1 INFO-VAX	Mon, 18 Jul 2005	Volume 2005 : Issue 397       Contents:D Re: ... show management at HP that the licenses are being well used.D Re: ... show management at HP that the licenses are being well used.+ Backup/image complains about file not found / Re: Backup/image complains about file not found / Re: Backup/image complains about file not found / Re: Backup/image complains about file not found  CXX bug? Re: CXX bug? Re: CXX bug? Re: CXX bug? Fall 2004 VMS SIG tapes  Licence management glitch ? D Re: Now the UK Has it's own date in the history of al quiada attacks1 Re: Offshore Telephone Technical support to India 1 Re: Offshore Telephone Technical support to India 2 problem with DEC-VAXVMS-TCPIP_ECO-V0503-184-4.PCSI" Reminder: OpenVMS Hobbyist ContestA Should VMS query seconday DNS server if primary is still online ?  Third Party VAX support / Re: VMS process priorities and system processes / Re: VMS process priorities and system processes / Re: VMS process priorities and system processes / Re: VMS process priorities and system processes / Re: VMS process priorities and system processes / RE: VMS process priorities and system processes / Re: VMS process priorities and system processes / Re: VMS process priorities and system processes / Re: VMS process priorities and system processes   F ----------------------------------------------------------------------  + Date: Sun, 17 Jul 2005 22:28:50 -0500 (CDT) * From: sms@antinode.org (Steven M. Schweda)M Subject: Re: ... show management at HP that the licenses are being well used. 2 Message-ID: <05071722285008_2020ACC1@antinode.org>      From: "Kenneth Farmer":  G > If you're using a hobbyist license please enter the contest.  Even if F > you're not doing anything special with the license it still helps to? > show management at HP that the licenses are being well used.    8    As I'm not doing anything particularly "cool" with myB Hobbyist-licensed systems, I couldn't justify a contest entry, butH speaking of people at HP (management or other), and of people generally,5 a thought struck me recently (rare enough in itself).   D    1.  When I look through my Web server logs, I occasionally find aE visitor who seems to be using some proxy server or other at hp.com or G compaq.com, so, while they're probably flunkies rather than management, ( there could be some visibility that way.  E    2.  I recently viewed an Ebay page with image loading enabled on a C browser (also rare), and saw a "Powered by Sun" logo image off in a  corner.   A    3.  To the casual visitor, there's nothing on my Web server to G suggest that it's running on a VMS system, or that that VMS system uses  a free Hobbyist license.  H    While it would drastically increase the graphics content on my serverD (from the current two images plus directory icons), I would considerH adding a small, HP-approved "Reliably powered by VMS" or similar logo toH the trailer on the majority of my served pages, and, similarly, I'd alsoF consider adding a small, HP-approved "Reliable _and_ Cheap" or similarE VMS Hobbyist logo.  They could even be links to appropriate Web sites B related to VMS and its Hobbyist program.  (Animation not allowed.)  H    If nothing else, it might teach a few people how to spell "hobbyist",F but my hopes are no higher there than they are of going a week without( seeing an it's-its error on comp.os.vms.  E    Just something for the marketing department to ponder, if there is  one.  H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode-org     Saint Paul  MN  55105-2547    ------------------------------  % Date: Mon, 18 Jul 2005 01:47:40 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> M Subject: Re: ... show management at HP that the licenses are being well used. , Message-ID: <42DB4251.5D235CEF@teksavvy.com>   "Steven M. Schweda" wrote:: >    As I'm not doing anything particularly "cool" with my@ > Hobbyist-licensed systems, I couldn't justify a contest entry,  L I don't view this as a "contest", more of a survey. So I would encourage youM to fill that survey. I assume Sue wants to have some idea of the various uses ? and whether the programme is being well used or abused etc etc.     C >    3.  To the casual visitor, there's nothing on my Web server to I > suggest that it's running on a VMS system, or that that VMS system uses  > a free Hobbyist license.  - http://www.vaxination.ca/vms/shark/index.html    ------------------------------  % Date: Sun, 17 Jul 2005 19:20:37 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 4 Subject: Backup/image complains about file not found, Message-ID: <42DAE7B2.FED4C0C3@teksavvy.com>  % BACKUP/IMAGE/NOALIAS/IGNORE=INTERLOCK   = (various messages about files open for write by another user)    %BACKUP-E_OPENIN, error opening B DUA0:[VMS$COMMON.SYSEXE]SYS$QUEUE_MANAGER.QMAN$JOURNAL;1 as input ! -SYSTEM-W-NOSUCHFILE no such file   = (various messages about files open for write by another user)    -----------------   K I understand BACKUP complains about file not found in the recording pass if N the file was deleted between the time it was copied and the time the recording3 pass executes. But during the main backup pass ????   
 Further more:   = DIR $1$DUA0:[VMS$COMMON.SYSEXE]SYS$QUEUE_MANAGER.QMAN$JOURNAL   % Directory $1$DUA0:[VMS$COMMON.SYSEXE]     SYS$QUEUE_MANAGER.QMAN$JOURNAL;1   Total of 1 file.   ----  L This is in a cluster with 2 nodes having each their own system disk, but theI second node points the QMAN$MASTER to the [SYSEXE] directory on the first ( system's disk (the one being backed up).    N Is it possible there was some sort of file access conflict which resulted in a3 "no such file" message ? It seems really odd to me.    ------------------------------  % Date: Sun, 17 Jul 2005 19:48:07 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 8 Subject: Re: Backup/image complains about file not found, Message-ID: <42DAEE22.99A30334@teksavvy.com>  . In the same backup, I got another strange one:  S %BACKUP-E-OPENIN error opening DUA0:[]SYS$QUEUE_MANAGER.QMAN$JOURNAL_OLD;1 as input ! -SYSTEM-E-NOSUCHFILE no such file   C How come BACKUP would try to backup a file without proper directory  specification ?    (This is VAX VMS 7.2 btw).  < ANA/DISK/REPAIR on the disk revealed no unexpected messages.   ------------------------------    Date: 17 Jul 2005 19:09:02 -0700$ From: "AEF" <spamsink2001@yahoo.com>8 Subject: Re: Backup/image complains about file not foundC Message-ID: <1121652542.874349.127660@g47g2000cwa.googlegroups.com>    JF Mezei wrote: ' > BACKUP/IMAGE/NOALIAS/IGNORE=INTERLOCK  > ? > (various messages about files open for write by another user)  > ! > %BACKUP-E_OPENIN, error opening C > DUA0:[VMS$COMMON.SYSEXE]SYS$QUEUE_MANAGER.QMAN$JOURNAL;1 as input # > -SYSTEM-W-NOSUCHFILE no such file  > ? > (various messages about files open for write by another user)  >  > -----------------  > M > I understand BACKUP complains about file not found in the recording pass if P > the file was deleted between the time it was copied and the time the recording5 > pass executes. But during the main backup pass ????  >  > Further more:  > ? > DIR $1$DUA0:[VMS$COMMON.SYSEXE]SYS$QUEUE_MANAGER.QMAN$JOURNAL  > ' > Directory $1$DUA0:[VMS$COMMON.SYSEXE]  > " > SYS$QUEUE_MANAGER.QMAN$JOURNAL;1 >  > Total of 1 file.  E Look at the creation date of this file. I'll bet you'll find it to be E later than the start time of your backup. The queue system re-creates A this file every once in a while. I just looked at it on one of my A systems and it was created at 1:00 GMT today! So BACKUP/IMAGE was E likely looking for a file with the file-id of the previous version of E this file. So the backup begins, it reads INDEXF.SYS and gets all the F filenames and ids, then this file was re-created, and hence there is aE new file-id for it, then the backup save pass runs and can't find the A old file because it is looking for the file with the old file-id.   F I don't think it's critical to save this file. See the online docs. ItD has some very useful information about this file and the rest of the queue database.   N > This is in a cluster with 2 nodes having each their own system disk, but theK > second node points the QMAN$MASTER to the [SYSEXE] directory on the first * > system's disk (the one being backed up). >  > P > Is it possible there was some sort of file access conflict which resulted in a5 > "no such file" message ? It seems really odd to me.   * No, I don't think that's the problem here.   ------------------------------  % Date: Sun, 17 Jul 2005 22:13:31 -0700  From: Z <Z@no.spam> 8 Subject: Re: Backup/image complains about file not found* Message-ID: <_TGCe.5315$wN3.3463@fe07.lga>   JF Mezei wrote: K > Many thanks. That seems to be it. So VMS queue managet does smart version N > management by ensuring it creates new files as ;1 to prevent version numbers > from going up and up and up. > N > And that also explains the "old" one (I assume it renames the journal to the, > "old" file which also generated a message.  F It would be nice for that feature to be built into DCL and system and F RTL calls. Most developers I've known addressed file versions and the E ;32767 limit with kludges like batch files that scan the entire disk   looking for problem files.  I Or, they ignore long term file version growth and suddenly, years later,  G things like $SUBMIT/LOG in a DCL script just stops working and without  ( so much as a clue as to what went wrong.   ------------------------------  % Date: Sun, 17 Jul 2005 17:00:28 -0400 + From: Chip Coldwell <coldwell@gmail.nospam>  Subject: CXX bug? A Message-ID: <Pine.LNX.4.61.0507171653390.18996@frank.harvard.edu>    Is this a bug in the compiler?  
