1 INFO-VAX	Tue, 05 Sep 2006	Volume 2006 : Issue 486       Contents: Re: Blue Ray Discs= Debugging image invoked by SYMBIONT via LIB$FIND_IMAGE_SYMBOL A Re: Debugging image invoked by SYMBIONT via LIB$FIND_IMAGE_SYMBOL & Free MicroVax 3100-80 on NJ Craigslist# Heads Up:  Intel to axe 20,000 jobs  Hobbyist site down?? Re: Hobbyist site down?? Re: Hobbyist site down??  Re: multinet / cluster questions( Re: OpenVMS 7.3-2 to 8.2 Upgrade Warning" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on Itanium" RE: OpenVMS and Windows on Itanium" Re: OpenVMS and Windows on ItaniumC Re: Please Loot the Inforamtion/Mainframe Material  at "MFRESOURCE" ? Please Loot the Inforamtion/Mainframe Material  at "MFRESOURCE" . Re: VMS STARTUP documentation mis-leading V7.3. Re: VMS STARTUP documentation mis-leading V7.3  F ----------------------------------------------------------------------  # Date: Tue, 05 Sep 2006 04:48:19 GMT 3 From: Vance Haemmerle <vance@toyvax.Glendale.CA.US>  Subject: Re: Blue Ray Discs < Message-ID: <nk7Lg.6413$yO7.2893@newssvr14.news.prodigy.com>   Neil Rieck wrote: 8 > "Tom Linden" <tom@kednos-remove.com> wrote in message % > news:op.te5whbcgtte90l@hyrrokkin...  > 4 >>Are these going to be available for OpenVMS Alpha? >>Tom  >  > N > I hope not. One of my colleagues has done quite a bit of research on BluRay M > vs. HD-DVD and BluRay might be a big mistake and a big waste of time (just  N > like BetaMax was). The biggest problem relates to data being written on the F > external surface of the disk making it vulnerable to destruction by N > scratching. (their solution to fixing this problem is the reintroduction of  > the caddy. Yuk)  > + > http://en.wikipedia.org/wiki/Blu-ray_Disc  > % > http://en.wikipedia.org/wiki/HD_DVD  >  > Neil Rieck > Kitchener/Waterloo/Cambridge,  > Ontario, Canada.# > http://www3.sympatico.ca/n.rieck/  >  >   >   BetaMax wasn't a mistake - for people who appreciated better= quality video and audio.  My decks are still running, they're < from 1987. I've gotten more than my money's worth from them.   Vance    ------------------------------  % Date: Mon, 04 Sep 2006 17:56:06 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> F Subject: Debugging image invoked by SYMBIONT via LIB$FIND_IMAGE_SYMBOL, Message-ID: <44FCA0EC.17DD62C7@teksavvy.com>  6 I have a very crude VMS MAIL foreign transport image.   D From within VMS Mail, it works fine, and is able to use C file io to" create a log file and write to it.  H However, for a message processed by the TCPIP symbiont, it fails with anF access violation. (symbiont survives, issues non-delivery notification+ and continues processing for next message).   ; the symbiont allegedly uses callable mail which itself used @ lib$find_image symbol to load my routine into memory and call itG repeatedly. The first time it is called, my routine is supposed to open 5 the log file. It doesn't appear to even get that far.   E The symbiont itself doesn't have any standard SYS$OUTPUT or SYS$ERROR G files, so I can understand statements such as printf failing.  But I've @ removed all printf's. And this image doesn't make use of any DCLF facilities. (It does use LIB$GET_LOGICAL though, but that is permitted= even for a process that is detached and without DCL, right ?)   D I've verified that the TCPIP$SMTP user account is able to create the3 file in its own directory (via decnet and a proxy).     C Any hints on how one can go about debugging an image (mine) that is A invoked by an image (callable mail) that is invoked by a symbiont  (TCPIP$SMTP_SYMBIONT) ?????   H Since it appears to work from DCL "MAIL" command, what differences wouldH there be when invoked by the callable mail interface ? Or is it the factD that it runs as part of a symbiont ? Shirley, the TCPIP symbiont hasA been able to process destinations that are foreign transports for 	 VMSMAIL ?   Q (The destination is actually a username who has a "SET FORWARD" to my transport).    ------------------------------  $ Date: Tue, 5 Sep 2006 06:59:30 +08003 From: "Richard Maher" <maher_rj@hotspamnotmail.com> J Subject: Re: Debugging image invoked by SYMBIONT via LIB$FIND_IMAGE_SYMBOL1 Message-ID: <ediaup$1c7$1@news-02.connect.com.au>    hi JF,  L I don't know if it would help in your particular circumstance but could yourG shareable that has been loaded by lib$fis call lib$signal to invoke the H debugger? I guess the main image has to be linked with /traceback and itJ would be nice to be able to $set display/create somewhere and these things& may be out of your control but FWIW: -  L If you look at your t3$examples directory for the file demo_tip.cob you will see the following example: -  I     call "lib$get_symbol" using by descriptor "T3$DEBUG", debug_on giving  sys_status. A     if sys_status not = ss$_normal call "lib$stop" using by value  sys_status.        if debug_on = "Y"          call "lib$signal" 2               using   by value        ss$_debug, 15                       by reference    debug_commands.    Regards Richard Maher     : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message& news:44FCA0EC.17DD62C7@teksavvy.com...7 > I have a very crude VMS MAIL foreign transport image.  > F > From within VMS Mail, it works fine, and is able to use C file io to$ > create a log file and write to it. > J > However, for a message processed by the TCPIP symbiont, it fails with anH > access violation. (symbiont survives, issues non-delivery notification- > and continues processing for next message).  > = > the symbiont allegedly uses callable mail which itself used B > lib$find_image symbol to load my routine into memory and call itI > repeatedly. The first time it is called, my routine is supposed to open 7 > the log file. It doesn't appear to even get that far.  > G > The symbiont itself doesn't have any standard SYS$OUTPUT or SYS$ERROR I > files, so I can understand statements such as printf failing.  But I've B > removed all printf's. And this image doesn't make use of any DCLH > facilities. (It does use LIB$GET_LOGICAL though, but that is permitted? > even for a process that is detached and without DCL, right ?)  > F > I've verified that the TCPIP$SMTP user account is able to create the5 > file in its own directory (via decnet and a proxy).  >  > E > Any hints on how one can go about debugging an image (mine) that is C > invoked by an image (callable mail) that is invoked by a symbiont  > (TCPIP$SMTP_SYMBIONT) ?????  > J > Since it appears to work from DCL "MAIL" command, what differences wouldJ > there be when invoked by the callable mail interface ? Or is it the factF > that it runs as part of a symbiont ? Shirley, the TCPIP symbiont hasC > been able to process destinations that are foreign transports for  > VMSMAIL ?  > G > (The destination is actually a username who has a "SET FORWARD" to my  transport).    ------------------------------   Date: 4 Sep 2006 12:59:01 -0700  From: paulinnj2002@yahoo.com/ Subject: Free MicroVax 3100-80 on NJ Craigslist C Message-ID: <1157399941.138082.211190@m79g2000cwm.googlegroups.com>   0 This isn't me -- just thought I'd pass it along:  2 http://newjersey.craigslist.org/zip/187328311.html  ' Reply to: sale-187328311@craigslist.org  Date: 2006-07-28, 12:13PM EDT    *Free* MicroVax System  D **We are a business and must be picked up between 8am and 5 pm M-F**   *MicroVax 3100-80 
 *40 mb memory  *Ka47-A V1.2-343-V4 CPU    Items included are: 
 1 - Micro Vax  1 - Storage Expansion  1 - 32 port Terminal Server  1 - T285 Tape Drive 	 1 - Delni * 1 - Sypher F880 3200/1600 bpi Reel to Reel 3 - VT420 Terminals   0 this is in or around Englewood Cliffs (07632) NJ   ------------------------------  % Date: Tue, 05 Sep 2006 01:41:13 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> , Subject: Heads Up:  Intel to axe 20,000 jobs, Message-ID: <44FD0DD2.617FE1BA@teksavvy.com>  C BBC reports that Intel is to announce it is axing up to 20,000 jobs  today. (20% of its workforce)   D One of the analysts interviewed by BBC mentioned that Itanium is one6 reason Intel fell behind and lost market share to AMD.  E Other analysts mention Intel's marketing flobs such as Viao and other  unknown branding attempts.    F While I doubt we would get the good news that IA64 is to be put out of@ its misery right away, I think that Intel will focus on the 8086G architecture and Ia64 isn't going to get an increase in human resources E anytime soon , which means that chip will lag behind at a faster rate . compared to the fast moving 8086 product line.  H I think Hoff will only be able to say "there are no plans to port VMS toD *****" for another few months, after that, he may need to change hisF tune and perhaps add the word "announced" or "public" or "official" to his statement.   ------------------------------  # Date: Mon, 04 Sep 2006 21:20:53 GMT 0 From: John Santos <john.santos@post.harvard.edu> Subject: Hobbyist site down?? > Message-ID: <MPG.1f6662be322cb536989725@news.bellatlantic.net>  ? Is the VMS Hobbyist site down or having problems?  (Maybe a DOS  attack?)  : Since yesterday, I can't register for licenses.  (Opera on@ Win2K and Safari on O/S X.)  The behaviour is a little flaky and> different on each (and I seem to be getting farther today than
 yesterday.)     D Yesterday it seemed totally dead.  Today, I can fill in the 1st pageG of the license registration form (name, address, email info, etc), but  E the response never comes back and eventually I get a "server stopped   responding" message.  A Ping seems intermittent.  From home, I'm getting about 30% packet B loss.  From work, nothing (but maybe work's firewalls are blocking@ ping...  Two different firewalls on 2 Verizon DSL lines at work,  and yet another VZ DSL at home.)  @ I just fired up my MV4000-300 for the first time since mid-july,? and discovered all my licenses had expired :-(  I have new VMS, @ LP and TCPware licenses for my Alpha (at work pending upgrade to= VMS 8.3) so I was able to update the LP licenses, but I don't = think I can fake out the VMS base license or TCPware license, 5 except by setting the date back to early August :-(        --   John   ------------------------------  % Date: Mon, 04 Sep 2006 18:14:36 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> ! Subject: Re: Hobbyist site down?? , Message-ID: <44FCA541.8C0E888E@teksavvy.com>   John Santos wrote: > A > Is the VMS Hobbyist site down or having problems?  (Maybe a DOS 
 > attack?)    
