1 INFO-VAX	Fri, 19 Dec 2003	Volume 2003 : Issue 700       Contents: Re: Alpha Server Road Map  Re: CWS 433au info ? Re: CWS 433au info ? Re: DE500-XA & full-duplex Re: DE500-XA & full-duplex equipment available  Re: equipment available ! Re: lbr$ routines and concurrency ! Re: lbr$ routines and concurrency ! Re: lbr$ routines and concurrency  mV3100 & DECNet? Re: mV3100 & DECNet? Re: mV3100 & DECNet? Re: OpenVMS 8.1 ships  Re: OpenVMS 8.1 ships J Re: OpenVMS Pearl - OpenVMS V8.0, first release on Itanium, ships today!!! Opteron's chances  Re: PID creation in a cluster  Re: PID creation in a cluster  Re: PID creation in a cluster  Re: PID creation in a cluster . Q: Convert quadword time to Ascii time in DCL?2 Re: Q: Convert quadword time to Ascii time in DCL?2 Re: Q: Convert quadword time to Ascii time in DCL?( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMS( Re: Singapore Exchange to run on OpenVMSJ Subject: Re: backing up VMS files to PC based CD writer - belated followup Re: Support for passive FTP  Re: Support for passive FTP  Re: Support for passive FTP  Re: Support for passive FTP  Re: Support for passive FTP   F ----------------------------------------------------------------------  % Date: Thu, 18 Dec 2003 22:11:43 -0600 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>" Subject: Re: Alpha Server Road Map6 Message-ID: <3FE27A7F.7D8C6039@NeOaSrPtAhMlNiOnWk.net>   Sue Skonetski wrote: > [snip]H > Please forgive my delay in getting back to you, I have been out with a% > surgery and just back 1/2 time now.   ( Wow! Hope you're doing better every day!  0 Be well, Sue! Happy Holidays to you and yours...   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  + Date: Thu, 18 Dec 2003 14:17:42 -0500 (EST) + From: Lord Isildur <isildur@andrew.cmu.edu>  Subject: Re: CWS 433au info ? H Message-ID: <Pine.LNX.4.58-035.0312181416520.4606@unix44.andrew.cmu.edu>   what info do you need? isildur     ; On Thu, 18 Dec 2003, Jan-Erik [iso-8859-1] Sderholm wrote:    > Hi. > > Just got me a new "hobby" box, a Personal Workstation 433au.; > Now, I have been looking around at HP to try to find some A > info on it, like a User Guide. All I found was some Golden Eggs : > and old SOC pages. Anyone having any other links/URLs to > more info on the PWS 433au ? > 
 > Jan-Erik4 > PS. This box will replace a A-Station 250 4/266... >    ------------------------------  % Date: Thu, 18 Dec 2003 21:04:15 +0100 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com>  Subject: Re: CWS 433au info ? ' Message-ID: <3FE2083F.F42E66D7@aaa.com>   ! The ones posted by David Grey :-)   
 Thanks !!!	 Jan-Erik.    Lord Isildur wrote:  >  > what info do you need?	 > isildur  > = > On Thu, 18 Dec 2003, Jan-Erik [iso-8859-1] Sderholm wrote:  >  > > Hi. @ > > Just got me a new "hobby" box, a Personal Workstation 433au.= > > Now, I have been looking around at HP to try to find some C > > info on it, like a User Guide. All I found was some Golden Eggs < > > and old SOC pages. Anyone having any other links/URLs to  > > more info on the PWS 433au ? > >  > > Jan-Erik6 > > PS. This box will replace a A-Station 250 4/266... > >    ------------------------------  % Date: Thu, 18 Dec 2003 22:14:32 +0100  From: Dirk Munk <munk@home.nl># Subject: Re: DE500-XA & full-duplex 2 Message-ID: <brt5rh$oer$1@news4.tilbu1.nb.home.nl>   Christoph Gartmann wrote:  > Hello, > O > now I tried a different DE500-XA in the workstation - same result: CRC errors M > when connected to the 3COM 4400 switch, no errors when connected to a cheap M > Soho switch. But the duplex LED on the Soho switch is flickering instead of  > being constantly on.    O That means the port of your Soho switch is in half-duplex mode, and the Duplex  M LED is now showing collisions. This LED had two functions, when it lights up  N constantly it means the port is in full-duplex mode. When it is (mainly) off, F the port is in half-duplex mode, and every time it flickers there are Q collisions. You can not have collisions in full-duplex mode, only in half-duplex   mode.     : > Nevertheless LANCP on the VMS side states "full duplex".  P Yes, because it is configurered that way. Now you are sure it is in full-duplex K mode. Of course this combination of full-duplex DE500 and half-duplex Soho  Q switch port is wrong, You should set the DE500 to half-duplex. You can't get the  N Soho switch port into full-duplex mode because it can not auto-negotiate with H the DE500. In this setting it will always fall back to half-duplex mode.   > Q > Next, I replaced the DE500-XA with a DE500-BA: no problem with the 4400. I have ) > both ends hard-coded to 100 FD, though.   3 Good, that is one way how it certainly should work.    > O > Unfortunately I cannot try the DE500-XA under Tru64-Unix, something in my box * > exploded when I tried to power it up :-(  7 I'm sorry to hear that, I can imagine your frustration.    > 
 > Regards, >    Christoph Gartmann  >    ------------------------------  % Date: Thu, 18 Dec 2003 22:14:06 -0600 @ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net># Subject: Re: DE500-XA & full-duplex 6 Message-ID: <3FE27B0E.A25A1926@NeOaSrPtAhMlNiOnWk.net>   Christoph Gartmann wrote:  >  > Hi,  > P > OpenVMS 7.3-1 on an AlphaStation 255. The station has a DE500-XA network card.K > The card is set to 100 MB/s and full-duplex. The same at the other end (a Q > switch). This setup worked for quite some time. Now we replaced the switch with M > one from 3Com (3C17204, a 4400). Now we get lots of CRC errors. The machine P > boots up and is working more or less all right. If we let the switch negotiateQ > the speed and duplex mode this results in 100 MB/s and half-duplex although the P > card is configured with full duplex. Setting the switch to full duplex results > in these CRC errors. > O > Next I used a Soho switch. This one detects 100 MB/s. In addition the LED for Q > full-duplex is flickering. Sometimes it is on for a few seconds, then goes off, O > goes on again and so on but no CRC errors. Is my card broken? Anyhting else I 
 > can try?  H I seem to recall some problems with DE500s and auto-negotiate. I thoughtG conventional wisdom was to force them to FAST-FD or whatever at the SRM  level...   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Thu, 18 Dec 2003 14:58:26 -0500 / From: "Hank Vander Waal" <hvanderw@comcast.net>  Subject: equipment available? Message-ID: <NHEHKAGLFGJGFJENEGDMKEOOCGAA.hvanderw@comcast.net>   J Anyone have a need/want for a Dec hub 90 with 1 DEC repeater and 3 90L+ ??  J Also have a Alpha workstation 255/233 with 96 MB memory & Cdrom (no disk )  - Located in West Michigan (Grand Rapids area)    K Please let me know - really don't want to dump them but space is limited :)    hvanderw-at-comcast#dot#net    ------------------------------  % Date: Fri, 19 Dec 2003 06:01:26 +0800 , From: Paul Repacholi <prep@prep.synonet.com>  Subject: Re: equipment available- Message-ID: <87hdzxiy4p.fsf@prep.synonet.com>   1 "Hank Vander Waal" <hvanderw@comcast.net> writes:   D > Anyone have a need/want for a Dec hub 90 with 1 DEC repeater and 3	 > 90L+ ??   B If you still have any more hubs, hang on to it so you have a spare
 Power Supply.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 18 Dec 2003 12:40:16 -08001 From: usenet_vms@lehrerfamily.com (Joshua Lehrer) * Subject: Re: lbr$ routines and concurrency= Message-ID: <477e0934.0312181240.4c1b1def@posting.google.com>    winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") wrote in message news:<00A2A898.B3EDBFB8@SSRL.SLAC.STANFORD.EDU>... > N > Huh.  What version of VMS?  What kind of i/o cacheing are you doing?  VIOC?    C++ 6.5 on 7.3-1, XFC.  G > EFC?  (I'd guess that once the compiler opened the text library that  O > EFC might suck the whole thing in if there was enough free memory; that would & > make subsequent access really fast.)  : I doubt that it is reading the entire library, it is HUGE.  & > The lock manager issue is confusing,L > unless the directory with all the headers is getting updated all the time,L > which might mean having to wait for the update to finish before getting toH > read the directory and get the next header file.  A text library would  D nope, once our directory is prepared and released to consumers there? is no writing to the directory, just reading via the compilers.   E > Failing that, how about a dedicated drive with nothing on in it but 4 > that, so you aren't contending with any other i/o?  - Yup, tried it on a dedicated disk, no faster.   D The funny thing is that it is really fast, up until the 10th flush. D We flush the TLB periodically.  When we didn't, writing the TLB onceD in the end was really really slow.  So now we periodically flush theD TLB.  With 2000 modules, we flush around 20 times.  