 $ cxx/version , Compaq C++ V6.5-004 for OpenVMS Alpha V7.3-1	 $ mms/ext   O define freetype DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE.FREETYPE.]  set default [.xpdf]  MMS/Extended_syntax  CXX /NOOPT/DEBUG/XREF/FLOAT=IEEE/EXTERN=STRICT_REFDEF/WARNINGS=(DISABLE=(NOSIMPINT))/INCLUDE=([-],[-.GOO],[-.FOFI],[-.SPLASH],"DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE]")/DEFINE=(HAVE_T1LIB_H,HAVE_PAPER_H,HAVE_FREETYPE_H) XSPLASHOUTPUTDEV.-9 CC+SYS$LIBRARY:T1LIB/LIBRARY+SYS$LIBRARY:LIBPAPER/LIBRARY v %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=000000000000001D, PC=00000000003F2220, PS=0000001B/ %TRACE-F-TRACEBACK, symbolic stack dump follows K    image    module    routine             line      rel PC           abs PC *   CXX$COMPILER  FUSE  note_source_positionP                                          45744 0000000000000630 00000000003F2220P   CXX$COMPILER  FUSE  push_fuse_scope    45595 0000000000000438 00000000003F2028.   CXX$COMPILER  FUSE  process_noted_referencesP                                          46243 0000000000002234 00000000003F3E24P   CXX$COMPILER  SCOPE_STK  pop_scope     56558 0000000000004B14 0000000000594954,   CXX$COMPILER  FUNC_DEF  scan_function_bodyP                                          54345 0000000000001264 00000000004E4FA4;   CXX$COMPILER  CLASS_DECL  inline_function_fixup_for_class P                                          50336 0000000000001558 0000000000450A789   CXX$COMPILER  CLASS_DECL  process_deferred_class_fixups P                                          50442 0000000000001770 0000000000450C90L   CXX$COMPILER  CLASS_DECL  process_deferred_class_fixups_and_instantiationsP                                          50463 0000000000000000 0000000000000000*   CXX$COMPILER  DECL_SPEC  class_specifierP                                          50797 0000000000003FC8 000000000049E158/   CXX$COMPILER  DECL_SPEC  report_bad_type_name P                                          53900 000000000000788C 00000000004A1A1C*   CXX$COMPILER  DECL_SPEC  decl_specifiersP                                          54267 0000000000000000 0000000000000000P   CXX$COMPILER  DECLS  declaration       59329 0000000000010CAC 000000000048FF8CP   CXX$COMPILER  DECLS  translation_unit  60592 0000000000012DC4 00000000004920A4P   CXX$COMPILER  CFE  edg_main            55917 0000000000000188 000000000044F3D8P   CXX$COMPILER  COMPILE  gem_xx_compile 171001 0000000000002214 00000000003F14F4P   CXX$COMPILER  GEM_CP_VMS  GEM_CP_MAIN   2603 00000000000018CC 0000000000684CACP   CXX$COMPILER                               0 0000000000984434 0000000000994434P                                              0 FFFFFFFF8028759C FFFFFFFF8028759CU %MMS-F-ABORT, For target XSPLASHOUTPUTDEV.OBJ, CLI returned abort status: %X1000000C.   D %MMS-F-ABORT, For target ALL, CLI returned abort status: %X10EE8034.  : I can send the sources that provoked this if it will help.   Chip   --   Charles M. "Chip" Coldwell Turn on, log in, tune out    ------------------------------  # Date: Mon, 18 Jul 2005 00:37:07 GMT 4 From: "Ed Vogel" <edward.vogel_stop_the_spam@hp.com> Subject: Re: CXX bug? 2 Message-ID: <TQCCe.8710$kP5.3725@news.cpqcorp.net>  9 "Chip Coldwell" <coldwell@gmail.nospam> wrote in message  ; news:Pine.LNX.4.61.0507171653390.18996@frank.harvard.edu...  >   > Is this a bug in the compiler? >  > $ cxx/version . > Compaq C++ V6.5-004 for OpenVMS Alpha V7.3-1 > $ mms/ext  >  > define freetype A > DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE.FREETYPE.]  > set default [.xpdf]  > MMS/Extended_syntax  > CXX  > /NOOPT/DEBUG/XREF/FLOAT=IEEE/EXTERN=STRICT_REFDEF/WARNINGS=(DISABLE=(NOSIMPINT))/INCLUDE=([-],[-.GOO],[-.FOFI],[-.SPLASH],"DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE]")/DEFINE=(HAVE_T1LIB_H,HAVE_PAPER_H,HAVE_FREETYPE_H)  > XSPLASHOUTPUTDEV.-; > CC+SYS$LIBRARY:T1LIB/LIBRARY+SYS$LIBRARY:LIBPAPER/LIBRARY > > %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual < > address=000000000000001D, PC=00000000003F2220, PS=0000001B  G     Without a doubt this is a compiler bug.  The compiler should never   crash.  2     If you can, create a preproces file by adding ! /PREPROCESS_ONLY/IMPLICIT_INCLUDE M     to the command line.  This should not crash, and should instead create a   .IXX file.  ThisJ     .IXX file should be able to reproduce the failure without needing any  include files.  If you couldJ     then put the .IXX file someplace where we can fetch it, we can have a  look.   L     Note that this newsgroup is not an official supprort channel, but we'll  try to take a look. H     If this is a serious problem for you, you should report it via your  problem-reporting 
     channels.        Ed Vogel#     HP/Compaq/DEC C/C++ Engineering    ------------------------------  % Date: Mon, 18 Jul 2005 09:39:41 +0800 + From: Tim E Sneddon <tesneddon@bigpond.com>  Subject: Re: CXX bug? B Message-ID: <1121650786.22c00bbe9e72d5bce9dd19c7d4f91b61@teranews>   Chip Coldwell wrote: >   > Is this a bug in the compiler? >  > $ cxx/version . > Compaq C++ V6.5-004 for OpenVMS Alpha V7.3-1  : Your compiler version is a bit old. I run V6.5-042 and I'm= sure that's out of date now as well. I can dig up the address 8 of where to pick the new C++ compilers if you want. Just mail me off-list.   
