0 INFO-VAX	Thu, 20 Jan 2005	Volume 2005 : Issue 40      Contents:) Re: An Interview With Mr. OpenVMS Himself ) Re: An Interview With Mr. OpenVMS Himself ) Re: An Interview With Mr. OpenVMS Himself ? Announcement: IBM WebSphere MQ V5.3 for the HP OpenVMS platform # Re: AS 1000 boot failure (battery?)  Re: Error in C++ compiler  Re: Error in C++ compiler  Re: Error in C++ compiler  Re: GnuPG in batch Re: GnuPG in batch' How do you batch a PGP decrypt command?  Re: job anyware?1 Re: Marcello predicts 500% perfomance improvement 1 Re: Marcello predicts 500% perfomance improvement 1 Re: Marcello predicts 500% perfomance improvement 1 Re: Marcello predicts 500% perfomance improvement 1 Re: Marcello predicts 500% perfomance improvement  Re: Multi threaded cookbook ?  Re: Multi threaded cookbook ?  Re: Multi threaded cookbook ? * Re: New Command Language (CL) for OpenVMS?* Re: pthread_create() call returning ENOMEM* Re: pthread_create() call returning ENOMEM0 Reading a VMS IMage backup tape form Windows NT. Re: satellite clusters Re: satellite clusters Using CONVERT/FDL  Re: Using CONVERT/FDL  Re: Using CONVERT/FDL  VMS Docs Re: vms versus solaris Re: vms versus solaris Re: vms versus solaris  F ----------------------------------------------------------------------  # Date: Thu, 20 Jan 2005 16:04:08 GMT & From: John Reagan <john.reagan@hp.com>2 Subject: Re: An Interview With Mr. OpenVMS Himself2 Message-ID: <YDQHd.6238$jO6.4924@news.cpqcorp.net>   Bill Todd wrote: > Kenneth Farmer wrote:  >  >> ShannonKnowsHPC.com: D >> http://www.shannonknowshpc.com/stories.php?story=05/01/18/0435366 >  > H > What a disappointment:  I thought you meant that Terry had snagged an  > interview with Cutler. >  > - bill  F There are many opinions and folklore on how much impact Cutler had on A OpenVMS as compared to people like Dick Hustvedt.  As a point of  I reference, we have a Hustvedt conference room here in the OpenVMS group,  I but we don't have a Cutler conference room.  Not to say that Dave didn't  G have a significant impact on OpenVMS, but he is just a very well known  = character from a much larger cast of lesser known characters.    --   John Reagan / HP Pascal/{A|I}MACRO for OpenVMS Project Leader  Hewlett-Packard Company    ------------------------------  % Date: Thu, 20 Jan 2005 11:44:30 -0500 ( From: Bill Todd <billtodd@metrocast.net>2 Subject: Re: An Interview With Mr. OpenVMS Himself= Message-ID: <06mdnbF7pe92QnLcRVn-vA@metrocastcablevision.com>    John Reagan wrote: > Bill Todd wrote: >  >> Kenneth Farmer wrote: >> >>> ShannonKnowsHPC.com:E >>> http://www.shannonknowshpc.com/stories.php?story=05/01/18/0435366  >> >> >>I >> What a disappointment:  I thought you meant that Terry had snagged an   >> interview with Cutler.  >>	 >> - bill  >  > H > There are many opinions and folklore on how much impact Cutler had on 3 > OpenVMS as compared to people like Dick Hustvedt.   C Those of us who were around at the time don't have to rely on such   things, however.      As a point ofK > reference, we have a Hustvedt conference room here in the OpenVMS group,  - > but we don't have a Cutler conference room.   G That may have as much to do with how people felt after Dick's accident  H as anything else.  Plus the fact that he stayed centrally involved much 1 more than Dave did after VMS V1 was out the door.       Not to say that Dave didn'tI > have a significant impact on OpenVMS, but he is just a very well known  ? > character from a much larger cast of lesser known characters.   G There's a reason why Dave was so well-known, and it's not just because  ! he was such a colorful character.   E It's because he was absolutely central to the development of VMS V1,  F more so than any other single individual.  VMS's structure and system D interface closely followed the designs of the earlier RSX operating C systems he had been central to developing (in some cases rejecting  G experience from other system groups which might actually have made VMS  D better), but if Dave's main contribution had merely been design and G coding (though he was incredibly prolific at the latter) he would have  A just been one of a half-dozen or more truly stellar contributors.   H However, as the leader/manager of the VMS V1 development effort he was, D more than any other individual, the one whose drive made it happen. F Couple that with his other contributions, and there really was no one G else who could be called the father of VMS (in somewhat the same sense  D that people call George Washington the 'father of our country' even > though he shared credit for its birth with many other amazing G individuals and pretty much retired after his term in office was over).    - bill   ------------------------------  # Date: Thu, 20 Jan 2005 17:11:24 GMT * From: "FredK" <fred.nospam@nospam.dec.com>2 Subject: Re: An Interview With Mr. OpenVMS Himself1 Message-ID: <0DRHd.6248$xO6.888@news.cpqcorp.net>    I agree entirely with Bill.   5 "Bill Todd" <billtodd@metrocast.net> wrote in message 7 news:06mdnbF7pe92QnLcRVn-vA@metrocastcablevision.com...  > John Reagan wrote: > > Bill Todd wrote: > >  > >> Kenneth Farmer wrote: > >> > >>> ShannonKnowsHPC.com:G > >>> http://www.shannonknowshpc.com/stories.php?story=05/01/18/0435366  > >> > >> > >>J > >> What a disappointment:  I thought you meant that Terry had snagged an > >> interview with Cutler.  > >> > >> - bill  > >  > > I > > There are many opinions and folklore on how much impact Cutler had on 5 > > OpenVMS as compared to people like Dick Hustvedt.  > D > Those of us who were around at the time don't have to rely on such > things, however. >  >    As a point ofL > > reference, we have a Hustvedt conference room here in the OpenVMS group,/ > > but we don't have a Cutler conference room.  > H > That may have as much to do with how people felt after Dick's accidentI > as anything else.  Plus the fact that he stayed centrally involved much 3 > more than Dave did after VMS V1 was out the door.  >   >    Not to say that Dave didn'tJ > > have a significant impact on OpenVMS, but he is just a very well knownA > > character from a much larger cast of lesser known characters.  > H > There's a reason why Dave was so well-known, and it's not just because# > he was such a colorful character.  > F > It's because he was absolutely central to the development of VMS V1,G > more so than any other single individual.  VMS's structure and system E > interface closely followed the designs of the earlier RSX operating D > systems he had been central to developing (in some cases rejectingH > experience from other system groups which might actually have made VMSE > better), but if Dave's main contribution had merely been design and H > coding (though he was incredibly prolific at the latter) he would haveC > just been one of a half-dozen or more truly stellar contributors.  > I > However, as the leader/manager of the VMS V1 development effort he was, E > more than any other individual, the one whose drive made it happen. G > Couple that with his other contributions, and there really was no one H > else who could be called the father of VMS (in somewhat the same senseE > that people call George Washington the 'father of our country' even ? > though he shared credit for its birth with many other amazing I > individuals and pretty much retired after his term in office was over).  >  > - bill   ------------------------------  % Date: Thu, 20 Jan 2005 18:41:31 +0100 0 From: Keith Cayemberg <keith.cayemberg@arcor.de>H Subject: Announcement: IBM WebSphere MQ V5.3 for the HP OpenVMS platformB Message-ID: <41efed4c$0$17618$9b4e6d93@newsread4.arcor-online.net>  H The IBM UK WebSphere MQ Development Team lead by Nick Allen has recentlyH announced the (electronic) GA** for WebSphere MQ V5.3 on the HP OpenVMS 	 platform.   # Congratulations to the entire Team!   . The announcement is available on the Web at...9 http://www-306.ibm.com/software/integration/wmq/v53/ovms/    and more here...n http://www-306.ibm.com/common/ssi/fcgi-bin/ssialias?infotype=an&subtype=ca&appname=GPA&htmlfid=897/ENUS205-003  + WebSphere MQ for HP OpenVMS, V5.3 includes:   @      * Functional parity with other WebSphere MQ, V5.3 platforms=      * Availability on the HP Alpha and Itanium (32-bit mode)         architecturesE      * Full enablement of the Java Message Service (JMS) application G        programming interface: embedded classes for Java and JMS support .        with improved performance and packaging#      * A Publish/Subscribe function D      * Performance, scalability, and usability improvements over the"        previous MQSeries offering:      * Improved MQ workload balancing (through clustering)H      * Improved national language support (for example, codepage GB18030(        for use in the Republic of China)F      * SSL security authentication facilities via HP's SSL offering (a        no-charge HP delivery)    ** = general availability    Cheers!    Keith Cayemberg ) IBM Business Services - Hannover, Germany    ------------------------------  + Date: Thu, 20 Jan 2005 17:02:52 +0000 (UTC) - From: klewis@OMEGA.MITRE.ORG (Keith A. Lewis) , Subject: Re: AS 1000 boot failure (battery?). Message-ID: <csoo7s$71v$1@newslocal.mitre.org>   "Ransom Fitch" <rlf_vms@earthlink.net> writes in article <RtwHd.131$r27.110@newsread1.news.pas.earthlink.net> dated Wed, 19 Jan 2005 17:08:01 GMT:7 >When I attempt a cold boot on a AS 1000 it fails with:  > - >EISA Data in non-volatile storage is corrupt  >EISA Configuration Error..... >  > I >After running ECU I can boot normally (requires setting date and time).    >Reboots (warm boots) work fine. > N >I suspect that this is due to a dead battery, and if so may have been caused J >by a KVM switch. The only thing that I find that looks like a battery is K >~0.5(w) x 0.375(h) x 2.0(l) inches that's marked "lithium battery...." on   >the main board(?).  >  >Any suggestions on:" >1) determining if battery is dead  L Measure the battery voltage while the battery is "in use", i.e. the power isG off.  If it's even .1 volts lower than the spec printed on it, replace.   % >2) finding a replacement (if needed)   N >3) installing a replacement (if needed).  I believe that this is (multi-)pin 4 >socket but may require some tool(s) for extraction.  K I purchased a replacement for my dead PWS battery at a camera counter.  But A it wasn't multi-pin, it just went into a clip on the motherboard.   0 --Keith Lewis              klewis {at} mitre.org> The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Thu, 20 Jan 2005 12:00:39 +0100 9 From: "Robert Trawiski" <robert.trawinski@softax.com.pl> " Subject: Re: Error in C++ compiler/ Message-ID: <cso328$sh6$1@atlantis.news.tpi.pl>    Thanks for your help.   M I've got Software Product Library  (December 2004). And ... there is version  M 6.5-004 on disc. Isn't it funny? We pay to have most up-to-date software and   we get something else.  H Anyway. Is it possible to dowload CXX 6.5-039 or 6.5-048  from internet.   Robert     ------------------------------  % Date: Thu, 20 Jan 2005 14:01:14 +0100 9 From: "Robert Trawiski" <robert.trawinski@softax.com.pl> " Subject: Re: Error in C++ compiler/ Message-ID: <csoa40$488$1@atlantis.news.tpi.pl>   I Uytkownik "Robert Trawiski" <robert.trawinski@softax.com.pl> napisa w  4 wiadomoci news:cso328$sh6$1@atlantis.news.tpi.pl... > J > Anyway. Is it possible to dowload CXX 6.5-039 or 6.5-048  from internet. >    OK. I've found all I need on:   4 ftp://ftp.compaq.com/pub/products/C-CXX/openvms/cxx/   Robert     ------------------------------  % Date: Thu, 20 Jan 2005 14:24:24 +0100 2 From: Ulrich Bellgardt <uli_bellgardt@hotmail.com>" Subject: Re: Error in C++ compiler7 Message-ID: <41efb10a$0$25917$9b622d9e@news.freenet.de>    Robert Trawiski wrote:  > Thanks for your help.  > O > I've got Software Product Library  (December 2004). And ... there is version  O > 6.5-004 on disc. Isn't it funny? We pay to have most up-to-date software and   > we get something else. > J > Anyway. Is it possible to dowload CXX 6.5-039 or 6.5-048  from internet. >    Yes. You find it on   4 ftp://ftp.compaq.com/pub/products/c-cxx/openvms/cxx/  E The patch directory on HP's site (ftp://ftp.itrc.hp.com) lists only a  patch for CXX V6.2.    Regards, -Uli Bellgardt   ------------------------------  % Date: Thu, 20 Jan 2005 09:06:32 +0000 % From: David Gray <police@spamcop.net>  Subject: Re: GnuPG in batch 8 Message-ID: <uqsuu0ddlmuaj01rdn6b6lqsg3k7thqo9k@4ax.com>   Hi,   C Cheers for that.  I've got it working just fine using the following 
 method...   @ $ pipe type password.txt | gpg --yes --passphrase-fd 0  --output( output.txt  --batch --decrypt input.asc    Dave.        E On Wed, 19 Jan 2005 22:01:54 -0600, "Tom M" <kryios@spam.comcast.net>  wrote:   > 3 >"David Gray" <police@spamcop.net> wrote in message 3 >news:ccssu09lrehgkq3a6rik3nvdluhblah07p@4ax.com... 
 >> Hi all, >> >> OpenVMS 7.3-2 >> GnuPG  v1.2.3 >>G >> Has anyone got an example of how to use the PASSPHRASE-FD command of F >> GnuPG when running in batch?  I've searched the net but cannot find& >> any examples of it's use under VMS. >>E >> I'm currently writing the GPG command into a temporay command file = >> with the passphrase on the second line but wondered if the 3 >> PASSPHRASE-FD would provide a 'neater' solution.  >> >> >> Thanks in advance >> Dave. > 5 >You can embed the passphrase in the command file as:  > 6 >$gpg --batch --passphrase-fd 0 --decryptfile 'pgpfile >passphrase  > / >Don't know if thats what you were looking for.  >  >    ------------------------------    Date: 20 Jan 2005 07:58:56 -0800 From: bob@instantwhip.com  Subject: Re: GnuPG in batch C Message-ID: <1106236736.672288.189000@f14g2000cwb.googlegroups.com>    Tom M wrote:4 > "David Gray" <police@spamcop.net> wrote in message4 > news:ccssu09lrehgkq3a6rik3nvdluhblah07p@4ax.com... > > Hi all,  > >  > > OpenVMS 7.3-2  > > GnuPG  v1.2.3  > > E > > Has anyone got an example of how to use the PASSPHRASE-FD command  ofG > > GnuPG when running in batch?  I've searched the net but cannot find ' > > any examples of it's use under VMS.  > > F > > I'm currently writing the GPG command into a temporay command file> > > with the passphrase on the second line but wondered if the4 > > PASSPHRASE-FD would provide a 'neater' solution. > >  > >  > > Thanks in advance 	 > > Dave.  > 6 > You can embed the passphrase in the command file as: > 7 > $gpg --batch --passphrase-fd 0 --decryptfile 'pgpfile  > passphrase > 0 > Don't know if thats what you were looking for.8 how about for PGP?  How can you automate a decrypt then?   ------------------------------    Date: 20 Jan 2005 08:23:15 -0800 From: bob@instantwhip.com 0 Subject: How do you batch a PGP decrypt command?C Message-ID: <1106238195.109780.199980@c13g2000cwb.googlegroups.com>   & million dollar question ... any ideas?