Somewhere aroundD the 10th flush, the flush command takes almost a minute.  All of the other ones are fast.   -josh    ------------------------------  # Date: Thu, 18 Dec 2003 22:04:07 GMT L From: winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr")* Subject: Re: lbr$ routines and concurrency6 Message-ID: <00A2A8F8.44509E5A@SSRL.SLAC.STANFORD.EDU>  q In article <477e0934.0312181240.4c1b1def@posting.google.com>, usenet_vms@lehrerfamily.com (Joshua Lehrer) writes:  >winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") wrote in message news:<00A2A898.B3EDBFB8@SSRL.SLAC.STANFORD.EDU>...  >>  O >> Huh.  What version of VMS?  What kind of i/o cacheing are you doing?  VIOC?   >  >C++ 6.5 on 7.3-1, XFC.  > H >> EFC?  (I'd guess that once the compiler opened the text library that P >> EFC might suck the whole thing in if there was enough free memory; that would' >> make subsequent access really fast.)  > ; >I doubt that it is reading the entire library, it is HUGE.   K Okay.  (You could probably tell by looking at how much memory the cache was K using before and then during the compilation, if your system was relatively  quiescent.)   E >The funny thing is that it is really fast, up until the 10th flush.  E >We flush the TLB periodically.  When we didn't, writing the TLB once E >in the end was really really slow.  So now we periodically flush the E >TLB.  With 2000 modules, we flush around 20 times.  Somewhere around E >the 10th flush, the flush command takes almost a minute.  All of the  >other ones are fast.   H Are you doing as JF suggested, preallocating the library file?  My firstI guess would be that at the 10th flush you're filling up the initial file  F allocation, triggering a file expansion (on what's now likely a prettyK fragmented disk), that you've got a large RMS extend quantity, and so right I then it goes out and gathers up the blocks which you end up using for the L subsequent flushes.  If that's right - and it may be totally off-base - thenG preallocate the file bigger.  (That may just move the long pause to the  beginning of the process.)   -- Alan  --  O =============================================================================== 0  Alan Winston --- WINSTON@SSRL.SLAC.STANFORD.EDUM  Disclaimer: I speak only for myself, not SLAC or SSRL   Phone:  650/926-3056 M  Paper mail to: SSRL -- SLAC BIN 99, 2575 Sand Hill Rd, Menlo Park CA   94025 O ===============================================================================    ------------------------------    Date: 18 Dec 2003 18:29:46 -08001 From: usenet_vms@lehrerfamily.com (Joshua Lehrer) * Subject: Re: lbr$ routines and concurrency= Message-ID: <477e0934.0312181829.548f2cb2@posting.google.com>    winston@SSRL.SLAC.STANFORD.EDU ("Alan Winston - SSRL Admin Cmptg Mgr") wrote in message news:<00A2A8F8.44509E5A@SSRL.SLAC.STANFORD.EDU>...J > Are you doing as JF suggested, preallocating the library file?  My firstK > guess would be that at the 10th flush you're filling up the initial file  H > allocation, triggering a file expansion (on what's now likely a prettyM > fragmented disk), that you've got a large RMS extend quantity, and so right K > then it goes out and gathers up the blocks which you end up using for the N > subsequent flushes.  If that's right - and it may be totally off-base - thenI > preallocate the file bigger.  (That may just move the long pause to the  > beginning of the process.) >   C Yeah, that was our first guess, too, but it turns out not to be the F case.  I am pre-setting the initial allocation to just slightly larger@ than I know the final file will be.  So, yes, it should take theF longest on the first flush, or the create, when it tries to extend and make the file.  D However, this appears to complete very quickly.  The file is createdC an has the full allocation very quickly.  I confirm this by doing a E "dir" from a different window and seeing the file and the allocation.   @ I wish I could debug the code somehow, and hit "stop" during the- offending flush, and see what code it was in.    -josh    ------------------------------  % Date: Thu, 18 Dec 2003 16:22:16 -0600 % From: Neil Cherry <njc@wolfgang.uucp>  Subject: mV3100 & DECNet? . Message-ID: <slrnbu4a4o.76u.njc@wolfgang.uucp>  E I have a MicroVax 3100 and the Hobbyist License for VMS. I decided to E load VMS 6.1 (I was in a hurry). I've entered everything and now find @ I can't bring up DECNet (I need Phase IV) because I don't have aC license. How do I get such a license for a 3100? URL's, pointers or 7 whatever it takes would be greatly appreciated. Thanks.    --  D Linux Home Automation         Neil Cherry        ncherry@comcast.net; http://home.comcast.net/~ncherry/               (Text only) = http://linuxha.sourceforge.net/                 (SourceForge) 8 http://hcs.sourceforge.net/                     (HCS II)   ------------------------------  # Date: Thu, 18 Dec 2003 22:44:11 GMT 4 From: brad@.gateway.2wire.net (Bradford J. Hamilton) Subject: Re: mV3100 & DECNet? 0 Message-ID: <%4qEb.425474$ao4.1359451@attbi_s51>  V In article <slrnbu4a4o.76u.njc@wolfgang.uucp>, Neil Cherry <njc@wolfgang.uucp> writes:F !I have a MicroVax 3100 and the Hobbyist License for VMS. I decided toF !load VMS 6.1 (I was in a hurry). I've entered everything and now findA !I can't bring up DECNet (I need Phase IV) because I don't have a D !license. How do I get such a license for a 3100? URL's, pointers or8 !whatever it takes would be greatly appreciated. Thanks.  F When you mention that you have the "Hobbyist License", did you get theL layered-product licenses, as well?  They are obtained at the same page whereJ you registered for the VAX license.  The layered product licenses contain:  !  DVNETEND                     DEC !  DVNETEXT                     DEC    which are used for DECnet.   !  !-- E !Linux Home Automation         Neil Cherry        ncherry@comcast.net < !http://home.comcast.net/~ncherry/               (Text only)> !http://linuxha.sourceforge.net/                 (SourceForge)9 !http://hcs.sourceforge.net/                     (HCS II)   J __________________________________________________________________________A Bradford J. Hamilton                    "All opinions are my own" K bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-'  0                                          with @"   ------------------------------  % Date: Thu, 18 Dec 2003 16:51:49 -0600 % From: Neil Cherry <njc@wolfgang.uucp>  Subject: Re: mV3100 & DECNet? . Message-ID: <slrnbu4bs4.76u.njc@wolfgang.uucp>  = On Thu, 18 Dec 2003 22:44:11 GMT, Bradford J. Hamilton wrote: X > In article <slrnbu4a4o.76u.njc@wolfgang.uucp>, Neil Cherry <njc@wolfgang.uucp> writes:H > !I have a MicroVax 3100 and the Hobbyist License for VMS. I decided toH > !load VMS 6.1 (I was in a hurry). I've entered everything and now findC > !I can't bring up DECNet (I need Phase IV) because I don't have a F > !license. How do I get such a license for a 3100? URL's, pointers or: > !whatever it takes would be greatly appreciated. Thanks. > H > When you mention that you have the "Hobbyist License", did you get theN > layered-product licenses, as well?  They are obtained at the same page whereL > you registered for the VAX license.  The layered product licenses contain: > # >  DVNETEND                     DEC # >  DVNETEXT                     DEC  >  > which are used for DECnet.  D Sorry to have bothered everyone, I rebooted and it worked (I thoughtA it failed). I guess I was just tired as I had to enter a bunch of F licenses by hand (until I found a cable I could hack to work). Now I'mB the happy hobbyist admin of an Alpha & 3100 VMS system. I wish the- 3400's had survived (die of neglect I guess).   % Thanks and sorry for all the trouble.    --  D Linux Home Automation         Neil Cherry        ncherry@comcast.net; http://home.comcast.net/~ncherry/               (Text only) = http://linuxha.sourceforge.net/                 (SourceForge) 8 http://hcs.sourceforge.net/                     (HCS II)   ------------------------------  % Date: Thu, 18 Dec 2003 19:51:23 -0600 , From: "Dave Gudewicz" <k9jdk@NOSPAMarrl.net> Subject: Re: OpenVMS 8.1 ships/ Message-ID: <vu4mcoh8qmej48@corp.supernews.com>   L Good news.  Now we'll have something to run on the rx2600 we're getting real	 soon now.    Dave...   > "Keith Parris" <keithparris_NOSPAM@yahoo.com> wrote in message7 news:cf15391e.0312180922.21bb5cce@posting.google.com...  > From: Gorham, Mark, > Sent: Thursday, December 18, 2003 10:17 AM > Subject: VMS 8.1 ship today! > A > I am delighted to announce the evaluation release of HP OpenVMS @ > Version 8.1 for Integrity servers based on the Intel ItaniumA > processor.  This release is designed for early adopter ISVs and A > customers who wish to port their applications in advance of the C > production-quality release.   