 Regards, Tim.    ------------------------------  % Date: Sun, 17 Jul 2005 23:37:36 -0400 4 From: Chip Coldwell <coldwell@frank.harvard.invalid> Subject: Re: CXX bug? A Message-ID: <Pine.LNX.4.61.0507172305100.20500@frank.harvard.edu>   $ On Mon, 18 Jul 2005, Ed Vogel wrote:   > : > "Chip Coldwell" <coldwell@gmail.nospam> wrote in message= > news:Pine.LNX.4.61.0507171653390.18996@frank.harvard.edu...  >>! >> Is this a bug in the compiler?  >> >> $ cxx/version/ >> Compaq C++ V6.5-004 for OpenVMS Alpha V7.3-1  >> $ mms/ext >> >> define freetypeB >> DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE.FREETYPE.] >> set default [.xpdf] >> MMS/Extended_syntax >> CXX >> /NOOPT/DEBUG/XREF/FLOAT=IEEE/EXTERN=STRICT_REFDEF/WARNINGS=(DISABLE=(NOSIMPINT))/INCLUDE=([-],[-.GOO],[-.FOFI],[-.SPLASH],"DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE]")/DEFINE=(HAVE_T1LIB_H,HAVE_PAPER_H,HAVE_FREETYPE_H) >> XSPLASHOUTPUTDEV.- < >> CC+SYS$LIBRARY:T1LIB/LIBRARY+SYS$LIBRARY:LIBPAPER/LIBRARY> >> %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual= >> address=000000000000001D, PC=00000000003F2220, PS=0000001B  > G >    Without a doubt this is a compiler bug.  The compiler should never  > crash. > 2 >    If you can, create a preproces file by adding# > /PREPROCESS_ONLY/IMPLICIT_INCLUDE M >    to the command line.  This should not crash, and should instead create a  > .IXX file.   I'm afraid it does crash:    $ mms/ext xsplashoutputdev.obj   CXX /XREF/PREPROCESS_ONLY/IMPLICIT_INCLUDE/FLOAT=IEEE/EXTERN=STRICT_REFDEF/WARNING=(DISABLE=NOSIMPINT)/INCLUDE=([-],[-.GOO],[-.FOFI],[-.SPLASH],DISK$USER:[COLDWELL.SOFTWARE.FREETYPE.2_1_10.INCLUDE])/DEFINE=(HAVE_T1LIB_H,HAVE_PAPER_H,HAVE_FREETYPE_H) - J XSPLASHOUTPUTDEV.CC+SYS$LIBRARY:T1LIB/LIBRARY+SYS$LIBRARY:LIBPAPER/LIBRARYv %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=000000000000001D, PC=00000000003F2220, PS=0000001B/ %TRACE-F-TRACEBACK, symbolic stack dump follows K    image    module    routine             line      rel PC           abs PC *   CXX$COMPILER  FUSE  note_source_positionP                                          45744 0000000000000630 00000000003F2220P   CXX$COMPILER  FUSE  push_fuse_scope    45595 0000000000000438 00000000003F2028.   CXX$COMPILER  FUSE  process_noted_referencesP                                          46243 0000000000002234 00000000003F3E24P   CXX$COMPILER  SCOPE_STK  pop_scope     56558 0000000000004B14 0000000000594954,   CXX$COMPILER  FUNC_DEF  scan_function_bodyP                                          54345 0000000000001264 00000000004E4FA4;   CXX$COMPILER  CLASS_DECL  inline_function_fixup_for_class P                                          50336 0000000000001558 0000000000450A789   CXX$COMPILER  CLASS_DECL  process_deferred_class_fixups P                                          50442 0000000000001770 0000000000450C90   CXX$COMPILER  CLASS_DECL  0 process_deferred_class_fixups_and_instantiationsP                                          50463 0000000000000000 0000000000000000*   CXX$COMPILER  DECL_SPEC  class_specifierP                                          50797 0000000000003FC8 000000000049E158/   CXX$COMPILER  DECL_SPEC  report_bad_type_name P                                          53900 000000000000788C 00000000004A1A1C*   CXX$COMPILER  DECL_SPEC  decl_specifiersP                                          54267 0000000000000000 0000000000000000P   CXX$COMPILER  DECLS  declaration       59329 0000000000010CAC 000000000048FF8CP   CXX$COMPILER  DECLS  translation_unit  60592 0000000000012DC4 00000000004920A4P   CXX$COMPILER  CFE  edg_main            55917 0000000000000188 000000000044F3D8P   CXX$COMPILER  COMPILE  gem_xx_compile 171001 0000000000002214 00000000003F14F4P   CXX$COMPILER  GEM_CP_VMS  GEM_CP_MAIN   2603 00000000000018CC 0000000000684CACP   CXX$COMPILER                               0 0000000000984434 0000000000994434P                                              0 FFFFFFFF8028759C FFFFFFFF8028759CU %MMS-F-ABORT, For target XSPLASHOUTPUTDEV.OBJ, CLI returned abort status: %X1000000C.   I The problem seems to be related to the /XREF option.  If I take that out  " of the command line, it runs fine.  I > If you could then put the .IXX file someplace where we can fetch it, we  > can have a look.  < I gathered together the whole kit and put it on an ftp site:  1 ftp://frank.harvard.edu/pub/coldwell/vms/cxx-bug/   G There are three archives there that you need (all created with vmstar):   8 freetype-2_1_10.tar-gz  headers.tar  xpdf-3_00pl3.tar-gz  H The first contains the complete freetype source code, but you only need K the headers.  The second contains two text libraries with headers from the  J t1lib and libpaper libraries.  The third is xpdf, the program I am trying  to compile and analyze.   I Put freetype and the two text libraries wherever you want, and then edit  J [.3_00pl3]descrip.def in the xpdf sources to point to where you put them.  Then   $ set def [.3_00pl3.xpdf]  $ mms/ext xsplashoutputdev.obj  I should reproduce the bug.  At the risk of being repetitive, my version is   
 $ cxx/version , Compaq C++ V6.5-004 for OpenVMS Alpha V7.3-1  J > Note that this newsgroup is not an official supprort channel, but we'll H > try to take a look.  If this is a serious problem for you, you should 0 > report it via your problem-reporting channels.  E I'm afraid that I'm a mere hobbyist, so I don't have recourse to the  F official problem-reporting channels.  But I am very grateful for your  attention to this issue.   Chip   --   Charles M. "Chip" Coldwell "Turn on, log in, tune out"    ------------------------------  % Date: Sun, 17 Jul 2005 23:09:54 -0400 ' From: Glenn Everhart <Everhart@gce.com>   Subject: Fall 2004 VMS SIG tapes, Message-ID: <SKednTbNTMROgEbfRVn-ow@rcn.net>   Folks - L The Fall 2004 VMS SIG tapes have now been sent to the top level of the tree.@ Some may have arrived; others are mailed. I will post a table of contents shortly.   G I expect spring 2005 collection pretty soon after; this one took longer  to get out than expected.   E Hopefully some of those receiving the collection will get it onto net  distribution points.   Glenn Everhart everhart@gce.com 156 clark farm rd  smynra, del 19977  302 659 0460   ------------------------------  % Date: Sun, 17 Jul 2005 16:24:28 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> $ Subject: Licence management glitch ?, Message-ID: <42DABE74.3C7CE685@teksavvy.com>  I The electric utility decided to cut power to about 100,000 customers this J afternoon. My all mighty adult microvax II had no choice but to shut down, deprived of its life support.   M When power came back, it rebooted apparently normally (along with its younger G brother, a vax station in a cluster, but with 2 separate system disks).   H During the long process, the MVII came to the part where it has to startN ALL-IN-1.  The startup procedure setups plenty of logicals, installs plenty ofJ images, and imvokes ALL-IN-1 a few times to do initialisation (and it also installs stuff itself).   B Today, ONE (and only one) such invocation of ALL-IN-1 resulted in:  F $ say " Running ALL-IN-1 to start local File Cabinet Server Processes">  Running ALL-IN-1 to start local File Cabinet Server Processes $ say ""   $ ALLIN1 /NOINIT /OVERRIDE  B %LICENSE-E-NOAUTH, DEC ALL-IN-1 use is not authorized on this nodeD -LICENSE-F-NOLICENSE, no license is active for this software product1 -LICENSE-I-SYSMGR, please see your system manager   B %DCL-W-SKPDAT, image data (records not beginning with "$") ignored $ RETURN8 $ if .not. a1$kipostinstall then gosub START_AIDA_SERVER $START_AIDA_SERVER:  $ say ""  > $ say " Running ALL-IN-1 to start local AIDA Server Processes"6  Running ALL-IN-1 to start local AIDA Server Processes $ say ""   $ ALLIN1 /NOINIT /OVERRIDE    OA$FLO_OPEN_LIB OA$LIB:MANAGER