- pipe fails ... redirect sys$input fails ... ?    ------------------------------    Date: 20 Jan 2005 06:44:23 -0800  From: "Barry" <dysert@gmail.com> Subject: Re: job anyware? C Message-ID: <1106232263.451954.204660@f14g2000cwb.googlegroups.com>   @ You could also get an account at decuserve.org and check out the EMPLOYMENT Notes conference.   ------------------------------  % Date: Thu, 20 Jan 2005 08:40:40 +0100 ( From: "Rudolf Wingert" <win@fom.fgan.de>: Subject: Re: Marcello predicts 500% perfomance improvement3 Message-ID: <002101c4fec3$56e71880$994614ac@wat153>    Hello,  < The performance may be good, but in our institute we do needH workstations with the compute and I/O power of an server. Is it possible5 to use a cheap (F&L) Integrity Server as workstation?    TIA and best regards R. Wingert    ------------------------------  % Date: Thu, 20 Jan 2005 02:35:51 -0500 ( From: Bill Todd <billtodd@metrocast.net>: Subject: Re: Marcello predicts 500% perfomance improvement= Message-ID: <YpmdnfH-Re_JwnLcRVn-uw@metrocastcablevision.com>    Dave Froble wrote:   ...   J > I always like it when some know-it-all predicts what's going to happen, K > based upon their wishes.  Rob sure called it right on HAMMER, huh?  Like  G > they say in football, when upsets happen.  "That's why they play the  , > game."  Well Rob, the game isn't over yet.  I Not worth the effort, Dave.  As we both know, Rob used to appear to be a  H moderately intelligent human being, but got replaced at the time of the F Alphacide by some kind of Usenet bot with an unswerving agenda rather 6 than anything resembling objective critical faculties.  I You can kind of understand, and even sympathize a bit with, those people  I who act that way because their jobs are on the line.  With Rob, it seems  I more to be an article of faith which must be tirelessly prosecuted until   the unfaithful are converted.   H Opening the door when such people ring the bell implicitly assumes that I they're playing by the same logical rules the rest of us try to use, and  C hence deserve the courtesy of an audience.  Bad premise, incorrect   conclusion.    - bill   ------------------------------  % Date: Thu, 20 Jan 2005 03:24:44 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> : Subject: Re: Marcello predicts 500% perfomance improvement, Message-ID: <41EF6AB9.FE87AC57@teksavvy.com>   Rudolf Wingert wrote: > > The performance may be good, but in our institute we do needJ > workstations with the compute and I/O power of an server. Is it possible7 > to use a cheap (F&L) Integrity Server as workstation?   D Despite the lack of mention on HP's web site, the VMS engineers haveF repeatedly said that IA64 small servers would continue to be availableF with workstation configurations. The one area where they cannot commit is support of sound card.   A Note that the HP roadmap for VMS makes no mention of graphics and  DECwindows support or upgrades.    ------------------------------  # Date: Thu, 20 Jan 2005 16:06:56 GMT * From: "FredK" <fred.nospam@nospam.dec.com>: Subject: Re: Marcello predicts 500% perfomance improvement2 Message-ID: <AGQHd.6239$5C6.1589@news.cpqcorp.net>  F Yes.  The rx2600 and rx1600 have pedestal mounts that can be purchasedH instead of rackmounting them.  They both have built-in 2D graphics usingC a ATI Radeon 7000 chip (either as part of the management option, or ! using a USB/Graphics combo card).   H VMS is finishing the process of also making the Radeon 7500 available on1 the 1-4 socket servers for 2D, 3D and multi-head.     3 "Rudolf Wingert" <win@fom.fgan.de> wrote in message - news:002101c4fec3$56e71880$994614ac@wat153...  > Hello, > > > The performance may be good, but in our institute we do needJ > workstations with the compute and I/O power of an server. Is it possible7 > to use a cheap (F&L) Integrity Server as workstation?  > ! > TIA and best regards R. Wingert  >    ------------------------------  % Date: Thu, 20 Jan 2005 11:34:31 -0500 # From: "John Smith" <a@nonymous.com> : Subject: Re: Marcello predicts 500% perfomance improvement, Message-ID: <ddqdncnQqYMHQHLcRVn-3A@igs.net>   FredK wrote:H > Yes.  The rx2600 and rx1600 have pedestal mounts that can be purchasedD > instead of rackmounting them.  They both have built-in 2D graphics@ > using a ATI Radeon 7000 chip (either as part of the management. > option, or using a USB/Graphics combo card). > G > VMS is finishing the process of also making the Radeon 7500 available 6 > on the 1-4 socket servers for 2D, 3D and multi-head.    K How about the 9000 or better Radeon's with the TV tuners? Then we could use  VMS as a PVR.  :-)   ------------------------------    Date: 20 Jan 2005 01:21:52 -0800 From: martinkirby12@yahoo.co.uk & Subject: Re: Multi threaded cookbook ?C Message-ID: <1106212912.529380.182700@c13g2000cwb.googlegroups.com>   A > Also, in terms of X windows, there are a few routines which are  usableE > from within an AST. (for instance, inserting an event). Would those : > automatically be usable from a different thread as well? > % Depends on the version of DECwindows.   @ With DECwindows versions before V1.3 the (unwritten?) rules are:, 1.      You can do anything from within AST.E 2.      If you use multiple threads you must implement a mechanism to ? make sure there are no concurrent calls from different threads.   A In practice, #2 is best done as "all calls are made from the main C thread" since all (user-mode) ASTs are executed on the main thread.   B If you mix threads and ASTs you also need to be aware of the C RTLE re-entrancy condition. In practice, I have never had problems with it F set to the default of TOLERANT, but I don't know if this is just luck.  F With DECwindows, from V1.3, you can either follow the earlier rules or2 enable the X Windows thread support. In that case:8 1.      The only function that can be called from AST is XtNoticeSignal. E 2.      The HP extensions XSelectAsyncInput and XSelectAsyncEvent are  not allowed.D 3.      Multiple threads can use X, X Extensions, Xt and ICE and theE code handles the locking for you. Motif and above are not thread-safe = and require user locking - using the XLockDisplay and similar 
 functions.F 4.      It is best to have upcalls enabled, if not enabled and you useF thread priorities there can be some problems with priority inversions.  B > If a thread does a SYS$DCLAST, will the AST be queued to execute under a ' > single AST queue in the main thread ?   G All user-mode ASTs are executed on the main thread. Inner-mode ASTs can " be executed on any kernel thread.    Martin Kirby   ------------------------------  % Date: Thu, 20 Jan 2005 05:37:27 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com> & Subject: Re: Multi threaded cookbook ?, Message-ID: <41EF89CD.37FBBAA2@teksavvy.com>    martinkirby12@yahoo.co.uk wrote:B > With DECwindows versions before V1.3 the (unwritten?) rules are:. > 1.      You can do anything from within AST.G > 2.      If you use multiple threads you must implement a mechanism to A > make sure there are no concurrent calls from different threads.      Are you 100% sure ?   F I was told that only a limited number of routines could be called fromF AST, since many cannot execute while another X/motif routine is in theF process of executing. (for instance, if the mainline is in the processF of XtManage_child a complex widget and midway is interrupted by an AST3 which also wants to XTmanage_child another widget).   A From a Motif point of view (or any other layered product for that C matter), is there a difference between code that executes in an AST H versus code that executes in a separate thread when the mainline is also3 running code that calls into that layered product ?    ------------------------------    Date: 20 Jan 2005 03:02:08 -0800 From: martinkirby12@yahoo.co.uk & Subject: Re: Multi threaded cookbook ?C Message-ID: <1106218928.742104.225920@c13g2000cwb.googlegroups.com>    JF Mezei wrote: " > martinkirby12@yahoo.co.uk wrote:D > > With DECwindows versions before V1.3 the (unwritten?) rules are:0 > > 1.      You can do anything from within AST.F > > 2.      If you use multiple threads you must implement a mechanism toC > > make sure there are no concurrent calls from different threads.  >  >  > Are you 100% sure ?  >   D I misunderstood/misspoke. You are supposed to be able to do anything> from within user-mode AST - as long as all your calls are from user-mode AST.  @ The X library read routine is set-up to work correctly if called; re-entrantly and to enable ASTs if it blocks. Therefore, my F understanding is that you can do most (?all) things from base code andF user-mode AST if you disable ASTs across each base mode call. However,* I don't know what is officially supported.  C > I was told that only a limited number of routines could be called  fromD > AST, since many cannot execute while another X/motif routine is in the @ > process of executing. (for instance, if the mainline is in the process D > of XtManage_child a complex widget and midway is interrupted by an AST 5 > which also wants to XTmanage_child another widget).  > C > From a Motif point of view (or any other layered product for that E > matter), is there a difference between code that executes in an AST E > versus code that executes in a separate thread when the mainline is  also5 > running code that calls into that layered product ?   F Lots of difference. The most obvious is that the code can disable ASTsA over critical regions to avoid re-entrancy but there is no way to F prevent the two threads swapping at any point. More subtle is that ASTF can interrupt the base code but the base code cannot interrupt the ASTA so some asymmetric mechanisms for consistency can be implemented.    ------------------------------  % Date: Thu, 20 Jan 2005 08:22:02 +0100 ( From: "Rudolf Wingert" <win@fom.fgan.de>3 Subject: Re: New Command Language (CL) for OpenVMS? 3 Message-ID: <002001c4fec0$bcec5f80$994614ac@wat153>    Hello   D Many thanks to all the answers. I did get my information without the4 headerline "this is a joke". Sorry for my questions.   Best regards R. Wingert    ------------------------------   Date: 20 Jan 2005 07:04:39 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow>3 Subject: Re: pthread_create() call returning ENOMEM ? Message-ID: <DTiotGxQ0bj6-pn2-8sdGJlnKs13C@dave2_os2.home.ours>   D On Thu, 20 Jan 2005 05:17:25 UTC, Dave Froble <davef@tsoft-inc.com>  wrote:   > Kannan wrote:  > K > > Could you give some pointers as to how to narrow down the source of the : > > leak. What can I look for in  the SDA for the process? > >  > >  > H > I'd agree with David, the first place to look is for some memory leak. > O > It's hard to give some help when we don't know the application and what it's  Q > doing.  I'd start by determining whether a thread will allocate any resources,  G > such as memory.  If so, then attempting to trace both allocation and   > deallocation of resources. > M > I've never had to do this.  Hopefully someone who has will offer some tips.  >  > Dave >   F At a guess when the threads exit they don't give up all their memory. F I have similar behaviour and I'm still confused how to do it properly.D i.e. how do join, delete etc play together. I have to go back to it  soon.    --   Cheers - Dave W.   ------------------------------    Date: 20 Jan 2005 01:29:52 -0800 From: martinkirby12@yahoo.co.uk 3 Subject: Re: pthread_create() call returning ENOMEM C Message-ID: <1106213392.848520.322680@f14g2000cwb.googlegroups.com>   2 As others have said, it is probably a memory leak.  G I suggest using the heap analyzer. It sounds like you should be able to D get a clear enough leak after about an hour  while running under theF analyzer. As long as you can run the main image in debug you should be? able to get useful callstacks to the leak without needing debug F versions of any libraries used. Does assume you have DECwindows client+ and a workstation or PC you can display to.    ------------------------------    Date: 20 Jan 2005 10:18:39 -0800! From: gribbinjf@npt.nuwc.navy.mil 9 Subject: Reading a VMS IMage backup tape form Windows NT. B Message-ID: <1106245119.071992.62260@c13g2000cwb.googlegroups.com>   Hi All,   9 PLease point me in the right direction.  I am looking for ; utilities to read a VMS BACKUP/IMAGE tape from a Windows NT  server.    ------------------------------  % Date: Thu, 20 Jan 2005 03:35:43 -0500 - From: JF Mezei <jfmezei.spamnot@teksavvy.com>  Subject: Re: satellite clusters , Message-ID: <41EF6D4C.D88CD394@teksavvy.com>   Dave Froble wrote:M > >  operations with full cluster functionality at distances up to 800 Km and + > >  up to 90,000 Km for disaster recovery"   R > That's too far out for a geo-sync orbit, which is what I'd choose if I was going* > off-planet for disaster tollerance.  :-)  H geo is 35,796km. So if you package an Alpha with solar arrays and properG cooling and its own telecom and data storage, you could cluster with it A since it would be well within this new extended limit fo 90,000km " according to the previous post :-)  D Oh, you need to protect it from solar radiation, extreme heat, solar% flares and possible micrometeor hits.   H And you might want to design disk drives that also double as CMGs (usingD gyroscopic properties of spinning platters to orient the satellite).  G The advantage is of security: no human would ever be able to access its C OPA0: and thus have the ability to bypass its security :-) Hardware  upgrades would be quite hard.    ------------------------------  % Date: Thu, 20 Jan 2005 06:30:38 -0500 ( From: Bill Todd <billtodd@metrocast.net> Subject: Re: satellite clusters = Message-ID: <9_ydneKjTO3HC3LcRVn-og@metrocastcablevision.com>    Dave Froble wrote: > reb wrote: >  >>9 >>  From the VMS info linked to from yesterday's Webcast:  >>K >> "Drive continuous operations for your local, national, and international L >>  operations with full cluster functionality at distances up to 800 Km and* >>  up to 90,000 Km for disaster recovery" >>K >> Looks like someone has been doing cluster comms bouncing off a satellite ? >> or two.  Does the 'same time zone' requirement still apply ?  >> >> >> reb >  > I > That's too far out for a geo-sync orbit, which is what I'd choose if I  4 > was going off-planet for disaster tollerance.  :-) >   D It's not too far off what the distance via geosynchronous satellite D from, say, somewhere on the Pacific rim to somewhere in the U.S. or G Europe would be.  And considerably short of the distance it would take  D to connect opposite sides of the globe (which would require a relay  through two such satellites).   