1 INFO-VAX	Tue, 20 Aug 2002	Volume 2002 : Issue 458       Contents:# Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) # Re: attn: Paddy (re: DCL procedure) < Re: Charon-VAX (was: [VAX] VMS to [Alpha] OpenVMS conversion< Re: Charon-VAX (was: [VAX] VMS to [Alpha] OpenVMS conversion4 Configuring webDAV for CSWS T1.3 for Dreamweaver MX?  DE101 - can it go to Full Duplex( Re: DECCXX 6.5 Unreachable code warnings( Re: DECCXX 6.5 Unreachable code warnings' Re: Fortune Magazine and a post-VMS rap ' Re: Fortune Magazine and a post-VMS rap P RE: forwarding email from system to system (was: RE: attn: Paddy	(re: DCL procedP RE: forwarding email from system to system (was: RE: attn: Paddy	(re: DCL procedP forwarding email from system to system (was: RE: attn: Paddy (re: DCL procedure)P Re: forwarding email from system to system (was: RE: attn: Paddy (re: DCL procedP Re: forwarding email from system to system (was: RE: attn: Paddy (re: DCL proced! Re: From Today's Network World... - Re: How to "forward declare" a function in c? & How to configure Reflection X-server ?* Re: How to configure Reflection X-server ?  HP buying EMC. Is that true ????" Re: HP-Compaq Merger Went Smoothly" Re: HP-Compaq Merger Went Smoothly" RE: HP-Compaq Merger Went Smoothly" Re: HP-Compaq Merger Went Smoothly" Re: HP-Compaq Merger Went Smoothly" Re: HP-Compaq Merger Went Smoothly" RE: HP-Compaq Merger Went Smoothly" Re: HP-Compaq Merger Went Smoothly" Re: HP-Compaq Merger Went Smoothly2 Re: Java plugin performance for interactive applet@ Re: Licenses (was Re: Charon-VAX (was: [VAX] VMS to [Alpha]...)) Memory Slot? Re: Memory Slot? Re: New missive from HP  Re: NewHp & VMS Licensing  Re: NewHp & VMS Licensing  Re: NewHp & VMS Licensing  Re: NewHp & VMS Licensing  Re: NewHp & VMS Licensing  RE: NewHp & VMS Licensing & Re: NFS Server configuration questions& Re: NFS Server configuration questions OpenVMS + Joint STARS  Re: OpenVMS + Joint STARS  Re: OpenVMS + Joint STARS  Re: OpenVMS + Joint STARS  RE: OT: UK Universities + Printing barcodes via postscript on OpenVMS  Simple cluster tps ? * Re: VMS upgrade with shadowed system disk.* Re: VMS upgrade with shadowed system disk.* Re: VMS upgrade with shadowed system disk.* Re: VMS upgrade with shadowed system disk.* Re: VMS upgrade with shadowed system disk.* Re: VMS upgrade with shadowed system disk.* Re: VMS upgrade with shadowed system disk.  F ----------------------------------------------------------------------    Date: 20 Aug 2002 00:40:38 -0700# From: dooleys@snowy.net.au (dooley) , Subject: Re: attn: Paddy (re: DCL procedure)= Message-ID: <1ca82fc6.0208192340.34be1e6b@posting.google.com>   V Michael Unger <unger@decus.de> wrote in message news:<00A12B33.D3FC958A.2@decus.de>... > David J. Dachtera wrote: >  > > Michael Unger wrote: > > > O > > > I wasn't able to send the new (beta test) release of the DCL procedure to  >  youQ > > > for more than one week now. Are there currently restrictions regarding mail + > > > delivery (size, contents, generally)?  > > K > > I'm fairly certain that Paddy know how to use ZIP and UNZIP for VMS, if  > > size is a problem. > & > Thanks for the pointer to ZIP/UNZIP. > N > Having tried a few times more I don't think size is really the problem; evenO > simple mails (short ASCII only text, about 10 to 15 lines) can not be sent to Q > him -- the error message I get back is exactly the same as with large(r) files.  > R > (Perhaps my sender address got blocked because of the many "$" characters within > the mail ... ;-) > 	 > Michael    Did you remove the zzz ?. you could also replace the apostrophe with %27 Phil   ------------------------------  % Date: Tue, 20 Aug 2002 18:40:38 +1000 1 From: Paddy O'Brien <paddy.o'brien@tg.nsw.gov.au> , Subject: Re: attn: Paddy (re: DCL procedure), Message-ID: <3D620086.5080708@tg.nsw.gov.au>   O.K., the story -- sad story.   F Our IT section is waging a vendetta (hmm, you can wage a war, but can I you wage a vendetta?) against VMS. Normal reasons for corporate to move,   lack of applications.   H I knew that my dedicated zzz sub-domain would eventually go, but our IT E (accountants and economists -- no IT background in sight) decided to   pull the plug on August 9th.  F Since we had had a meeting with them where they assured me they would F give me notice, I assumed it was one of their regular b*lls-ups.  No, I apparently it was intentional and it took them till this morning to tell   me/admit that.  ? I am stuck with a bl**dy Outlook account which they believe is  G god-given.  At least I am connecting from a VMS box via Mozilla.  It's  I like leaving the champagne drinking crowd and having to imbibe sulphuric   acid.   A I have already apologised to Mark Berryman for the imposition on  E info-VAX for all the bounces, which were beyond my control.  For the  F several people I have enjoyed private mail with, my address no longer   has zzz. in it (as shown above).  G Michael, I am still working on getting my set-up correct and trying to  D connect this stupidity with my previous VMS mail, and I shall be in  touch ASAP.   B A further disadvantage of this Gatesian world is the overly long, G duplicated, verbose rubbish that will come at the bottom of any mail I  H send.  IANAL, but it seems to have virtually no foundation in law, just 9 verbosity and a conning trick -- yep, that's BG all over.    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."   G ***********************************************************************    ------------------------------  % Date: Tue, 20 Aug 2002 18:47:05 +1000 1 From: Paddy O'Brien <paddy.o'brien@tg.nsw.gov.au> , Subject: Re: attn: Paddy (re: DCL procedure), Message-ID: <3D620209.7020108@tg.nsw.gov.au>  
 dooley wrote: X > Michael Unger <unger@decus.de> wrote in message news:<00A12B33.D3FC958A.2@decus.de>... >  >>David J. Dachtera wrote: >> >> >>>Michael Unger wrote:  >>> M >>>>I wasn't able to send the new (beta test) release of the DCL procedure to  >>>  >> you >>O >>>>for more than one week now. Are there currently restrictions regarding mail ) >>>>delivery (size, contents, generally)?  >>> J >>>I'm fairly certain that Paddy know how to use ZIP and UNZIP for VMS, if >>>size is a problem.  >>& >>Thanks for the pointer to ZIP/UNZIP. >>N >>Having tried a few times more I don't think size is really the problem; evenO >>simple mails (short ASCII only text, about 10 to 15 lines) can not be sent to Q >>him -- the error message I get back is exactly the same as with large(r) files.  >>R >>(Perhaps my sender address got blocked because of the many "$" characters within >>the mail ... ;-) >>	 >>Michael  >  >  > Did you remove the zzz ?0 > you could also replace the apostrophe with %27 > Phil  F See my previous note (if it got through).  zzz used to be valid until A 9th August thanks to a load of IT illiterates who manage our IT.  * Michael's mail got through till that time.  H The apostrophe works for many/most and it alleviates (to zero) spam for G me; the escape that I understood, if required, was \s.  This goes back  0 to what Steve Lionel found many, many years ago.   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."   G ***********************************************************************    ------------------------------  % Date: Tue, 20 Aug 2002 18:58:41 +1000 1 From: Paddy O'Brien <paddy.o'brien@tg.nsw.gov.au> , Subject: Re: attn: Paddy (re: DCL procedure), Message-ID: <3D6204C1.7080304@tg.nsw.gov.au>  D When I wrote this, the formatting was about 70 characters per line. G When the return came from info-VAX, it's about double that, yet when I  , pick this up to reply, I'm back to about 70.  E What sort of stupid trick is BG playing ..... what sort of editor is  ? behind this?  I am totally (proudly and happily) a PC/BG world  5 non-savant.  Have I lost all/any control over sanity?   H What I would love for Christmas (but preferably sooner) would be my VMS  mail :-)  The impossible dream.    Regards, Paddy     Paddy O'Brien wrote: > O.K., the story -- sad story.  > H > Our IT section is waging a vendetta (hmm, you can wage a war, but can K > you wage a vendetta?) against VMS. Normal reasons for corporate to move,   > lack of applications.  > J > I knew that my dedicated zzz sub-domain would eventually go, but our IT G > (accountants and economists -- no IT background in sight) decided to   > pull the plug on August 9th. > H > Since we had had a meeting with them where they assured me they would H > give me notice, I assumed it was one of their regular b*lls-ups.  No, K > apparently it was intentional and it took them till this morning to tell   > me/admit that. > A > I am stuck with a bl**dy Outlook account which they believe is  I > god-given.  At least I am connecting from a VMS box via Mozilla.  It's  K > like leaving the champagne drinking crowd and having to imbibe sulphuric   > acid.  > C > I have already apologised to Mark Berryman for the imposition on  G > info-VAX for all the bounces, which were beyond my control.  For the  H > several people I have enjoyed private mail with, my address no longer " > has zzz. in it (as shown above). > I > Michael, I am still working on getting my set-up correct and trying to  F > connect this stupidity with my previous VMS mail, and I shall be in 
 > touch ASAP.  > D > A further disadvantage of this Gatesian world is the overly long, I > duplicated, verbose rubbish that will come at the bottom of any mail I  J > send.  IANAL, but it seems to have virtually no foundation in law, just ; > verbosity and a conning trick -- yep, that's BG all over.  >  > Regards, Paddy >  >  > I > ***********************************************************************  > E > "This electronic message and any attachments may contain privileged K > and confidential information intended only for the use of the addressees  E > named above.  If you are not the intended recipient of this email,  9 > please delete the message and any attachment and advise D > the sender.  You are hereby notified that any use, dissemination, 9 > distribution, reproduction of this email is prohibited.  > C > If you have received the email in error, please notify TransGrid  C > immediately.  Any views expressed in this email are those of the  I > individual sender except where the sender expressly and with authority  ; > states them to be the views of TransGrid.  TransGrid uses @ > virus scanning software but excludes any liability for viruses > contained in any attachment."  > I > ***********************************************************************         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."   G ***********************************************************************    ------------------------------  + Date: Tue, 20 Aug 2002 11:31:45 +0000 (UTC)  From: david20@alpha2.mdx.ac.uk, Subject: Re: attn: Paddy (re: DCL procedure)+ Message-ID: <ajt9b1$q6g$1@aquila.mdx.ac.uk>   ` In article <3D620086.5080708@tg.nsw.gov.au>, Paddy O'Brien <paddy.o'brien@tg.nsw.gov.au> writes: >O.K., the story -- sad story. > C >A further disadvantage of this Gatesian world is the overly long,  H >duplicated, verbose rubbish that will come at the bottom of any mail I I >send.  IANAL, but it seems to have virtually no foundation in law, just  : >verbosity and a conning trick -- yep, that's BG all over. >  >Regards, Paddy  >   L Not actually anything to do with Gates. The lawyers have been foisting these7 stupid disclaimers on companies for the last few years.   4 See http://www.goldmark.org/jeff/stupid-disclaimers/  G Note. That by posting this message you have forced lots of law abiding  J news server providers (and the google archive) to inadvertently contraveneJ this disclaimer by disseminating, distributing and reproducing this email.     >  > H >*********************************************************************** > D >"This electronic message and any attachments may contain privileged? >and confidential information intended only for the use of the  C >addressees named above.  If you are not the intended recipient of  D >this email, please delete the message and any attachment and adviseC >the sender.  You are hereby notified that any use, dissemination,  8 >distribution, reproduction of this email is prohibited. > B >If you have received the email in error, please notify TransGrid B >immediately.  Any views expressed in this email are those of the > >individual sender except where the sender expressly and with D >authority states them to be the views of TransGrid.  TransGrid uses? >virus scanning software but excludes any liability for viruses  >contained in any attachment." > H >*********************************************************************** >   
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Tue, 20 Aug 2002 08:19:05 -0400 6 From: "John.Malmberg" <Malmberg@dskwld.zko.dec.compaq>, Subject: Re: attn: Paddy (re: DCL procedure)4 Message-ID: <3D6233B9.3040102@dskwld.zko.dec.compaq>   Paddy O'Brien wrote:F  > When I wrote this, the formatting was about 70 characters per line.G  > When the return came from info-VAX, it's about double that, yet when 1  > I pick this up to reply, I'm back to about 70.   B It is a "feature" of MOZILLA.  Use the VIEW -> SOURCE to pull up a' window on how your post looks normally.   C For some reason MOZILLA rewraps the plain text message to fit your  . screen when the text is not preceded by a ">".  G I have been meaning to look up in BUGZILLA to see if anything has been  C logged about this feature, and several other minor issues with the   mail-news client.N  F BUGZILLA allows you to give feedback to the people working on MOZILLA,1 there is a link to it from the MOZILLA home page.D   -Johnu! malmberg@dskwld.zko.dec.compaq.hpe Personal Opinion Only    ------------------------------  % Date: Tue, 20 Aug 2002 08:27:08 -0400 6 From: "John.Malmberg" <Malmberg@dskwld.zko.dec.compaq>, Subject: Re: attn: Paddy (re: DCL procedure)4 Message-ID: <3D62359C.1080702@dskwld.zko.dec.compaq>   John.Malmberg wrote: >  > Paddy O'Brien wrote:H >  > When I wrote this, the formatting was about 70 characters per line.I >  > When the return came from info-VAX, it's about double that, yet whenX3 >  > I pick this up to reply, I'm back to about 70.h > D > It is a "feature" of MOZILLA.  Use the VIEW -> SOURCE to pull up a) > window on how your post looks normally.g > E > For some reason MOZILLA rewraps the plain text message to fit your w0 > screen when the text is not preceded by a ">".  > Further investigation shows that Mozilla is tagging it's text:  9 Content-Type: text/plain; charset=us-ascii; format=flowedf  B So that is what is causing the text to be set to the width of the F window.  I do not know how to get Mozilla not to set that last tag of  "format=flowed".  I > I have been meaning to look up in BUGZILLA to see if anything has been iE > logged about this feature, and several other minor issues with the   > mail-news client.t > H > BUGZILLA allows you to give feedback to the people working on MOZILLA,3 > there is a link to it from the MOZILLA home page.a   -JohnW! malmberg@dskwld.zko.dec.compaq.hpW Personal Opinion Onlym   ------------------------------  % Date: Tue, 20 Aug 2002 09:54:01 -0400 ; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com> , Subject: Re: attn: Paddy (re: DCL procedure)$ Message-ID: <3d624a42$1@news.si.com>  ? >I am stuck with a bl**dy Outlook account which they believe iseG >god-given.  At least I am connecting from a VMS box via Mozilla.  It's I >like leaving the champagne drinking crowd and having to imbibe sulphurici >acid.  L My complete sympathy.  In a couple of short months, I'll be in the positioinK in which you've found yourself: having to use Outlook.  Moreover, our EmailR9 will be outsourced to HP!  No more internal mail routers.   K If anyone can tell me how to get Exchange to deliver mail to me VMS accountcC and still appear to be from the person who actually sent it, I'd bep	 grateful.s --A Brian Tillman                   Internet: tillman_brian at si.comMA Smiths Aerospace                          tillman at swdev.si.com = 3290 Patterson Ave. SE, MS      Addresses modified to preventM< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Tue, 20 Aug 2002 19:58:08 +0200R$ From: Michael Unger <unger@decus.de>, Subject: Re: attn: Paddy (re: DCL procedure)+ Message-ID: <00A12C0C.5E3A0BE1.13@decus.de>d   [...]i  I > I knew that my dedicated zzz sub-domain would eventually go, but our IT F > (accountants and economists -- no IT background in sight) decided to > pull the plug on August 9th.   [...]e  B > I have already apologised to Mark Berryman for the imposition onF > info-VAX for all the bounces, which were beyond my control.  For theG > several people I have enjoyed private mail with, my address no longero" > has zzz. in it (as shown above). > H > Michael, I am still working on getting my set-up correct and trying toE > connect this stupidity with my previous VMS mail, and I shall be in 
 > touch ASAP.,   [...]   O I tried again using Paddy's new address (i.e., without "zzz") and didn't get antM immediate reply from "postmaster" so I guess the mail got delivered to him. IbG don't know how he can manage to "forward" the mail (DCL procedure) fromeO "Outlook" (my rather old dictionary describes this as "philosophical conceptionh$ of the world") to the VMS system ...   Michael,   ------------------------------  % Date: Tue, 20 Aug 2002 11:20:53 -0400.* From: "Stanley F. Quayle" <stan@stanq.com>E Subject: Re: Charon-VAX (was: [VAX] VMS to [Alpha] OpenVMS conversion . Message-ID: <3D622615.30426.FEB3096@localhost>  ) On 19 Aug 2002 at 14:18, Nic Clews wrote: E > Hum. Err I need to be delicate saying this, because we've a special E > evaluation of this product going ahead at the moment, but the power G > supply in the "system who's maker shall remain nameless, but could ben( > relatively easy to guess" has failed !  B That's why I recommend "server class" PC's.  Dual power supplies, F hardware RAID, etc.  If the client buys an unbrand clone, it could be  a problem...  His problem.