OpenVMS Version 8.1 includes native C > compilers, and a substantial suite of functionality that ISVs and F > customers are eager to use, including clustering and a wide range ofA > development tools and integration technologies.  We are working E > closely with industry-leading software vendors such as Oracle, BEA, = > BMC Software, and Computer Associates to fully port OpenVMS G > environments and application portfolios to Integrity servers.  The HP F > OpenVMS ISV partners have already committed to port more than 550 ofF > their applications to the OpenVMS Integrity platform.  Selected ISVsD > including MVP, Accucorp, Cadture, Attunity and TECSys have alreadyG > ported their applications and have them running on evaluation release 8 > of OpenVMS Version 8.0 on the Integrity rx2600 server. > D > OpenVMS Version 8.1 is a significant achievement and contains muchF > more functionality than was originally planned for this release.  WeG > are running mixed-architecture clusters of over a dozen systems, have @ > TCP/IP and DECnet network stacks, and have four-way SMP system? > support.  We are even near completion of an OpenVMS Integrity D > Superdome demo!  Several customers who have seen the mixed-clusterE > demo have been so excited that they are accelerating their adoption , > plans for OpenVMS on HP Integrity servers. > G > We have met all of our scheduled OpenVMS release dates, and we are on G > track to release OpenVMS Version 8.2, a production-quality release of F > OpenVMS for the HP Integrity server and AlphaServer platforms in theG > second half of 2004.  This and subsequent releases of OpenVMS will be = > based on common source code, so that non-hardware dependent G > enhancements to the operating system will be available simultaneously G > on both AlphaServer and Integrity systems.  It will also make it very G > cost-efficient for ISVs and for customers with custom applications to ; > support both platforms - recompile, relink, test, and go.  > G > Great technology and great business practices come together in the HP H > Alpha RetainTrust program.   As OpenVMS customers, you will be able toG > plan the integration of OpenVMS on Integrity servers into AlphaServer E > environments as your business needs dictate.  In the near term, you C > can continue to upgrade your AlphaServer systems and when your IT H > plans call for it, you can simply add HP Integrity servers to existing > OpenVMS clusters.  > G > The excitement about Integrity servers is accelerating.  Our partners F > are moving forward and we are on or ahead of our OpenVMS schedule inG > all areas.  I have never been more thrilled to be part of the OpenVMS D > business.  And I hope that you will feel the same excitement aboutB > using this tremendous technology and about HP as your technology
 > partner. > 
 > Regards,
 > Mark Gorham * > Vice President, OpenVMS Systems Division > ( > For ordering information, please visit' > www.hp.com/go/openvms-integrity-ready  > / > DSPP members, please visit www.hp.com/go/dspp    ------------------------------  % Date: Thu, 18 Dec 2003 19:54:17 -0600 6 From: "Craig A. Berry" <craigberry@mac.com.spamfooler> Subject: Re: OpenVMS 8.1 shipsT Message-ID: <craigberry-0D3178.19541718122003@dsl081-159-101.chi1.dsl.speakeasy.net>  = In article <cf15391e.0312180922.21bb5cce@posting.google.com>, 3  keithparris_NOSPAM@yahoo.com (Keith Parris) wrote:    > From: 	Gorham, Mark   , > Sent:	Thursday, December 18, 2003 10:17 AM > Subject:	VMS 8.1 ship today! > A > I am delighted to announce the evaluation release of HP OpenVMS @ > Version 8.1 for Integrity servers based on the Intel ItaniumA > processor.  This release is designed for early adopter ISVs and A > customers who wish to port their applications in advance of the   > production-quality release.     G I'm delighted to announce that it only took a couple of hours tweaking  H configuration scripts to get the latest development snapshot of Perl up G and running on the v8.1 hp testdrive system today.  I'm not announcing  H a kit at the moment, just mentioning that the porting effort for a very D large open source project in C was a bit of a yawn.  Perl has about G 500,000 lines of C (if you can call it that, since it's really more of  B a pseudo-language implemented in C macros).  No code changes were D necessary, only modifications to one DCL script and one Perl script  used to configure the build.    H There were a couple of small gotchas that may be of interest here.  The ; hardware model number of the testdrive system surprised me:   ' $ write sys$output f$getsyi("HW_MODEL")  0   A and this caused a minor hiccup or two since we had been assuming  ) anything with HW_MODEL <= 1024 was a VAX.   D The other gotcha was that by default on Alpha we replace the CRTL's C signals with a home-grown implementation based on the undocumented  G system service SYS$SIGPRC.  This provides more modern functionality to  G VMS v6.x systems than is available from their CRTLs and works around a  I CRTL bug on 7.x (and now 8.x) systems that precludes the ability to call  G kill() from within a signal handler.  But what we do on Alpha causes a  F stack corruption on Itanium, so for now we retreat to the bug-for-bug < compatible (but 99% functional) signal handling in the CRTL.  G So there are really no surprises here in moving to Itanium.  If you do  F something wacky and unsupported it might break, but otherwise there's F not much to it.  We keep hearing it's going to be compile and go, but F seeing is believing and now I've seen it.  Congratulations to all the  folks who made this happen.    ------------------------------    Date: 18 Dec 2003 12:02:37 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) S Subject: Re: OpenVMS Pearl - OpenVMS V8.0, first release on Itanium, ships today!!! = Message-ID: <cf15391e.0312181202.73f016fe@posting.google.com>   2 "David J. Dachtera" <djesys.nospam@fsi.net> wrote:3 > Any chance this might be available for hobbyists?  > ! > Low-cost (less than $50US) CDs?   F At the page pointed to by http://www.hp.com/go/openvms-integrity-ready we see:   , "Here's How to Order the Version 8.1 SDK Kit  E QA-73CAA-H8 OpenVMS I64 EVAL SDK CDROM KIT is available for USD 75.00 # + any locally applicable charges. "    ------------------------------    Date: 18 Dec 2003 23:50:53 -0600+ From: young_r@encompasserve.org (Rob Young)  Subject: Opteron's chances3 Message-ID: <BVXcuNHElN2d@eisner.encompasserve.org>   = 	Off-topic?  To make it topical, let's just say Opteron won't > 	be in the tier-1 4-way space for a while.  And when/if it is,? 	will it be Sun cannablizing Sun servers with 2nd and 3rd tier  C 	manufactured boxes?   HPC for them too.  Since there isn't tier-1  = 	db/web support, Opteron is relegated to HPC (IBM) and Intel  C 	can scheme their schemes on transitioning to Itanium with minimal  ! 	bumps (arguably an open debate).    	Earlier I wrote:   \ http://groups.google.com/groups?selm=dC%24EWa68vOTE%40eisner.encompasserve.org&output=gplain  + From: young_r@encompasserve.org (Rob Young)  Newsgroups: comp.os.vms 5 Subject: Re: OT: Why Scott McNeally is not worried...   Date: 21 Oct 2003 10:57:06 -0600  C 	Is it an accident that IBM offers a 2-way Opteron ONLY?  Wouldn't  E 	be much of a stretch to assume IBM would lose a steep Xeon discount  D 	if they shipped a 4-way Opteron, would it?  Why no 4-way Opteron's  	from IBM?     ---   < 	We see a good picture of Opteron's chances at Anand's site.9 	Supporting my guess at IBM's Opteron strategy/direction:   / http://www.anandtech.com/IT/showdoc.html?i=1935   N "The launch of the Opteron processor gave AMD a much needed breath of new lifeH and energy, especially with the announcement that IBM would be producingK servers based on the new Opteron platform. Unfortunately, IBMs designs are, L once again, targeted at the HPC market and left the web and database servers( for Intel and IBM processors to handle."    B 	If Opteron can't break into tier-1 database/web space, it will be1 	relegated to niche status.  HPC being its niche.    				Rob    ------------------------------  + Date: Thu, 18 Dec 2003 20:15:29 +0000 (UTC) 7 From: moroney@world.std.spaamtrap.com (Michael Moroney) & Subject: Re: PID creation in a cluster( Message-ID: <brt1t1$c4g$1@pcls4.std.com>  * brandon@dalsemi.com (John Brandon) writes:  J >I was curious about Process ID (PID) creation in a cluster and standalone	 >servers.   F The format of a PID is not documented, but as it currently exists goes somewhat like this:   D The highest bits are a unique cluster value.  It is the same for allG processes on a given node, but different from the same field on another G node in the same cluster.  