   etc etc etc     L You'll note that immediatly after the failure of the invocation, the startupN procedure invokes ALL-IN-1 again and this time it works. Prior to the failure,G ALL-IN-1 had already been invoked. I had never seen this happen before.   M Multiple concurrent invocations of ALL-IN-1 are permitted by the licence. The @ licence database is unique to this node (not shared to cluster).  M Are there situations where the LMF itself would lock out and refuse access to L a layered product due to various circumstances ? i am curious as to why this
 happened ?    L Note that this happens a very long time after the system has done it initialI boot where licences are loaded and activated.  And at the time this runs, M there isn't much other activity on the system since this is runned as a batch M job in a /JOB_LIMT=1 queue and I generally wait for all to be complete before  I login to that node.    ------------------------------    Date: 17 Jul 2005 21:46:23 -0700$ From: "AEF" <spamsink2001@yahoo.com>M Subject: Re: Now the UK Has it's own date in the history of al quiada attacks C Message-ID: <1121661983.088652.308680@g49g2000cwa.googlegroups.com>    John Smith wrote:  > Bill Todd wrote: > > AEF wrote: > >> > >> John Smith wrote: > >  > > ...  > > I > >>> Who said I wanted good mileage? I have customers with lots of money 	 > >>> :-)  > >> > >> > >>B > >> This just supports my statement that it is odd to have higher4 > >> mileage-optimal speeds as a technological goal. > > J > > Anecdotal evidence seldom supports anything all that much.  Of course,H > > since John did not actually say that he was *not* interested in goodI > > gas mileage (just observed that no statement one way or the other had G > > yet been made about that point - including what I interpreted as an ; > > ironic smiley), it wouldn't support that thesis anyway.  > > G > > Optimal mileage at the speeds their cars are typically driven at is B > > almost certainly a goal at BMW:  they can pass pretty well forG > > perfectionists in most aspects of vehicle performance, not just the F > > screechy-tires kinds.  Back in the automotive Iron Age they had toJ > > compromise a lot between power at full throttle and efficiency at partJ > > throttle, but with modern feedback-controlled carefully-metered engine6 > > components such compromise is no longer necessary. > >  > >   And if you don't care 8 > >> about mileage, why have you bothered to measure it? > > H > > Even if he doesn't actually care, techies are often into such things. > > simply as a matter of technical curiosity. > >  > >   And how did you C > >> measure it? Are you sure you took all influencing factors into A > >> account? Changes in wind, grade, steadiness of your foot(!)?  > >> Personal bias?  > > J > > That's a better question, but still omits the crucial one:  whether heJ > > made equally-careful measurements at a range of speeds under identical= > > conditions so that legitimate comparisons could be drawn.  > >  > >>H > >> I'm almost surprised that these luxury/power cars even have mileageJ > >> meters. Those who are interested in mileage wouldn't even be shopping > >> for such cars.  > > J > > You really need to learn to speak only for yourself when talking aboutE > > personal preferences:  I'm interested in mileage (along with most C > > other aspects of vehicle performance), my interest in it is not H > > primarily financial (but rather one of social responsibility), and IF > > would very definitely be interested in BMWs if they weren't such aF > > conspicuous extravagance in other ways (which have less to do withI > > social responsibility than with personal inclinations:  we native New F > > Englanders often don't cotton to conspicuous extravagance all that
 > > much). > > > > >   And I can't see a luxury car shopper saying, "Well, thisI > >> one has a mileage meter! I'll get it!" I guess it's just part of the , > >> total "lots of bells and whistles" bit. > > H > > Somehow I get the feeling that you're not a car enthusiast, in which> > > case your complete lack of comprehension of the species is > > understandable.  > > E > > Of course, not all BMW buyers are car enthusiasts, either.  And I J > > don't even know whether John falls into that category, though from his< > > statements I suspect he has at least a toe in that camp. > >  > >> > >> > >>E > >>>>> The 540 has a pretty good Cd (coefficient of drag), somewhere  > >>>>> aroundI > >>>>> 0.30 IIRC, where a 1.0 represents something akin to a 4'x8' sheet $ > >>>>> of plywood face to the wind. > >>>>G > >>>> That helps overall mileage, but I doubt it changes the fact that ! > >>>> drag increases with speed.  > >>> D > >>> Physics is physics, at least at the Newtonian scale of things. > >> > >>I > >> At least? Are you saying that non-Newtowian physics isn't or somehow 7 > >> might not be physics? That doesn't make any sense.  > > F > > Without wishing to put words into John's mouth, my reading is thatF > > he's referring to the unpredictability of individual events at theH > > quantum level and/or the incestuous interaction of measurements with  > > what one is measuring there. >  >  > Bingo. > F > Yes to auto enthusiast. No to having time to worry about whether theL > ignition point is 3 or 4 degress before TDC or whether Bilsteins or Koni's > are better than the other. > L > Heisenberg doesn't really apply on the scale of measurements being done toH > quantify gas mileage or Cd at 70 vs. 71 mph. We're not talking nuclearI > cross-sections of  1EE-29 and c here....it's more like 1/4 mile in 12.5 
 > seconds.  G "Heisenberg" is still physics. Measurements in the macrosopic realm are F STILL subject to errors. Even the measuring process can produce error.E Measure the voltage between two points in a circuit and you draw some D current. That will affect the measurement. But, in classical physicsD you can in principle measure things as precisely as you wish. Things< have real properties independent of the measurement process.  E The difference in the quantum mechanics is that EVEN IN PRINCIPLE you B cannot measure something as precisely as you wish without creatingC uncertainty in a conjugate variable. Experiments support the notion < that a particle doesn't even have a complete definite set ofE properties. There are wave functions for which some of the properties D aren't even "real" until a measurement is done (or something happensD that can, in principle, be used to make a measurement -- there is noG need for an actual experimenter to be present or the cause of the thing F that happens). Look up the GHZ paradox in quantum mechanics for a good example.  E Getting back to classical physics (Newtonian, if you wish): There can G still be errors. Also, digital readouts are not necessarily accurate to F the precision indicated. If it says 23.4 MPG, due to measurement errorE it may be a little higher or lower. Of course one can build a display D that only shows properly to the correct significant figures that theE device can measure down to, but I wouldn't be the least bit surprised E if these digitial readouts aren't so accurate. Precision and accuracy D are not the same thing. Accuracy is how close you get to the correct? value. Precision is how well specified the value is without any E reference to the actual value. As a crude example, if you had a scale B that reads down to 0.1 pound increments, but tended to read 10% toG high, your precision is to the nearest 0.1 pounds, but your accuracy is B considerably less. Also, the readings must be consistent, though I, suspect that's not a big worry in this case.  N > Besides, I'm more interested in getting VMS advertised and marketed - that'sM > what would allow me to afford a new M5-series w/ 500 bhp and a V-10 engine.   C And if you don't get this car it will be the end of the world? This  should be your biggest problem.    ------------------------------  % Date: Sun, 17 Jul 2005 21:47:48 -0600 " From: GreyCloud <cumulus@mist.com>: Subject: Re: Offshore Telephone Technical support to India( Message-ID: <42DB2664.F8CE66EA@mist.com>   JF Mezei wrote:  > L > The local squirrels and racoons get fed with peanuts from china. The localO > convenience store has bottled water from the region, as well as bottled water O > from Fiji.   In New Zealand, I saw bottled water from Qubec. In Australia, I ! > saw bottled water from Iceland.  >   
