed, 17 Nov 2004 22:04:56 -0500s2 From: Chip Coldwell <coldwell@physics.harvard.edu>% Subject: Re: t1lib, %SYSTEM-F-HPARITHmH Message-ID: <Pine.LNX.4.44.0411172158440.16385-100000@frank.harvard.edu>   On Wed, 17 Nov 2004, Z wrote:m   > John Laird wrote:m8 > >>%SYSTEM-F-HPARITH, high performance arithmetic trap,P > >>Imask=00000000, Fmask=00040000, summary=02, PC=000000000006DD14, PS=0000001BR > >>-SYSTEM-F-FLTINV, floating invalid operation, PC=000000000006DD14, PS=0000001B > >  > > H > > Both these messages are consistent with an instruction referencing aP > > floating-point value that was not a valid one.  (Not all bit representationsP > > are valid floating-point numbers.)  The Fmask suggests the value in registerP > > F18 was the bad one.  Examination of the compiler listing, together with theP > > stack dump, should help narrow this down to a line of code and hopefully the > > errant variable. > > O > > This is often the result of passing in the address of something that is not2N > > a genuine float or double variable, or has perhaps never been assigned but, > > instead contains a previous stale value. > F > Good point. The code uses a typedef (DOUBLE) for dx and dy. But the * > actual function declares them as double. > - > What's the definition of type DOUBLE, Chip?l >  > (It'll be in your .LIS file)  > Well, I should beg your forgiveness for indulging my knee jerkH blame-the-compiler reaction to what has turned out to be a rather run ofI the mill array index out of bounds bug.  The problem is that in the line:   a   FindStems( currx, curry, currx-ppoints[numppoints-2].x, curry-ppoints[numppoints-2].y, dx, dy);m  H "numppoints-2" is -1.  I was able to get the example program to work by . simply prefixing this line with one that reads     if(numppoints > 2)  I Unfortunately, this exact construct appears in quite a few places in the -F source code, so some fixup will be required.  I'm in contact with the F author of this library, so hopefully this fix will be included in the  5.0.2 release of the T1LIB.i   Chip   --   Charles M. "Chip" Coldwell System Administrator Harvard Physics Department 617-495-3388   ------------------------------  % Date: Wed, 17 Nov 2004 16:09:18 -0500h From: norm.raphael@metso.com0 Subject: Re:  Upgrade base/path for V8.2 releaseQ Message-ID: <OF8F41EA6B.82BBCE4E-ON85256F4F.00742A89-85256F4F.00743D1C@metso.com>^  I Will the new (nonSDK) OpenVMS V8.2 VAX require V7.3 before upgrade or can^4 it be upgraded from V7.2 or (which) earlier version?I Will the new (nonSDK) OpenVMS V8.2 Alpha require V7.3-2 before upgrade or : can it be upgraded from V7.3-1 or (which) earlier version?  0 I've seen this, but the question(s) still stand.   Software Support:e  ' Current and Prior Version Support (PVS)D   Rule of Thumb:  - Current Release (CVS)9  - Upgrade Window from Immediately Previous Release (CVS) ;  - Last Maintenance Release of Previous Minor Release (PVS)d;  - Last Maintenance Release of Previous Major Release (PVS)e+  o CVS and PVS Algorithm at OpenVMS WebsitekA  o OpenVMS Alpha V7.3-2 and OpenVMS VAX V7.3 will be PVS releases0   ------------------------------  # Date: Wed, 17 Nov 2004 21:25:17 GMT 3 From: hammond@not@peek.ssr.hp.com (Charlie Hammond)d/ Subject: Re: Upgrade base/path