H Kind of a long way for synchronous replication, though (which is likely ; why they differentiate between conventional clustering and   disaster-recovery).    - bill   ------------------------------    Date: 20 Jan 2005 02:15:14 -0800* From: renedeliege@yahoo.com (Rene Deliege) Subject: Using CONVERT/FDL< Message-ID: <adbbdfee.0501200215.8f133bb@posting.google.com>  / I use CONVERT to copy files from a remote node. 1 For files with a large record size CONVERT fails:     >convert NODE::TEMP.TMP TEMP.TMP: %CONV-F-READERR, error reading NODE::$USER:[xyz]TEMP.TMP;2= -RMS-E-NETBTS, network buffer too small for 20243 byte record   > When I first give SET RMS/NETWORK_BLOCK_COUNT=64 it copies ok.  # Thus I want to achieve the same by:   . >convert/fdl=sys$input NODE::TEMP.TMP TEMP.TMP NETWORK          BLOCK_COUNT     64 Ctrl/Z: %CONV-F-READERR, error reading NODE::$USER:[xyz]TEMP.TMP;2= -RMS-E-NETBTS, network buffer too small for 20243 byte record    Why doesn't this work??   F It seems it does not use the parameter at all, even if I enter invalid values it does not complain.   ------------------------------  % Date: Thu, 20 Jan 2005 12:12:20 +0000 - From: John Laird <nospam@laird-towers.org.uk>  Subject: Re: Using CONVERT/FDL8 Message-ID: <ns5vu0hthdb09fuqedtpj6utdh9fo430ja@4ax.com>  J On 20 Jan 2005 02:15:14 -0800, renedeliege@yahoo.com (Rene Deliege) wrote:  0 >I use CONVERT to copy files from a remote node.2 >For files with a large record size CONVERT fails: > ! >>convert NODE::TEMP.TMP TEMP.TMP ; >%CONV-F-READERR, error reading NODE::$USER:[xyz]TEMP.TMP;2 > >-RMS-E-NETBTS, network buffer too small for 20243 byte record > ? >When I first give SET RMS/NETWORK_BLOCK_COUNT=64 it copies ok.  > $ >Thus I want to achieve the same by: > / >>convert/fdl=sys$input NODE::TEMP.TMP TEMP.TMP  >NETWORK >        BLOCK_COUNT     64  >Ctrl/Z ; >%CONV-F-READERR, error reading NODE::$USER:[xyz]TEMP.TMP;2 > >-RMS-E-NETBTS, network buffer too small for 20243 byte record >  >Why doesn't this work??  L If you were converting *to* a remote file, it would work.  FDL describes theI output file structure and optionally how to access it when it is created.   J As the network block count is not a file property, you have to change yourD process (or system) quota instead, as you have found.  Issuing a setK rms/network/system command will persist until the next reboot, or you could F use sysgen and make the change permanent (the set rms command actuallyK modifies the parameter in active use, which is dynamic).  I'll let you find 5 the associated parameter - exercise is good for you !    --  5 Despite the high cost of living, it remains popular.     Mail john rather than nospam...    ------------------------------  % Date: Thu, 20 Jan 2005 14:57:26 +0100 0 From: Keith Cayemberg <keith.cayemberg@arcor.de> Subject: Re: Using CONVERT/FDLB Message-ID: <41efb8c7$0$17605$9b4e6d93@newsread4.arcor-online.net>   John Laird wrote:wL > On 20 Jan 2005 02:15:14 -0800, renedeliege@yahoo.com (Rene Deliege) wrote: >  > 1 >>I use CONVERT to copy files from a remote node.S3 >>For files with a large record size CONVERT fails:0 >> >>" >>>convert NODE::TEMP.TMP TEMP.TMP >>< >>%CONV-F-READERR, error reading NODE::$USER:[xyz]TEMP.TMP;2? >>-RMS-E-NETBTS, network buffer too small for 20243 byte recordd >>@ >>When I first give SET RMS/NETWORK_BLOCK_COUNT=64 it copies ok. >>% >>Thus I want to achieve the same by:% >> >>0 >>>convert/fdl=sys$input NODE::TEMP.TMP TEMP.TMP >>	 >>NETWORKn >>       BLOCK_COUNT     64% >>Ctrl/Z< >>%CONV-F-READERR, error reading NODE::$USER:[xyz]TEMP.TMP;2? >>-RMS-E-NETBTS, network buffer too small for 20243 byte record  >> >>Why doesn't this work??  >  > N > If you were converting *to* a remote file, it would work.  FDL describes theK > output file structure and optionally how to access it when it is created.p > L > As the network block count is not a file property, you have to change yourF > process (or system) quota instead, as you have found.  Issuing a setM > rms/network/system command will persist until the next reboot, or you could-H > use sysgen and make the change permanent (the set rms command actuallyM > modifies the parameter in active use, which is dynamic).  I'll let you find 7 > the associated parameter - exercise is good for you !3 >   G Although CONVERT may be already sufficient for you, you might find the  