 I haven't.  K > Globalisation is there. live with it. Learn to profit from it. If a small O > qubec bottler was able to export water to New Zealand (which has no shortage = > of water), then it shows there are opportunities to be had.  >   ; Clean drinking water, unfortunately, will be in high demand  in the not so distant future. : There are opportunities is creating filtration systems for! home use that not only filter out 3 harmful viruses or bacteria but also harmful exotic ( chemicals that eventually seeps into the/ public wells that current filters can't handle.   5 But globalization is whose idea??  Is there an agenda ' attached here somewhere and who exactly  benefits from it??  9 Getting your degree in any particular field today will no $ longer guarantee a continuing future: career let alone any kind of hopes for a safe retirement. # That's why a lot of the peoples in  7 Europe aren't voting for the EU constitution along with  other reasons.  P > Moving a service to India isn't wrong per say. What is wrong is moving it to aP > whole bunch of folks who just got trained and don't yet have enough experienceM > to provide the level of technical expertise that used to be provided. Given S > time, they will acquire such experience, but meanwhile, customers won't be happy.  >   ; It is fine as long as the caller for service can understand  the support staff on the phone. 7 If you can't understand them, where else can you call?? : Example:  HP all-in-one printers are supported in India by! one guy.  No other way to contact - HP on the phone.  Basically it is a dead end.    N > pre outsourcing, your support infrastructure had a relatively small turnoverA > with new folks coming in and working with experienced folks, so . > knowledge/experience was quickly transfered. > N > But when you train just a few very quickly and send them to India to train aN > whole bunch of people who take over all support calls next month, then thereK > is no way that they can provide the depth of knowledge and experience the 2 > former support organisation was able to provide. > J > This would happen irrespective to where you're moving it. As a canadian,P > dealing with someone with an indian accent isn't much harder than dealing withS > an american with deep southern accent, or a newfoundlander with deep nfld accent.  > G > If the knowledge is good, you don't complain about the accent. If the J > knowledge/service isn't good, then you complain and add complaints about > accent to the lot.  ; If the accent is far enough alien, then the service is poor  for the customer.    ------------------------------  % Date: Mon, 18 Jul 2005 01:44:38 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> : Subject: Re: Offshore Telephone Technical support to India, Message-ID: <42DB419C.39FD54DD@teksavvy.com>   GreyCloud wrote:< > There are opportunities is creating filtration systems for# > home use that not only filter out 5 > harmful viruses or bacteria but also harmful exotic   L Yep, and that is what Coke and Pepsi are using to make their "bottled water"P (Dassani and Aquafina) from normal tap water and sell it in stores at a premium.  H point is that if you make a product efficiently enough, by investigatingJ various transportation options, you find that you can ship your product to0 some distant countries and still be competitive.    7 > But globalization is whose idea??  Is there an agenda ) > attached here somewhere and who exactly  > benefits from it??    @ Globalisation is happening because the world is getting smaller,J telecomminication costs dropping, and more and more transportation options exist.    J India's outsourcing success would not be possible if it were not for cheapM telecommunications where a long distance call to India doesn't cost much more - than an intra-state long distance in the USA.   ; > Getting your degree in any particular field today will no & > longer guarantee a continuing future; > career let alone any kind of hopes for a safe retirement.     L Lets be blunt here: The standard of living of many is unsustainable and willA have to drop, while the standard of living of the poor will rise.   J This worked even before globalisation, but on a local scale. DEC won salesN because they were cheaper than IBM. Then SUN cam in and the PCs came in , eachK cheaper than the other and killed DEC because DEC was unwilling to make its 3 own product compete head to head and reduce prices.   M You expect a home, 2.5 BMW/SUVs, TVs, VCR/DVD,IPOD, trip to europe every year K etc etc. To sustain your expectations, you set salary demands and since you L know fellow workers set the same demands, you know you have a good chance of getting what you want.  J Well now, your co-workers are in India, China etc  and set much lower wageK demands. You have a choice of lowering your wage demands, or lose your job.  It's that simple.   K If you want to maintain your high wages, you must offer unique products and I services. That means writing software for instance. (And in that vein, HP C would be far better off focusing on software than on wintel boxes).   N It isn't something people are willing to discuss or even consider. But look atH Wallmart in the USA. They don't care about sustaining a high standard ofL living with high waged, they care about selling at the lowest possible cost.    K Ask yourself this: If your customer/employer came to you and said that they E have another contractor willing to work at half the price and similar N experience. Would you be willing to accept such a wage cut ? Probably not, andK your initial reaction would be to pooh pooh that competitor and try to make L yourself look much better than your competitor. But in the end, the employerB hires the cheaper one and after a while, things just run smoothly.    N You also have to look at it from a trends point of view. PHB rtead their tradeM rags and read that outsourcing is great. So, to keep up with the Jones', they L decide to also outsource.  They send out an RFP, and get bids from the largeJ USA based outsourcing/consulting firms with high prices, and they get bidsK from Indian firms that are way lower. Which do you think they will choose ?   H What is needed in the western world is the IT equivalent of the low costM airlines. The big fat consulting firms charge ridiculous rates because in the M past, the big fat customers such as banks actually wanted to pay huge amounts L for a report so that they could to to their superiors and show the report asE justification for a decision, backed by some really serious expensive  consulting firm.    I Look at VMS which historically has been unwilling to compete head to head K against PCs and PCs/Unix stole the marketplace from under VMS. Its the same 