 --Stan Quayleo! President, Quayle Consulting Inc.i  
 ----------C Stanley F. Quayle, P.E.   N8SQ   +1 614-868-1363   Fax: +1 614 868-  16711 8572 North Spring Ct. NW, Pickerington, OH  43147 = Preferred address:  stan@stanq.com       http://www.stanq.comn   ------------------------------   Date: 20 Aug 2002 15:47:03 GMT, From: bill@gw5.cs.uofs.edu (Bill Gunshannon)E Subject: Re: Charon-VAX (was: [VAX] VMS to [Alpha] OpenVMS conversiont6 Message-ID: <ajto9n$1eeu7m$1@ID-135708.news.dfncis.de>  K In article <rdeininger-1908022203190001@1cust150.tnt2.nashua.nh.da.uu.net>,o5 	rdeininger@mindspring.com (Robert Deininger) writes:nP > In article <ajqlvb$1cqirq$2@ID-135708.news.dfncis.de>, bill@cs.uofs.edu wrote: >  > G >>I don't know how to break this to you, but if your talking Win2K hereeG >>it isn't stable when it has nothing running on it.  I have a lab full G >>of them. After the students left back in June, they were just sitting E >>there all by themselves.  Inside of 30 days every one of them had a*G >>message on the screen informing someone that they had run dangerously F >>low of Virtual Memory.  Within the next 15 days, they all ran out ofG >>virtual memory and locked up.  A memory leak that big hardly fits then! >>definition of a stable machine.e > I > If you turn off the power, Win2K won't leak memory.  It's actually veryd > stable in that configuration.r   I stand corrected.  8 I have tested and verified that this is in fact correct.  G Now, if I could just convince people here.........                  :-)y   bill   -- mJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>   l   ------------------------------    Date: 20 Aug 2002 02:44:17 -0700" From: ccc_crg@yahoo.com (Chris G.)= Subject: Configuring webDAV for CSWS T1.3 for Dreamweaver MX?e= Message-ID: <5437ff2f.0208200144.6bcc45c4@posting.google.com>o  B I'm attempting (and failing miserably) to get webDAV working under	 CSWS T1.2   = I have included the module_dav  and with defining the logicaleF APACHE$CONFIG_VERBOSE shows the dav modules as being loaded at startup time.   0 Here is what I have put into my httpd.conf file:   ### WEBDAV conf>" DAVLockDB /apache$root/var/DAVLock DAVMinTimeout 600e #L$ <Location /apache$virtualhosts/xxxx>
     DAV On     AuthType Basic'       AuthName "OpenVMS authentication"i       AuthOpenVMSUser On       require valid-user     AllowOverride None     Options NoneB     <Limit PUT POST DELETE PROPFIND PROPPATCH MKCOL COPY MOVE LOCK UNLOCK>t         Require user chris     </Limit> </Location>s #r" ##################################/ (xxxx is the folder where my website is setup.)x Now...D I have created an apache$root:[var] folder, but nothing has appeared	 in there.a= When I try to use Dreamweaver MX to connect it responds with:r Connecting to xxxx+ Retrieving remote folder information for /.>$ Could not understand server response9 An error occurred - cannot get remote folder information.e   In my error.log file I have: ---yC [Tue Aug 20 21:12:04 2002] [error] [client xxx.xx.xx.xxx] File does + not exist: /apache$virtualhosts/xxx/_notes/  ---e   In my access.log file I have:r ---tD xxx.xx.xx.xxx - - [20/Aug/2002:21:41:07 +1200] "PROPFIND / HTTP/1.1" 200 587) ---   - (where xxx.xx.xx.xxx is my client Ip address)e  E I have to admit, I don't know much about CSWS (I've only been running F it for a few weeks), and know nothing at all about webDAV. I have read5 the CSWS T1.3 Release Notes and checked their link tosE http://www.webdav.org/mod_dav/ but I'm still pretty much in the dark,r I'm afraid.e  + Any asistance would be greatly appreciated!o   Chris G.   ------------------------------    Date: 20 Aug 2002 09:36:47 -07001 From: glenmark@utxvms.cc.utexas.edu (Glen Martin)u) Subject: DE101 - can it go to Full Duplexa= Message-ID: <6e2f14f4.0208200836.22267f6f@posting.google.com>   E I have a couple of old AlphaServer 2100a systems running OpenVMS v6.2mD which appear to have DE101 NICs (at least that is what is stamped on? the outside of the cards). CLUE CONFIG lists them only as TuliprD adapters. I've been trying to set them to Full-Duplex mode, both viaD LANCP and via the EWA0_MODE environment variable within SRM, but theD cards aren't going to full (the switch to which they connect are setD to Full Duplex). Am I fighting an uphill battle? Do these cards evenD support Full Duplex? Thus far, I've been unable to find this info at Compaq's website....   TIAt Glen   ------------------------------    Date: 20 Aug 2002 09:51:20 -0700* From: krblock@computer.org (Kenneth Block)1 Subject: Re: DECCXX 6.5 Unreachable code warningso= Message-ID: <d3999c0b.0208200851.1ba89eb0@posting.google.com>t  C The reason you are not getting the diagnostic for the "BAD" case isn thatB define template will not instantiate an inline member function. An inlineF member function will get inlined when it is used, so there is no point toB also instantiate it seperately. You will get the diagnostic if you; were to actually use it. Add the following to your program:    #ifdef REALLY_BAD  void main()t {    Mango<char[1]> a("a");   Mango<char[2]> b("b");     a.input_str(cin,a);a   b.input_str(cin,b);a }a #endif   and then compile:a  # $ cxx/define=(bad,really_bad) t.cxxa           x = 1;	 ........^ ; %CXX-W-CODCAUUNR, statement either is unreachable or causest unreachablea           code7 at line number 56 in file C2$:[BLOCK.TEST.PLAY4]T.CXX;3u           x = 0;	 ........^ ; %CXX-W-CODCAUUNR, statement either is unreachable or causese unreachable            code7 at line number 54 in file C2$:[BLOCK.TEST.PLAY4]T.CXX;3     B As for why the diagnostic comes out sometimes, but not others mostD likely has to do with it being hueristically driven. I think that itD is fair to say that you are more interested in seeing the diagnosticD always off, then understanding why it is sometimes on. Atleast I am.  - A problem report (CPP#5707) has been entered.   _ "John E. Malmberg" <wb8tyw@qsl.network> wrote in message news:<3D619D61.7040506@qsl.network>...u > Stewart Van Kirk wrote:c< > > "John E. Malmberg" <wb8tyw@qsl.network> wrote in message, > >   news:<3D5D2B58.3040902@qsl.network>... > > 
 > >><snip>I > >>sizeof() is evaluated at compile time.  Therefore the code path that n > >>follows will never change. > > E > > OK, I accept that because I'm using the pragma to define specifice/ > > instances of the template, e.g. T=char [1].l > >   G > > Would you have an insight as to why the compiler does not issue theII > > same warnings in the other two cases, BAD and UGLY?  The UGLY case issD > > especially perplexing because the code is modified in a separateH > > control flow path; as such, I would not expect any difference in the > > compiler output. > K > I really do not know why the CXX compiler is not issuing diagnostics for e > the BAD and UGLY variants. > D > One thing to do with each variant is to compile with machine code J > generated in the listings, and compare the machine code.  This may help ; > show why the compiler is not giving the same diagnostics.b > C > The inline modifier may have something to do with it, or may not.  >  > No offense intended but:C > The long all caps title on your post may have caused some people  K > browsing this newsgroup/mailing list to miss that this was a C++ posting.i > G > What shows up on many newsreaders is "REQUEST FOR CLARIFICATION" and tH > almost all posts with all capital letters in the subject are spam, so J > they are generally ignored or discarded by spam filters.  That specific I > title is currently very similar to one used for the Nigerian 419 scams.* > I > Some C/CXX developers scan this newsgroup from time to time, so I have *; > retitled this note to see if it will get their attention.n >  > -Johne > wb8tyw@qsl.network > Personal Opinion Onlyf > < > ------------- Original Post ------------------------------ >  > E > The DECCXX 6.5-004 compiler on OpenVMS 7.3 is confusing me with itsaD > warning messages about unreachable code.  I do not know whether to > take them seriously or not.c > A > The following code illustrates that compiled one way, I get two B > unreachable code warnings and if I slightly modify the code [notH > changing the operational meaning] in three (probably more) other ways,H > I do not get the warnings.  In fact, two of the three "other ways" areE > functionally-isomorphic to the code that produces the warnings.  IsfH > there a substantive problem here, or is the compiler just a little too > smart? >  > The compiler output:D > ------------------------------------------------------------------ > - > $cxx mango.cxx/define=Bad      ! no problem ( > $cxx mango.cxx/define=Ugly     ! ditto$ > $cxx mango.cxx/define=Good     ! "' > $cxx mango.cxx                 ! oops8 >  >          x = 0;O > ........^.= > %CXX-W-CODCAUUNR, statement either is unreachable or causes 
 > unreachabled >            code 5 > at line number 53 in file DKA100:[FRUIT]MANGO.CXX;1o >  >          x = 1;m > ........^o= > %CXX-W-CODCAUUNR, statement either is unreachable or causes 
 > unreachablen >            codeo5 > at line number 55 in file DKA100:[FRUIT]MANGO.CXX;1  > D > ------------------------------------------------------------------ > The code: D > ------------------------------------------------------------------ > #include <iostream.h>m >  > template <class T>
 > class Mango  > {b	 > public:h >    Mango(char *nt_str);l >  > #ifdef BAD@ >    inline static istream& input_str(istream& is, Mango<T>& s); > #elsee9 >    static istream& input_str(istream& is, Mango<T>& s);o > #endif > ; >    friend istream& operator>>(istream& is, Mango<T>& s) {* >      return input_str(is, s);  >    } >  > protected: >    T value_; > }; >  > template <class T>- > Mango<T>::Mango(char *str) { *this = str; }  >  > template <class T>9 > istream&  Mango<T>::input_str(istream& is, Mango<T>& s)t > {m >    int x = 0;e >    if (!is.eof()) {t >      if (is.peek() == '"') { >        int i;n! >        char tmpchar = is.get();,
 > #ifdef UGLYo >        for (i=0;5 > 	   i < sizeof(T) && (tmpchar = is.get()) != EOF && 4 >                  '\n' != tmpchar && '"' != tmpchar > 	   ; i++) > 	s.value_[i] = tmpchar;t > #elseh( >        for (i=0 ; i<sizeof(T) ; i++) { > 	tmpchar = is.get();+ > 	if (tmpchar == EOF || tmpchar == '\n' ||r > 	    tmpchar == '"') break;s > 	s.value_[i] = tmpchar;u
 >        } > #endif >      }
 >      else {n
 > #ifdef GOOD % >        x = sizeof (T) <= 1 ? 0 : 1;* > #else* >        if (sizeof(T) <= 1)	 > 	x = 0;e
 >        elsex	 > 	x = 1;r > #endif >      } >    }
 >  return is;- > }- > * > #pragma define_template Mango<char[ 1 ]>* > #pragma define_template Mango<char[ 2 ]>   ------------------------------    Date: 20 Aug 2002 09:59:46 -0700* From: krblock@computer.org (Kenneth Block)1 Subject: Re: DECCXX 6.5 Unreachable code warnings = Message-ID: <d3999c0b.0208200859.5952a542@posting.google.com>e  O You can always turn the message off using pragma message if you do not like it:p   $ cxx t2
 $ type t2.cxx  template <class T> void input_str(T& s) { 
    int x = 0;l  &    for (int i=0 ; i<sizeof(T) ; i++) {      if (s > 0) break;    }   #pragma message save! #pragma message disable codcauunre    if (sizeof(T) <= 1)      x = 0;     elseb      x = 1;o #pragma message restoree }f   void main()e { 	   char x;o	   int  y;t     input_str(x);h   input_str(y);i }t   ------------------------------  # Date: Tue, 20 Aug 2002 13:30:22 GMTe1 From: "Terry C. Shannon" <terryshannon@attbi.com>m0 Subject: Re: Fortune Magazine and a post-VMS rap. Message-ID: <Ovr89.110758$me6.15386@sccrnsc01>  7 "James Cameron" <james.cameron@hp.com> wrote in message:3 news:pan.2002.08.20.13.16.07.884790.24834@hp.com...z5 > On Tue, 20 Aug 2002 03:54:18 +1000, JF Mezei wrote:dI > > Since Compaq is gone, and HP has its own user group, wouldn't it make H > > sense to return everything back to normal and call a spade a spade : > > DECUS ?  >r
 > Hear, hear.. > C > I'm a DECUS member in Australia.  Was so before I joined Digital.eC > I'd like to keep the name!  As far as I know, we are still DECUS.   I We in Encompass mulled over the idea. Reason we scuttled DECUS is that to L some of the Big Dogs in Houston, DECUS = VMS, which was not a Good Thing. (IH think VMS is a good thing, a lot of folks in Houston certainly did not!)   cheers,    terry so   ------------------------------    Date: 20 Aug 2002 09:08:10 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)s0 Subject: Re: Fortune Magazine and a post-VMS rap3 Message-ID: <LBA2NeTG41Jg@eisner.encompasserve.org>   b In article <Ovr89.110758$me6.15386@sccrnsc01>, "Terry C. Shannon" <terryshannon@attbi.com> writes: > 9 > "James Cameron" <james.cameron@hp.com> wrote in messageh5 > news:pan.2002.08.20.13.16.07.884790.24834@hp.com... 6 >> On Tue, 20 Aug 2002 03:54:18 +1000, JF Mezei wrote:J >> > Since Compaq is gone, and HP has its own user group, wouldn't it makeI >> > sense to return everything back to normal and call a spade a spade :- >> > DECUS ? >> >> Hear, hear. >>D >> I'm a DECUS member in Australia.  Was so before I joined Digital.D >> I'd like to keep the name!  As far as I know, we are still DECUS. > K > We in Encompass mulled over the idea. Reason we scuttled DECUS is that to N > some of the Big Dogs in Houston, DECUS = VMS, which was not a Good Thing. (IJ > think VMS is a good thing, a lot of folks in Houston certainly did not!)  / Perhaps the thinking in Cupertino is different.o   ------------------------------  % Date: Tue, 20 Aug 2002 17:54:06 +0100e* From: Andrew Robinson <arobinson@hspg.com>Y Subject: RE: forwarding email from system to system (was: RE: attn: Paddy	(re: DCL procedoM Message-ID: <CDA4BAD1E10ED41181AC00508B6051D3C3E8CA@grumpy.internal.hspg.com>o  G We have an Exchange server as our Mail backbone (Sigh). How to direct =  mail= to the OVMS box depended on how the OVMS box is addressed.=20 H We use MX mail on our OVMS boxes http://www.madgoat.com/ configured as = anI internal sub domain of our main domain. MX records were created for the =R mailE alias of the internal domain in our internal DNS server (OVMS Clustery	 members).-I In the Internet Connection within Exchange, the sub domain was defined, =h soF any mail for the sub domain was sent to the MX alias. The mail going = out-F from OVMS just used the Exchange Box, which in turn sends it out and = about.  C If you want to have all OVMS users and Exchange users on the same =i domain, / then it gets only slightly more complicated.=20uD You still need to do the above except you create a totally private = internal sub domain like vmsbox.xyz.comF You would have to create a custom internet recipient on Exchange for = everyvG user on the OVMS system, create a second internet Address with the OVMSs8 users public address and set it to be the reply address.D On the OVMS server you need to use an address rewriter table for allC outgoing emails so the internal address is replaced by the public =e address. (Again MX mail does this)e  G BTW - Don't use full exchange mailboxes, otherwise you have to create =I jobsF etc just to clear down all the duplicate emails, which can appear in = bothF exchange & OVMS, if you rely on forwarding rules, these can go wrong = during 'patches / service packs'.  0 The Political problem I can't help you with ....   Regards-   Andrew Robinson-   -----Original Message-----2 From: Jan-Erik S=F6derholm [mailto:aaa@aaa.com]=20 Sent: 20 August 2002 15:34 To: Info-VAX@Mvb.Saic.ComlE Subject: Re: forwarding email from system to system (was: RE: attn: => Paddy14 (re: DCL procedure)) DCL procedure)) DCL procedure))    9 I'v seen MAJOR problems, both political and technical, tol9 get mail "down" to VMS systems where one have selected tooD use Exchange as mail backbone. Usualy ends up with creating ExchangeB mailboxes for each VMS user, and adding forwarding rules in each =