It will also change upon reboot so the field = will be different from an "old" incarnation of the same node. - This field is 0 if the node is non-clustered. C Since the first node in a cluster has PIDs of 2020XXXX, the next at B 2040XXXX, etc. this field must be bits 21-30. (bit 31 is always 0)  I There are MAXPROCESSCNT 'slots', or processes that can exist at one time. D The lowest bits are which 'slot', a range from 0 to MAXPROCESSCNT-1.G Slot 0 is not used (used to be for the NULL process way back when), and  Slot 1 is always the SWAPPER.   H The middle bits are a sequence number on the "use" of a particular slot.J Starts at 1 for first use, increments each time.  SWAPPER will always have> this field as 1 since it is the first and only user of Slot 1.  D The "boundary" between the low and middle bits is different for eachG system, and depends on what MAXPROCESSCNT is set to.  (only enough bits = are used for the 'low' bits to be able to encode the value of  MAXPROCESSCNT-1)  H >I have a standalone server and the first process PID starts at 00001001 >(SWAPPER).   G MAXPROCESSCNT on this system must be somewhere between 801 hex and 1000 E hex (2049 through 4096).  Quite large, must be a fairly large server.   F >  I assume that during the startup process that there are a number ofO >processed that are created (starting with 00000000 or 00000001?) and obviously L >after startup has completed those processes have self-terminated.  Hence my >first process of 00001001.   K No, SWAPPER is the first process, hence it gets to be first user of Slot 1. I It creates a second process which kicks off all the rest.  (you'll rarely J "catch" or see this process, it has a blank/null process name and I forget which image it runs)  P >In a CLUSTER environment I would assume the same at cluster startup, however ifL >a cluster member reboots it starts with the next PID in the cluster and not
 >at 00000000.   J No, the upper bits will be nonzero for a cluster member and will be uniqueD for each member of the cluster (including departed nodes and earlierD instances of a cluster member).  This makes PIDs unique throughout a cluster.  E Again, all this is undocumented, don't write code that depends on it. 0 Not complete, either, not sure how threads work.   --   -Mike    ------------------------------  % Date: Fri, 19 Dec 2003 05:59:40 +0800 , From: Paul Repacholi <prep@prep.synonet.com>& Subject: Re: PID creation in a cluster- Message-ID: <87oeu5iy7n.fsf@prep.synonet.com>   * brandon@dalsemi.com (John Brandon) writes:  @ > I was curious about Process ID (PID) creation in a cluster and > standalone servers.   @ > I have a standalone server and the first process PID starts atD > 00001001 (SWAPPER).  I assume that during the startup process thatA > there are a number of processed that are created (starting with B > 00000000 or 00000001?) and obviously after startup has completedB > those processes have self-terminated.  Hence my first process of > 00001001.   F > In a CLUSTER environment I would assume the same at cluster startup,D > however if a cluster member reboots it starts with the next PID in" > the cluster and not at 00000000.  @ > This is my observation from my existing cluster and standalone
 > servers.  . > Am I missing something?  Just being curious.  ? The SWAPPER is not even a real process now, it is a placeholder < for all sorts of data, used to double map other processes so? as to get them contiguous in VA for swapping, and kicks off the ? shells to create new processes. It is, I suspect fairly subject  to change in the port.  C Get the Internals and Data Structures books, the chapter on startup / and on process creation have a good runthrough.   ? The PIDs are different in a cluster and a non-clustered system. A See the book for the current details. The low <n> bits of the PID > are an index into the processes, and will cycle. the high bits> can be a index into the cluster number. Again, see the book as! there are several `ifs' in there.   C Once, swapper was a real process, real enough that you could delete + it! This was, as they say, not a good idea.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Thu, 18 Dec 2003 23:06:24 +0100 " From: Didier Morandi <no@spam.com>& Subject: Re: PID creation in a cluster3 Message-ID: <3fe22532$0$1168$636a55ce@news.free.fr>    Michael Moroney wrote:  , > brandon@dalsemi.com (John Brandon) writes: >  > K >>I was curious about Process ID (PID) creation in a cluster and standalone 
 >>servers. >  > H > The format of a PID is not documented, but as it currently exists goes > somewhat like this:   J The format of the PID is documented in the OpenVMS AXP Internals and Data P Structures v1.5 (Goldenberg, Saravanan) pages 930 seq. where it is very clearly K explained how a PID has actually two parts, the internal PID, IPID and the  Q External one, EPID, together with the differences when you have a VMS Cluster or n not.   D.
 www.vaxus.orga   ------------------------------    Date: 18 Dec 2003 16:14:58 -0800# From: dooleys@snowy.net.au (dooley) & Subject: Re: PID creation in a cluster= Message-ID: <1ca82fc6.0312181614.5fbfe166@posting.google.com>n  a brandon@dalsemi.com (John Brandon) wrote in message news:<03121812243922@dscis6-0.dalsemi.com>...0K > I was curious about Process ID (PID) creation in a cluster and standalone 
 > servers. > I > I have a standalone server and the first process PID starts at 00001001iQ > (SWAPPER).  I assume that during the startup process that there are a number ofeP > processed that are created (starting with 00000000 or 00000001?) and obviouslyM > after startup has completed those processes have self-terminated.  Hence my  > first process of 00001001. > Q > In a CLUSTER environment I would assume the same at cluster startup, however if M > a cluster member reboots it starts with the next PID in the cluster and notp > at 00000000. > I > This is my observation from my existing cluster and standalone servers.e > . > Am I missing something?  Just being curious. >  >  > J*o*h*n B*r*a*n*d*o*nt > VMS Systems Administratorr, > firstname.lastname.spam.me.not@dalsemi.com: There is an explanation of the contents of PID and EPID at; http://h71000.www7.hp.com/doc/72final/6466/6466pro_006.htmle
 (section 6.3)a Phil   ------------------------------  % Date: Thu, 18 Dec 2003 14:55:43 -0800-, From: Ken Fairfield <My.Full.Name@intel.com>7 Subject: Q: Convert quadword time to Ascii time in DCL? + Message-ID: <brtb9f$63e$1@news01.intel.com>O  @      I'm sure I've seen this done as a "DCL Dirty Trick", but...  @      I'm reading a (sequential, fixed record length, implied CC)? file that has several VMS quadword times in each record.  I canaA read the records into a DCL symbol and extract the arious fields,H= e.g., the 12 (ascii) character Username field, as well as thea! 8 character (binary) time fields.i  B      I also know how to use, e.g., F$CVUI to convert the extracted. time sub-fields into a local (integer) symbol.  B      To be concrete, suppose I read a record and do the following:   $ atime = F$Extract(34, 8, hdr)  $d $ ttime = F$Cvui( 0, 32, hdr)u $ dtime = F$Cvui(32, 32, hdr) ) $ ftime = F$Fao ("!XL !XL", dtime, ttime)1 $0( $ Write Sys$Output "Date/Time: ''ftime'"  < where I'm making the assumption that the 1st 32 bits contain; the time field, and the 2nd 32 bits contain the date field.y  This gives the following output:   Date/Time: 50324152 0000F0F1   If tried using, e.g,  & $ Write Sys$Output F$Fao("!%D", dtime)  9 but that just gives an access violation.  I've also triedn variations on:  