 for workers.    L There is a new kid on the block competing with lower wages. That is going toI push out the big fat cats and keep only hungry workers willing to work at J lower wages and show at Wallmart and fly Southwest instead of expecting to3 shop at Saks 5th Avenue and fly United First Class.   = > It is fine as long as the caller for service can understand ! > the support staff on the phone.   F In the USA, you are perhaps not used to dealing with people having nonK american accents. But over time, americans should adapt to this. Elsewhere, J people are more used to dealing with accents so speaking to someone havingJ indian accent isn't so traumatic. (however, there are limits , of course).    = > If the accent is far enough alien, then the service is poor  > for the customer.     L That I agree with. There are limits to how far from "baseline" an accent can% be when providing worldwide support.    K And a worldwide corporation must also be able to cater to people who do not M speak english. For instance, a company serving all of north america should be L able to provide support in english, french and spanish. India can't do that.   ------------------------------  + Date: Sun, 17 Jul 2005 19:32:27 +0000 (UTC) 0 From: helbig@multiCLOTHESvax.de (Phillip Helbig); Subject: problem with DEC-VAXVMS-TCPIP_ECO-V0503-184-4.PCSI 1 Message-ID: <dbebob$1kpk$1@fiasco.xenopsyche.net>    This is strange.  D After doing a fresh install of VMS 7.3 on a VAXstation 4000/60 (fromB the same CD from which I did an UPGRADE from 7.2 to 7.3 on anotherB VAXstation 4000/60 and a VAX 4000/100A) and installing ALL patchesC (as well as DECwindows, TCPIP and C), I tried to install the latest F TCPIP ECO, which I have successfully installed on the two other VAXes,C also running 7.3.  (The only difference---apart from upgrade versus C fresh install---is that on the other two machines I first had TCPIP F 5.1 under 7.3, then later went to 5.3, whereas on the new machine with7 the fresh install, I installed TCPIP 5.3 from scratch.)    I get this problem:    $ product install *   ( The following product has been selected:E     DEC VAXVMS TCPIP_ECO V5.3-184          Patch (maintenance update)    Do you want to continue? [YES]    Configuration phase starting ...  N You will be asked to choose options, if any, for each selected product and forO any products that may be installed to satisfy software dependency requirements.   B DEC VAXVMS TCPIP_ECO V5.3-184: Compaq TCP/IP Services for OpenVMS.  >     (c) Compaq Computer Corporation 2004. All Rights Reserved.       Compaq Computer Corporation   F     COMPAQ TCP/IP Services for OpenVMS offers several license options.  7 * This product does not have any configuration options.    Execution phase starting ...  7 The following product will be installed to destination: H     DEC VAXVMS TCPIP_ECO V5.3-184          DISK$VAXSYS_220:[VMS$COMMON.]   Portion done: 0%...10%  P %PCSI-E-READERR, error reading DSA220:[SYS0.][SYSUPD.20050717]DEC-VAXVMS-TCPIP_E CO-V0503-184-4.PCSI;1 # -DDIS-E-TNF, invalid element syntax " %PCSI-E-OPFAILED, operation failedE Terminating is strongly recommended.  Do you want to terminate? [YES] P %PCSI-E-CANCEL_WIP, termination resulted in an incomplete modification to the sy stem/ %PCSI-E-S_OPCAN, operation cancelled by request M %PCSIUI-E-ABORT, operation terminated due to an unrecoverable error condition    Any idea what's going on?   E I expanded the downloaded file again, same problem.  Since it expands D OK, I assume that the .PCSI file itself is OK.  Also, the read error= looks like a "logical", rather than a "physical", read error.    Some questions:    Has anyone else seen this?  5 Why the difference between upgrade and fresh install?    What does the error above mean?    What will happen if I continue?   F I haven't noticed any problems since aborting.  I rebooted for anotherE reason since aborting.  TCPIP 5.3 with no patches seems to be running C OK, i.e. doesn't appear to be affected by the aborted installation.   D The upgrade itself and subsequent patch installation had no problemsD at all.  I installed all the patches in the 7.3 area except a couple% for stuff I am not running (COM etc).   E Do I need some sort of PCSI patch?  If so, where can I get it?  It is J not in the 7.3 area.  (On the other hand, though, though the other patchesL were with VMSINSTAL, other stuff such as TCPIP and DECwindows were installed with PCSI with no problems.  \    ------------------------------  # Date: Mon, 18 Jul 2005 02:47:35 GMT 6 From: "Kenneth Farmer" <kfarmer@NOSPAM.spyderbyte.com>+ Subject: Reminder: OpenVMS Hobbyist Contest < Message-ID: <bLECe.6824$oj4.224970@twister.southeast.rr.com>  9 Just a reminder to everyone using the hobbyist license...   K The OpenVMS Hobbyist Team is having a contest for the five coolest uses of  K the VMS Hobbyist licenses.  The team will pick the five top contenders and  + everyone will then be allowed to rank them.   M If you're using a hobbyist license please enter the contest.  Even if you're  C not doing anything special with the license it still helps to show  7 management at HP that the licenses are being well used.    OpenVMS Hobbyist Contest* http://www.openvms.org/pages.php?page=Cool   View current top runners1 http://www.openvms.org/pages.php?page=Contestants   L Once again, please enter the contest if you're using the license.  The more  the better.    Ken   % _____________________________________ " Kenneth R. Farmer <>< 336-736-7376# OpenVMS.org: http://www.OpenVMS.org ' dba.OpenVMS.org: http://dba.OpenVMS.org ( dcl.OpenVMS.org: http://dcl.OpenVMS.org    ------------------------------    Date: 17 Jul 2005 17:49:28 -0700" From: freestyle_london@yahoo.co.ukJ Subject: Should VMS query seconday DNS server if primary is still online ?C Message-ID: <1121647768.711620.100790@g49g2000cwa.googlegroups.com>    Hi.   = >From a windows background so thanks in advance for any help.   D Our VMS server is connected to 2 DNS servers. I'm told that VMS does= not have primary or secondary DNS servers only a lookup list.   G I need it to ALWAYS query server1 first, and ONLY server2 if server1 is  not responding...   $ How do you fellows have yours setup.  0 By the way the DNS BIND services on windows 2003   ------------------------------  % Date: Sun, 17 Jul 2005 23:01:38 -0700 . From: "Barry Streets" <berrys2552@comcast.net>  Subject: Third Party VAX support0 Message-ID: <c7mdne6Ki745oUbfRVn-3A@comcast.com>  J I work for a company that has a legacy software application that only runsK on VAX's and can't be ported to Alpha. I know about the Charron-VAX option, L but are there any third party companies out there doing hardware maintenance on the VAX's dropped from HP.   @ Looking for several used 4000-96 workstations cheap any ideas???  * Please respond to the e-mail below; thanks    
 Barry Streets  The Echo Group bstreets@echoman.com   ------------------------------  % Date: Sun, 17 Jul 2005 19:55:52 +0200 , From: Albrecht Schlosser <ajs856@tiscali.de>8 Subject: Re: VMS process priorities and system processes, Message-ID: <e36ebd.mtu.ln@news.hus-soft.de>   JF Mezei wrote:   P > If your system runs just one application, then why not make your batch defaultO > priority 0, and then you can use 1 2 3 for lower priority processes and 4 for 2 > your highest and interactrive system management.  J yes, good idea, I thought of something like this, too. In fact we lowered N interactive (maintenance user) base priority already from 4 to 3, and we have G batch queues with priority 1 and 3 (now better 1 and 2 ...). We'll see.    ------------------------------  % Date: Sun, 17 Jul 2005 20:16:49 +0200 , From: Albrecht Schlosser <ajs856@tiscali.de>8 Subject: Re: VMS process priorities and system processes, Message-ID: <na7ebd.30v.ln@news.hus-soft.de>   Main, Kerry schrieb:   >  >>-----Original Message-----6 >>From: Albrecht Schlosser [mailto:ajs856@tiscali.de]  >>Sent: July 17, 2005 10:07 AM >>To: Info-VAX@Mvb.Saic.Com : >>Subject: Re: VMS process priorities and system processes >> >  > 
 > [snip..] >  > F >>>As others have stated, you can get into trouble by purposely mixing% >>>priorities on specific priorities.  >>* >>I don't really understand this sentence. >> >  > G > Oops, my apologies - slippery fingers. The statement should have read E > "you can get into trouble by manually mixing priorities on specific 