for V8.2 release 1 Message-ID: <1lPmd.3064$jh3.225@news.cpqcorp.net>a  p In article <OF55B49540.B236F81D-ON85256F4F.006657F8-85256F4F.00669F25@metso.com>, norm.raphael@metso.com writes: >  >o >n > J >Will the new (nonSDK) OpenVMS V8.2 VAX require V7.3 before upgrade or can5 >it be upgraded from V7.2 or (which) earlier version?iJ >Will the new (nonSDK) OpenVMS V8.2 Alpha require V7.3-2 before upgrade or; >can it be upgraded from V7.3-1 or (which) earlier version?m)                                          wM The official answer to this question will be in the Upgrade and Installation nI manual.  This is still in review and could change, but as of now it lookscK like we will support upgrading from OpenVMS Alpha V7.3-1 or V7.3-2 to V8.2.w  E We "support" these paths because we [will] have tested them.  UpgradesE from older versions may work, but we do not test the and they are notbI recommended; certainly not for "production systems".  That said, I expecttD that an upgrade from V6.1 and newer may actually work -- but at YOURG risk!  NOT recommended and certainly NOT supported!  (You will probablycF save time by following a SUPPORTED upgrade path, rather than having to6 do your own validation of an unsuported upgrade path.)  F As of this point, OpenVMS VAX V8.2 does not exist; it is not expected H to ship with Alpha and I64 V8.2.  Therefor I cannot answer this question w.r.t OpenVMS VAX.  A You did not ask about OpenVMS I64 V8.2, but we will ONLY support i@ NEW installations.  There ARE various problems in upgrading fromC OpenVMS I64 E8.2 (field test).  And I suggest you don't even think m, about upgrading OpenVMS I64 from 8.1 or 8.0!   --  J       Charlie Hammond -- Hewlett-Packard Company -- Ft Lauderdale  FL  USAF           (hammond@not@peek.ssr.hp.com -- remove "@not" when replying)J       All opinions expressed are my own and not necessarily my employer's.   ------------------------------  % Date: Wed, 17 Nov 2004 17:05:13 -0500p- From: JF Mezei <jfmezei.spamnot@teksavvy.com>r/ Subject: Re: Upgrade base/path for V8.2 release 8 Message-ID: <1100728638.xZ55UcqOEf7Wnq34vC+Hhg@teranews>   Charlie Hammond wrote:N > The official answer to this question will be in the Upgrade and InstallationK > manual.  This is still in review and could change, but as of now it looks M > like we will support upgrading from OpenVMS Alpha V7.3-1 or V7.3-2 to V8.2.   N Wasn't 7.2 declared a "landing zone" version ? Aren't there a lot of customersK at 7.2 who have been waiting to see with all the turmoil from merger, alphat murder etc ?  A Avoiding support for 7.2 to 8.2 would be a mistake in my opinion.t   ------------------------------  # Date: Thu, 18 Nov 2004 02:25:09 GMTt# From: Beach Runner <bob@nospam.com>f/ Subject: Re: Upgrade base/path for V8.2 releasee8 Message-ID: <9KTmd.6858$Oc.3949@tornado.tampabay.rr.com>   JF Mezei wrote:- > Charlie Hammond wrote: > N >>The official answer to this question will be in the Upgrade and InstallationK >>manual.  This is still in review and could change, but as of now it looksnM >>like we will support upgrading from OpenVMS Alpha V7.3-1 or V7.3-2 to V8.2.: >  > P > Wasn't 7.2 declared a "landing zone" version ? Aren't there a lot of customersM > at 7.2 who have been waiting to see with all the turmoil from merger, alphaw > murder etc ?    