 I can get to: ( http://www.montagar.com/htbin/ohp_hobreg  E which asked to confirm that we agree to the terms and conditions. But 8 pressing "REGISTER LICENSE" yields nothing (no response)    E Pressing the button first yiels a connection to montagar, followed by E (probably a reditect) a connection attempt to www.openvmshobbyist.com  which has no answer.    G Going to the www.openvmshobbyist.com directly also yields no answer. (I H tend to always go to www.montagar.com first because it is an easier name9 to remember than that long name for the hobbyist system).   > <BREAKING NEWS> Connection to www.openvmshobbyist.com has been? established, and after a long pause, the page finally appeared.     F Perhaps Mr Cathay has fired up ALLIN1 on the vms hobbyists site and it was temporarly overloaded :-)     3 Seems to be working now. (fixed a few seconds ago).    ------------------------------  # Date: Mon, 04 Sep 2006 23:00:00 GMT 0 From: John Santos <john.santos@post.harvard.edu>! Subject: Re: Hobbyist site down?? > Message-ID: <MPG.1f6679fd9ff373ca989727@news.bellatlantic.net>  - In article <44FCA541.8C0E888E@teksavvy.com>,  $ jfmezei.spamnot@teksavvy.com says... > John Santos wrote: > > C > > Is the VMS Hobbyist site down or having problems?  (Maybe a DOS  > > attack?) >  >  > I can get to: * > http://www.montagar.com/htbin/ohp_hobreg > G > which asked to confirm that we agree to the terms and conditions. But : > pressing "REGISTER LICENSE" yields nothing (no response) >  > G > Pressing the button first yiels a connection to montagar, followed by G > (probably a reditect) a connection attempt to www.openvmshobbyist.com  > which has no answer.   > I > Going to the www.openvmshobbyist.com directly also yields no answer. (I J > tend to always go to www.montagar.com first because it is an easier name; > to remember than that long name for the hobbyist system).  > @ > <BREAKING NEWS> Connection to www.openvmshobbyist.com has beenA > established, and after a long pause, the page finally appeared.  >  > H > Perhaps Mr Cathay has fired up ALLIN1 on the vms hobbyists site and it  > was temporarly overloaded :-)  >  > 5 > Seems to be working now. (fixed a few seconds ago).   A Got'em.  Once it accepted the 2nd page (with model and SN), I got - the licenses back almost instantly.  Yippee!!   ! (And thanks to whoever fixed it.)        --   John   ------------------------------  % Date: Mon, 04 Sep 2006 20:35:16 -0400 ' From: Rich Whalen <whalenr@process.com> ) Subject: Re: multinet / cluster questions 8 Message-ID: <9dhpf2h7os6au4tmrq4n0rq06bn528eh34@4ax.com>  D On Mon, 04 Sep 2006 15:05:47 +0200, rejoc <rejoc@FREEfree.fr> wrote:   >Rich Whalen a crit :G >> On Mon, 04 Sep 2006 10:51:08 +0200, rejoc <rejoc@FREEfree.fr> wrote:  >>  > >>> Multinet 4.4 in a cluster where ip-cluster-aliases defined >>> E >>> 2/ For one specific outgoing service and on the node holding the  M >>> cluster's address, I would like to have the cluster's IP address used as  . >>> the outgoing IP@ (and not the interface @) >>>  >>   >> Can't be done. % >Is a newer version able to do that ?  > I >Is it the same if I define a pseudo interface (pd0) and want to use its  0 >address as the outgoing address for a service ?  A I thought about this a bit more, and there may be a way to do it. ' Your program needs to do the following:  socket()& bind(socket, desired outgoing address)$ connect(socket, destination address) exchange packets
 close(socket)   ! Note that I have not tested this.  >  >>> TiA    ------------------------------   Date: 4 Sep 2006 20:51:03 -0200 6 From: peter@langstoeger.at (Peter 'EPLAN' LANGSTOEGER)1 Subject: Re: OpenVMS 7.3-2 to 8.2 Upgrade Warning , Message-ID: <44fc91b7$1@news.langstoeger.at>  Y In article <4m2nf5F47jc4U1@individual.net>, Martin Vorlaender <mv@pdv-systeme.de> writes:  >Neil Rieck wrote:N >> In the past 24 hours I've updated two Alphas from OpenVMS-7.3-2 (Update 7) M >> to OpenVMS-8.2 and the existing TCPware-5.7-2 stack was rendered unusable   >> both times. [...]  I I run V5.7-2 on OpenVMS Alpha V8.2 (and before on V7.3-2, for some hours) G without any such problem. (My only problem is incoming DECnet-over-IP).   M In case you haven't done: You need a reinstallation of TCPware after VMS upg. . But this have been the case for decades now...  O >> If you're doing a similar upgrade and want to save some time, make sure you  O >> disable TCPware startup in "sys$manager:systartup_vms.com" before you start   >> your upgrade.  5 ...only until you reinstalled TCPware on OpenVMS V8.2   E >Not sure whether this message on the TCPware-Info list from Richard  6 >Whalen of Process also applies to 8.2, but here goes:  $ No, it applies to OpenVMS Alpha V8.3   ><quote># >DRIVERS_V572P021 contains the fix. I >It is necessary to install this kit after upgrading to VMS V8.3 because  H >the kit uses the version of VMS that it is running on to determine the 1 >variant of the UCX$IPC_SHR image that is needed. 	 ></quote>  >  >You can get the patch from 8 >ftp://ftp.process.com/support/57_2/drivers_v572p021.zip > F >As this is a rank 1 patch ("Recommended; system crash possible"), you >should install it anyway...  