 mailbox to" have the mail sent back and forth.   Don't know Lotus...c   Jan-Erik S=F6derholm.u   Phillip Helbig wrote:  >=20I > > If anyone can tell me how to get Exchange to deliver mail to me VMS =   F > > account and still appear to be from the person who actually sent = it,=20 > > I'd be grateful. >=20 > Ditto for Lotus Notes.  :-)    ------------------------------  % Date: Tue, 20 Aug 2002 18:04:13 +0100-* From: Andrew Robinson <arobinson@hspg.com>Y Subject: RE: forwarding email from system to system (was: RE: attn: Paddy	(re: DCL procedXM Message-ID: <CDA4BAD1E10ED41181AC00508B6051D3C3E8CB@grumpy.internal.hspg.com>w  F Having just read the original email (Ignore previous email on subject)   <snip>   Phillip Helbig wrote:d > H > > If anyone can tell me how to get Exchange to deliver mail to me VMS I > > account and still appear to be from the person who actually sent it, r > > I'd be grateful. >  > Ditto for Lotus Notes.  :-)n  G If you have an existing exchange mailbox, delete it and create a Custom.J internet recipient pointing to your OVMS mailbox. Create a second internetK address on the exchange recipient with the correct external address and sets to be the reply address.H On the OVMS box set your mail reply address to be your external address.L All incoming mail will show who sent it, and all outgoing mail will have the right address.   Regardse   Andrew Robinsone   ------------------------------  + Date: Tue, 20 Aug 2002 15:57:29 +0100 (MET)l9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>cY Subject: forwarding email from system to system (was: RE: attn: Paddy (re: DCL procedure)n; Message-ID: <01KLIYNNV78U970ARJ@sysdev.deutsche-boerse.com>-  E > If anyone can tell me how to get Exchange to deliver mail to me VMSoJ > account and still appear to be from the person who actually sent it, I'd > be grateful. a   Ditto for Lotus Notes.  :-)a   ------------------------------  % Date: Tue, 20 Aug 2002 16:33:55 +0200-9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>.Y Subject: Re: forwarding email from system to system (was: RE: attn: Paddy (re: DCL proced ' Message-ID: <3D625353.D8473684@aaa.com>g  9 I'v seen MAJOR problems, both political and technical, toD9 get mail "down" to VMS systems where one have selected tod; use Exchange as mail backbone. Usualy ends up with creatingw; Exchange mailboxes for each VMS user, and adding forwardingi; rules in each mailbox to have the mail sent back and forth.>   Don't know Lotus...t   Jan-Erik Sderholm.e   Phillip Helbig wrote:a > G > > If anyone can tell me how to get Exchange to deliver mail to me VMSnL > > account and still appear to be from the person who actually sent it, I'd > > be grateful. >  > Ditto for Lotus Notes.  :-)P   ------------------------------  % Date: Tue, 20 Aug 2002 09:20:39 -0700r( From: Alan Frisbie <Abuse@NelsonUSA.com>Y Subject: Re: forwarding email from system to system (was: RE: attn: Paddy (re: DCL procedo- Message-ID: <3D626C57.88344946@NelsonUSA.com>r   Jan-Erik Sderholm wrote:n  ; > I'v seen MAJOR problems, both political and technical, tod; > get mail "down" to VMS systems where one have selected toa= > use Exchange as mail backbone. Usualy ends up with creatingc= > Exchange mailboxes for each VMS user, and adding forwardingo= > rules in each mailbox to have the mail sent back and forth.t  ; I got lucky here.   Just inside the firewall is a Linux box/; running Sendmail.   It acts as a gateway for all mail, bothx< incoming and outgoing.   I just edited the /etc/aliases file: to forward all e-mail addressed to my Exchange account, to: my VMS account.   Check out what you have there; you might
 get lucky.   Alan   ------------------------------  % Date: Tue, 20 Aug 2002 12:49:32 +0100h% From: Alan Greig <a.greig@virgin.net>h* Subject: Re: From Today's Network World...8 Message-ID: <j2b4mucbsnlp032cl6vj9nmdjgal7csd8g@4ax.com>  4 On Mon, 19 Aug 2002 22:59:13 GMT, "Terry C. Shannon" <terryshannon@attbi.com> wrote:-   >a > L >But it will never happen, Art. Just ask all the "experts" in this group who, >are still whining about the death of Alpha. >s( >See ya in a less-contentious newsgroup!  B Hey, you are the one telling us that Capellas is so unstable he isC likely to cancel VMS if customers complain to him that they fear hei will do just that.     -- Alan   ------------------------------  % Date: Tue, 20 Aug 2002 09:35:58 -0400r2 From: Atlant Schmidt <atlantnospam@mindspring.com>6 Subject: Re: How to "forward declare" a function in c?. Message-ID: <3D6245BE.A54D91B7@mindspring.com>   JF Mezei wrote:e   > Lyndon Bartels wrote:., > > How do you "forward declare" a function? >v; > int cook_cake(short temperature, unsigned long minutes) ;h >i9 > int cook_cake(short temperature, unsigned long minutes)- > {-* > printf("Waiting for oven to warm up\n");. > printf("Oven warmed up, starting baking\n");( > printf("Time elapsed, cake cooked\n"); > return(1); > }D  5 Expanding on JF's example, in the forward declarationh6 of the function, you *should* use parameter names as a6 sort of self-documentation of what the parameters are.  4 But I'll note that while it's conventional and "good0 practice" to use the same parameter names in the4 forward declaration as in the real declaration, it's* not enforced anywhere. So this also works:   /* Forward routines */  0 int cook_cake(short deg_f, unsigned long mins) ;  7 int cook_cake(short temperature, unsigned long minutes).  {,     printf("Waiting for oven to warm up\n");0     printf("Oven warmed up, starting baking\n");*     printf("Time elapsed, cake cooked\n");     return(1);   }     5 Also, as I hinted above, it's a good idea to indicate 4 with a comment that these are, in fact, just forward> declarations. It's also nice to use the "Forward declarations"4 section as a module Table-of-Contents, listing *ALL*5 the routines, even if they aren't used ahead of theirO5 real declarations. Placed near the top of the module,o7 this provides another nice bit of "self documentation".w   Atlant    : P.S.: By the way, JF, "unused parameters in function!" :-)   ------------------------------    Date: 20 Aug 2002 01:50:05 -0700' From: amits_77@yahoo.com (Amit Sawhney)I/ Subject: How to configure Reflection X-server ?r= Message-ID: <71939fd6.0208200050.26287378@posting.google.com>e   Dear VMS Friends,0  F I m trying to configure reflections-x on my W2kPro.I m able to see theE reflection X root window but not able to see anything more than that.tA Do i need to run any application on my VMS server before startingmF Reflections-X ? or is there any problem in configuration of reflection x suite.   thanks in advance  Amit   ------------------------------  % Date: Tue, 20 Aug 2002 11:11:03 +0200h% From: "Jakob Erber" <erberj@yahoo.de> 3 Subject: Re: How to configure Reflection X-server ? % Message-ID: <3d6207a7$1@news.post.ch>   C > Do i need to run any application on my VMS server before starting  > Reflections-X ?    NoI In the Reflexion X main window, you should see on the left side a list oftH Operating system templates, which contains also an entry for VMS. Modify this one for your needs.L You can also adapt one of the session manager entries, to display a complete8 VMS Session Manager on your PC screen, thats, what I do.   regardsm   Jakobm   ------------------------------  + Date: Tue, 20 Aug 2002 06:57:14 -0700 (PDT)n. From: Fabio Cardoso <fabiopenvms@yahoo.com.br>) Subject: HP buying EMC. Is that true ????o@ Message-ID: <20020820135714.23956.qmail@web20201.mail.yahoo.com>  " A Field Technician from HP/Compaq  told me about HP buying EMC.   Is this true ??s   Regardsi   FC     =====s ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazile fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?& HotJobs - Search Thousands of New Jobs http://www.hotjobs.com   ------------------------------  % Date: Tue, 20 Aug 2002 09:34:13 +0100eU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>w+ Subject: Re: HP-Compaq Merger Went Smoothly 0 Message-ID: <ajsuvl$o4o$1@new-usenet.uk.sun.com>   Rob Young wrote:  p > In article <wfc89.79057$Fw3.3731014@bin2.nnrp.aus1.giganews.com>, "Bill Todd" <billtodd@metrocast.net> writes: > K >>Sun's only major low-end cost/performance competition is the same as it'siJ >>been all along:  IA32, augmented soon by Hammer (both of which of courseN >>compete with the other 64-bit entrants at least as effectively).  With AlphaM >>and PA-RISC at least somewhat off the table, Sun now only need compete withhI >>POWER4 and Itanic2 (well, I guess MIPS should at least be mentioned) ateK >>higher levels - and so far Sun has appeared more than capable of offeringj, >>attractive cost/performance in that arena. >> >> > B > 	Depends on how and what you measure.  For instance, if for some@ > 	reason you are inclined to purchase equivalent performing SunC > 	versus IBM box to run Oracle, you are going to spend a whole lotc@ > 	more money on Sun kit (and will you be able to get there with? > 	a single U15K?).  If however, you decide a "32 CPU box" fitscA > 	the bill, you will get a much more powerful IBM solution.  As f > 	a for instance: >       E > http://www.tpc.org/tpcc/results/tpcc_result_detail.asp?id=102081501i > @ > 	32 CPUs does over 400K tpmC.  Sun will let you guess at theirH > 	numbers as they have decided tpmC is no longer important (they can't  > 	compete basically). >     < But then do you really think that TPC-C is a usefull measure? that can be used to work out how big a system you might requireh9 from two competing vendors to run your DBMS application ?c  ; If you can provide a reasoned and thought out answer to thet< question, TPC-C is it a usefull measure and why then I might< be more inclined to listen to you however it very much doubt you can.   Answers on a post card      = > 	Now before Andrew calls me an IBM apologist , keep in mind D > 	Marvel will deliver very good tpmC numbers (similar to Regatta no
 > 	doubt). >     > Actually I was thinking no S**T Rob's making yet another TPC-C; prediction for a yet to be released AlphaServer, how dumb !       B > 	I'm not quite sure what Sun excels at today compared to IBM, ifC > 	there is something, I haven't found it (and/or they probably use > > 	twice as many CPUs to get there and if third-party softwareC > 	charges per-CPU - i.e. Oracle - they are quite unattractive with A > 	that tagging along).  How could Sun get "better?"  Ship a more @ > 	powerful CPU.  They really are quite underpowered compared to# > 	Itanium 2, Power4 and EV7 vapor.h >   F Well look at the other big TPC benchmark TPC-H. And there you go againE Itanium 2 has no DBMS benchmarks to its name that anyone considers to*D be usefull except possible you and SPARC and Power 4 are pretty much@ even stevens when it comes to DBMS benchmarks. We don't do TPC-CA they don't do TPC-H (for understandable system design issues) etct etc.  @ The claim that SPARC is unatractive on DBMS licensing standpoint= because you claim we need 2x the CPU's is one that you cannotj> justify, remember it is the claim you have always made and you@ have never been able to justify it yet, so cut the BS its boring# and makes you look serially stupid.a     > N >>In sum, Sun does not appear to be at any sustained cost/performance, or evenL >>absolute performance (in multi-thread server loads), disadvantage comparedM >>with anything except IA32 and Hammer (too bad Alpha doesn't make that list,n >>for non-technical reasons).  >> > ; > 	Sun is at a considerable disadvantage versus IBM on manyd: > 	per-CPU sensitive metrics (again, Oracle for instance). >     : Which ones. SPECint and fp well if you only use 1 of the 88 CPU's in a Regretta MCM module to get access to all that9 extra cache then yup it does great SPECint and SPECfp butt6 nobody is going to buy an 8 CPU Regretta let alone a 8% CPU Regretta with only 1 running CPU.o  8 TPC-C wow who cares it doesn't measure anything usefull.   Regards.   Andrew Harrisong   ------------------------------  % Date: Tue, 20 Aug 2002 10:37:53 +0100sU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>n+ Subject: Re: HP-Compaq Merger Went Smoothly 0 Message-ID: <ajt2n1$pcc$1@new-usenet.uk.sun.com>   Fred Kleinsorge wrote:  8 > Andrew Harrison SUNUK Consultancy wrote in message ... >  >>1 >>Which is of course why Sun's stock is currentlya3 >>at 4-5 dollars rather than 20. We had 15 years ofi2 >>profit and revenue growth followed by 2 quarters! >>of revenue declines and losses.n >> >> >    > ? >>The revenue declines you have seen so far are probably mostly-= >>customer sentiment not the fact that they can no longer buym3 >>the product range that they used to buy from you.R >>< >>Wait for that before announcing that the merger has been a	 >>sucess.m >> >> > K > This is simply random guessing - "I think you are doing badly because youy@ > walk funny, your mother dresses you bad, and calls you Sally". >     8 Well perhaps you can come up with an explanation for the5 larger than market declines in your server businessesJ@ with a pretty much intact product ranges, support organistations$ and technical/engineering resources.    K > I can "guess" that Sun is going into a stock decline and has taken a huge C > hit on it's revenue and profits because Linux is eating into youreL > traditional technical UNIX market, the UNIX market future is in doubt withK > pressures from both Linux and Windows, the future of Sparc is making yourCM > customers skittish - since it's unlikely that you will fix your performance M > problems soon, your margins have noplace to go but down, the customers thatyK > drove your growth in recent years are part of the bubble that burst, etc,-K > etc, etc. - causing you to "restructure" - i.e. fire employees to containb > costs. >     G Windows has been a "pressure" on the UNIX market since the introductiont@ of NT. The fact that it hasn't show any signs of erroding UNIX'sC dominance of the datacenter is mostly because Windows 2000 is stillrA as far away from the capabilities of any of the commercial UNIX'soH now as NT was when it was introduced. In addition some of the incursionsA that Windows made into the datacenter on the back of applicationsdC vendors like SAP and Seibel have been reversed by these vendors uselE of web services based technolgies. Seibel apps server and client onlyeD used to run on Win32 with a DBMS on virtually anything. Seibel 7 andE later runs in a browser with an apps server tier that runs on SolarissD Win32 and AIX and a DBMS server tier which is supported on Solaris,  Win32, AIX etc.o  6 I have been involved in 3 very large Seibel 7 projectsA all of which have involved deploying UNIX apps servers to replaceyF win32 apps servers that were being used for the older Seibel versions.  ? In all 3 cases the decision to deply on UNIX for the apps layer @ was made without any input from Sun, IBM etc but on the basis of= the customers experiences with their existing infrastructure.r  @ People on this group have made the point that it is difficult to= wean people of Windows and this is very true for the desktop.r< However the rash of Virii, denials of service etc which have@ infected Windows server products such as IIS and exchange server< have made people more reluctant to use it as a datacenter OS where there are substitutes.  ; As for Linux, it currently as the market share numbers show % has a very small share of the market.i  ? In addition what ever you may surmise about Sun losing business : to Linux in our "traditional" technical market the fact is< that Sparc Solaris's market share is still increasing as are: revenues after the downturn. If Linux is stealing business< from Sun and we do have our own Linux boxes it isn't obvious> from the overall numbers which suggest that only SPARC/Solaris> and Linux are growing market share at the expense of the other vendors commercial OS's.  > In many ways this is exactly what happened when NT entered the9 server space, everyone said it would kill Solaris/AIX etcu6 but instead the 3 major commercial UNIX's Solaris, AIX: and HP-UX grew share and revenues as did NT at the expenseA of the smaller players such as Tru64, NCR, Pyramid, OpenVMS, IRIXr etc.  > So if we as some analysts predict end up with Solaris, another< commercial UNIX and Linux plus Windows as the players in the: server market then this will not be suprising and Sun will not be badly positioned.  9 HP's problem is trying to ensure that the 2and commercial-5 UNIX is HP-UX and not AIX something that currently ism8 very much in doubt. HP isn't a Linux technology provider6 though you do bend tin that happens to run Linux, this6 without HP-UX in the frame would put you very low down1 in the Linux food chain, not a great place to be.j      M > The genesis of this note was that someone said that HP's stock decline is aeI > result of it's merger activity, and that Sun's stock was moving higher.mM > BOTH statements are misleading or false.  I *will* agree with you that IMHO M > it is *too early* to determine how well the merger has worked.  The successhM > or failure of the merger is probably not a major factor in the price of thepH > stock since the merger.  Nor has Sun stock performed well in that same > timeframe. >     < Funnily enough I agree with you. HP's stock has outperformedB Sun's. Though you do have a highly profitable printing consumables< unit which we don't and for that reason the numbers look OK.  = And you are right the merger really hasn't started to pan outw> from a product standpoint. The web sites have been merged, the? sales forces all have the same cards etc but with the exceptionv? of the PC's which was always the easiest product merge very fews? of the product ranges have been rationalised and only when thisK0 happens will the merger proper be in full swing.  9 At the moment you are in the honeymoon period, people canm3 still call HP any buy the products they used to buys8 from either HP or Compaq, wait until that isn't the case5 anymore. The fact that they arn't in the numbers that  they used to is interesting.  4 HP also appear to have made a huge technology gamble8 in IA-64 which to date does not appear to be paying off.  2 Getting off that horse if it stumbles will be very6 costly and for that fact alone I would mark HP's stock down.,  7 Sun on the other hand doesn't have any of these issues.s  < Persistent claims of huge CPU performance gaps between SPARC/ and other CPU's have made by you and the choir.x  D Which you will also recollect that you and the choir have never beenI able to substantiate, although you do appear to beleive your own retoric.r    < Sun's market share is growing in servers but also in storage  : as well and our cladding rate for services and software on9 our servers is also increasing. Despite claims that Linuxd5 is eating Solaris share in fact Linux and Solaris are : eating AIX/HP-UX/Tru64 share, not something that is of any comfort to IBM or HP.l  ; I am not a stock analyst but given the fundamentals of both 6 busineses I know which one I would bet on for the next 6-12 months.   Regards5 Andrew HarrisonU   ------------------------------  % Date: Tue, 20 Aug 2002 07:58:01 -0400o' From: "Main, Kerry" <Kerry.Main@hp.com>o+ Subject: RE: HP-Compaq Merger Went SmoothlyaT Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF402660941@kaoexc01.americas.cpqcorp.net>   Andrew,M  D >>> Persistent claims of huge CPU performance gaps between SPARC and+ other CPU's have made by you and the choir.   D Which you will also recollect that you and the choir have never been@ able to substantiate, although you do appear to believe your own retoric.>>>>  E It simply means when evidence of this is presented to you, you always , need to find a way to introduce a yeahbut ..   As an example:= http://ideasinternational.com/benchmark/spec/specjbb2000.html2( HP Superdome number 1 rating - 614358=20% Sun Fire 15K number 2 rating - 602270F  A Now, on the surface, this seems like a small difference until oneaG notices that the Superdome number was done with 64 Cpu's, while the Sun  15K was done with 104 cpu's.   :-)a   Regardsi  