G 7.2 is no longer under support, no is it available under prior version tH support.  Of course Colorado Support will provide best efforts and knownI fixes, but elevations to engineering aren't supported.  7.3-2 is current rG and 7.3-1 is also under support.   XFC alone is a good reason to be at g' 7.3x, as is better lan and SAN support.k   > C > Avoiding support for 7.2 to 8.2 would be a mistake in my opinion.    ------------------------------  % Date: Wed, 17 Nov 2004 21:20:45 -0600s2 From: David J Dachtera <djesys.nospam@comcast.net>/ Subject: Re: Upgrade base/path for V8.2 releasee+ Message-ID: <419C150D.94E11757@comcast.net>    JF Mezei wrote:s >  > Charlie Hammond wrote:P > > The official answer to this question will be in the Upgrade and InstallationM > > manual.  This is still in review and could change, but as of now it looksaO > > like we will support upgrading from OpenVMS Alpha V7.3-1 or V7.3-2 to V8.2.  > P > Wasn't 7.2 declared a "landing zone" version ? Aren't there a lot of customersM > at 7.2 who have been waiting to see with all the turmoil from merger, alpha  > murder etc ?  E I believe that's V7.2-2 (a roll-up of the patches to V7.2 and V7.2-1,v9 plus) as the "landing zone", as I believe V7.3-2 will be.h  C > Avoiding support for 7.2 to 8.2 would be a mistake in my opinion.-  F Guess we'll have to watch for the Upgrade and Installation Manual that comes out with V8.2.   --   David J Dachtera dba DJE Systemsi http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page:g" http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/e  " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/l   ------------------------------   End of INFO-VAX 2004.640 ************************ as this book needs updating, we REALLY need updates to Ken Bates O > >> "VAX I/O Subsystem: Optimizing Performance" [set in the days of the HSC50] 7 > >> and Roy Davis "VAXCluster Principles" [pre Alpha].  > > ! > > Note to Sue, Hoff, whoever...  > H > Well, I certainly qualify as "whoever" and I have heard the publishers > speak to this issue. > B > > How receptive are the publishers to new/updated book concepts? > A > Regarding suggestions foۭ"gtCᑌ{}F lcw*bn9dDV8Zh=
bJ٤"$!UC_1@Y>Iu|[<!2lg4/=`$y?f{݆xodL%:cP<uuin	i	6siu
}!QW;Oub&1h(9Dz?rq2J
_&^mƟE_DxM];>?5#fzG?yCLh
7kZg6υb(9=([J$ϢWOUR{#wF7C8+̏h_{|~2hڣ[x܊m'vk{}(jpܴ0tL+:,mXGjoZtl=G/.,mjӺF*o\vJ֜Re\	ߔ}T_O]	T:Cr 3ZQ*j`;pwbo7t|v&σ;p+mP2EkawE[w	R36iZxa]4s☝VzZk87阱gH֡iszj sDURLwm3PY"CWbG"{Ê<höj8?s
<tDQ7}'Wkpm~0c2Xwd^>)6\
x𰰏d2f6-%#,xxV$[`uryUzJ*N	>y\bG)\b>?
Wa7}&)Gz*`m'^߅[1?O;~ނ; ߷Mz:3:a_zg[jkТC:c"Cǥg"+V!
fFh_#{H=!EaKqu#OِvXyoFBQa|!=?
}8].?'Zyf~5[!:V
$
QWзh{!"$ـ{dvGf8ÑC}	pFgW@=&Pq=b
+O؃XNӽz)Dc&\vZ8q٥7%bEүp&VǢN
9lgxYvovEH*nsۺYEhWh=[hm%k6jW"\׶hq6FjX0V^e9ssL?|p2ә \Ơx<}bN29)w2FqсKט}gf):Sa:'9Cՠo"=}ޡ:pq>҅ѽ!ö3iNѪG,"4;b?eVp.=!1X8}\V'e+x4!ȹSbA*g@)גZ
kp'I]4;@ND&[jCۚ
+6mTd8z]8f79'[>*t?wg.+ >tG\x\!Ӌ2&?GG_$hfldA{TO{Ik4g\Xٯ__?yyYbEEDɒG/C-=1il5HVE>Zl{@%{HkiDy3Ix NP>DlQ#ǌ*색rPNC!s(0|Wl,Vy?E:kiۧ'o;
W1S[g o=
vkK/xzVFm??|t+b=)m8щ73.oe0oUoўb$l}VtSxMYӞݡǋZ)7.\;B&XLx-۱動D`DF7?`#։x}Zs+lEҞ ĶjWgFl[)Տ{6t*nu0+~ÌO|~Gb@}]u|?etBSL7179v%޳cĤh]/8z
S\	fT|/N6,U!=k}$@S	8?,hƟWV]>2|i`i 8L'7ߓ7vwR^`&30LcJ3/uX.}wPt^Y3e\$̪1L9&8/r)^MmE1=x(]þXe }KO]vqG<N~[P<:̴(	nd6`o<΂{xG
<K