 > processes."  > G > Fwiw, whenever I give talks about virtualization and/or consolidation G > (on any platform), workload management (WLM - not to be confused with C > the prod on HP-UX) is one of the main key's to making application J > stacking work. Workload management is certainly not a new concept and inJ > fact, the mainframe folks live and breathe this stuff all the time. Some6 > may know WLM as another term ie. "class scheduler".  > I > In a nutshell, you decide what processes should be part of a scheduling H > class and then decide what service level objectives (SLO) are for thatG > group i.e. "it is low priority, so that group of processes should not H > take more than 15% of the overall cpu resources between x and y hrs ofJ > the day" while another group might be "medium priority, hence this group" > can take up to 35% of the CPU.."  Q Well, I understand the concept, but it's not that easy, because I can't say that  P some users should have a certain amount of cpu time "between x and y hrs of the P day". I don't want to limit the use of resources that are available - /windfall  comes to mind...  B > While OpenVMS's priority and multi-tasking design has made it anI > extremely stable multi-tasking OS, it is not immune from the occasional B > process doing something silly (ask those who implemented earlierI > versions of WEBES diag stuff), so WLM or the class scheduler is one way E > to further minimize the risk of the "single process doing something   > silly" problem from occurring. > I > And the class scheduling components are part of the OS, so there are no  > additional charges.   O I'm not sure if this concept will help. "Partioning" the CPU in a way that the  M sum of classes uses 100% seems to be a waste of resources if not all classes  I have that demand at a certain time. Would it be okay to give 70+20+20+30  O percent? What would happen if there would be more than 100% demand? Scheduling  O by priority? I couldn't find much about the class scheduler in the performance  6 management manual, but I'll try to find more tomorrow.   Thanks for your help   Albrecht   ------------------------------  + Date: Sun, 17 Jul 2005 19:03:07 +0000 (UTC) 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 8 Subject: Re: VMS process priorities and system processes? Message-ID: <dbea1a$fiv$1@nwrdmz02.dmz.ncs.ea.ibs-infra.bt.com>    Hi Albrecht,  K I think I'm in agreement with you regarding class schedulers (and Row Cache J memory schedulers) and all other forms of bollocks nanny-state Dr, ZhivagoL "far more equitable comrade" socialist interference. (Oh here come all thoseL wankers from the other thread(s)) "The some of the parts is a shit-load lessE than the whole!" Kerry would do well to put that above his desk. It's . motivated numerous Rdb engineers for years :-)  K My product has Network communication servers running, by default, at a base H of 6 (expecting your usual interactive users to run at 4. My TransactionH Manager runs at one higher than the communication server. (Leaving a gapJ till the real-time priorities) I think it makes sense! Look at some of theH servers running on your system and see the base priorities they run at - they're all doing it.   H In my case, if the system manager doesn't like it, he simply changes theC base priority in the UAF for TIER3$SERVER. The lovely documentation  discusses priorities.    Regards Richard Maher.  9 "Albrecht Schlosser" <ajs856@tiscali.de> wrote in message & news:na7ebd.30v.ln@news.hus-soft.de... > Main, Kerry schrieb: >  > >  > >>-----Original Message-----7 > >>From: Albrecht Schlosser [mailto:ajs856@tiscali.de]   > >>Sent: July 17, 2005 10:07 AM > >>To: Info-VAX@Mvb.Saic.Com < > >>Subject: Re: VMS process priorities and system processes > >> > >  > >  > > [snip..] > >  > > H > >>>As others have stated, you can get into trouble by purposely mixing' > >>>priorities on specific priorities.  > >>, > >>I don't really understand this sentence. > >> > >  > > I > > Oops, my apologies - slippery fingers. The statement should have read G > > "you can get into trouble by manually mixing priorities on specific  > > processes."  > > I > > Fwiw, whenever I give talks about virtualization and/or consolidation I > > (on any platform), workload management (WLM - not to be confused with E > > the prod on HP-UX) is one of the main key's to making application L > > stacking work. Workload management is certainly not a new concept and inL > > fact, the mainframe folks live and breathe this stuff all the time. Some7 > > may know WLM as another term ie. "class scheduler".  > > K > > In a nutshell, you decide what processes should be part of a scheduling J > > class and then decide what service level objectives (SLO) are for thatI > > group i.e. "it is low priority, so that group of processes should not J > > take more than 15% of the overall cpu resources between x and y hrs ofL > > the day" while another group might be "medium priority, hence this group$ > > can take up to 35% of the CPU.." > I > Well, I understand the concept, but it's not that easy, because I can't  say thatJ > some users should have a certain amount of cpu time "between x and y hrs of theG > day". I don't want to limit the use of resources that are available - 	 /windfall  > comes to mind... > D > > While OpenVMS's priority and multi-tasking design has made it anK > > extremely stable multi-tasking OS, it is not immune from the occasional D > > process doing something silly (ask those who implemented earlierK > > versions of WEBES diag stuff), so WLM or the class scheduler is one way G > > to further minimize the risk of the "single process doing something " > > silly" problem from occurring. > > K > > And the class scheduling components are part of the OS, so there are no  > > additional charges.  > L > I'm not sure if this concept will help. "Partioning" the CPU in a way that the F > sum of classes uses 100% seems to be a waste of resources if not all classes J > have that demand at a certain time. Would it be okay to give 70+20+20+30E > percent? What would happen if there would be more than 100% demand? 
 SchedulingD > by priority? I couldn't find much about the class scheduler in the performance 8 > management manual, but I'll try to find more tomorrow. >  > Thanks for your help > 
 > Albrecht >  >    ------------------------------  % Date: Sun, 17 Jul 2005 15:12:54 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 8 Subject: Re: VMS process priorities and system processes, Message-ID: <42DAADB3.54415C45@teksavvy.com>   Richard Maher wrote:M > My product has Network communication servers running, by default, at a base J > of 6 (expecting your usual interactive users to run at 4. My Transaction; > Manager runs at one higher than the communication server.   K Don't forget that for short bursts of work, priority doesn't make much of a G difference if all your processes do short bursts of work at a time. Low J priority processes get their priority boosted while waiting for IO so that> when IO completes, they have a chance to execute now and then.  J Where it matters is when you have a process which will sometimes go into aK long compute process (for instance, encrypting some large in-memory buffer. G Giving that process a high priority will essentially jam lower priority  processes during the process.   M If you give users a lower priority, it means that while the background server I process is eating CPU cycles by the dozen, users will see severe delay in 3 echoing of characters in a screen/GUI application.    M If the background server process is mostlty IO bound, then its priority won't K matter much because at each IO, it gives another process a chance. And when ' its IO completes, it is given a chance.    ------------------------------  + Date: Sun, 17 Jul 2005 19:20:53 +0000 (UTC) 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 8 Subject: Re: VMS process priorities and system processes? Message-ID: <dbeb2l$334$1@nwrdmz01.dmz.ncs.ea.ibs-infra.bt.com>    Hi JF,  L I always thought the base prority played a part in how big a boost that idle' process would get when work came along?    