 Kerry Main Senior Consultantk Hewlett-Packard Canada! Consulting & Integration Serviceso Voice: 613-592-4660. Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----' From: Andrew Harrison SUNUK Consultancye7 [mailto:andrew_nospam.harrison_remove_this@sun#.com]=20e Sent: August 20, 2002 5:38 AM  To: Info-VAX@Mvb.Saic.Com-+ Subject: Re: HP-Compaq Merger Went Smoothly5         Fred Kleinsorge wrote:  8 > Andrew Harrison SUNUK Consultancy wrote in message ... >=20 >>1 >>Which is of course why Sun's stock is currentlyn3 >>at 4-5 dollars rather than 20. We had 15 years ofw2 >>profit and revenue growth followed by 2 quarters! >>of revenue declines and losses.t >> >> >=20   >=20H >>The revenue declines you have seen so far are probably mostly customer  I >>sentiment not the fact that they can no longer buy the product range=20t! >>that they used to buy from you.e >>D >>Wait for that before announcing that the merger has been a sucess. >> >> >=20J > This is simply random guessing - "I think you are doing badly because=20D > you walk funny, your mother dresses you bad, and calls you Sally". >=20    8 Well perhaps you can come up with an explanation for the5 larger than market declines in your server businessesdD with a pretty much intact product ranges, support organistations and  technical/engineering resources.    I > I can "guess" that Sun is going into a stock decline and has taken a=20 H > huge hit on it's revenue and profits because Linux is eating into your  J > traditional technical UNIX market, the UNIX market future is in doubt=20G > with pressures from both Linux and Windows, the future of Sparc is=20eH > making your customers skittish - since it's unlikely that you will fix  H > your performance problems soon, your margins have noplace to go but=20H > down, the customers that drove your growth in recent years are part of  H > the bubble that burst, etc, etc, etc. - causing you to "restructure" -  ' > i.e. fire employees to contain costs.u >=20    G Windows has been a "pressure" on the UNIX market since the introduction:@ of NT. The fact that it hasn't show any signs of erroding UNIX'sF dominance of the datacenter is mostly because Windows 2000 is still asH far away from the capabilities of any of the commercial UNIX's now as NTC was when it was introduced. In addition some of the incursions thatlD Windows made into the datacenter on the back of applications vendorsB like SAP and Seibel have been reversed by these vendors use of webF services based technolgies. Seibel apps server and client only used toG run on Win32 with a DBMS on virtually anything. Seibel 7 and later runscH in a browser with an apps server tier that runs on Solaris Win32 and AIX8 and a DBMS server tier which is supported on Solaris,=20 Win32, AIX etc.f  6 I have been involved in 3 very large Seibel 7 projectsG all of which have involved deploying UNIX apps servers to replace win32s@ apps servers that were being used for the older Seibel versions.  H In all 3 cases the decision to deply on UNIX for the apps layer was madeE without any input from Sun, IBM etc but on the basis of the customerso/ experiences with their existing infrastructure.e  E People on this group have made the point that it is difficult to weanhD people of Windows and this is very true for the desktop. However theH rash of Virii, denials of service etc which have infected Windows serverH products such as IIS and exchange server have made people more reluctant9 to use it as a datacenter OS where there are substitutes.i  ; As for Linux, it currently as the market share numbers show % has a very small share of the market.   H In addition what ever you may surmise about Sun losing business to LinuxF in our "traditional" technical market the fact is that Sparc Solaris'sG market share is still increasing as are revenues after the downturn. IfoF Linux is stealing business from Sun and we do have our own Linux boxesA it isn't obvious from the overall numbers which suggest that only:F SPARC/Solaris and Linux are growing market share at the expense of the other vendors commercial OS's.  E In many ways this is exactly what happened when NT entered the server-D space, everyone said it would kill Solaris/AIX etc but instead the 3F major commercial UNIX's Solaris, AIX and HP-UX grew share and revenuesC as did NT at the expense of the smaller players such as Tru64, NCR,t Pyramid, OpenVMS, IRIX etc.e  > So if we as some analysts predict end up with Solaris, anotherC commercial UNIX and Linux plus Windows as the players in the servery@ market then this will not be suprising and Sun will not be badly positioned.d  9 HP's problem is trying to ensure that the 2and commerciall5 UNIX is HP-UX and not AIX something that currently isf8 very much in doubt. HP isn't a Linux technology provider6 though you do bend tin that happens to run Linux, this6 without HP-UX in the frame would put you very low down1 in the Linux food chain, not a great place to be.v      H > The genesis of this note was that someone said that HP's stock decline  I > is a result of it's merger activity, and that Sun's stock was moving=20 J > higher. BOTH statements are misleading or false.  I *will* agree with=20I > you that IMHO it is *too early* to determine how well the merger has=20:J > worked.  The success or failure of the merger is probably not a major=20J > factor in the price of the stock since the merger.  Nor has Sun stock=20( > performed well in that same timeframe. >=20    C Funnily enough I agree with you. HP's stock has outperformed Sun's.oF Though you do have a highly profitable printing consumables unit which1 we don't and for that reason the numbers look OK.c  D And you are right the merger really hasn't started to pan out from aH product standpoint. The web sites have been merged, the sales forces allD have the same cards etc but with the exception of the PC's which wasD always the easiest product merge very few of the product ranges haveF been rationalised and only when this happens will the merger proper be in full swing.  9 At the moment you are in the honeymoon period, people canc3 still call HP any buy the products they used to buyrE from either HP or Compaq, wait until that isn't the case anymore. The E fact that they arn't in the numbers that they used to is interesting.   4 HP also appear to have made a huge technology gamble8 in IA-64 which to date does not appear to be paying off.  2 Getting off that horse if it stumbles will be very6 costly and for that fact alone I would mark HP's stock down.w  7 Sun on the other hand doesn't have any of these issues.h  F Persistent claims of huge CPU performance gaps between SPARC and other% CPU's have made by you and the choir.b  D Which you will also recollect that you and the choir have never been@ able to substantiate, although you do appear to beleive your own retoric.    < Sun's market share is growing in servers but also in storage  : as well and our cladding rate for services and software on9 our servers is also increasing. Despite claims that Linux 5 is eating Solaris share in fact Linux and Solaris arepE eating AIX/HP-UX/Tru64 share, not something that is of any comfort tou
 IBM or HP.  G I am not a stock analyst but given the fundamentals of both busineses Iu7 know which one I would bet on for the next 6-12 months.k   Regardst Andrew Harrison>   ------------------------------  # Date: Tue, 20 Aug 2002 14:51:26 GMTv5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>s+ Subject: Re: HP-Compaq Merger Went Smoothlye2 Message-ID: <OHs89.16$WV6.551071@news.cpqcorp.net>  6 Andrew Harrison SUNUK Consultancy wrote in message ... >a >n< >I am not a stock analyst but given the fundamentals of both7 >busineses I know which one I would bet on for the nextn
 >6-12 months.  >w  K Your confusing fundamentals with your personal opinion of the company, it's J products, and it's direction.  From a pure stock play perspective, Sun hasI shown little evidence that it will make it's slight profit prediction forrK the fiscal year ending in 2003.  Nor has the tech market shown any evidenceeI of a turn around in general.  I would expect most tech stocks to flounder2; until something starts driving a new growth spurt in sales.l  H The longer the industry slump continues, the higher the probability thatI only the strong (balance sheets) will survive.  Sun's current stock price K and market capitalization may make it a target for a takeover - who knows -eJ maybe Fujitsu will buy you - or some enterprising Defense industry company	 might :-)e   ------------------------------  % Date: Tue, 20 Aug 2002 16:12:50 +0100uU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>u+ Subject: Re: HP-Compaq Merger Went Smoothlyt0 Message-ID: <ajtmb3$2ik$1@new-usenet.uk.sun.com>   Fred Kleinsorge wrote:  8 > Andrew Harrison SUNUK Consultancy wrote in message ... >  >>= >>I am not a stock analyst but given the fundamentals of boths8 >>busineses I know which one I would bet on for the next >>6-12 months. >> >> > M > Your confusing fundamentals with your personal opinion of the company, it'snL > products, and it's direction.  From a pure stock play perspective, Sun hasK > shown little evidence that it will make it's slight profit prediction foroM > the fiscal year ending in 2003.  Nor has the tech market shown any evidence K > of a turn around in general.  I would expect most tech stocks to flounder = > until something starts driving a new growth spurt in sales.X >     @ Fundamentals are revenue growth and market share growth. We have@ them you don't. Not cutting R&D and other longer term investment= does have a short term impact but most experts agree that thec> companies that can do this and grow revenue and share during a; downturn are the ones that come out of it in the best shapea when the recovery starts.,    J > The longer the industry slump continues, the higher the probability thatK > only the strong (balance sheets) will survive.  Sun's current stock priceeM > and market capitalization may make it a target for a takeover - who knows -NL > maybe Fujitsu will buy you - or some enterprising Defense industry company > might :-)m >     ? I doub't it Sun's stock is low because of the economic downturn < very few companies have either the cash or the guts during a< downturn to stump up the kind of money needed for a purchase8 of this type in these market conditions. Nice FUDDY spin though.f  7 As an alternative how about spinning out the profitablee2 printing business in HP and ditching the rest in a7 garage sale now that just might give your stock holdersa a nice return.   Regardss Andrew Harrison3   ------------------------------  % Date: Tue, 20 Aug 2002 15:57:30 +0100XU From: Andrew Harrison SUNUK Consultancy <andrew_nospam.harrison_remove_this@sun#.com>o+ Subject: Re: HP-Compaq Merger Went Smoothly 0 Message-ID: <ajtleb$275$1@new-usenet.uk.sun.com>   Main, Kerry wrote:  	 > Andrew,s >  > D >>>>Persistent claims of huge CPU performance gaps between SPARC and >>>>- > other CPU's have made by you and the choir.r > F > Which you will also recollect that you and the choir have never beenB > able to substantiate, although you do appear to believe your own > retoric.>>>> > G > It simply means when evidence of this is presented to you, you alwaysc. > need to find a way to introduce a yeahbut .. >  > As an example:? > http://ideasinternational.com/benchmark/spec/specjbb2000.html ( > HP Superdome number 1 rating - 614358 ' > Sun Fire 15K number 2 rating - 602270e > C > Now, on the surface, this seems like a small difference until one I > notices that the Superdome number was done with 64 Cpu's, while the Sune > 15K was done with 104 cpu's. >     < How did a Java benchmark intrude into a discussion about the8 relative performance of SPARC CPU's and others and their( effects on DBMS licensing costs ????????  < Personnally I don't see the point in scaling SPECjbb to more? than about 24 CPU's at the maximum. Most people are not runninge: java apps with this kind of footprint and if they are then: they are doing so in a J2EE apps server like BEA, iPlanet,> Oracle or WebSphere all of which balance loads across multiple5 smaller servers rather than running one large server.d  > What you are also seing with SPECjbb is mostly Java JVM tuning: and OS tuning which causes the results to increase without recourse to faster hardware.  9 IBM for example introduced the idea of using a 64 bit JVM : and effectively turning off garbage collection, its on but? the heap sizes are so large that it in effect never runs duringr> the bnehcmark runs. This produced a very large jump in SPECjbb: numbers which everyone else then copied. How indicative of, how real Java apps run this is it debatable.  : If you want an example of what is a more realistic test of9 java/apps server performance then look at ecperf which is,9 a J2EE apps server/DBMS server performance test using thee1 horizontal scaling delivered by the apps servers.e  : The best HP-PA config delivers 37791 operations per minute' using 16 CPU's in the apps server tier.l  : The best SPARC config delivers 61862 operations per minute' using 24 CPU's in the apps server tier.g  > Including hardware costs, OS and apps and DBMS server licenses1 the Sun config costs $28/op vs $36/op for the HP.   = But then you wouldn't want to use ecperf because it disproves 9 your conjecture and horror of horrors both in performance B terms and because it has the temerity to include the configuration costs as well. :):):)        > :-)  > 	 > Regardsh >  > Kerry Main > Senior Consultanto > Hewlett-Packard Canada# > Consulting & Integration Servicesa > Voice: 613-592-4660e > Fax   : 613-591-4477 > Email: Kerry.Main@hp.com >  >  > -----Original Message-----) > From: Andrew Harrison SUNUK Consultancye7 > [mailto:andrew_nospam.harrison_remove_this@sun#.com] e > Sent: August 20, 2002 5:38 AMt > To: Info-VAX@Mvb.Saic.Com - > Subject: Re: HP-Compaq Merger Went Smoothlyr >  >  >  >  > Fred Kleinsorge wrote: >  > 8 >>Andrew Harrison SUNUK Consultancy wrote in message ... >> >>2 >>>Which is of course why Sun's stock is currently4 >>>at 4-5 dollars rather than 20. We had 15 years of3 >>>profit and revenue growth followed by 2 quarterst" >>>of revenue declines and losses. >>>d >>>o >>>e > I >>>The revenue declines you have seen so far are probably mostly customert >>>d > H >>>sentiment not the fact that they can no longer buy the product range " >>>that they used to buy from you. >>>sE >>>Wait for that before announcing that the merger has been a sucess.' >>>e >>>  >>>uH >>This is simply random guessing - "I think you are doing badly because D >>you walk funny, your mother dresses you bad, and calls you Sally". >> >> >  > : > Well perhaps you can come up with an explanation for the7 > larger than market declines in your server businessesaF > with a pretty much intact product ranges, support organistations and" > technical/engineering resources. >  >  > G >>I can "guess" that Sun is going into a stock decline and has taken a oH >>huge hit on it's revenue and profits because Linux is eating into your >> > H >>traditional technical UNIX market, the UNIX market future is in doubt E >>with pressures from both Linux and Windows, the future of Sparc is  H >>making your customers skittish - since it's unlikely that you will fix >> > F >>your performance problems soon, your margins have noplace to go but H >>down, the customers that drove your growth in recent years are part of >> > H >>the bubble that burst, etc, etc, etc. - causing you to "restructure" - >> > ' >>i.e. fire employees to contain costs.  >> >> >  > I > Windows has been a "pressure" on the UNIX market since the introductioneB > of NT. The fact that it hasn't show any signs of erroding UNIX'sH > dominance of the datacenter is mostly because Windows 2000 is still asJ > far away from the capabilities of any of the commercial UNIX's now as NTE > was when it was introduced. In addition some of the incursions thateF > Windows made into the datacenter on the back of applications vendorsD > like SAP and Seibel have been reversed by these vendors use of webH > services based technolgies. Seibel apps server and client only used toI > run on Win32 with a DBMS on virtually anything. Seibel 7 and later runsiJ > in a browser with an apps server tier that runs on Solaris Win32 and AIX8 > and a DBMS server tier which is supported on Solaris,  > Win32, AIX etc.U > 8 > I have been involved in 3 very large Seibel 7 projectsI > all of which have involved deploying UNIX apps servers to replace win32uB > apps servers that were being used for the older Seibel versions. > J > In all 3 cases the decision to deply on UNIX for the apps layer was madeG > without any input from Sun, IBM etc but on the basis of the customers 1 > experiences with their existing infrastructure.g > G > People on this group have made the point that it is difficult to wean0F > people of Windows and this is very true for the desktop. However theJ > rash of Virii, denials of service etc which have infected Windows serverJ > products such as IIS and exchange server have made people more reluctant; > to use it as a datacenter OS where there are substitutes.  > = > As for Linux, it currently as the market share numbers showa' > has a very small share of the market.s > J > In addition what ever you may surmise about Sun losing business to LinuxH > in our "traditional" technical market the fact is that Sparc Solaris'sI > market share is still increasing as are revenues after the downturn. If H > Linux is stealing business from Sun and we do have our own Linux boxesC > it isn't obvious from the overall numbers which suggest that onlydH > SPARC/Solaris and Linux are growing market share at the expense of the  > other vendors commercial OS's. > G > In many ways this is exactly what happened when NT entered the serverrF > space, everyone said it would kill Solaris/AIX etc but instead the 3H > major commercial UNIX's Solaris, AIX and HP-UX grew share and revenuesE > as did NT at the expense of the smaller players such as Tru64, NCR,a > Pyramid, OpenVMS, IRIX etc.t > @ > So if we as some analysts predict end up with Solaris, anotherE > commercial UNIX and Linux plus Windows as the players in the servereB > market then this will not be suprising and Sun will not be badly