N I second this recommendation. Don't run w/o ECOs (VMS, OSI, TCPIP, TCPware...)# Many ECOs fixes serious problems...    --   Peter "EPLAN" LANGSTOEGER % Network and OpenVMS system specialist  E-mail  peter@langstoeger.atF A-1030 VIENNA  AUSTRIA              I'm not a pessimist, I'm a realist   ------------------------------  * Date: Mon, 4 Sep 2006 18:17:48 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)+ Subject: Re: OpenVMS and Windows on Itanium $ Message-ID: <edhqkc$3tq$1@online.de>  5 In article <44FB2C68.FD2289CD@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:    J > So when VMS will run as a virtual machine, it is essentially a glorifiedI > process/application running on a real instance of HP-UX. The neat thing I > is that the VMS "process" will run in little endian mode while the host & > HP-UX runs in ist native big endian.  C Are the plans that some day VMS will run ONLY on a virtual machine?    ------------------------------   Date: 4 Sep 2006 12:18:07 -0700 % From: "Bart" <bart.deboeck@gmail.com> + Subject: Re: OpenVMS and Windows on Itanium B Message-ID: <1157397487.293556.204660@74g2000cwt.googlegroups.com>   Hi Hein,  ?   Thanks for the book link and elaboration on the HP-UX option.    Regards, Bart    Hein RMS van den Heuvel schreef:  
 > Bart wrote:  > > Thanks for the reply!  > > B > > So, it might be an option to run HP-UX natively and (modified)I > > instances of OpenVMS and Windows as "processes" on HP-UX? Any idea of  > > the impact on performance? > >  > > Regards, > > Bart > 
 > Roughly. > + > The OpenVMS would NOT be modified (much). D > The same version would run as a client or on its own, but supposed > needs to be there. >  > Each CPU would be a thread. B > Yes you could have more virtual CPUs than physical CPUs over all
 > clients.F > The HP-UX host would (IMHO) only be used as a host, to keep it clean > and safe. > > If you also wanted HPUX on the box you'd run it as a client. > B > Recommended read: "The HP Virtual Server Environment: Making the: > Adaptive Enterprise Vision a Reality in Your Datacenter"" > by Dan Herington & Bryan Jacquot8 > http://www.hp.com/hpbooks/prentice/ptr_0131855220.html > D > Because it is supposed to be unmodified, it unfortunately (for me) > requires Montecito. C > The current crop of Itanium chips does not have enough protection  > layers to map VMS. > I > Personally I would have been very happy and excited if OpenVMS had been E > rebuild with reduced memory protection layers. The host runs in its  > own, innermost, shell. > G > Combining DCL (Super) and RMS (Exec)into a single layer (with RDB and I > what have you not) would be just dandy by me. While you are at it, also G > drop RMS's usage of UREW pages. If DCL or RMS were to behave badly it I > should crash the process (and optionally the system). No need (IMHO) to E > protect them from each other, but the this is not at all an area of  > expertise for me.  > H > I would love to be backup to run a HPUX host on my dual CPU rx2620 andF > guests for HPUX and multiple OpenVMS instances all at the same time.F > As a developer/consultant this would allow me to do CLUSTER testing,G > Mixed-environment Oracle client/server testing, and so on, all from a ( > single box (single heat/noise source). > / > I can understand the Engineering perspective. E > The existing Itanium base is just small stuff, relatively speaking. ) > Montecito will be the volume real soon.  >  > Hein.  > HvdH Performance Consulting  > I > (UREW page protection is used by RMS to support LOCATE mode, a marginal A > (best I can tell !?) performance option notably used by Fortran H > applications for simply sequential file unshared read, with a required+ > fallback with move mode already in place. G > No application woule break, some would run slower, most would benefit G > from better memory usage by RMS no longer fragmenting its memory with 7 > all the page rounding and such that comes with that).    ------------------------------   Date: 4 Sep 2006 12:09:43 -0700 % From: "Bart" <bart.deboeck@gmail.com> + Subject: Re: OpenVMS and Windows on Itanium B Message-ID: <1157396982.954350.58830@p79g2000cwp.googlegroups.com>  	 Hi Kerry,   '   Thanks for the hard partitioning tip!   F   For the case I'm confronted with, the application running on OpenVMS> is *very* resource intensive and should be finished as soon asE possible. On average, this app runs twice a week. The apps running on = Windows are not resource intensive, but should be continously 