Regards Richard Maher   : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message& news:42DAADB3.54415C45@teksavvy.com... > Richard Maher wrote:J > > My product has Network communication servers running, by default, at a baseL > > of 6 (expecting your usual interactive users to run at 4. My Transaction= > > Manager runs at one higher than the communication server.  > K > Don't forget that for short bursts of work, priority doesn't make much of  a I > difference if all your processes do short bursts of work at a time. Low L > priority processes get their priority boosted while waiting for IO so that@ > when IO completes, they have a chance to execute now and then. > L > Where it matters is when you have a process which will sometimes go into aE > long compute process (for instance, encrypting some large in-memory  buffer. I > Giving that process a high priority will essentially jam lower priority  > processes during the process.  > H > If you give users a lower priority, it means that while the background serverK > process is eating CPU cycles by the dozen, users will see severe delay in 4 > echoing of characters in a screen/GUI application. > I > If the background server process is mostlty IO bound, then its priority  won't H > matter much because at each IO, it gives another process a chance. And when) > its IO completes, it is given a chance.    ------------------------------  % Date: Sun, 17 Jul 2005 16:20:50 -0400 ' From: "Main, Kerry" <kerry.main@hp.com> 8 Subject: RE: VMS process priorities and system processesR Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB65134B@tayexc19.americas.cpqcorp.net>   > -----Original Message-----= > From: Richard Maher [mailto:maher_rj@hotspamnotmail.com]=20  > Sent: July 17, 2005 3:03 PM  > To: Info-VAX@Mvb.Saic.Com : > Subject: Re: VMS process priorities and system processes >=20 > Hi Albrecht, >=20A > I think I'm in agreement with you regarding class schedulers=20  > (and Row Cache7 > memory schedulers) and all other forms of bollocks=20  > nanny-state Dr, Zhivago B > "far more equitable comrade" socialist interference. (Oh here=20 > come all thoseB > wankers from the other thread(s)) "The some of the parts is a=20 > shit-load lessG > than the whole!" Kerry would do well to put that above his desk. It's 0 > motivated numerous Rdb engineers for years :-) >=20  H Of course, using the above analogy, the following could also be stated -F "The some of the parts is a shit-load less than the whole - unless oneE of the parts does something stupid in which the sum of what the other  parts can do is zero"!   :-)    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: Sun, 17 Jul 2005 16:31:19 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 8 Subject: Re: VMS process priorities and system processes, Message-ID: <42DAC00F.A90D0604@teksavvy.com>  J As an example of why you need one interactive process with great priority:  M If you are logged in to priority 4 on a slow machine, and some hacker tries a I gazillion FTP user/password combinations rapidly, you will find that your K system grinds to a halt, and from the $ sign, you can't do much because the K system is "so slow". this is because FTP session management is done at 4 or = higher priority, and that include process creation attempts.    I For instance, i would have great difficulty in telneting to my router and I entering the menus to add that particular IP to the list of IPs to block.   M So no matter what you do, make sure that one username/password on your system N is given a priority that is high enough that you can do debugging work from itK should the application processes go nuts in some infinite loop or otherwise R hig system resources to a point where normal priority processes can't act quickly.   ------------------------------  % Date: Sun, 17 Jul 2005 16:49:49 -0500 2 From: David J Dachtera <djesys.nospam@comcast.net>8 Subject: Re: VMS process priorities and system processes+ Message-ID: <42DAD27D.8827A1C3@comcast.net>    Albrecht Schlosser wrote:  > [snip]Q > The highest priority process is a kind of database server that is I/O bound and O > uses little CPU, but should get the CPU ASAP, because it serves (potentially) R > all other processes in the application system. (You may say that's a bad design,> > but it is as it is and can't be changed in the near future). > [snip]  E Remember also that processes which voluntarily wait for I/O receive a G priority boost, I think it's five points above their base priority. So, D attempting to circumvent that might be more harmful than beneficial.  C I read at another point in this thread that you had already lowered G certain interactive priorities below DEFPRI (4, by default, but you can 4 change that (*CAUTION*)). I shuddered at that point.  @ Is the performance of this system such that you are getting this( granular trying to forestall an upgrade?  H I can certainly appreciate what you're explaining here, but I don't know@ as twiddling priorities is going to get you to where need to be.  E Is this based on studies of performance data? If so, what bottlenecks @ were observed/identified? What is the rationale behind trying toG prioritize processes in this way? Some of it seems a bit paradoxical to  me.    --   David J Dachtera dba DJE Systems  http://www.djesys.com/  ) Unofficial OpenVMS Hobbyist Support Page: " http://www.djesys.com/vms/support/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/   " Unofficial OpenVMS-IA32 Home Page: http://www.djesys.com/vms/ia32/    Coming soon:& Unofficial OpenVMS Marketing Home Page   ------------------------------  % Date: Sun, 17 Jul 2005 18:03:57 -0400 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> 8 Subject: Re: VMS process priorities and system processes, Message-ID: <42DAD5BF.18122BC6@teksavvy.com>   David J Dachtera wrote: B > Is the performance of this system such that you are getting this* > granular trying to forestall an upgrade?  L There are some application designs that can benefit from priority twiddling.M But I'd say this should be seen as a temporary solution and the real solution D is to retrcucture the application so that it is no longer a problem.  F You can look at interprocess communications, the use of ASTs so that aL transaction can be accepted by the process while the mainline of the process* is processing another transaction etc etc.   ------------------------------   End of INFO-VAX 2005.397 ************************                                                                                                                                                              p.ײތo%jU`N6WbQAѪU,x,#UZrK!ZE,3VZ磦0MO6Hbtgչ)?՟/NMޢxvHj:u8ހŤ
}<-6Mic
3?UOU
%FS_2le_b

4`9(y{kP2[X i+S$[l'ՍԳKpk 6ᑯ̀/=([e\ٽo1pGo?ĥٞbXyDç(uI}]Y8t1x]yŋJ?ڬ:姞O$<r_Ϊo,ϋ|坌V,	Aqzub e@Eui:<l f5}ӱfa_7}K\U`r(Rp*IA?@*@'-J^%: +g9{J*1^).ݯ7U{|1鰋Vsm@Vdu?bb!v1Lb"%[b]sm1XjBJ9$"G7jY} 6EF$^(	|
jUU9(:W\KPv|n|Ն|U!_]|1a6:#_KW2|ƷBǐ6[j㫤W
uX#HAzCmר4blas0VT5Yj'IfkDuH6&C4UJrt6џ1af=Ct2LVM_#t41ÞQkqVmʘVk=c%6J{MMDҹOqF*{M1H樓B{ƷwJX^Nb@6L;ևQ-6XHxh9N;T+;#c;;ګd"?Gvڄ=|b`{UWaFb.`n̭XlAX
:O+Zu^KS/IVM,MuJ8=t6ݹi[L**%g{y5a<̰O߆.Ƕ!<pFz`.`.K<]):fIpl8o{pn>2I49u&*v:S.J)F"M4]t"M"``y[5Jvk1ڍt!]}b]>L(JG=Ra+~
Sq݉|+goX?9sMǿʱ'>BryuS9B}TϹ;Tކ*)6!^N=Yu5
\m,,^b>俫u_VNʗ,8ױ݅</Z,BYMKN'h,y]1PwY@.g"dO<v:	NMz6ؼ4Mx*/<CPAғG-1> !.skh

҅|зs=c
V-s^8^Gr^;}ԐZ9M73KYL۶K pw?Mahi:,9Y-%@#w~-uY;?)Wz]ȯ/oaŲWB^#EgS
,ķә,xq53hbV~5o~3fmESG.eE~ 9p $:DFM \z
DMCܪ_URZo"SN~"f5~B݋Gq돎wG1xe˱L0db[xX)t(Ƌd@D@N{օ&֒ۓ?665-B19QBN(CV7'483}wߛgbI<@,yfېMȱdA9$XH{5h6:_
>_*ecJˈ/2WN4IINn L) |y+$#Q=bC	ݔc/}kR[]/yQ׷{gXR2S>|Ц-TO*Tz_SeF
ɇ/W-A;u7e2&^s%FjhS˦Хjܒ#M0ٷV*za.m6 49M&Mh2Hf,|Ã$EIX3kE>Sϧ\uO4v\}Tu(fŶPIn>7-m˰@JeΤSר;N)~w(rPmk, _}{"oc^`zȼv@1'g@.󅮇t+'Wx7_\GG>=mlJ9}sz^5?d:![s>floz.2Gʌ[SXXޢ)x,yL^N%/KVf"7
'u:s!p-}oK<π/mphbdO@[߰,kKet/99ՎpGElV
E2Έ~m>4`8=&-v,
jʆ/ekǲlD4:{٣/>?<?)͂˵n=
Jr@r 8Ⱦnro˷/Ϡg̜%<[qYaz_bcQf<8Qw0RJɼC?_-En:7\j;urME~Pw\xHG-TWp˓\vTw[ާ'܉R[B7[ީ/"w!A},ft/\jkTGU/
?e%/\Z|FuOŶ
/Y>g%eVBXGڡv(5ڣi]{*{61fu'b"w(٢rJuxr"oӂw/h"HnoDy\N*צRH)"?yF		A`=б`-FŬ)>4~\8k\x36[G}̿dexҗ_XCA9a gj
JXCaMyhV	~"kQvRа^3B;bm7;8Smymb5-肙6	Z93g~LsmL{ux/<7Kxtף뤖M!4E\!w2dpWP;ܻ|XUj