 > positioned.n > ; > HP's problem is trying to ensure that the 2and commercial 7 > UNIX is HP-UX and not AIX something that currently isi: > very much in doubt. HP isn't a Linux technology provider8 > though you do bend tin that happens to run Linux, this8 > without HP-UX in the frame would put you very low down3 > in the Linux food chain, not a great place to be.b >  >  >  > H >>The genesis of this note was that someone said that HP's stock decline >> > G >>is a result of it's merger activity, and that Sun's stock was moving cH >>higher. BOTH statements are misleading or false.  I *will* agree with G >>you that IMHO it is *too early* to determine how well the merger has cH >>worked.  The success or failure of the merger is probably not a major H >>factor in the price of the stock since the merger.  Nor has Sun stock ( >>performed well in that same timeframe. >> >> >  > E > Funnily enough I agree with you. HP's stock has outperformed Sun's. H > Though you do have a highly profitable printing consumables unit which3 > we don't and for that reason the numbers look OK.p > F > And you are right the merger really hasn't started to pan out from aJ > product standpoint. The web sites have been merged, the sales forces allF > have the same cards etc but with the exception of the PC's which wasF > always the easiest product merge very few of the product ranges haveH > been rationalised and only when this happens will the merger proper be > in full swing. > ; > At the moment you are in the honeymoon period, people cann5 > still call HP any buy the products they used to buysG > from either HP or Compaq, wait until that isn't the case anymore. TheaG > fact that they arn't in the numbers that they used to is interesting.d > 6 > HP also appear to have made a huge technology gamble: > in IA-64 which to date does not appear to be paying off. > 4 > Getting off that horse if it stumbles will be very8 > costly and for that fact alone I would mark HP's stock > down.r > 9 > Sun on the other hand doesn't have any of these issues.T > H > Persistent claims of huge CPU performance gaps between SPARC and other' > CPU's have made by you and the choir.a > F > Which you will also recollect that you and the choir have never beenB > able to substantiate, although you do appear to beleive your own
 > retoric. >  > > > Sun's market share is growing in servers but also in storage > < > as well and our cladding rate for services and software on; > our servers is also increasing. Despite claims that Linuxd7 > is eating Solaris share in fact Linux and Solaris areeG > eating AIX/HP-UX/Tru64 share, not something that is of any comfort to  > IBM or HP. > I > I am not a stock analyst but given the fundamentals of both busineses In9 > know which one I would bet on for the next 6-12 months.  > 	 > Regardsn > Andrew Harrisona >  >  >  >    ------------------------------  % Date: Tue, 20 Aug 2002 13:03:15 -0400k' From: "Main, Kerry" <Kerry.Main@hp.com>a+ Subject: RE: HP-Compaq Merger Went Smoothly T Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF402660944@kaoexc01.americas.cpqcorp.net>   Andrew,   " And in response to your yeahbut ..  H >> How did a Java benchmark intrude into a discussion about the relative? performance of SPARC CPU's and others and their effects on DBMSe licensing costs ???????? <<<  H The response was in reply to your generic statement that was not related to db licensing - G "Persistent claims of huge CPU performance gaps between SPARC and others& CPU's have made by you and the choir."  H >>> If you want an example of what is a more realistic test of java/appsG server performance then look at ecperf which is a J2EE apps server/DBMSeE server performance test using the horizontal scaling delivered by thei apps servers.<<<  @ Oh my goodness, I thought for a moment you were going to say the= "clustering" word that you seem to dislike so much .. I guess 1 "horizontal scaling" is an easier word to say eh?o   :-)n  H So, you admit there is a place for clustering. Wow. And since many, manyE of our Customers are looking at adopting J2EE / Java Apps, I guess weaH can say clustering is the way to go for these Customers - and right from a Sun Consultant.s  G Course, the Application Level Clustering (term used by vendors like BEA G etc) concept you are outlining is application specific, so if one had alC number of different applications on a server, then each applicationnC would require their own application clustering strategy as well.=20s  @ Now, I wonder if there is an OS out there that provides a nativeG clustering solution that would mean the Application vendors do not havehG to worry about Application clustering? And wouldn't it be nice if thesedF Java/J2EE apps did not have to change anything to become cluster aware with this OS clustering?   :-)    Regardsn  
 Kerry Main Senior Consultanta Hewlett-Packard Canada! Consulting & Integration Servicesi Voice: 613-592-4660t Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----' From: Andrew Harrison SUNUK Consultancy 7 [mailto:andrew_nospam.harrison_remove_this@sun#.com]=20a Sent: August 20, 2002 10:58 AM To: Info-VAX@Mvb.Saic.Com + Subject: Re: HP-Compaq Merger Went Smoothlyw         Main, Kerry wrote:  	 > Andrew,- >=20 >=20D >>>>Persistent claims of huge CPU performance gaps between SPARC and >>>>- > other CPU's have made by you and the choir.S >=20I > Which you will also recollect that you and the choir have never been=20UE > able to substantiate, although you do appear to believe your own=20  > retoric.>>>> >=20J > It simply means when evidence of this is presented to you, you always=20. > need to find a way to introduce a yeahbut .. >=20 > As an example:=20 ? > http://ideasinternational.com/benchmark/spec/specjbb2000.htmlo' > HP Superdome number 1 rating - 614358e' > Sun Fire 15K number 2 rating - 602270r >=20F > Now, on the surface, this seems like a small difference until one=20H > notices that the Superdome number was done with 64 Cpu's, while the=20" > Sun 15K was done with 104 cpu's. >=20    E How did a Java benchmark intrude into a discussion about the relative ? performance of SPARC CPU's and others and their effects on DBMSn licensing costs ????????  G Personnally I don't see the point in scaling SPECjbb to more than aboutfH 24 CPU's at the maximum. Most people are not running java apps with thisG kind of footprint and if they are then they are doing so in a J2EE apps-H server like BEA, iPlanet, Oracle or WebSphere all of which balance loadsE across multiple smaller servers rather than running one large server.:  E What you are also seing with SPECjbb is mostly Java JVM tuning and OS.F tuning which causes the results to increase without recourse to faster	 hardware.m  9 IBM for example introduced the idea of using a 64 bit JVMf: and effectively turning off garbage collection, its on butC the heap sizes are so large that it in effect never runs during theaH bnehcmark runs. This produced a very large jump in SPECjbb numbers whichH everyone else then copied. How indicative of how real Java apps run this is it debatable.  D If you want an example of what is a more realistic test of java/appsG server performance then look at ecperf which is a J2EE apps server/DBMSdE server performance test using the horizontal scaling delivered by thee
 apps servers.i  C The best HP-PA config delivers 37791 operations per minute using 16o CPU's in the apps server tier.  C The best SPARC config delivers 61862 operations per minute using 24  CPU's in the apps server tier.  F Including hardware costs, OS and apps and DBMS server licenses the Sun) config costs $28/op vs $36/op for the HP.   B But then you wouldn't want to use ecperf because it disproves yourF conjecture and horror of horrors both in performance terms and becauseF it has the temerity to include the configuration costs as well. :):):)       > :-)  >=20	 > Regardsl >=20 > Kerry Main > Senior Consultante > Hewlett-Packard Canada# > Consulting & Integration Servicesr > Voice: 613-592-4660c > Fax   : 613-591-4477 > Email: Kerry.Main@hp.com >=20 >=20 > -----Original Message-----, > From: Andrew Harrison SUNUK Consultancy=206 > [mailto:andrew_nospam.harrison_remove_this@sun#.com] > Sent: August 20, 2002 5:38 AMy > To: Info-VAX@Mvb.Saic.Com.- > Subject: Re: HP-Compaq Merger Went Smoothly  >=20 >=20 >=20 >=20 > Fred Kleinsorge wrote: >=20 >=208 >>Andrew Harrison SUNUK Consultancy wrote in message ... >> >>2 >>>Which is of course why Sun's stock is currently4 >>>at 4-5 dollars rather than 20. We had 15 years of3 >>>profit and revenue growth followed by 2 quartersr" >>>of revenue declines and losses. >>>r >>>_ >>>e >=20C >>>The revenue declines you have seen so far are probably mostly=20  >>>customer@ >>>s >=20G >>>sentiment not the fact that they can no longer buy the product rangea" >>>that they used to buy from you. >>>SE >>>Wait for that before announcing that the merger has been a sucess.h >>>o >>>l >>> G >>This is simply random guessing - "I think you are doing badly because D >>you walk funny, your mother dresses you bad, and calls you Sally". >> >> >=20 >=20I > Well perhaps you can come up with an explanation for the larger than=20rH > market declines in your server businesses with a pretty much intact=20E > product ranges, support organistations and technical/engineering=205 > resources. >=20 >=20 >=20F >>I can "guess" that Sun is going into a stock decline and has taken aH >>huge hit on it's revenue and profits because Linux is eating into your >> >=20G >>traditional technical UNIX market, the UNIX market future is in doubtiG >>with pressures from both Linux and Windows, the future of Sparc is=20nH >>making your customers skittish - since it's unlikely that you will fix >> >=20E >>your performance problems soon, your margins have noplace to go butxH >>down, the customers that drove your growth in recent years are part of >> >=20H >>the bubble that burst, etc, etc, etc. - causing you to "restructure" - >> >=20' >>i.e. fire employees to contain costs.  >> >> >=20 >=20? > Windows has been a "pressure" on the UNIX market since the=20 H > introduction of NT. The fact that it hasn't show any signs of erroding  I > UNIX's dominance of the datacenter is mostly because Windows 2000 is=20lE > still as far away from the capabilities of any of the commercial=20fI > UNIX's now as NT was when it was introduced. In addition some of the=20fD > incursions that Windows made into the datacenter on the back of=20I > applications vendors like SAP and Seibel have been reversed by these=20fJ > vendors use of web services based technolgies. Seibel apps server and=20H > client only used to run on Win32 with a DBMS on virtually anything.=20G > Seibel 7 and later runs in a browser with an apps server tier that=20 B > runs on Solaris Win32 and AIX and a DBMS server tier which is=20' > supported on Solaris, Win32, AIX etc.e >=208 > I have been involved in 3 very large Seibel 7 projectsF > all of which have involved deploying UNIX apps servers to replace=20H > win32 apps servers that were being used for the older Seibel versions. >=20H > In all 3 cases the decision to deply on UNIX for the apps layer was=20E > made without any input from Sun, IBM etc but on the basis of the=20o; > customers experiences with their existing infrastructure.s >=20J > People on this group have made the point that it is difficult to wean=20I > people of Windows and this is very true for the desktop. However the=20aF > rash of Virii, denials of service etc which have infected Windows=20J > server products such as IIS and exchange server have made people more=20E > reluctant to use it as a datacenter OS where there are substitutes.n >=20H > As for Linux, it currently as the market share numbers show has a very   > small share of the market. >=20G > In addition what ever you may surmise about Sun losing business to=20:G > Linux in our "traditional" technical market the fact is that Sparc=20>I > Solaris's market share is still increasing as are revenues after the=20hH > downturn. If Linux is stealing business from Sun and we do have our=20D > own Linux boxes it isn't obvious from the overall numbers which=20J > suggest that only SPARC/Solaris and Linux are growing market share at=203 > the expense of the other vendors commercial OS's.l >=20J > In many ways this is exactly what happened when NT entered the server=20I > space, everyone said it would kill Solaris/AIX etc but instead the 3=20>H > major commercial UNIX's Solaris, AIX and HP-UX grew share and revenues  H > as did NT at the expense of the smaller players such as Tru64, NCR,=20 > Pyramid, OpenVMS, IRIX etc.u >=20C > So if we as some analysts predict end up with Solaris, another=20yH > commercial UNIX and Linux plus Windows as the players in the server=20E > market then this will not be suprising and Sun will not be badly=20>
 > positioned.h >=20F > HP's problem is trying to ensure that the 2and commercial UNIX is=20I > HP-UX and not AIX something that currently is very much in doubt. HP=20eJ > isn't a Linux technology provider though you do bend tin that happens=20I > to run Linux, this without HP-UX in the frame would put you very low=20r8 > down in the Linux food chain, not a great place to be. >=20 >=20 >=20 >=20H >>The genesis of this note was that someone said that HP's stock decline >> >=20F >>is a result of it's merger activity, and that Sun's stock was movingJ >>higher. BOTH statements are misleading or false.  I *will* agree with=20I >>you that IMHO it is *too early* to determine how well the merger has=20tJ >>worked.  The success or failure of the merger is probably not a major=20J >>factor in the price of the stock since the merger.  Nor has Sun stock=20( >>performed well in that same timeframe. >> >> >=20 >=20H > Funnily enough I agree with you. HP's stock has outperformed Sun's.=20H > Though you do have a highly profitable printing consumables unit which  3 > we don't and for that reason the numbers look OK.S >=20I > And you are right the merger really hasn't started to pan out from a=20sI > product standpoint. The web sites have been merged, the sales forces=20 I > all have the same cards etc but with the exception of the PC's which=20eH > was always the easiest product merge very few of the product ranges=20F > have been rationalised and only when this happens will the merger=20 > proper be in full swing. >=20I > At the moment you are in the honeymoon period, people can still call=20iG > HP any buy the products they used to buy from either HP or Compaq,=20 H > wait until that isn't the case anymore. The fact that they arn't in=20/ > the numbers that they used to is interesting.n >=206 > HP also appear to have made a huge technology gamble: > in IA-64 which to date does not appear to be paying off. >=204 > Getting off that horse if it stumbles will be very8 > costly and for that fact alone I would mark HP's stock > down.d >=209 > Sun on the other hand doesn't have any of these issues.c >=20H > Persistent claims of huge CPU performance gaps between SPARC and other  ' > CPU's have made by you and the choir.t >=20I > Which you will also recollect that you and the choir have never been=20 E > able to substantiate, although you do appear to beleive your own=20u
 > retoric. >=20 >=20> > Sun's market share is growing in servers but also in storage >=20H > as well and our cladding rate for services and software on our servers  J > is also increasing. Despite claims that Linux is eating Solaris share=20D > in fact Linux and Solaris are eating AIX/HP-UX/Tru64 share, not=200 > something that is of any comfort to IBM or HP. >=20J > I am not a stock analyst but given the fundamentals of both busineses=20; > I know which one I would bet on for the next 6-12 months.t >=20	 > Regardsc > Andrew Harrison  >=20 >=20 >=20 >=20   ------------------------------  % Date: Tue, 20 Aug 2002 10:51:07 -0700 & From: Greg Cagle <gregc@gregcagle.com>+ Subject: Re: HP-Compaq Merger Went SmoothlyN, Message-ID: <3D62818B.9070509@gregcagle.com>   Fred Kleinsorge wrote:  M > Reading your SEC filings, it also appears that while wanting to hold R&D athM > 10% of net revenue, you also plan on reducing your headcount 9%.  Not good. M > Are you too fat in management?  Too many salesmen?  Too many support staff?aJ > Too many engineers?  Where is the 9% comming from?  You are also closing