 $ x = "  " $ x[0,32]  = %X0000F0F1S $ x[32,32] = %X50324152 " $ Write Sys$Output F$Fao("!%D", X)  < but that seems to give the same (wrong) value independent of< what values X holds.  I've also tried ":=" in the above with no joy.h  <      I'm reasonably sure I've seen this done before.  Anyone got the solution?7   	Thanks, Ken -- s6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfielde! D1C Automation VMS System Support " who:   kenneth dot h dot fairfield where: intel dot com   ------------------------------  + Date: Thu, 18 Dec 2003 23:43:09 +0000 (UTC)o, From: lewis@mazda.mitre.org (Keith A. Lewis); Subject: Re: Q: Convert quadword time to Ascii time in DCL?e. Message-ID: <brte2d$l19$1@newslocal.mitre.org>   Ken Fairfield <My.Full.Name@intel.com> writes in article <brtb9f$63e$1@news01.intel.com> dated Thu, 18 Dec 2003 14:55:43 -0800:lA >     I'm sure I've seen this done as a "DCL Dirty Trick", but...- >-A >     I'm reading a (sequential, fixed record length, implied CC)@@ >file that has several VMS quadword times in each record.  I canB >read the records into a DCL symbol and extract the arious fields,> >e.g., the 12 (ascii) character Username field, as well as the" >8 character (binary) time fields. >eC >     I also know how to use, e.g., F$CVUI to convert the extracted / >time sub-fields into a local (integer) symbol.h > C >     To be concrete, suppose I read a record and do the following:r >f  >$ atime = F$Extract(34, 8, hdr) >$ >$ ttime = F$Cvui( 0, 32, hdr) >$ dtime = F$Cvui(32, 32, hdr)* >$ ftime = F$Fao ("!XL !XL", dtime, ttime) >$) >$ Write Sys$Output "Date/Time: ''ftime'"W > = >where I'm making the assumption that the 1st 32 bits containy< >the time field, and the 2nd 32 bits contain the date field.! >This gives the following output:n >l >Date/Time: 50324152 0000F0F1t >e >If tried using, e.g,e >a' >$ Write Sys$Output F$Fao("!%D", dtime)c >i: >but that just gives an access violation.  I've also tried >variations on:s >m >$ x = "  "u >$ x[0,32]  = %X0000F0F1 >$ x[32,32] = %X50324152# >$ Write Sys$Output F$Fao("!%D", X)  >f= >but that seems to give the same (wrong) value independent of = >what values X holds.  I've also tried ":=" in the above withs >no joy. >i= >     I'm reasonably sure I've seen this done before.  Anyonec >got the solution?  I From what I've observed, the 64-bit time is in 10,000,000ths of seconds.  H (i.e. divide the 64-bit number by 1e7 to get seconds since 17-NOV-1858.)  L It's not at all easy to manipulate using 32-bit variables.  An increase of 1G to the field you call dtime represents 429.4967296 seconds, or roughly   00:07:09.50.    K If I had to do it for a DCL script I'd write a short C program which called-* $ASCTIM and stored the result in a symbol.  0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Thu, 18 Dec 2003 16:22:41 -0800a, From: Ken Fairfield <My.Full.Name@intel.com>; Subject: Re: Q: Convert quadword time to Ascii time in DCL?c+ Message-ID: <brtgci$8j2$1@news01.intel.com>d   Ken Fairfield wrote:  A >     I'm sure I've seen this done as a "DCL Dirty Trick", but...e  < Bad form following-up my own post, but I found the following> posted to c.o.v back in May 1992 by Wolfgang Moeller to be the key:? ---------------------------------------------------------------e ...  Newsgroups: comp.os.vmsi1 Subject: RE: Re-formatting F$GETJPI(pid,"CPUTIM") 4 Message-ID: <9205050736.AA09780@ucbvax.Berkeley.EDU> Date: 5 May 92 07:20:29 GMTC ...r. Fred Stevens <STEVENS@RAD.HUP.UPENN.EDU> asks:B  > Is there a VMS lexical which will re-format the CPUTIM returned7  > by F$GETJPI("","CPUTIM") into an HH:MM:SS.CC string?   F Not really. But there is a neat & unobvious method, which does respectD the fact that the number returned by $GETJPI is actually *unsigned*:   $ ticks = f$getjpi("","cputim")i $! $ L[ 0,32] = ticks& $ Q[ 0,32] = -100000 * f$cvui( 0,11,L)! $ if ticks.eq.0 then Q[0,32] = -1r8 $ Q[11,32] = -100000 * f$cvui(11,11,L) + f$cvsi(11,21,Q)8 $ Q[22,32] = -100000 * f$cvui(22,10,L) + f$cvsi(22,21,Q) $ Q[54,10] = -1- $!> $ write SYS$OUTPUT f$fao("!%D",f$cvui(32,32,f$fao("!AD",8,Q)))  < Wolfgang J. Moeller, GWDG, D-3400 Goettingen, F.R.Germany |  Disclaimer ...@ ----------------------------------------------------------------  8 I've trimmed the original header and e-mail addresses to protect Herr Moeller. :-)    So in my example:5  ; > but that just gives an access violation.  I've also triedi > variations on: >  > $ x = "  " > $ x[0,32]  = %X0000F0F1' > $ x[32,32] = %X50324152c$ > $ Write Sys$Output F$Fao("!%D", X)  . I just needed to modify the last line like so:  C $ Write Sys$Output F$Fao("!%D", F$Cvui(32, 32, F$Fao("!AD", 8, X)))y   to get     6-JAN-**** 15:13:51.29  : (which is kind of odd in its own right...reversing the twoA longwords gives an equally incorrect, 19-SEP-1859 14:52:02.32...)e  @ Now I need to go back a determine whether I've got an off-by-one= error in extracting the time field, or whether the time is ine( a different format that I thight it was.     	-Ken  --  6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken FairfieldK! D1C Automation VMS System Support " who:   kenneth dot h dot fairfield where: intel dot com   ------------------------------    Date: 18 Dec 2003 15:20:46 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)h1 Subject: Re: Singapore Exchange to run on OpenVMS-< Message-ID: <cf15391e.0312181520.87a4551@posting.google.com>   Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<brsn01$728$1@new-usenet.uk.sun.com>... = > If that is true why are OpenVMS Server revenues declining ?C   and   5 > Because HP's OpenVMS server revenues are declining.N  C According to IDC data, the trend for OpenVMS servers is upward, notnC declining. The IDC Quarterly Server Tracker data shows that OpenVMSeC server customer revenues for Q3CY2003 (HP's fiscal Q4) increased by E more than 7% over the same quarter a year ago, and units shipped were  up by more than 23%.  E For this same period, Sun's server customer revenues DECLINED by moreG	 than 11%.   F It's quite clear from the IDC data that OpenVMS server revenues took aE significnat hit after 9/11 when the economy tanked, and hit their low C point in Q4CY2002, but the trend has been upward since then (servergC customer revenues for Q3CY2003 were up more than 15% since that lowc7 point in Q4CY2002, and units shipped up more than 36%).r  & > Is ~60-70 million a quarter enough ?  B $75M in the latest quarter's estimate from IDC.  Note that this isC purely server revenue. I doubt if it includes workstations, I don't/F know if it includes operating system license revenue, and it certainly? does not include layered software, storage, services, and othere  revenue associated with OpenVMS.   ------------------------------  % Date: Fri, 19 Dec 2003 00:48:58 +0100d* From: Paul Sture <nospam@sture.homeip.net>1 Subject: Re: Singapore Exchange to run on OpenVMSh0 Message-ID: <3FE24AFA.52CD5180@sture.homeip.net>   Bill Todd wrote: >  >f   <snip>  L > Bullshit.  The problem is clear:  HP isn't willing to step up to the plateM > and support VMS.  If people want to be 'pro-active', they should beat on HPuH > to make the kind of visible commitment to VMS that would allow them toJ > encourage others to use it as you suggest without being irresponsible in > doing so.s >   C Team Christmas lunch today. Extremely enjoyable do, but the generalFG consensus around the table was that since HP do not promote VMS, we areA* all off to other things as fast as we can.  F It's an incredible shame. The best and most competent team I have everA worked with breaking up. Not just VMS though, we had a much wider 9 variety of skills, but got put under the VMS only banner.C  ; Not needed any more, because, well "VMS is dead, isn't it?"    -- u
 Paul Sture   ------------------------------  % Date: Fri, 19 Dec 2003 01:21:05 +0100e* From: Paul Sture <nospam@sture.homeip.net>1 Subject: Re: Singapore Exchange to run on OpenVMSh0 Message-ID: <3FE25281.601983FE@sture.homeip.net>  ( Andrew Harrison SUNUK Consultancy wrote: >  > Rob Young wrote:   A load of whatever.a   Andrew,l  C Please stop cluttering up this newsgroup with attacks on Rob Young.i  G We can do it all by ourselves since he tried calling people here liars.a  C My personal opiniom is that he has turned into a complete and utterm plonker.   Please let _us_ go at him :-)c  G But meanwhile Andrew, Merry Christmas to you and yours, and a Happy Newo Year.    ------------------------------    Date: 18 Dec 2003 18:52:26 -0600+ From: young_r@encompasserve.org (Rob Young) 1 Subject: Re: Singapore Exchange to run on OpenVMSx3 Message-ID: <N60USFsJfrJp@eisner.encompasserve.org>s  ] In article <3FE25281.601983FE@sture.homeip.net>, Paul Sture <nospam@sture.homeip.net> writes:h* > Andrew Harrison SUNUK Consultancy wrote:	 > Andrew,u > E > Please stop cluttering up this newsgroup with attacks on Rob Young.  > I > We can do it all by ourselves since he tried calling people here liars.p >   : 	That isn't true.  You're acting like Andrew.  Andrew just9 	today trots out VMS revenue decline (from a company with-: 	declining revenues for the last 10 quarters - sheesh) and* 	Keith pointed out that simply isn't true.  G 	What I said is they are either lying or ill-informed.  This is in the k* 	context of the non-sensical statement of:  @ "VMS is dead, Compaq killed it last year, didn't you know that!"   	What I said was this:  N > Peter's customer believes what they believe because they see nothing from HP > to the contrary. n   	Not at all.  It is either:    		1)  ill-informed 		2)  a deliberate lie   ----  = 	By the way, if VMS is dead, why are its revenues increasing?M   				Robi   ------------------------------  # Date: Fri, 19 Dec 2003 03:15:28 GMTo# From: "John Smith" <a@nonymous.com>e1 Subject: Re: Singapore Exchange to run on OpenVMStH Message-ID: <k3uEb.89586$ea%.75401@news01.bloor.is.net.cable.rogers.com>   Rob Young wrote:= > In article <3FE25281.601983FE@sture.homeip.net>, Paul Sture # > <nospam@sture.homeip.net> writes:s+ >> Andrew Harrison SUNUK Consultancy wrote:h