 command...  % 	$ EXCHANGE/NETWORK/TRANSFER_MODE/FDLp  C to be a more robust solution, depending on the remote volume, file nG system and file characteristics you are dealing with. CONVERT and COPY eH really expects to be working with ODS and RMS conform devices, which of G course can also be emulated by the network or middleware you are using.   F EXCHANGE was designed to deal with foreign (to VMS) file systems, and C I've been pleased with it's flexibility to even deal with new file  = systems not yet existing when the exchange tool was designed.e     Cheers!r   Keith Cayembergr   ------------------------------  % Date: Thu, 20 Jan 2005 11:08:14 -0500a$ From: "Noce, Guy" <gnoce@towson.edu> Subject: VMS DocseH Message-ID: <957B7AB3AFF9EB4182E46BD96F48B77B01A881B8@helium.towson.edu>  G I have several complete brand spanking new sets of VMS documentation in:F boxes in my office.  They're pretty much v7.3 and above (we're sitting; on 7.2-1 until the machines die, it seems.)  I have all thei! documentation I need, for now.=20e  G If they have any value to anyone, I'll ship them for a nominal fee pluss	 shipping.s  * Please reply to me here:  gnoce@towson.edu   --Guyt  E Former VMS Systems Admin, now "Business Server (NT Goatboy) Engineer"  Towson Universityb Towson, MD  21252r  F "University politics is so vicious precisely because the stakes are so small."  --Henry Kissinger   ------------------------------   Date: 20 Jan 2005 07:04:37 GMT2 From: "Dave Weatherall" <djw-nothere@nospam.nohow> Subject: Re: vms versus solarisH? Message-ID: <DTiotGxQ0bj6-pn2-ME4ZYQWgLyxw@dave2_os2.home.ours>n  ? On Wed, 19 Jan 2005 16:17:06 UTC, Kilgallen@SpamCop.net (Larry i Kilgallen) wrote:a  v > In article <DTiotGxQ0bj6-pn2-Nk38g9oIUdJI@dave2_os2.home.ours>, "Dave Weatherall" <djw-nothere@nospam.nohow> writes: > G > >        am I parsing this right when I take 'decisive plus'  in the iC > > context of 'bringing new customers to the VMS platform'.  As a eG > > record-oriented user/programmer, it _is_ a decisive plus but then, n? > > perhaps, I fall into the 'locked in' category (mentally or s > > philosophically).r > G > The languages Cobol, Pascal and Ada all have the notion of record-IO, C > and probably PL/I does also.  So I would say you are locked in by B > programming paradigms that have been found to actually be needed > in real applications.e  @ Well they're mainly written in Fortran and Macro-32 (very small 2 amounts of Pascal and C) but it's a fair analysis.  F We have tended to concentrate on RMS in its file i/o guise (record v. C byte). There is also the other system  aspect (search, parse, etc)  D that need to be remembered. Sticky defaults, input/output file name F parsing etc. I have _no_ idea how well this aspect of RMS is reflectedE in other OS's, apart from DOS where I can remember having to roll my  = own. That leads us into Keith's 3D spider"s web difficulties.s   --   Cheers - Dave W.   ------------------------------  % Date: Thu, 20 Jan 2005 03:13:44 -0500G4 From: "Mark Buda" <buda@SPAMTripleBreakProducts.com> Subject: Re: vms versus solarist1 Message-ID: <0bOdnenpEIFR-nLcRVn-jw@adelphia.com>m  6 "Bill Todd" <billtodd@metrocast.net> wrote in message 7 news:0LudneELkKGboXPcRVn-iw@metrocastcablevision.com...0F > The fact that RMS is bundled and supported as part of the OS may be K > convenient and reassuring, but is hardly of *major* significance to most sM > users (or applications, though not having a third-party dependency in this s# > area is part of the convenience).F  M I will kindly disagree.  It is ONE of the big reasons that one can have such  H seemless clustering.  A file on one machine can be accessed with record J locking to another cluster member without having to done a lot of te work 	 yourself.   E If you think about, VMS heavily uses RMS and record locking and many g applications do also.w   mark   ------------------------------  % Date: Thu, 20 Jan 2005 06:23:00 -0500o( From: Bill Todd <billtodd@metrocast.net> Subject: Re: vms versus solarise= Message-ID: <XamdnSMc8tcNCXLcRVn-tw@metrocastcablevision.com>    Mark Buda wrote:8 > "Bill Todd" <billtodd@metrocast.net> wrote in message 9 > news:0LudneELkKGboXPcRVn-iw@metrocastcablevision.com...n > F >>The fact that RMS is bundled and supported as part of the OS may be K >>convenient and reassuring, but is hardly of *major* significance to most  M >>users (or applications, though not having a third-party dependency in this d# >>area is part of the convenience).  >  > O > I will kindly disagree.  It is ONE of the big reasons that one can have such uJ > seemless clustering.  A file on one machine can be accessed with record L > locking to another cluster member without having to done a lot of te work  > yourself.   E This is not *all* that different from the ability to use distributed yI byte-range locking plus NFS to accomplish the same goals.  Leaving aside iI those Unices which have developed real cluster file systems of their own gG (at least AIX [GPFS], Solaris [which already had at least one prior to pH its new ZFS], IRIX [CXFS], Tru64 of course, and arguably Linux [GFS and L others]) plus those using one from Veritas or some lesser-known third-party.   > G > If you think about, VMS heavily uses RMS and record locking and many e > applications do also.b  G The 'records vs. bytes' discussion has already occurred.  Leaving that ,E aside, locking, including distributed locking, simply isn't anything e like a VMS exclusive any more.   - bill   ------------------------------   End of INFO-VAX 2005.040 ************************