 > facilities.u  G My friends over at the Starcat plant in Beaverton say they are the onlypD group within Sun that is making money - and yet they are getting cut back 10% as well. Hmmm.x   -- h
 Greg Cagle gregc at gregcagle dot com   ------------------------------  % Date: Tue, 20 Aug 2002 14:05:58 -0400c- From: JF Mezei <jfmezei.spamnot@videotron.ca>v+ Subject: Re: HP-Compaq Merger Went Smoothlyh, Message-ID: <3D628505.7A711C28@videotron.ca>   "Main, Kerry" wrote:  _ > Do you really think those at the top that signed the contract know the difference between an aN >Alpha or a SPARC or a Power4 or what the difference between HP-UX and Solaris
 and Tru64 is?e  K But they know that Compaq had committed to long term Alpha development, and L that Compaq broke that commitment months later. By now, they would also knowM that Compaq, as a corporation, was fully aware that it was lying to customerstM when it was making such commitments since it already knew that Alpha would ber cancelled shortly.  M Furthermore, Compaq lied to customers on June 25 when it announced that Tru64eL was to be ported to IA64 since Compaq, as a corporation, was already engagedN into the merger talks with Carly with detailed discussions on product roadmaps already in progress.    b > So what do you say to Customers that bought lots of IBM disk drives in the months leading up to > >the point where IBM sold its disk drive business to Hitachi?   L Disk drives are not an "architecture". You can replace drives without havingL to change your software, renegotiate licence agreements/transfers, recertifyK your application and mount a large scale migration in-house project. In thenK case of the Qubec government, they were embarking on a long term migrationtK from IBM to Alpha-Tru64 and had the rug pulled from under them months afterm the project began.   ------------------------------  # Date: Tue, 20 Aug 2002 12:58:02 GMT ! From: Andy <acs@fcgnet.works.net>f; Subject: Re: Java plugin performance for interactive appletu> Message-ID: <Xns92705AEC041C8acsfcgnetworksnet@216.166.71.232>  ; lewis@spyder.mitre.org (Keith A. Lewis) enlightened us withy' news:ajrtrj$dmi$1@newslocal.mitre.org: i  4 > jordan@ccs4vms.com (Rich Jordan) writes in article@ > <cc5619f2.0208140731.1e63981a@posting.google.com> dated 14 Aug > 2002 08:31:46 -0700:    v? >> Both the custom client and Powerterm Java work well under IEo? >> 5.5 with Java, and under IE and Netscape 6.2 on a Mac, so atu9 >> least it doesn't appear to be caused by writing to ther  >> MS-bastardized Java changes.  > = > I'm nowhere near enough of a Java expert to solve this, bute9 > since nobody else is saying anything I'll throw in some  > thoughts.  > B > Comparing PC-native and Mac-native versions of Java to X-windowsB > versions is not fair.  X-windows gives you the added flexibility? > of running on one machine and displaying on another, and with0$ > that comes a performance penalty.     > I assumep= > that OS X includes X-windows, but I haven't verified this.    < No. Mac OS X does it's own, non-X Windows thing called Aqua.@ There's a port of XFree86 called XDarwin for those who want X onB Mac OS X. I have yet to set it up to my satisfaction to work with ' VMS but that's a whole different story.l   -Andy- -- e   ------------------------------  % Date: Tue, 20 Aug 2002 12:09:15 +0100 % From: Alan Greig <a.greig@virgin.net>cI Subject: Re: Licenses (was Re: Charon-VAX (was: [VAX] VMS to [Alpha]...))v8 Message-ID: <te74mu0804o9hl99adcars24quthmjvb9h@4ax.com>  E On 18 Aug 2002 21:20:59 -0700, whohe@whoever.com (DL Phillips) wrote:i  5 >The DEC2000-300 in question was purchased before the 7 >UCX-TCP/IP or NET-xxx licenses were bundled. Actually,r  > My recollection is that all Alphas sold in standard configs by@ DEC/Compaq came with some form of network integration license. I? certainly recall our first machines (DEC 3000-400 April 1993 orrE earlier) came with a license which allowed the use of UCX even thoughaD Alpha UCX was not yet out of beta test. Oh how our Unix support guysD laughed when they discovered that DEC had released Alpha/VMS with noA TCPIP support. IIRC Multinet (and/or TCPWARE) was out before UCX.o  E If you have double-checked ths then I'd advise a triple-check. Do your have the original paperwork?  6 >as I understand, they are not even now really part of3 >VMS but are just usually "packaged with" and stillh2 >have their own PAKs. Many new systems "come with"1 >a single Adv-Server license, too, but that, too,y >has its own PAK.c >t. >Many times, the "bundling" is/was done by the+ >reseller, but if you read the SPD's you'lle  A There were standard DEC part numbers which included UCX licenses.e  / >see that they are not actually "included" witha2 >VMS. It seems there were many convuluted packages  F The SPD for VMS doesn't say the hardware is included for free but some= part number did include the hardware, the O/S and integration:> packages. It's the description of these part numbers you need.  1 >that DECpq has put together over the years. None:1 >of the deals seem to last very long, though, and 4 >try to find a sales person who understands anything >about what you get with what. >s. >The fact that my client has kept his software/ >support agreement current and if he bought thea2 >system today he'd probably get the TCP/IP license/ >with it doesn't seem to matter. He can't get ae- >PAK without buying it (and it'll probably be0 >backordered:-)n >h4 >The 2 user WinXP W/S license would be more closely 1 >related to the Advanced Server licenses than the=4 >to VMS User license, though, I'd say, and the fact 1 >that you must buy the Base license, then add thee6 >interactive users, and then add network user licenses6 >on top of that makes this quite expensive for a small3 >user. It doesn't matter how much more reliable ande. >secure and easily managed OurVMS is; it isn't >competitively priced. >w >  -Doug >. >-----------------] >John Santos <JOHN@egh.com> wrote in message news:<1020816181810.415M-100000@Ives.egh.com>...o, >> On Fri, 16 Aug 2002, Brian Tillman wrote: >> ?J >> > >Well, I've found prices, but here's a little example of why everyoneL >> > >in the world isn't/hasn't jumped on the VMS bandwagon. I have a clientJ >> > >with a small Alpha (DEC 2000/300) who has kept his hardware/softwareH >> > >agreement current. He wants to add 2 interactive users, 3 Advanced0 >> > >Server users, and TCP/IP TELNET & FTP now. >>  ...snip...4 >> > >  QL-0M2AE-AA  TCP/IP Client license. $1750.00 >> >  N >> > Chances are, this last license is unnecessary, since TCP/IP licenses haveJ >> > been bundled with most Alphas for some time now.  He should check his >> > shipping memo.v >> uC >> The DEC 2000/300 is old enough that it might not include TCP/IP. A >> However, our DEC 3000 Model 300, delivered at the beginning ofuC >> November, 1993, did come with an NAS-250 license, which includedp >> UCX (TCPIP).v >> hA >> The OP should also check SHOW LICENSE and LICENSE LIST/FULL top= >> see if there are any loaded or disabled UCX or NET-APP-SUPmA >> licenses.  (They might have been added and then disabled sincepB >> they weren't being used, but still might be present and valid.) >> eD >> Later systems have EIS (Enterprise Integration Services) bundles,C >> but I think they were a bunch of separate PAKs for each product,e> >> rather than a bundled PAK like NAS.  I think TCPIP ServicesB >> (new name for UCX) still uses UCX PAKs, but maybe it uses TCPIP >> now instead?o >> c= >> Also, the two-user license seems expensive, but compare itt0 >> to two Windows XP Pro workstation licenses...   -- Alan   ------------------------------  % Date: Tue, 20 Aug 2002 15:01:04 +0800i) From: "Steven Xie" <r33300@email.mot.com>) Subject: Memory Slot? + Message-ID: <ajspgi$3e8$1@newshost.mot.com>u  
 Hi everybody?-  J Is there a easy way to get the information regarding current system memoryI slots? I'm now using a com file to extract from "show mem/slot" output. IiF think there must be some easy way to get it, such as lexical function?   Thanks,s Steven   ------------------------------  # Date: Tue, 20 Aug 2002 14:25:13 GMTVF From: lederman@star.enet.dec.DISABLE-JUNK-EMAIL.com (Bart Z. Lederman) Subject: Re: Memory Slot?e2 Message-ID: <djs89.14$pU6.506925@news.cpqcorp.net>  4 I'm not sure how much 'easier' it's going to be, but& you can get information via a lexical.   value = F$GETSYI ("BAL_SLOTS") value = F$GETSYI ("PROC_SLOTS")e  2 What comes back is one long string with all of the. values in HEX, which you will have to separate and convert, as in  )  WRITE SYS$OUTPUT F$GETSYI ("PROC_SLOTS")   00000005000000280000005300000080   This correspondts to:e   $ sho mem/slot@               System Memory Resources on 20-AUG-2002 10:15:17.69  L Slot Usage (slots):                Total        Free    Resident     SwappedL   Process Entry Slots                128          83          40           5L   Balance Set Slots                  126          88          38           0  ? You can see how the value returned has the figures for Swapped,f Resident, Free and Total.a   -- c(  B. Z. Lederman   Personal Opinions Only  8  Posting to a News group does NOT give anyone permission8  to send me advertising by E-mail or put me on a mailing  list of any kind.  5  Please remove the "DISABLE-JUNK-EMAIL" if you have a 5  legitimate reason to E-mail a response to this post.    ------------------------------  % Date: Tue, 20 Aug 2002 12:27:17 +0100a% From: Alan Greig <a.greig@virgin.net>r  Subject: Re: New missive from HP8 Message-ID: <ol94mu48eu81tkff3avhu8221nj8tkamtv@4ax.com>  F On Mon, 19 Aug 2002 10:04:22 +0100, Nic Clews <sendspamhere@127.0.0.1> wrote:   >JF Mezei wrote: >> lN >> It is ironic that when Compaq bought Digital, some of its Armani consultantP >> told it to shed the "Digital" trademark ASAP because it was a hindrance. Yet,N >> I think that today, "Digital" would have much better image than "Compaq" in5 >> terms of image with regards to enterprise systems.s >eE >I agree; I think it would be better to see "Digital, part of the newoE >HP", the name Compaq does little for the greatness that lies within.n  B Four years after the Compaq merger, HP senior staff were amused toC find delegates to the recent European MANMAN conference all walking E around with large badges saying DEC/MANMAN. Computer Associates nevere bothered renaming the product.  F >Again it comes down to marketing and perceptions. The name digital isG >not forgotten, it is still known what they stood for, sort of. I'm notl2 >going to state the glaringly obvious about the C. >  >Personal opinion, as ever.l   -- Alan   ------------------------------  % Date: Tue, 20 Aug 2002 06:41:45 -0400I) From: "Neil Rieck" <n.rieck@sympatico.ca>p" Subject: Re: NewHp & VMS Licensing: Message-ID: <V1p89.15371$_75.709402@news20.bellglobal.com>  F "Webb, William W Raleigh, NC" <wwebb1@email.usps.gov> wrote in messageE news:BA52530E3149734A9BAABDBBFA808E49028BE9FC@rlghncst964.usps.gov...n > 9 > I suspect that corporate taxation is the driving factore= > behind the convoluted pathway as opposed to tariffs betweenb > Point A and Point B. > > > (After all, that's why we signed a document called the North> > American Free Trade Agreement that's over two thousand pages> > long- if it were really free trade couldn't they have put it! > on the front side of one page?)o >i > WWWebb >o  E Don't get me started on NAFTA. Something originally set up by lawyersIK consulting for governments, then later argued by those same lawyers who aretH now consulting to companies affected by it, just has to be a bad idea. IK wonder how much documentation is required between countries joining the EU?.  
 Neil Rieck Kitchener/Waterloo/Cambridge,s Ontario, Canada.! http://www3.sympatico.ca/n.rieck/r   ------------------------------  % Date: Tue, 20 Aug 2002 06:56:05 -0400s) From: "Neil Rieck" <n.rieck@sympatico.ca>t" Subject: Re: NewHp & VMS Licensing: Message-ID: <lfp89.15483$_75.711286@news20.bellglobal.com>   Kerry,  * Are PAKs still issued in Kanata (or Hull)?  I On a related issue, has the Compaq sign at 100 Hertzberg Road (in Kanata)0  been changed to an H/P sign yet?  
 Neil Rieck Kitchener/Waterloo/Cambridge,i Ontario, Canada.! http://www3.sympatico.ca/n.rieck/=   >  >    ...snip...a >i0 >Re: licensing in Canada - not much has changed. > E >Imho, the chances of any invoice going to a Canadian Customer in any ' >non-Canadian currency is slim to none.e  G >As an additional fyi - language issues also dictate that some invoiceso, >in Canada must be in French in some area's.   >Reference:2. >http://www.compaq.com/products/software/info/? >http://www.compaq.com/products/software/info/swl_contacts.htmli >g >Regards >  >Kerry Main2 >Senior Consultant >Hewlett-Packard Canadar" >Consulting & Integration Services >Voice: 613-592-4660 >Fax   : 613-591-4477  >Email: Kerry.Main@hp.com  >d   ------------------------------  # Date: Tue, 20 Aug 2002 11:47:37 GMTt1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie)i" Subject: Re: NewHp & VMS Licensing= Message-ID: <t%p89.313546$q53.10167207@twister.austin.rr.com>=  ( Neil Rieck (n.rieck@sympatico.ca) wrote: : G : Don't get me started on NAFTA. Something originally set up by lawyersfJ : consulting for governments, then later argued by those same lawyers who L : are now consulting to companies affected by it, just has to be a bad idea.H : I wonder how much documentation is required between countries joining 	 : the EU?r :   ; Mexico is now experiencing the "benefits" of globalization:f  F    http://www.washingtonpost.com/wp-dyn/articles/A14093-2002Jun19.html7    Mexican Workers Pay for Success (washingtonpost.com)I  @   "TIJUANA, Mexico -- Cesiah Ruiz Brena came to Tijuana in 1989,E    deliriously happy to get a job at a new Japanese factory. Her worknG    space was grand, the lights were bright and the pay was unimaginablyh    good: $100 a week to start.     F    But after 13 years during which her wages rose to $200 a week, RuizF    Brena lost her job on June 1. Her Canon inkjet printer factory shutD    down. She and her co-workers shared a cake, snapped photos of oneG    another and said goodbye. The factory, they were told, was moving toaE    Thailand and Vietnam, where wages are as low as $15 a week -- lessm     than what she earns in a day.    oA    All along the Mexican border with the United States, once-busyfI    factories are closing. Since the end of 2000, tearful farewell parties I    have been held for 250,000 factory workers in Mexico. Some of the samenC    jobs that left North Carolina textile plants and Ohio auto-partsIE    assembly lines for Mexico in the 1980s are now moving to Asia. Thef(    reason is the same: cheaper labor..."      y2 --Jerry Leslie   (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for emailh   ------------------------------  % Date: Tue, 20 Aug 2002 14:16:13 +0200n9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>t" Subject: Re: NewHp & VMS Licensing' Message-ID: <3D62330D.92C18CC2@aaa.com>n   Neil Rieck wrote:a5 > I wonder how much documentation is required between  > countries joining the EU?'    9 Well, when Sweden was going to join EU, the swedish state < dept. responsible for farming, had to hire a couple of extra; baracs just to house all the copy machines. The local Xeroc  salesman was very happy...  5 Only the part about having milk-cows is well over thea 2000 pages mentioned earlier.    Jan-Erik Sderholm.r   ------------------------------  % Date: Tue, 20 Aug 2002 13:24:39 +0100u( From: Nic Clews <sendspamhere@127.0.0.1>" Subject: Re: NewHp & VMS Licensing) Message-ID: <3D623507.EF304473@127.0.0.1>0   Neil Rieck wrote:- >  > Kerry, > , > Are PAKs still issued in Kanata (or Hull)? > K > On a related issue, has the Compaq sign at 100 Hertzberg Road (in Kanata)o" > been changed to an H/P sign yet?  H Kelvin Close in Warrington is still Compaq, at least one of the signs on the many 100's of roundabouts.  F OT: Driving round Birchwood and Risley in Warrington UK, reminds me ofE "you are in a maze of twisty passages, all alike" not that I ever gots lost driving there ;o)   --  ? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Scienceso nclews at csc dot comu   ------------------------------  % Date: Tue, 20 Aug 2002 10:38:57 -0400p' From: "Main, Kerry" <Kerry.Main@hp.com> " Subject: RE: NewHp & VMS LicensingT Message-ID: <BE56C50EA024184DAF48F0B9A47F5CF402660942@kaoexc01.americas.cpqcorp.net>   Neil,-  D PAKs are primarily issued from Toronto. Has been that way for awhileE now. Customer Support Center in Hull can issue temp paks on emergencye basis as I recall.  H Re: Herzberg Rd sign - well, I was in that office a few days ago and didH not notice it, so I am assuming it still says Compaq. I do know they are4 looking at getting it changed pretty quickly though.   Regardsa  
 Kerry Main Senior Consultant  Hewlett-Packard Canada! Consulting & Integration Servicesl Voice: 613-592-4660m Fax   : 613-591-4477 Email: Kerry.Main@hp.com     -----Original Message-----1 From: Neil Rieck [mailto:n.rieck@sympatico.ca]=20i Sent: August 20, 2002 6:56 AMa To: Info-VAX@Mvb.Saic.Com " Subject: Re: NewHp & VMS Licensing     Kerry,  * Are PAKs still issued in Kanata (or Hull)?  A On a related issue, has the Compaq sign at 100 Hertzberg Road (in4( Kanata) been changed to an H/P sign yet?  