 available.>   So, if it is possible to run OpenVMS and Windows on the sameD Integrity, it would be nice to have as much resources for OpenVMS as possible/necessary.   B   Is it correct that hard partitioning does not allow to move CPUsD among hard partitions, while virtual partitions and virtual machines allow to reassign CPUs?   D   Is it correct that in my case, hard partitioning is not an option?  E   On the other hand HPVM seems to be no option due to the performance @ constraint. Does this imply that virtual partitions are the only9 option? Or is a new machine with Windows a better option?    Thanks!  Bart   Main, Kerry schreef:   > > -----Original Message-----. > > From: Bart [mailto:bart.deboeck@gmail.com]$ > > Sent: September 3, 2006 12:23 PM > > To: Info-VAX@Mvb.Saic.Com + > > Subject: OpenVMS and Windows on Itanium  > >  > > Hi,  > > C > >   I'm an OpenVMS/Itanium newbie and trying to understand the HP  > > virtualization solutions.  > > I > >   Is it correct that only the future version of "HP Integrity Virtual H > > Machines" will allow to run OpenVMS and Windows simultaneously on an" > > Itanium (HP Integrity server)? > > J > >   Is it correct that there does not exist a virtual machine which runs@ > > on OpenVMS and supports Windows as a guest operating system? > > ) > >   All feedback is highly appreciated!  > >  > > Thanks,  > > Bart > >  > H > With hard partitioning on some of the Integrity servers that currentlyG > support it, you can run OpenVMS, HP-UX, Windows and Linux on the same J > server today - even with OpenVMS V8.2. This is similar to what you couldF > do with Alpha servers that supported hard partitioning about 5 years > ago. > I > What is coming is a way (HPVM) to consolidate lower performance OpenVMS I > systems and run these in virtual partitions on an Integrity server that  > supports it. > H > Like all emulated environments, the performance will not be as good asG > native performance on the target system, but the emulation could very E > well be better than the current source server which is running on a ( > much, much slower server architecture. > J > For those that understand other options out there like VMware, it is theG > same thing. These virtual instances are typically not meant for heavy < > CPU or heavy IO systems, but as a means to consolidate HW. > E > As an example, with HPVM, you could consolidate 3 or 4 lightly used F > OpenVMS systems with direct attached storage onto a single IntegrityJ > server on a SAN and share the IO FC adapters - hence, you do not need to1 > buy dual FC adapters for each system/partition.  > I > Keep in mind that OpenVMS has always supported Application stacking, so H > another option to consolidating OpenVMS servers is to simply move themJ > and run the app's on the same system. This has the advantage of reducingG > the number of OS instances as well as the HW (HPVM does not reduce OS 
 > instances).  > H > App stacking on Windows and some flavours of UNIX has been problematicH > for technical and cultural issues, but typically, this is not an issue > with OpenVMS.  >  >  > 	 > Regards  >  > Kerry Main > Senior Consultant  > HP Services Canada > Voice: 613-592-4660  > Fax: 613-591-4477  > kerryDOTmainAThpDOTcom > (remove the DOT's and AT)  > 6 > OpenVMS - the secure, multi-site OS that just works.   ------------------------------  % Date: Mon, 04 Sep 2006 15:49:10 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> + Subject: Re: OpenVMS and Windows on Itanium , Message-ID: <44FC8336.19CD3D30@teksavvy.com>   Tom Linden wrote: I > > native performance on the target system, but the emulation could very G > > well be better than the current source server which is running on a * > > much, much slower server architecture. >  > This is progress?     F I think it is more a question of HP needing to be able to say "we have
 that too".  C In fairness, during a DECUS presentation last spring, they provided D interesting examples of how this virtualisation worked and there are! instances where this makes sense.   A Remember that IA64 is being promoted as a big iron machine. So to E replace all those neat small PaRisc and Alpha systems around, you can G put all your eggs in one big basket and run multiple instances on a big  IA64 anvilHHHHHiron.   C Consider Stallard's original memo where he hoped to convert all VMS H customers over to HP-UX. In such a scenario, being able to run VMS as anA application on HP-UX would be great since customers could migrate H applications one by one and run both HP-UX (at full performance) and VMS- (on virtualised machine) on the same machine.   G And this technology may be a foundation to add architecture emulator to G allow HP-UX to host instances that run on other architectures. (PaRisc, H Alpha). And in the future, to allow an 8086 based HP-UX instance to host$ older IA64 based instances of HP-UX.   ------------------------------  % Date: Mon, 04 Sep 2006 16:04:08 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> + Subject: Re: OpenVMS and Windows on Itanium , Message-ID: <44FC86B7.765E46F7@teksavvy.com>   Tom Linden wrote: E > both windows and Linux, I'll get two [ boxes | blades ]  VM370, for 