 >> Andrew, >>F >> Please stop cluttering up this newsgroup with attacks on Rob Young. >>C >> We can do it all by ourselves since he tried calling people herea	 >> liars.  >> >g; > That isn't true.  You're acting like Andrew.  Andrew justt: > today trots out VMS revenue decline (from a company with; > declining revenues for the last 10 quarters - sheesh) andl+ > Keith pointed out that simply isn't true.  >,G > What I said is they are either lying or ill-informed.  This is in the + > context of the non-sensical statement of:e >,B > "VMS is dead, Compaq killed it last year, didn't you know that!" >e > What I said was this:c >.G >> Peter's customer believes what they believe because they see nothinge >> from HP to the contrary.  >i > Not at all.  It is either: >, > 1)  ill-informed > 2)  a deliberate lie     Back at it again Rob?n  H If the company in question is ill-informed, then only HP is to blame for that state of affairs.  K 'Oh', says Rob, 'There's HP's web site chock full of "plans of record" that4 they can refer to.'c  J Yet the people at the company in question see only ever decreasing numbersF of VMS customers in their area, a lack of marketing effort that occursH outside of the web site, promises and 'solemn commitments' which aren't.    L As to the question of 'a deliberate lie', whom are you saying is lying - HP, or the customer?    L When are you going to get it through your reptilian brain that all those whoH frequent c.o.v. (except Andrew and those others who don't think that VMSF needs to be advertised) want VMS to thrive at sales levels higher than currently exist.     > ---- > > > By the way, if VMS is dead, why are its revenues increasing?   Possible causes are:, 1) HP increased prices to increase revenues.J 2) Some *small* number of existing customers increased their investment in VMS systems.7 3) Some *small* number of new customers bought systems.iJ 4) The stated increase in revenues is a deliberate lie. It wouldn't be theD first time revenues from a product area at HP were, we shall we say, 'creatively manipulated'.o  K Among the points above, that 2) & 3) above are true is not in dispute. ThatcH the numbers of new VMS sales represented by 2) & 3) are small is a crime against shareholder interests.   ------------------------------  # Date: Fri, 19 Dec 2003 03:18:00 GMT,# From: "John Smith" <a@nonymous.com>e1 Subject: Re: Singapore Exchange to run on OpenVMSwH Message-ID: <I5uEb.89602$ea%.76254@news01.bloor.is.net.cable.rogers.com>   Paul Sture wrote:  > Bill Todd wrote: >> >> >X > <snip> > G >> Bullshit.  The problem is clear:  HP isn't willing to step up to theaB >> plate and support VMS.  If people want to be 'pro-active', theyG >> should beat on HP to make the kind of visible commitment to VMS thatg@ >> would allow them to encourage others to use it as you suggest+ >> without being irresponsible in doing so._ >> >_E > Team Christmas lunch today. Extremely enjoyable do, but the generaliE > consensus around the table was that since HP do not promote VMS, we-0 > are all off to other things as fast as we can. >'H > It's an incredible shame. The best and most competent team I have everC > worked with breaking up. Not just VMS though, we had a much widere; > variety of skills, but got put under the VMS only banner.s >-= > Not needed any more, because, well "VMS is dead, isn't it?"2    K Sorry to hear that Paul. At this rate, pretty soon the only people who willnI be left in c.o.v. will be the cheerleading section - Keith & Rob, and the  battlebots - Bob and Andrew.   ------------------------------  % Date: Thu, 18 Dec 2003 21:58:33 -0600n@ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>1 Subject: Re: Singapore Exchange to run on OpenVMS 5 Message-ID: <3FE27769.688A7FB@NeOaSrPtAhMlNiOnWk.net>t   Keith Parris wrote:p >  > Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<brsn01$728$1@new-usenet.uk.sun.com>...e? > > If that is true why are OpenVMS Server revenues declining ?c >  > andl > 7 > > Because HP's OpenVMS server revenues are declining.  > E > According to IDC data, the trend for OpenVMS servers is upward, notrE > declining. The IDC Quarterly Server Tracker data shows that OpenVMSIE > server customer revenues for Q3CY2003 (HP's fiscal Q4) increased byhG > more than 7% over the same quarter a year ago, and units shipped wereI > up by more than 23%. > G > For this same period, Sun's server customer revenues DECLINED by moree > than 11%.) > H > It's quite clear from the IDC data that OpenVMS server revenues took aG > significnat hit after 9/11 when the economy tanked, and hit their lowhE > point in Q4CY2002, but the trend has been upward since then (server E > customer revenues for Q3CY2003 were up more than 15% since that low 9 > point in Q4CY2002, and units shipped up more than 36%).1  ? Now: compare that to the same numbers prior to the Alphacide....   -- u David J. Dachtera8 dba DJE Systems- http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Thu, 18 Dec 2003 22:07:59 -0600O@ From: "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net>1 Subject: Re: Singapore Exchange to run on OpenVMSe6 Message-ID: <3FE2799F.C08593E8@NeOaSrPtAhMlNiOnWk.net>   Rob Young wrote: > [snip]= >         To claim a product is "dead" without even a cursoryeG >         investigation is either 1) , 2) - take your pick.  To suggest.G >         it is borne out of lack of advertising is a pitiful argument.i  @ Ah, yes. PollyAnna is truly alive and well and living at hp/VMS.  F Of course, out here in the real world where the rubber meets the road,B she's an alcoholic, drug-addicted two-bit whore with AIDS, Herpes,2 Syphilis, Gonorrhea, SARS and who knows WHAT else.  C I'd truly love to know what kind of sunshine is being "blown up youlH guys' skirts", and what ex-KGB agents were hired to brainwash the lot o' yaz...  H ...and before anyone goes quoting any more IDC numbers to this group, goD back to the days before GQ Bob and compare todays VMS numbers to the; numbers from back then. Should be a major attention-getter.l  G Perhaps if you burst your own bubble, we won't have to do it for you to, snap you back to reality.y   -- C David J. Dachtera  dba DJE Systemsi http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/0   ------------------------------    Date: 18 Dec 2003 22:48:10 -0600+ From: young_r@encompasserve.org (Rob Young)r1 Subject: Re: Singapore Exchange to run on OpenVMSn3 Message-ID: <NOLwzBGf6WgJ@eisner.encompasserve.org>,  n In article <k3uEb.89586$ea%.75401@news01.bloor.is.net.cable.rogers.com>, "John Smith" <a@nonymous.com> writes: > Rob Young wrote:> >> In article <3FE25281.601983FE@sture.homeip.net>, Paul Sture$ >> <nospam@sture.homeip.net> writes:, >>> Andrew Harrison SUNUK Consultancy wrote: >>> Andrew,p >>> G >>> Please stop cluttering up this newsgroup with attacks on Rob Young.r >>>eD >>> We can do it all by ourselves since he tried calling people here
 >>> liars. >>>l >>< >> That isn't true.  You're acting like Andrew.  Andrew just; >> today trots out VMS revenue decline (from a company withe< >> declining revenues for the last 10 quarters - sheesh) and, >> Keith pointed out that simply isn't true. >>H >> What I said is they are either lying or ill-informed.  This is in the, >> context of the non-sensical statement of: >>C >> "VMS is dead, Compaq killed it last year, didn't you know that!"o >> >> What I said was this: >>H >>> Peter's customer believes what they believe because they see nothing >>> from HP to the contrary. >> >> Not at all.  It is either:a >> >> 1)  ill-informed, >> 2)  a deliberate liel >  >  > Back at it again Rob?I >   D 	You bet!  You my friend are Johnny Smith Come Lately.  I've been at? 	this since 1987 (posting to Usenet).  Point?  I not only don'tp= 	give up easily but am arguably tireless (how else to explainnA 	multiple postings in the 11 p.m. to 3 a.m. timeframe on numerous  	occasions?)  J > If the company in question is ill-informed, then only HP is to blame for > that state of affairs.  . 	No.  I haven't checked today to see if Sybase> 	is still being manufactured/shipped.  I haven't seen a Sybase6 	ad in months (that I can recall, are there any?).  Of= 	the 300+ servers I walk past on occasion, not one of them ise@ 	running Sybase.  Therefore, Sybase is Dead.  Am I ill-informed ? 	or lying?  Wouldn't anyone with a bit of interest in a productl; 	take 5 minutes to see if it is still around?  Would anyoneS< 	blame Sybase if a number of computer industry professionals= 	that Sybase wasn't around?  Shouldn't Sybase advertise more?m  M > 'Oh', says Rob, 'There's HP's web site chock full of "plans of record" thats > they can refer to.'r   	Whatever this means.s   > L > Yet the people at the company in question see only ever decreasing numbersH > of VMS customers in their area, a lack of marketing effort that occursJ > outside of the web site, promises and 'solemn commitments' which aren't. >   A 	There are commitments all over the place.  