 Neil Rieck Kitchener/Waterloo/Cambridge,w Ontario, Canada.! http://www3.sympatico.ca/n.rieck/    >o >    ...snip...w > 0 >Re: licensing in Canada - not much has changed. >dH >Imho, the chances of any invoice going to a Canadian Customer in any=20' >non-Canadian currency is slim to none.r  J >As an additional fyi - language issues also dictate that some invoices=20, >in Canada must be in French in some area's.   >Reference:y. >http://www.compaq.com/products/software/info/? >http://www.compaq.com/products/software/info/swl_contacts.htmle >e >Regards >o >Kerry Main  >Senior Consultant >Hewlett-Packard Canadah" >Consulting & Integration Services >Voice: 613-592-4660 >Fax   : 613-591-4477- >Email: Kerry.Main@hp.com, >    ------------------------------    Date: 20 Aug 2002 08:48:40 -0600- From: frey@encompasserve.org (Sharon Guthrie)e/ Subject: Re: NFS Server configuration questionst3 Message-ID: <jSAPwvMQ3s+d@eisner.encompasserve.org>L  j In article <3D6135E7.2F9E9A68@youknow.where>, "John Gemignani, Jr." <John.Gemignani@youknow.where> writes:  E > 	The example here shows you that you define ucx$nfs and ucx$nobody. J > 	The example fred is just a regular user (nothing special, but the entryK > 	for vmsuser is for a PC.  fred is a user who may be on one or morehosts.nH > 	The PC user "vmsuser" usually comes in from one PC (PCs *usually* but > 	not always) have one user.s   	Ah, thanks, that clears it up.f  E > 	When you authenticate yourself using a username and password usinggH > 	PCNFS, you get a UID and GID back as a result.  NFS exchanges uid andF > 	gid information, not username and password.  They do not need to beF > 	in the same group unless you want them to be able to share files toG > 	some degree -- similar rules that you would apply to your VMS users.m  I 	It's hard to diagram this with plaintext, but I need to try.  Let's see   if I understand:  K 1.  PC tries to connect to fileshare by sending a username/password pair.  r< Say for example, "user1" "passwd", and the pc name is "pc1".  I 2.  UCX tries to match that up with a proxy entry.  It finds a proxy for -M username "user1" on host "pc1" (or asterisk, but let's not use wildcards for 0O this example).  It sends, via PCNFS, the uid/gid pair to the PC's NFS client.  - Let's say that pair is 52,43..  M 3.  Further communication along this particular stream are identified by the 2F uid/gid pair (52,43) and the proxy database is not referenced.  Right?  P Question:  how does the uid/gid pair grant access to a particular export path?  K Is it only by the /host=xyz that you attach to the export?  So if you have  D wildcarded the export, then all of your clients can map it, correct?  1 >>  Do I REALLY need to have a proxy for each pc?i > K > 	No, that's what /host=* is for.  Each PC could obtain a uid and gid fromaI > 	PCNFS using the same username and password.  Define a proxy record forg > 	that user with /host=*.  G 	Ok, that's clear.  So all my pc clients would obtain the same uid/gid TI pair, since they all send the same username/password, right?  Would this iL confuse the data streams at all?  In my case, I want them all to access the K same directory.  They'll each use a separate data file, though, and that's h# controlled in the PC application.  aF 	The applications on both ends that write to the data files will keep P the data files straight, no problem.  But I don't want problems with sockets or L streams getting hung up and preventing the applications from writing to the  files.  Thus my concern with:s  F >> the value of the cache timer:  sysconfig -r nfs tcp_idle_timeout=30R >> Coworkers who've done this before tell me this works, but I haven't found it toL >> work on any of the versions of UCX I've been fighting with.  (5.0A, 5.1)  > L > 	The tcp timer is not available in V5.0 because tcp support for NFS wasn'tH > 	released until V5.1.  The idle timer is used to close TCP connectionsJ > 	that haven't sent NFS requests in that period of time.  This forces the< > 	client to reconnect when it wants to make an NFS request. > O >> cache timer is apparently set high and causes a problem for our application.a > J > 	Cache timer?  Which one?  Are you talking about the vfs vnode_age timer6 > 	which is set to 120 seconds to keep files accessed?  G 	I meant the tcp timer, but have problems with the terminology.  Sorry   about that. G 	When testing this whole thing on my development machine, I noticed an  K occasional problem where my VMS application updated the data file, but the rP changes weren't visible on the PC end.  This was a major problem.  I'm not sure C if it was the tcp timer or just that I had NFS way misconfigured.  l  K > 	In my experience I find that users who are not familiar with Unix often  K > 	can't feel comfortable with NFS. The whole Unix authentication stuff,forsK > 	example, compared to VMS is close but no cigar.  You really have tothink2J > 	it out -- it's not as simple to just add entries without thinking aboutL > 	the values that you are using, as the numbers have particularsignificanceL > 	on Unix systems.  In your case you have two non-Unix systems duking itout6 > 	using NFS, each trying to accomodate NFS' Unixisms.  G 	Yeah, I'm roughly familiar with unix, but not enough to translate the PK functionality between it and VMS very well.  I think the pcnfs part really e confused me also.    > 	I hope this helps.s   	Much, thank you!a   ------------------------------  # Date: Tue, 20 Aug 2002 16:54:04 GMTm: From: "John Gemignani, Jr." <John.Gemignani@youknow.where>/ Subject: Re: NFS Server configuration questions - Message-ID: <3D62398B.7CF6F514@youknow.where>M   Sharon Guthrie wrote:e > l > In article <3D6135E7.2F9E9A68@youknow.where>, "John Gemignani, Jr." <John.Gemignani@youknow.where> writes: > L > >       The example here shows you that you define ucx$nfs and ucx$nobody.Q > >       The example fred is just a regular user (nothing special, but the entrydR > >       for vmsuser is for a PC.  fred is a user who may be on one or morehosts.O > >       The PC user "vmsuser" usually comes in from one PC (PCs *usually* butu$ > >       not always) have one user. > ( >         Ah, thanks, that clears it up. > L > >       When you authenticate yourself using a username and password usingO > >       PCNFS, you get a UID and GID back as a result.  NFS exchanges uid andtM > >       gid information, not username and password.  They do not need to besM > >       in the same group unless you want them to be able to share files to2N > >       some degree -- similar rules that you would apply to your VMS users. > Q >         It's hard to diagram this with plaintext, but I need to try.  Let's seee > if I understand: > K > 1.  PC tries to connect to fileshare by sending a username/password pair. > > Say for example, "user1" "passwd", and the pc name is "pc1". > J > 2.  UCX tries to match that up with a proxy entry.  It finds a proxy forN > username "user1" on host "pc1" (or asterisk, but let's not use wildcards forO > this example).  It sends, via PCNFS, the uid/gid pair to the PC's NFS client.t > Let's say that pair is 52,43.9 > N > 3.  Further communication along this particular stream are identified by theH > uid/gid pair (52,43) and the proxy database is not referenced.  Right?  B 	The request is sent with AUTH_UNIX which says that you are 52,43.D 	The NFS server then uses that information to match you in the proxyB 	database to a VMS user.  Using VMS protection checking of the VMSD 	user (UIC, privileges and identifiers) access to various objects on% 	the server (i.e. files) are checked.-   > P > Question:  how does the uid/gid pair grant access to a particular export path?L > Is it only by the /host=xyz that you attach to the export?  So if you haveF > wildcarded the export, then all of your clients can map it, correct?    F 	The exports list is qualified by the host.  When a mount operation isF 	performed, the client does a "get mount entry" (which is only checked? 	by host), then it usually does a "file system status" and "get  attributes"t> 	of the mount point.  "File system status" does not need to be
 protectionF 	checked, but the "get attributes" is, so you need to have read access 	against the directory file.   > 3 > >>  Do I REALLY need to have a proxy for each pc?2 > > R > >       No, that's what /host=* is for.  Each PC could obtain a uid and gid fromP > >       PCNFS using the same username and password.  Define a proxy record for! > >       that user with /host=*.  > O >         Ok, that's clear.  So all my pc clients would obtain the same uid/gid J > pair, since they all send the same username/password, right?  Would thisM > confuse the data streams at all?  In my case, I want them all to access theaL > same directory.  They'll each use a separate data file, though, and that's# > controlled in the PC application.   H 	There should be no confusion.  The requests are qualified by the host's: 	address and the XID (transaction ID) of the request.  The authentication: 	information (uid, gid, etc) are arguments to the request.  N >         The applications on both ends that write to the data files will keepQ > the data files straight, no problem.  But I don't want problems with sockets or M > streams getting hung up and preventing the applications from writing to thei > files.  Thus my concern with:l > H > >> the value of the cache timer:  sysconfig -r nfs tcp_idle_timeout=30T > >> Coworkers who've done this before tell me this works, but I haven't found it toM > >> work on any of the versions of UCX I've been fighting with.  (5.0A, 5.1). > >@S > >       The tcp timer is not available in V5.0 because tcp support for NFS wasn't0O > >       released until V5.1.  The idle timer is used to close TCP connections Q > >       that haven't sent NFS requests in that period of time.  This forces thevC > >       client to reconnect when it wants to make an NFS request.  > >iQ > >> cache timer is apparently set high and causes a problem for our application.  > >AQ > >       Cache timer?  Which one?  Are you talking about the vfs vnode_age timerg= > >       which is set to 120 seconds to keep files accessed?l > O >         I meant the tcp timer, but have problems with the terminology.  Sorrye
 > about that. O >         When testing this whole thing on my development machine, I noticed angL > occasional problem where my VMS application updated the data file, but theQ > changes weren't visible on the PC end.  This was a major problem.  I'm not suretC > if it was the tcp timer or just that I had NFS way misconfigured.c  F 	I see.  The NFS server does not share files with the VMS applicationsE 	all that well because it does not use RMS to access the files.  I am6 workingsE 	on code for the V5.4 timeframe to improve and, I expect, solve this.L 	You have two options:  B 		1) reduce your vnode_age to keep files open as short as possibleE 		2) use the Unix approach of a lockfile, which I believe should workg   > Q > >       In my experience I find that users who are not familiar with Unix oftencR > >       can't feel comfortable with NFS. The whole Unix authentication stuff,forR > >       example, compared to VMS is close but no cigar.  You really have tothinkQ > >       it out -- it's not as simple to just add entries without thinking aboutuS > >       the values that you are using, as the numbers have particularsignificance S > >       on Unix systems.  In your case you have two non-Unix systems duking itout/= > >       using NFS, each trying to accomodate NFS' Unixisms.p > O >         Yeah, I'm roughly familiar with unix, but not enough to translate thesL > functionality between it and VMS very well.  I think the pcnfs part really > confused me also.: >  > >       I hope this helps. >  >         Much, thank you!   ------------------------------  + Date: Tue, 20 Aug 2002 03:55:06 -0700 (PDT)O. From: Fabio Cardoso <fabiopenvms@yahoo.com.br> Subject: OpenVMS + Joint STARS@ Message-ID: <20020820105506.97140.qmail@web20201.mail.yahoo.com>   From The Inquirer...   Click'  C http://www.capitol.northgrum.com/press_releases/joint_stars_11.htmlo   Regardsr   FC m   =====n ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazil  fabiopenvms@yahoo.com.br ==========================  2 __________________________________________________ Do You Yahoo!?& HotJobs - Search Thousands of New Jobs http://www.hotjobs.com   ------------------------------  # Date: Tue, 20 Aug 2002 14:38:53 GMTs5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>e" Subject: Re: OpenVMS + Joint STARS2 Message-ID: <1ws89.15$AT6.492702@news.cpqcorp.net>  I Ah...  That's a year old story.  Not to say it wasn't good, or that we dotH not continue to work with NG in Melbourne on the next generation planes.     Fabio Cardoso wrote in message7 <20020820105506.97140.qmail@web20201.mail.yahoo.com>...J >From The Inquirer...  >n >Click > D >http://www.capitol.northgrum.com/press_releases/joint_stars_11.html >e >Regards >a >FCi >o >===== >==========================a >Fbio dos Santos Cardoson >OpenVMS System Managere >Rio de Janeiro - Brazil >fabiopenvms@yahoo.com.brs >==========================g >e3 >__________________________________________________  >Do You Yahoo!?o' >HotJobs - Search Thousands of New Jobsw >http://www.hotjobs.comn   ------------------------------    Date: 20 Aug 2002 08:19:51 -0700% From: Alan Greig <a.greig@virgin.net>t" Subject: Re: OpenVMS + Joint STARS) Message-ID: <ajtmmn02jjl@drn.newsguy.com>g  @ In article <1ws89.15$AT6.492702@news.cpqcorp.net>, "Fred says... >rJ >Ah...  That's a year old story.  Not to say it wasn't good, or that we doI >not continue to work with NG in Melbourne on the next generation planes.,  0 Just don't mention the anti-grav engines ok? :-)   >? >Fabio Cardoso wrote in messageu8 ><20020820105506.97140.qmail@web20201.mail.yahoo.com>... >>From The Inquirer... >> >>Clicka >>E >>http://www.capitol.northgrum.com/press_releases/joint_stars_11.htmln >>	 >>Regards  >> >>FC >> >>=====o >>========================== >>Fbio dos Santos Cardoso >>OpenVMS System Manager >>Rio de Janeiro - Brazilo >>fabiopenvms@yahoo.com.br >>========================== >>4 >>__________________________________________________ >>Do You Yahoo!?( >>HotJobs - Search Thousands of New Jobs >>http://www.hotjobs.com >- >-   ------------------------------  # Date: Tue, 20 Aug 2002 16:55:15 GMTs5 From: "Fred Kleinsorge" <kleinsorge@star.zko.dec.com>i" Subject: Re: OpenVMS + Joint STARS2 Message-ID: <Tvu89.20$bX6.579606@news.cpqcorp.net>   Alan Greig wrote in message ...oA >In article <1ws89.15$AT6.492702@news.cpqcorp.net>, "Fred says...z >>K >>Ah...  That's a year old story.  Not to say it wasn't good, or that we dorJ >>not continue to work with NG in Melbourne on the next generation planes. >o1 >Just don't mention the anti-grav engines ok? :-)a >n  J Mentioning the anti-grav is OK, as are the dilithium crystals - that's allG unclassified/declassified.  But *never* mention the trans-warp conduit.a   ------------------------------  % Date: Tue, 20 Aug 2002 12:15:33 +0200-7 From: "Dijk, Jeroen van" <Jeroen.vanDijk@Getronics.com>   Subject: RE: OT: UK UniversitiesO Message-ID: <2795B75EF003D311801A00A0C906B511011C723F@cucexec.gbc.getronics.nl>A  V In Holland we know a little about floods and water so we are unless the rest of europe$ prepared for a little bit of water. b So we don't have any problem with those so called floods, btw there are no floods in Britain also.  m You can also go to Holland, our entry levels are a little bit higher then some of the british 'universities',su but the language entry levels are lower. If you speak and understand English that is enough. The rest of the languagee you will learn along the way.r       \ In article <3D6015A8.2781@rswheeldon.com>, Richard Wheeldon <richard@rswheeldon.com> writes: >Fabio Cardoso wrote:O3 >> I am really thinking to go to Europe - even withk >> the floods - after October. >,2 >Just remember to go somewhere with hills then ;-)   Yes, to get avalanches...   E In times of increasing environment problems you need luck to find thes8 least harmful places on earth. Vienna is one of them ;-)   -- c Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialistd E-mail  peter@langstoeger.atP A-1030 VIENNA  AUSTRIA              I'm looking for (a) Network _and_ VMS Job(s)   ------------------------------  % Date: Tue, 20 Aug 2002 17:12:06 +0200C, From: "Toine Dirven" <tdirven@volvocars.com>4 Subject: Printing barcodes via postscript on OpenVMS4 Message-ID: <ajtm85$3c236@eccws12.dearborn.ford.com>   Hello,  D I want to print barcodes on our laserprinters via our Alpha servers.G I want to develop an application on OpenVMS that creates the postscripto file.h* I will print this file to an laserprinter.- Has anyone examples or experience with this ?t0 Where can I find postscript fonts for barcodes ? Please help me.4    
 Best regards,.   Toine Dirven Volvo Cars Gent-   ------------------------------  % Date: Tue, 20 Aug 2002 10:57:30 -0700r. From: "David D Miller" <ddmiller@raytheon.com> Subject: Simple clustertF Message-ID: <OF81D5D9BD.634A99EC-ON07256C1B.0061F8C2@rsc.raytheon.com>   Cluster experts:  I All this talk about updating a system shadow volume got me thinking.  I'm F trying to imagine the simplest (and cheapest) redundant Alpha cluster. Will this work?   K Two Alphas on Ethernet and each has a disk.  The two disks make up a system C shadow volume.  Each Alpha has a vote.  Is a quorum disk necessary?"  $ Are there other, better suggestions?  