 > example,A > was mostly used in a development environment, not in production   F VM was used extensively in serious production shops that were based onE DOS/VSE. DOS/VSE did not have proper file system, not any interactive @ abilities, so VM allowed interactive use thourgh VM/CMS (a smallE interactive operating system instance for each user that gave users a B file system.  (when VSAM was ported to DOS/VSE, the later acquiredH access to proper file system where the bacth jobs didn't have to specify; a range of tracks on a specific drive that had their data).   B MVS quickly gained TSO (time sharing option) and then SPF (variousB acrynyms over time) which gave ALLIN1-like menus and customisationD abilities). The need for VM dropped considerably. However, for largeH shops like banks that had some application requiring CICS V2.3 and otherA applicatiosn requiring CICS 3.2, then VM came in handy to provide D multiple instances. A bank I worked for had 14 instances of MVS on aA couple of  physical machines (by that time, hard partitioning had & arrived, so they no longer needed VM).    H With the increases awareness over clustering and high availability, I amC not so sure that serious shops are really interested in putting all > their eggs in one basket, especially since it runs as an HP-UXF application and that means that if HP-UX fails, all the rest goes down@ too. People may want to retain multiple separate physical boxes.  A In the windows world, it is different because they often have one D application per instance and that becomes harder to manage. However,E virtualising them onto one big and expensive system may not be better 2 than keeping each on el-cheapo commodity machines.  G I am sure there are cases where it does make sense. But I doubt that it C is "mainstream" even though the vendors would like you to think so.   A In fact, when you look at galaxy for VMS, I have to wonder if the G investment in that technology and software ever paid off, I am sure the < Engineers had plenty of rewarding inlellectual challenges inH implementing this. But was the market really big enough to warrant thoseB developments when you consider all the bit of VMS that have fallen> behind and not gotten any development since the last century ?   ------------------------------  * Date: Mon, 4 Sep 2006 20:24:46 +0000 (UTC)P From: helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply)+ Subject: Re: OpenVMS and Windows on Itanium $ Message-ID: <edi22e$ia1$1@online.de>  5 In article <44FC8336.19CD3D30@teksavvy.com>, JF Mezei ' <jfmezei.spamnot@teksavvy.com> writes:    K > > > native performance on the target system, but the emulation could very I > > > well be better than the current source server which is running on a , > > > much, much slower server architecture. > >  > > This is progress?  > H > I think it is more a question of HP needing to be able to say "we have > that too". > E > In fairness, during a DECUS presentation last spring, they provided F > interesting examples of how this virtualisation worked and there are# > instances where this makes sense.   E > Consider Stallard's original memo where he hoped to convert all VMS  > customers over to HP-UX.    ' I really think that he meant just that.   1 > In such a scenario, being able to run VMS as an C > application on HP-UX would be great since customers could migrate J > applications one by one and run both HP-UX (at full performance) and VMS/ > (on virtualised machine) on the same machine.   - I hear cries of "VMS?  That's SLOW!" already.   B My feeling is that running an OS on an emulator is like sex with aE blow-up doll.  (Since I have no experience with either, I can make a  A fair comparison.)  Even if the emulation were very good, I still   wouldn't like the concept.   ------------------------------  % Date: Mon, 04 Sep 2006 16:15:58 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> + Subject: Re: OpenVMS and Windows on Itanium , Message-ID: <44FC897A.840B61A9@teksavvy.com>   "Main, Kerry" wrote:F > The challenge with this philosophy is the gross under-utilization of9 > servers that almost all med-large companies face today.   D One needs to look at how linear the cost/performance curve is. ThereB comes a point where you start to pay a big premium for large scaleF systems because they come with more than just horsepower (and you ay a premium to buy a big box).    F Buying 30 small boxes at Wallmart/CDW may end up costing far less thanD buying a superdome from HP with the virtualisation software/kits and putting 30 instances on it.   > So, even if your boxes are underutilised, it is still cheaper.    D The one area where virtualisation makes sense is when you have largeD seasonal variations in usage.  Say the IRS runs on Windows, and theyH have obviously the instance that does the tax returns in the spring thatC requires a lot of CPU, while in the rest of the year, they have the D instance that analises data to find deadbeats and cheaters that also requires a lot of CPU.  G During the peak season fo each application, you need a large horsepower G machine. You could buy 2 great big machines, and during spring, the one C running the analysis would be iddle, and outside of spring, the one E running the IRS returns would be iddle. But during each applicatioN,s / peak season, they would be at 100% utilisation.   E With virtualisation, you can by a single great big machine, and shift D processing power on a seasonal basis to give the returns or analysisN instances more hosrsepower as needed. In such a case, you end up saving money.    B But in cases where your wintel instances are very happy in a smallE el-cheapo wallmart special machine which is fast enough to handle its C peak season, then it woudln't pay to combine everything into a very  expensive machine.   ------------------------------  # Date: Mon, 04 Sep 2006 21:54:59 GMT 0 From: John Santos <john.santos@post.harvard.edu>+ Subject: Re: OpenVMS and Windows on Itanium > Message-ID: <MPG.1f666ac2fc333176989726@news.bellatlantic.net>  E In article <edhqkc$3tq$1@online.de>, helbig@astro.multiCLOTHESvax.de   says... 7 > In article <44FB2C68.FD2289CD@teksavvy.com>, JF Mezei ) > <jfmezei.spamnot@teksavvy.com> writes:   > L > > So when VMS will run as a virtual machine, it is essentially a glorifiedK > > process/application running on a real instance of HP-UX. The neat thing K > > is that the VMS "process" will run in little endian mode while the host ( > > HP-UX runs in ist native big endian. > E > Are the plans that some day VMS will run ONLY on a virtual machine?  >  >   A Not that I've heard.  VAX V7.3 is forever :-)  (or at least until C December 31, 9999)  (BTW, can the 4-digit year issues be fixed with B the date and time international formatting stuff (LIB$DATE_FORMAT*# logical names, LIB$DT_FORMAT, etc.)   ! Same with Alpha, V8.3 (or later.)   H They would have to take some stuff out of VMS to break it on non-virtualC machines, which historically they haven't done.  However, maybe the " port to 8086 will require this ;-)     --   John   ------------------------------  $ Date: Mon, 4 Sep 2006 18:50:51 -0400' From: "Main, Kerry" <Kerry.Main@hp.com> + Subject: RE: OpenVMS and Windows on Itanium T Message-ID: <FA60F2C4B72A584DBFC6091F6A2B8684019BCB88@tayexc19.americas.cpqcorp.net>   > -----Original Message-----9 > From: JF Mezei [mailto:jfmezei.spamnot@teksavvy.com]=20 ! > Sent: September 4, 2006 4:16 PM  > To: Info-VAX@Mvb.Saic.Com - > Subject: Re: OpenVMS and Windows on Itanium  >=20 > "Main, Kerry" wrote:H > > The challenge with this philosophy is the gross under-utilization of; > > servers that almost all med-large companies face today.  >=20F > One needs to look at how linear the cost/performance curve is. ThereD > comes a point where you start to pay a big premium for large scaleH > systems because they come with more than just horsepower (and you ay a > premium to buy a big box).=20  >=20H > Buying 30 small boxes at Wallmart/CDW may end up costing far less thanF > buying a superdome from HP with the virtualisation software/kits and > putting 30 instances on it.  >=20@ > So, even if your boxes are underutilised, it is still cheaper. >=20   Don't forget to:H - add 30 x2 =3D 60 FC adapters to the cost to connect these servers to = the G SAN. Each FC adapter is around $4K. If not important you could get away  with a single FC x30. G - add Oracle per cpu licensing for all of those 30 servers. With larger D partitionable server, you can license only on the actual CPU's being used.    >=20F > The one area where virtualisation makes sense is when you have largeF > seasonal variations in usage.  Say the IRS runs on Windows, and theyA > have obviously the instance that does the tax returns in the=20 
 > spring that E > requires a lot of CPU, while in the rest of the year, they have the F > instance that analises data to find deadbeats and cheaters that also > requires a lot of CPU. >=20A > During the peak season fo each application, you need a large=20  > horsepower< > machine. You could buy 2 great big machines, and during=20 > spring, the one E > running the analysis would be iddle, and outside of spring, the one G > running the IRS returns would be iddle. But during each applicatioN,s 1 > peak season, they would be at 100% utilisation.  >=20G > With virtualisation, you can by a single great big machine, and shift F > processing power on a seasonal basis to give the returns or analysisB > instances more hosrsepower as needed. In such a case, you end=20 > up saving money. >=20 >=20D > But in cases where your wintel instances are very happy in a smallG > el-cheapo wallmart special machine which is fast enough to handle its E > peak season, then it woudln't pay to combine everything into a very  > expensive machine. >=20  C As long as you remember that you have to apply the monthly security @ patches to 30 servers, monitor 30 security log files, monitor 30% separate local disk environments etc.   H Again, the base HW costs is one of the smaller components of the overall Server TCO.    Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  4 OpenVMS - the secure, multi-site OS that just works.   ------------------------------   Date: 4 Sep 2006 20:05:51 -0700 < From: "Hein RMS van den Heuvel" <heinvandenheuvel@gmail.com>+ Subject: Re: OpenVMS and Windows on Itanium B Message-ID: <1157425551.933267.275780@74g2000cwt.googlegroups.com>   Main, Kerry wrote: > Don't forget to:O > - add 30 x2 = 60 FC adapters to the cost to connect these servers to the SAN.   = Ayup, it is pretty neat how the host OS can present seemingly F individual storage connections, witht he same device name if you like,E using a single interface to all guest OS instances. The HPUX host can A figure out the alternate paths and keep it simple for the guests.   C Ditto in the network space. The host can allow the guest to share a C single nic, or use the hpux lan aggregation to spread the load over  multiple available nic's. C Meanwhile all guests can talk to each other over a 'virtual' hub at 