There is even DII/COEtF 	commitments.  Some are much harder (arguably impossible) than others @ 	to break out of and obviously more important.  Tell us how they 	break the DII/COE commitments.r  N > As to the question of 'a deliberate lie', whom are you saying is lying - HP, > or the customer?   	Lying?  Not necessarily.y  @ 	Pay close attention ... you might miss the subtlety here.  Here! 	is re-hash of a recent go-round:l  - > Exactly who are you accusing of lying here?  >   ?  Someone that would say VMS is dead.  That statement is either:       1)  ill-informeda    2)  a deliberate lie    ---   8 	Note the "either" in the above.  That is a logical OR.    > N > When are you going to get it through your reptilian brain that all those whoJ > frequent c.o.v. (except Andrew and those others who don't think that VMSH > needs to be advertised) want VMS to thrive at sales levels higher than > currently exist. >   A 	Where am I arguing against VMS doing better?  I'm saying lack oft; 	advertising doesn't lead one to conclude a product is deadt9 	or no longer around.  I can't recall the last time I sawwC 	an ad for the brand of carpet I bought this past summer.  In fact,VC 	I don't even know what brand it is nor do I care to know.  When I C, 	jumped in this thread, here is what I said:  ?   Why not bring them in?  Carly is interested in OpenVMS sales:o  2 http://www.techweb.com/wire/story/TWB20031212S0009  O "Shannon said HP is heavily committed to the Alpha line and noted that HP's CEO8C Carly Fiorina has made several customer calls to close Alpha sales"a    #  Sharpen your knives you whiners.     "  Here let me help you get started:  ,  "But they should be advertising it more."  &  "Without it being pushed it is dead."  n  2 	I stand by that.  Whine on ya collective whiners.  B 	I'm all for advertising VMS and have had a few suggestions.  Here 	is one:    + From: young_r@encompasserve.org (Rob Young)  Newsgroups: comp.os.vmso# Subject: VMS advertising suggestione  Date: 14 Jun 2002 14:35:58 -0600    O         There is a full page ad in the back of yesterday's Wall Street Journal -H         Marketplace section page B12.  It features a success story using7         IBM's iSeries server (OS/400) and Omaha Steaks..  H         This is the type of ad that would work very well for VMS.  A bigG         VMS success story.  Maybe co-branding with Etrade for instance.i  #                                 Robe   ---b   >> ----y >>? >> By the way, if VMS is dead, why are its revenues increasing?  >  > Possible causes are:. > 1) HP increased prices to increase revenues.  @ 	Possible?  Not likely.  Revenue is up 7% with increase in unitsC 	shipped by 23%.  An industry trend.  Server margins are shrinking.g  L > 2) Some *small* number of existing customers increased their investment in > VMS systems.9 > 3) Some *small* number of new customers bought systems.sL > 4) The stated increase in revenues is a deliberate lie. It wouldn't be theF > first time revenues from a product area at HP were, we shall we say, > 'creatively manipulated'.  > M > Among the points above, that 2) & 3) above are true is not in dispute. ThateJ > the numbers of new VMS sales represented by 2) & 3) are small is a crime  > against shareholder interests. >   E 	They aren't growing VMS fast enough to satisfy?  You have all sorts b= 	of suggestions to improve their VMS business?  Trot out some/; 	anonymous suggestions, if tenable I'll forward it for you.a  2 	But before you do, you might want to review this:  S http://groups.google.com/groups?selm=921fl3%24pdv%241%40pyrite.mv.net&output=gplaint  9 	to avoid overlap.  But you are already well aware of it:c  s http://groups.google.com/groups?selm=59PJ8.152197%24ah_.136743%40news01.bloor.is.net.cable.rogers.com&output=gplaini  # From: "John Smith" <a@nonymous.com>r Newsgroups: comp.os.vmsa Subject: Re: Open Letter to HP# Date: Fri, 31 May 2002 18:19:45 GMTo  8 "Rob Young" <young_r@encompasserve.org> wrote in message- news:9Kfk36OXT5G5@eisner.encompasserve.org...d >hA > I agree with much of what you say, but I do believe advertisingrC > should be targetted.  It is for other similar high-end OSes, i.e.tA > AS/400, mainframes and NSK and when was the last time you saw aeD > Rolls-Royce, Lamborghini, etc. commercial on television?  They run+ > ads in targetted publications - I assume.n >aA > A very good attempt to sway opinion was put together by a groupp? > that worked months formulating ideas.  That piece - penned byl@ > Bill Todd with great amount of input from others - is found at? > the Google link below.  Advertizing ideas among other things.0 >U > Robr >  >oL http://groups.google.com/groups?selm=921fl3%24pdv%241%40pyrite.mv.net&output =gplain     F As I see it, nothing has changed very much for the positive since thatG letter was penned - Alpha is dead, Itanic is slowly sinking through the H warming polar ice cap, and Sun/IBM are still eating ComHpaq's lunch. TheK letter to Capellas should be resurrected, minor revisions made, given a newiG date, search-and-replace 'Compaq' for 'HP', address it to Carly and thetJ Board of Directors (one copy individually addressed to each so it can't be/ easily buried), and one copy to Walter Hewlett.    ---n    				Rob   ------------------------------    Date: 18 Dec 2003 23:06:34 -0600+ From: young_r@encompasserve.org (Rob Young)e1 Subject: Re: Singapore Exchange to run on OpenVMS.3 Message-ID: <DQkLaAo1+Phg@eisner.encompasserve.org>   y In article <3FE2799F.C08593E8@NeOaSrPtAhMlNiOnWk.net>, "David J. Dachtera" <djesys.nospam@NeOaSrPtAhMlNiOnWk.net> writes:0 > Rob Young wrote:	 >> [snip] > >>         To claim a product is "dead" without even a cursoryH >>         investigation is either 1) , 2) - take your pick.  To suggestH >>         it is borne out of lack of advertising is a pitiful argument. > J > ...and before anyone goes quoting any more IDC numbers to this group, goF > back to the days before GQ Bob and compare todays VMS numbers to the= > numbers from back then. Should be a major attention-getter.s >   < 	Numbers?  You want numbers across quarters?  I got numbers:   						All numbers in thousands  ' PERIOD ENDING 			30-Jun-03 		30-Jun-01    ' Total Revenue 			11,434,000		18,250,000(  P0 Operating Income or Loss 	(2,724,000)		1,311,000  ? 	The darling of Silicon Valley is still shrinking - 10 straighta 	quarters and counting.g  @ 	At least VMS is increasing revenues.  Sun certainly isn't.  But6 	their friends are desperately trying to pump them up:  1 http://news.com.com/2003-12_3-0.html?tag=nefd_twka  % Sun plots plan for services business e Sun fuels cosmic research  New angles for Sun 0( Sun giving away Solaris for test drives   Sun prepares Java server update   ? 	Puff piece after puff piece for Sun.  I guess the hope is with A 	enough free advertising Sun will return to profitability?  Maybep. 	even increase revenues?  Shocka!  News at 11!  I > Perhaps if you burst your own bubble, we won't have to do it for you toe > snap you back to reality.t  = 	Oh come on - a real tangible point with facts and figures to0 	back it up, got any?s   				Robn   ------------------------------  + Date: Fri, 19 Dec 2003 06:26:28 +0000 (UTC)l From: YYYlmhYYY@cco.caltech.edu S Subject: Subject: Re: backing up VMS files to PC based CD writer - belated followupi. Message-ID: <bru5mk$61e$1@sue.its.caltech.edu>  K  Some time ago, I had inquired about backing up VMS files to a PC based CD M writer.-  ,  I would like to thank the following people    Hoff (Stephen) Hoffman Wayne Sewell -Johnr Eberhard "Tom Linden" Barry Treahy, Jr