 TIA, dave.   ------------------------------  # Date: Tue, 20 Aug 2002 14:02:01 GMT . From: ">>> ^P" <plj@NOSPAM.byron.ext.telia.se> Subject: tps ?9 Message-ID: <3D624BF6.43C7791C@NOSPAM.byron.ext.telia.se>N   Hi VMS-fellows,d  B Does anyone know the exakt definition and it's origin of the "tps"G (Transaction Per Second?) used to describe performance figures on Alphao	 Systems ?e  > Seems like it's tending towards SPECint and SPECfp these days.  	 >>> ^P.Lj    ------------------------------  + Date: Tue, 20 Aug 2002 11:16:09 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>p3 Subject: Re: VMS upgrade with shadowed system disk.a; Message-ID: <01KLIOSO2PF69LVCGS@sysdev.deutsche-boerse.com>t  I Basic plan sounds fine (after a quick reading).  To be on the safe side, a7 use MOUNT/OVERRIDE=SHADOW_MEMBERSHIP where appropriate:.  C    Allows you to override the write protection of former shadow set'J    members. Applicable only if you have the volume shadowing option. Refer$    to Volume Shadowing for OpenVMS.     aD    When you mount a volume with this qualifier, the volume shadowingI    generation number is erased. If you attempt to remount the volume in a J    shadow set, the volume is considered an unrelated volume and receives a9    full copy operation from a current shadow set member. t   ------------------------------  % Date: Tue, 20 Aug 2002 10:30:05 +0100,( From: Nic Clews <sendspamhere@127.0.0.1>3 Subject: Re: VMS upgrade with shadowed system disk. ( Message-ID: <3D620C1D.16A3A2C@127.0.0.1>   JF Mezei wrote:s > $ > rob.buxton@wcc.spam.govt.nz wrote:7 > > >2. Shutdown and reboot from the DKC0: system disk. ; > > >   This will give me a DSA10: with just DKC0:, right ?  > 
 > Correct. > H > But if you wish to be paranoid, you could also disable the system diskL > shadowing in sysgen and set the default boot device to the particular hardK > drive. Not sure about this particular upgrade, but consider some that dideL > multiple reboots, so having the more "vanilla" system drive would simplifyN > things should problems occur during upgrade. Not sure such paranoia would be > worth the extra time/trouble.   @ Agreed here, this then makes the following steps redundant. JF'sD paranoia is warranted, with no shadowing code loaded, you reduce the risk to almost 0.t  G Typically, with *both* (or all) of the shadow set members mounted, I'llhH issue a SET SHADOWING 0 and WRITE CURRENT from SYSGEN, so conversational< booting to re-enable is a manual task. BEWARE of settings inD MODPARAMS.DAT when an AUTOGEN may override your wishes. When this is8 resolved, then I'll dissolve the system disk shadow set.  O > In any event, commenting out any of the mount commands that add drives to thee- > system drive shadowset would be a big must.i > 7 > > >   Now, will re-mounting DKC0: into DSA10: start aoH > > >   shadow copy B => C so I'll end up with a shadowed 7.2-1 system ? > O > Should be. As you crete DSA10 with the 7.2 system, that new shadowset becomes P > more current than the contents of the 7.3 hard drive that was offline, so whenK > the later is added to DSA10, it becomes a target of a copy, not a source.t  B Shadowing's prime directive is to ensure that the data between two0 volumes is identical, not necessarily 'correct'.  A In the case of a system disk, simply booting it successfully as aaF shadowset member makes it the master and therefore the source for copyE operations, regardless of the contents of the Shadow Control Block onrE any other disks which were, or are to be, members in this shadow set.-  G So the added drive becomes a copy target merely because of the command,eC nothing to do with the currency of the contents. If booting the oldo$ volume, it could be easy to miss the  @ %SYSINIT-I-INCSHAMEM, system shadow set membership inconsistency  E message on startup, and any disks in the /SHADOW string not currentlye! mounted will become copy targets.l  C The throwback is the ease and speed with which you can back out thef upgrade.  C (Data disks are slightly more deterministic about source and target- selection).-  H Good luck with the upgrade, I don't think you'll need it but I wanted to1 clear up possible confusion and misunderstanding.o -- r? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciences  nclews at csc dot come   ------------------------------  % Date: Tue, 20 Aug 2002 11:56:53 +0200 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>x3 Subject: Re: VMS upgrade with shadowed system disk.n' Message-ID: <3D621265.1E1ECD91@aaa.com>t   OK.l Many thanks all !e  - Now, my modified check-list looks like this :r  2 - Comment out the following lines in MODPARS.DAT :!   !!! SHADOWING               = 2s!   !!! SHADOW_SYS_DISK         = 1u"   !!! SHADOW_SYS_UNIT         = 10  8 (I suppose the ALLOCLASS param can be left "enabled" !?)  : - In SYSGEN, set SHADOWING and SHADOW_SYS_DISK both to "0"   and WRITE CURRENT   : - Comment out some application specific code and data-disk4   mounts in SYSTARTUP (just to get a faster reboot).  : (Now both copies of the system disk have all this changes,? so it doesn't matter which disk I use for the upgrade, right ?)   8 - Reboot to get a non-shadowed system with just a single*   system disk running and no "data" disks.   - Upgrade VMS !a   - Of OK.5   Re-enable the MODPARAMS params and SYSTARTUP edits..   Reboot and enjoy !6   (A shadow copy will start when the system is booted)   - Of *not* OK.+   Reboot from the second (old) system disk. 5   Re-enable the MODPARAMS params and SYSTARTUP edits.t+   Reboot and get back to the drawing-board.t5   (A shadow copy will start that overwrites the newly     upgraded disk)s     Another thing.> In a few of the replies, there was a warning about any MOUNT's7 of members into the system shadow set from the startup.iA I'v no such MOUNT's in my startup. VMS does this by itself, not ?p@ I think I just made a *manual* MOUNT of the second disk into theC system shadow set, and after that, VMS has taken care of it at eache reboot.f   Another "another thing"...7 There is no problem with booting the system from a diskk3 that formely was a shadow set member ? The message sB "%SYSINIT-I-INCSHAMEM, system shadow set membership inconsistency"8 is just a "-I-", and doesn't mean any real problem, or ?    Now back to the Release Notes...   Best Regards< (and sorry for interupting the usual c.o.v threads with some *technical* issues :-) )   Jan-Erik Sderholm.a   ------------------------------  + Date: Tue, 20 Aug 2002 13:17:58 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>y3 Subject: Re: VMS upgrade with shadowed system disk.d; Message-ID: <01KLISX1TFPI970ARJ@sysdev.deutsche-boerse.com>o  # >   !!! SHADOWING               = 2o# >   !!! SHADOW_SYS_DISK         = 1u$ >   !!! SHADOW_SYS_UNIT         = 10  = Actually, I think that if you have SHADOW_SYS_UNIT set, then oC SHADOW_SYS_DISK is set automatically, but it is better to set both t explicitly.=  E Any reason to use 10?  I use things like 1236, where 1 and 3 are the  E allocation classes of the machines to which the disks are physically hH attached, and 2 and 6 are the corresponding SCSI addresses.  This makes I bookkeeping a bit easier.  (No application, procedure etc would ever use e3 the DSA name of course, but rather a logical name.)a  : > (I suppose the ALLOCLASS param can be left "enabled" !?)   Yes, no problem there.  < > - Comment out some application specific code and data-disk6 >   mounts in SYSTARTUP (just to get a faster reboot).  H If you didn't do this, then you would have to change the mount commands G if you changed the allocation class (or, if you are using logicals for )B the disk names---as should be the case---would have to change the C definition of the logicals)---unless you are using just the "short h. names" to mount them (or define the logicals).  < > (Now both copies of the system disk have all this changes,A > so it doesn't matter which disk I use for the upgrade, right ?)    Right.  @ > In a few of the replies, there was a warning about any MOUNT's9 > of members into the system shadow set from the startup.hC > I'v no such MOUNT's in my startup. VMS does this by itself, not ?   C Right.  There is an explicit warning against this in the shadowing l> documentation, including an example of why it is bad practice.  B > I think I just made a *manual* MOUNT of the second disk into theE > system shadow set, and after that, VMS has taken care of it at eachr	 > reboot.    Right.   ------------------------------  + Date: Tue, 20 Aug 2002 11:08:38 +0000 (UTC)i From: david20@alpha2.mdx.ac.uk3 Subject: Re: VMS upgrade with shadowed system disk.h+ Message-ID: <ajt7vm$pog$1@aquila.mdx.ac.uk>h  c In article <3D615C36.9FF91760@aaa.com>, Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:  >Hi all.* >Just to check that I'm "on track", is the >following scenario OK ? > , >- AS1200 system with shadowed system disk :F >  DSA10:     Mounted              0  ALPHASYS       4934070   613   18 >  $1$DKB0:   ShadowSetMember      0  (member of DSA10:)8 >  $1$DKC0:   ShadowSetMember      0  (member of DSA10:) >t7 >DKB0: is the BOOTDEV_DEF (or DEFBOOT_DEV or whatever).w4 >DKC0: is the "second" disk in the DSA10 shadow set. >a  < >Goal : to upgrade one of the disk and still having a way of6 >backing out to the old version without going to tape. >o >My current plan : >m   0)  & Edit modparams and set shadowing to 0.   Autogen  savparams setparams  I Unless something has changed very recently you cannot upgrade a shadowed r disk.         . >1. Dismount DKC0: from the DSA10: shadow set. >n After this do a    MOUNT/OVER=(ID,SHAD) DKC0: DISMOUNT DKC0:  F This will get rid of the write protection of former shadowset members.      3 >2. Shutdown and reboot from the DKC0: system disk.s7 >   This will give me a DSA10: with just DKC0:, right ?t >    No.t  ( shutdown and Boot from the OS upgrade CD     >3. Upgrade DKC0: to 7.3-1.t >    Yes   C When completed boot system and turn shadowing back on in modparams.y   (Paranoia -     mount/over=(id,shad) dkb0:   dismount dkb0:    E This will insure that dkb0 won't be automatically added back into the 
 shadowset.   )l   Reboot    : >4. If all is OK with 7.3-1 => Re-mount DKB0: into DSA10:.2 >   This with start a schadow copy DKC0: --> DKB0:< >   and end up with a finished, shadowed 7.3-1 system, not ? >    Yes     7 >5. If problems with 7.3-1 => Shutdown and reboot usinge= >   DKB0: (The "old" 7.2-1 system). This will create a DSA10:f& >   shadowset with just DKB0:, right ?  F No. Since you autogened in step 0 to set shadowing 0 you will need to A do a conversational boot and turn shadowing back on in the rebootr9 process. Then once up you can alter modparams back again.   C  3 >   Now, will re-mounting DKC0: into DSA10: start atD >   shadow copy B => C so I'll end up with a shadowed 7.2-1 system ?   Yeso   I'd tend to do a =  @ MOUNT/SYS DSA10:  /shad=(dkb0:,dkc0:) ALPHASYS ALPHASYS /CONFIRM  J Just to be sure. Then it will tell you what disk it is going to use as the@ source and which disk it will use a the destination of the copy.    = >   Or will the shadow logic deside that the "old" DKB0: disk ? >   is older then the new (7.3-1) DKC0: disk and start the copy7 >   B <= C instead ? >h > 
 >Best Regardso >Jan-Erik Sderholm   
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Tue, 20 Aug 2002 12:54:34 +01002( From: Nic Clews <sendspamhere@127.0.0.1>3 Subject: Re: VMS upgrade with shadowed system disk. ) Message-ID: <3D622DFA.900453BB@127.0.0.1>e   Jan-Erik Sderholm wrote:  >  > OK.a > Many thanks all != > / > Now, my modified check-list looks like this :n > 4 > - Comment out the following lines in MODPARS.DAT :# >   !!! SHADOWING               = 2=# >   !!! SHADOW_SYS_DISK         = 1_$ >   !!! SHADOW_SYS_UNIT         = 10 > : > (I suppose the ALLOCLASS param can be left "enabled" !?)   But of course.   < > - In SYSGEN, set SHADOWING and SHADOW_SYS_DISK both to "0" >   and WRITE CURRENT< > < > - Comment out some application specific code and data-disk6 >   mounts in SYSTARTUP (just to get a faster reboot).  ) Wouldn't you be doing a MINIMUM startup ?.  B Arguably, I put SYSTARTUP_P1 = "MIN" in MODPARAMS.DAT which at theD 'final' reboot stops the system wandering into a full startup before' you've had chance to do a few checks...e  < > (Now both copies of the system disk have all this changes,A > so it doesn't matter which disk I use for the upgrade, right ?)m   Nope.i  : > - Reboot to get a non-shadowed system with just a single, >   system disk running and no "data" disks. >  > - Upgrade VMS != > 
 > - Of OK.7 >   Re-enable the MODPARAMS params and SYSTARTUP edits.   < You're saved editing SYSTARTUP if doing SYSTARTUP_P1 = "MIN"  B You'll need to use SYSGEN to enable shadowing, SHADOW_SYS_DISK and  SHADOWING to the correct values.     >   Reboot and enjoy !8 >   (A shadow copy will start when the system is booted)  . You'll need to manually add the "copy target".  E From then on, VMS will 'remember' which disks formed this shadow set.c  i > - Of *not* OK.- >   Reboot from the second (old) system disk.77 >   Re-enable the MODPARAMS params and SYSTARTUP edits.o  1 Or SYSTARTUP_P1 = "", and the SYSGEN settings :-)   - >   Reboot and get back to the drawing-board.o7 >   (A shadow copy will start that overwrites the newlyt >    upgraded disk)i  G Arguably you could delay the *manual* merge and do a post mortem on thei failed upgrade disk.  n > Another thing.@ > In a few of the replies, there was a warning about any MOUNT's9 > of members into the system shadow set from the startup.cC > I'v no such MOUNT's in my startup. VMS does this by itself, not ?1B > I think I just made a *manual* MOUNT of the second disk into theE > system shadow set, and after that, VMS has taken care of it at each1	 > reboot.1  D System disks, when mounting as shadow sets see the documentation forC /INCLUDE, this is the type of behaviour you see, with slightly morea precedent on the booted disk.o  F For data disks, a MOUNT/INCLUDE should form the shadow sets correctly.  B > Another "another thing"...9 > There is no problem with booting the system from a disk 4 > that formely was a shadow set member ? The messageD > "%SYSINIT-I-INCSHAMEM, system shadow set membership inconsistency": > is just a "-I-", and doesn't mean any real problem, or ?  H Well it is informational as you point out. The problem is it is not in 6D foot high red flashing letters with alarm bells sounding. What it isH telling you is that the contents of the SCB may not be all as they seem.F I booted a removed member of a shadowed system disk to get this error,! and the system continued to boot.g  H To carry on what could happen, a MOUNT/SHAD including both members wouldF make this potentially stale disk the copy source for shadowing, losingF the hard work on what you thought was the primary. In fact I did this,; while perhaps you'd hope it would complain, it doesn't, thed& informational message was the warning.  D I've spent some time experimenting with a test system with scenarios. you'd be nervous to try in a live environment.   This is a good Storage article:g  A [SHADOW] Volume Shadowing Phase II Overview, Questions and Common  Problems  F File System Internals [McCoy] documents the contents of the SCB, among$ nearly everything else in ODS2 land.  G I have either missed it, or it isn't documented, as to how the /INCLUDEeB processing operates, where the data is taken from on the volume toH reform the shadow set. I'll have a root in the source listings when I've time...G   --  ? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciences  nclews at csc dot comM   ------------------------------    Date: 20 Aug 2002 07:12:18 -0700. From: SPAMSINK2001@YAHOO.COM (Alan E. Feldman)3 Subject: Re: VMS upgrade with shadowed system disk. = Message-ID: <343f30ae.0208200612.37b453ee@posting.google.com>t  O rob.buxton@wcc.spam.govt.nz wrote in message news:<3d6188c4.1815319040@news>... . > On Mon, 19 Aug 2002 22:59:34 +0200, Jan-Erik3 > =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> wrote:p > 
 > >Hi all., > >Just to check that I'm "on track", is the > >following scenario OK ? > >v. > >- AS1200 system with shadowed system disk :H > >  DSA10:     Mounted              0  ALPHASYS       4934070   613   1: > >  $1$DKB0:   ShadowSetMember      0  (member of DSA10:): > >  $1$DKC0:   ShadowSetMember      0  (member of DSA10:) > >a9 > >DKB0: is the BOOTDEV_DEF (or DEFBOOT_DEV or whatever).r6 > >DKC0: is the "second" disk in the DSA10 shadow set. > >t> > >Goal : to upgrade one of the disk and still having a way of8 > >backing out to the old version without going to tape. > >e > >My current plan : > > 0 > >1. Dismount DKC0: from the DSA10: shadow set. > >s5 > >2. Shutdown and reboot from the DKC0: system disk.d9 > >   This will give me a DSA10: with just DKC0:, right ?s > >o > >3. Upgrade DKC0: to 7.3-1.  > >(< > >4. If all is OK with 7.3-1 => Re-mount DKB0: into DSA10:.4 > >   This with start a schadow copy DKC0: --> DKB0:> > >   and end up with a finished, shadowed 7.3-1 system, not ? > >m9 > >5. If problems with 7.3-1 => Shutdown and reboot usingt? > >   DKB0: (The "old" 7.2-1 system). This will create a DSA10:o( > >   shadowset with just DKB0:, right ?5 > >   Now, will re-mounting DKC0: into DSA10: start auF > >   shadow copy B => C so I'll end up with a shadowed 7.2-1 system ?? > >   Or will the shadow logic deside that the "old" DKB0: diskfA > >   is older then the new (7.3-1) DKC0: disk and start the copy- > >   B <= C instead ?    F When you manually add a disk to a shadow set, it will be the target ofD a shadow copy. Use the /CONFIRM qualifier when you do it and it willD tell you that. You only have to worry about this when you "create" aB shadow set. What I mean by create is the first MOUNT command after dissolving the shadow set.  @ $ DISM DSA0:    ! dissolve the set      ! (contained $1$dka0 and $1$dka1); $ MOUNT DSA0:/SHADOW=($1$DKA0,$1$DKA1)  ! create shadow sets  A Adding a disk to a mounted shadow set always overwrites that diskN@ because that disk will always be older than the already-monunted disks.    B > No, I'm pretty sure the Logic understands that the device you've# > booted from is the latest member.     D I'd do a standalone backup or its equivalent of the system disk that9 is still running when you shut down just as a precaution.a    FD > So, the plan is sound.  Just check there are no mount commands for > DSA10 anywhere.0    @ You should not have commands for adding members to a system disk5 shadow set in the startup. From the shadowing manual:m  
 [begin quote]rB The boot process automatically locates all the members of a systemE disk shadow set. You should not add system disk shadow set members insE startup procedures as formerly recommended when phase I shadowing wase
 supported.  (                                  Caution  E Do not add members to a system disk shadow set in startup procedures. F Doing so can result in loss of data under the following circumstances:  A A system is operating normally with a multiple member system disk. shadow set.   F The original boot device is removed from the shadow set but remains as a functioning disk.   1 The system continues with the remaining members.    % The system is shut down or it fails.    C The system is rebooted using the original boot device (which is now  out of date). t  B The boot process determines that the boot device is not consistentC with the other shadow set members and, therefore, does not add themr into the shadow set.  B This behavior preserves the up-to-date data on the other members.   B A MOUNT command in the startup procedure adds the other shadow set& members to the system disk shadow set.  E A copy operation from the boot device to the other shadow set members ' is initiated, thereby overwriting them.n   [end quote]n      Disclaimer: JMHO, i.e., OCICBWYK Alan E. FeldmanS- afeldmanNonospam gfigroupNonospam NonospamComi   ------------------------------   End of INFO-VAX 2002.458 ************************