 memory speed.   E > As long as you remember that you have to apply the monthly security B > patches to 30 servers, monitor 30 security log files, monitor 30' > separate local disk environments etc.   9 Hmm... don't think that will be different for hosts OS's. ! They are not aware of each other. C They each their own things, have their own securiy files and so on. ! VMS clusters on teh other hand...    Cheers,  Hein van den Heuvel  HvdH Performance Consulting    ------------------------------  # Date: Mon, 04 Sep 2006 18:25:01 GMT , From: Frank Clarke <m5srexx@tampabay.rr.com>L Subject: Re: Please Loot the Inforamtion/Mainframe Material  at "MFRESOURCE"8 Message-ID: <hprof25a8asjghs77i7d5j10l64r82jenb@4ax.com>  D On 4 Sep 2006 11:17:32 -0700, "Sri" <sridhargandla@gmail.com> wrote:6 <1157393852.096815.126370@i3g2000cwc.googlegroups.com>  4 >MFRESOURCE -New Yahoo Group for Mainframe Resources) >MFRESOURCE is Mainframe Resources Group:  > 7 >Group useful for Mainframe Materials & Mainframe Jobs.  > 8 >Related Link: http://groups.yahoo.com/group/mfresource/) >Post message: mfresource@yahoogroups.com 0 >Subscribe: mfresource-subscribe@yahoogroups.com > : >Mainframe Job Postings in INDIA & US and other Countries. > G >This Group is Useful for people who are attending Mainframe interviews F >and Professionals working in Mainframe Skills specifically working on >the following Skills: > @ >COBOL JCL VSAM CICS,DB2 IMS DB/DC IDMS ADABAS/NATURAL, ENDEVOR,G >FILE-AID, XPEDITOR COOLGEN ADABAS EZYTREIVE FOCUS MQSERIES NATURAL REX A >SAS TSO DATACOM ADSO IDMS ABENDAID CA7 CHANGEMAN FILEAID INFOMAN * >INTERTEST PANAVALET STROBE XMODS XPEDITOR > = >JOBS MVS,OS/390,Z/OS,S390,OS390,ZOS,JCL,VSAM, COBOL,PL1,CICS   P ...and it will be a series of marginally-educated subcontinentals all wanting toI know how they can get the date and time embedded into the dataset name...     0 (change Arabic number to Roman numeral to email)   ------------------------------   Date: 4 Sep 2006 11:17:32 -0700 % From: "Sri" <sridhargandla@gmail.com> H Subject: Please Loot the Inforamtion/Mainframe Material  at "MFRESOURCE"B Message-ID: <1157393852.096815.126370@i3g2000cwc.googlegroups.com>  3 MFRESOURCE -New Yahoo Group for Mainframe Resources ( MFRESOURCE is Mainframe Resources Group:  6 Group useful for Mainframe Materials & Mainframe Jobs.  7 Related Link: http://groups.yahoo.com/group/mfresource/ ( Post message: mfresource@yahoogroups.com/ Subscribe: mfresource-subscribe@yahoogroups.com   9 Mainframe Job Postings in INDIA & US and other Countries.   F This Group is Useful for people who are attending Mainframe interviewsE and Professionals working in Mainframe Skills specifically working on  the following Skills:   ? COBOL JCL VSAM CICS,DB2 IMS DB/DC IDMS ADABAS/NATURAL, ENDEVOR, F FILE-AID, XPEDITOR COOLGEN ADABAS EZYTREIVE FOCUS MQSERIES NATURAL REX@ SAS TSO DATACOM ADSO IDMS ABENDAID CA7 CHANGEMAN FILEAID INFOMAN) INTERTEST PANAVALET STROBE XMODS XPEDITOR   < JOBS MVS,OS/390,Z/OS,S390,OS390,ZOS,JCL,VSAM, COBOL,PL1,CICS   Thanks & Regards,  Group Moderator.   ------------------------------  % Date: Mon, 04 Sep 2006 13:22:25 -0500 + From: brandon@dalsemi.com (BRANDON, JOHN M) 7 Subject: Re: VMS STARTUP documentation mis-leading V7.3 1 Message-ID: <06090413222557@dscis6-0.dalsemi.com>    Volker Halle writes:G > Please note that the first manual referenced (Table 5-1 Site-Specific H > Startup Command Procedures) explicitly claims, that it shows the files > in the order executed.   Agreed.   H > You second reference (5.6.1 OpenVMS Startup Procedures) does not claimE > the order of execution, but just gives a list command files invoked C > during startup. Those files seem not to be listed in any specific 7 > order, neither aplhabetically nor in execution order.   O My point - well sort of - yes, I know the manual does not list in order however O I would agrue that it should be in execution order otherwise why bother to list N them at all?  If it is a discussion of startup execution then I would at least9 expect them to be in the executed order.  Not hap-hazard.      John "REBOOT" Brandon  VMS Systems Administrator * firstname.lastname.spam.me.not@dalsemi.com   ------------------------------  % Date: Mon, 04 Sep 2006 13:25:28 -0500 + From: brandon@dalsemi.com (BRANDON, JOHN M) 7 Subject: Re: VMS STARTUP documentation mis-leading V7.3 1 Message-ID: <06090413252894@dscis6-0.dalsemi.com>   
 Safir writes:  > Hi,  > H > You can check the flow of startup by creating a "non-execute" copy of 
 > startup.com  > F > (it doesn't start anything) by setting (see below) stdrv$execute to  > stdrv$false and  >  > stdrv$verbose to stdrv$true :   O Thanks, however I am well aware of what takes place at startup - I question the 6 integrity of the VMS documentation concerning startup.     John "REBOOT" Brandon  VMS Systems Administrator * firstname.lastname.spam.me.not@dalsemi.com   ------------------------------   End of INFO-VAX 2006.486 ************************