 John Laird WWWebb Zane Martin Vorlaenders Z  David M. Smith  H who responded with a number of useful suggestions, which can be roughly  summarized as:  % ... Use zip with the options "-V" ...e* ... Or hook the CD-R device to OpenVMS ...H ... If you have access to a unix box use 'dd' to copy the raw device ...1 ... creating a backup save-set if the files will r>     ultimately be restored on a VMS system, then zip that. ...4 ... Exabyte 8mm drives were QUITE particular- only a2     very few firmware revs would work with VMS ...I ... DCL command procedure named RESET_BACKUP_SAVESET_FILE_ATTRIBUTES.COM a;     on the Freeware disks thatfixes things quite nicely ...sH ... I'd suggest acquiring one of the various cheap (used) Alpha systems  ...o    E First, an expansion of what I wanted.  We have three ties to VAX/VMS:sH 1) An old lab machine controlled by a VAXStation 3100 mod 30/38 running M VMS 5.3.  The data acquistion program and RS-232 data interface will not work  properly under VMS 5.5.  mK This machine has been used rarely in the last 5 years (not at all by me).  eK I believe a PC version of the control program is available.  We have parts eM of several computers and I would like to keep the operating disk backed up - o* we are still using the original 1989 disk.K 2) There is a VAXStation 4000-60 (and we just picked up a spare which may, eO or not, run the instrument and free us from the 1.07 GB system disk limitation lM of the 3100) running VMS 5.5-2? (upgraded from 5.3 for a database now hosted bI elsewhere) and Multinet 3.?, on which one retired individual uses TeX ande! some home-grown FORTRAN programs.tI 3) Some of our data is stored on 8mm tapes.  It would be nice to have it  C available; however, we have survived for some years without access.    I have no luck at all withF 1) Zip. The VMS executable complains about missing some LIB$SHR (sic) /  file. I have no C compiler for the zip source.eA 2) Changing file attributes on ftp'ed save_set.  There is no set  I  file/attribute command in VMS 5.x.  I have fiddled with FDL to no avail  D  (I have changed file attributes on text files by including them in 2  files which already have the desired attributes).I 3) Attaching SCSI CD.  I downloaded driver and patch a few years ago but  1  could not get any of my SCSI CD devices to work.   * Current reasonably satisfactory situation:K  By mixing parts from our three Exabyte tape drives, I was able to get one  I to work and have transferred most of the archived tape data to the VAX.   N I downloaded CYGWIN to get the dd command to transfer a disk image to the PC. B [This also worked in the opposite direction, but only if the disk K image is the same size as the disk - I don't know fine points of this.]  I DI installed simh and can run the system copied via the CYGWIN dd command.   7 E-term32 terminal emulator is a good way of connecting -H to simh.  Simh is somewhat flaky but works a lot of the time.  Simh can K then use ftp to copy user files from the real VAX'es.  [Simh disks seem to 15 be limited in size so our 4 GB user disk image won't nD mount.  Also simh can not print, so it is not a replacement for the H retired guy's VAX.]  I can then zip the disk images on the PC and write < them to CD for long term backup.  I only back up our single F VAX user's files once or twice a year; perhaps I will start up a more ) frequent backup to disk with a batch job.t   thanks again larryo   ------------------------------  # Date: Thu, 18 Dec 2003 18:35:32 GMTd( From: Alder <PGDEHMKOKIMD@spammotel.com>$ Subject: Re: Support for passive FTP, Message-ID: <UrmEb.107908$bC.73575@clgrps13>   Joseph Huber wrote::  X > In article <yv7Eb.101007$bC.6943@clgrps13>, Alder <PGDEHMKOKIMD@spammotel.com> writes: >   > Which VMS and TCPIP versions ? >   I The FTP servers I have access to at the moment are HGFTP 3.1 and the one n# included in HP TCP/IP Services 5.4.s   Thanks,i Alders   ------------------------------    Date: 18 Dec 2003 21:16:37 +0100' From: huber@mppmu.mpg.de (Joseph Huber)n$ Subject: Re: Support for passive FTP+ Message-ID: <xntzElxD43rd@vms.mppmu.mpg.de>h  W In article <UrmEb.107908$bC.73575@clgrps13>, Alder <PGDEHMKOKIMD@spammotel.com> writes:- > Joseph Huber wrote:: > Y >> In article <yv7Eb.101007$bC.6943@clgrps13>, Alder <PGDEHMKOKIMD@spammotel.com> writes:o >>  ! >> Which VMS and TCPIP versions ? K > The FTP servers I have access to at the moment are HGFTP 3.1 and the one s% > included in HP TCP/IP Services 5.4.   3 I don't claim to know much about the protocol, but p -since nobody else answered-D my understanding of the Passive FTP mode is, that the data transfer 2 goes over the same socket as the command transfer,< i.e. the FTP server does not open another port for the data.. I don't think You have to do anything special 6 on the firewall: if the FTP (command-)port 21 is open,L then passive FTP goes through, and does not open a different -blocked- port.   -- e>    Joseph "Sepp" Huber, Muenchen   http://www.huber-joseph.de/   ------------------------------    Date: 18 Dec 2003 15:35:55 -0600 From: briggs@encompasserve.org$ Subject: Re: Support for passive FTP3 Message-ID: <8fNFZ8eFQfIl@eisner.encompasserve.org>h  U In article <xntzElxD43rd@vms.mppmu.mpg.de>, huber@mppmu.mpg.de (Joseph Huber) writes: Y > In article <UrmEb.107908$bC.73575@clgrps13>, Alder <PGDEHMKOKIMD@spammotel.com> writes:  >> Joseph Huber wrote::" >>  Z >>> In article <yv7Eb.101007$bC.6943@clgrps13>, Alder <PGDEHMKOKIMD@spammotel.com> writes: >>> " >>> Which VMS and TCPIP versions ?L >> The FTP servers I have access to at the moment are HGFTP 3.1 and the one & >> included in HP TCP/IP Services 5.4. > 5 > I don't claim to know much about the protocol, but h > -since nobody else answered-F > my understanding of the Passive FTP mode is, that the data transfer 4 > goes over the same socket as the command transfer,> > i.e. the FTP server does not open another port for the data.  D That understanding is not correct.  Note that there is a distinction= between port and socket.  A socket is an interface associated0F with an end to end connection.  That connection is uniquely identified= by a four-tuple:  source IP, dest IP, source port, dest port.2  C Regardless of whether you are using passive or active mode FTP, thewB server _ALWAYS_ uses a different socket for the data transfer.  It$ _NEVER_ uses the control connection.  > With active mode FTP, the server allocates a fresh socket data? structure and uses source port 20 (or one less than the port onhA which it accepted the control connection if that was on something < other than port 21) and whatever destination port the client8 told it to use in its PORT command.  It attempts to open> an outbound connection using that socket.  (Some software may < use a random high port as the source of the non-passive modeD data connection.  Not all firewalls tolerate this misbehavior well).  : With passive mode FTP, the server allocates a fresh socket< data structure and an ephemeral port number and communicates? that ephemeral port number to the client in its response to the A client's PASV command.  It listens for inbound connections to thenG allocated ephemeral port number.  When the client makes the connection,2 the server accepts it.  @ In the complete absence of negotiation (no PORT command, no PASV@ command), the default is for the server to allocate a socket and@ open a connection from source port 20 to the port from which the> client initiated the control connection.  That may be what you> were thinking of.  But this default is never used in practice.A The problem is that if you use this default, you end up using thelA same four-tuple of source IP, dest IP, source port, dest port fort8 every data connection.  And connection re-use before the? MSL (maximum segment lifetime) has expired is problematic undera= TCP -- you have the possibility of stale packets from the old.@ connection interfering with fresh packets on the new connection.A So every client in existence (barring pathologically stupid ones)b= uses either the PORT or PASV command to negotiate non-default4 ports for the data connections.   0 > I don't think You have to do anything special 8 > on the firewall: if the FTP (command-)port 21 is open,N > then passive FTP goes through, and does not open a different -blocked- port.  A This is not the case.  The command connection uses port 21.  The -A (passive mode) data connection will use a random high port.  Some07 firewalls monitor traffic involving TCP port 21 looking:: for dialogue that identifies the negotiated data ports and4 automatically opens pinholes to permit that traffic.   	John Briggs   ------------------------------  # Date: Fri, 19 Dec 2003 00:26:07 GMTm( From: Alder <PGDEHMKOKIMD@spammotel.com>$ Subject: Re: Support for passive FTP, Message-ID: <zArEb.108739$bC.28629@clgrps13>   Joseph Huber wrote::  Y > In article <UrmEb.107908$bC.73575@clgrps13>, Alder <PGDEHMKOKIMD@spammotel.com> writes:t >  >>Joseph Huber wrote:: >>5 > I don't claim to know much about the protocol, but e > -since nobody else answered-F > my understanding of the Passive FTP mode is, that the data transfer   D Nope.  Here's a link to a decent explanation of what I'm up against:  # http://slacksite.com/other/ftp.htmll   Cheers,r Aldern   ------------------------------  # Date: Fri, 19 Dec 2003 00:28:51 GMTr( From: Alder <PGDEHMKOKIMD@spammotel.com>$ Subject: Re: Support for passive FTP, Message-ID: <7DrEb.108756$bC.46133@clgrps13>    briggs@encompasserve.org wrote::   > Some9 > firewalls monitor traffic involving TCP port 21 looking)< > for dialogue that identifies the negotiated data ports and6 > automatically opens pinholes to permit that traffic. >   5 Unfortunately, not mine.  ZoneAlarm Pro 4.0 on a W2K.0   Regards, Aldere   ------------------------------   End of INFO-VAX 2003.700 ************************