1 INFO-VAX	Wed, 17 Sep 2003	Volume 2003 : Issue 515       Contents: Re: " USP for OpenVMS "  Re: " USP for OpenVMS " + Re: "HP breakup on the way" - Merrill Lynch + Re: "HP breakup on the way" - Merrill Lynch + Re: "HP breakup on the way" - Merrill Lynch + Re: "HP breakup on the way" - Merrill Lynch  RE: Block size in show dev? # Re: Booting hobbyist CD on PWS433au ! Choosing between ASTs and Threads % Re: Choosing between ASTs and Threads % Re: Choosing between ASTs and Threads % Re: Choosing between ASTs and Threads % Re: Choosing between ASTs and Threads # CSwing and OpenVMS v7.3-1, Again... , Re: FTP on AlphaServer running OpenVMS 7.2-1) How to use shareable image to share data? - Re: How to use shareable image to share data? - Re: How to use shareable image to share data? - HP extends lead in open storage area networks 1 Re: HP extends lead in open storage area networks  IA64 to be dual core by 2005  Re: IA64 to be dual core by 2005  Re: IA64 to be dual core by 2005  Re: IA64 to be dual core by 2005  Re: IA64 to be dual core by 20055 init/queue/start/processor=ucx$telnetsym/on=... hangs  KZPBA Question Re: KZPBA Question" Re: Looking for TK50s in the UK... Re: Montagar website?  Re: OpenVMS Security- Re: OT: Verisign messes with DSN root servers - Re: OT: Verisign messes with DSN root servers - Re: OT: Verisign messes with DSN root servers - Re: OT: Verisign messes with DSN root servers - Re: OT: Verisign messes with DSN root servers - Re: OT: Verisign messes with DSN root servers - Re: OT: Verisign messes with DSN root servers  Re: personal name in TO: ? Re: personal name in TO: ? Re: Reimport VMS mail problems Re: Reimport VMS mail problems# Remove satellite node from cluster? ' Re: Remove satellite node from cluster? < Re: Strange DCL Problem:  DCL Can't Find My GOSUB Labels....' Re: Summary: KVMs, LK461 and DECwindows ' Re: Summary: KVMs, LK461 and DECwindows  Re: VMS and MP3s  F ----------------------------------------------------------------------  + Date: Tue, 16 Sep 2003 12:07:14 -0700 (PDT) . From: Fabio Cardoso <fabiopenvms@yahoo.com.br>  Subject: Re: " USP for OpenVMS "@ Message-ID: <20030916190714.93466.qmail@web20203.mail.yahoo.com>   Richard   4 Is there any other way to encode this data in OVMS? 6 I tought I needed a product ! I didnt find any utility7 under OVMS ! I matched references about this product !  / But HP is discontinuing it ! So I want to know  ) if there is an alternative if they decide  to stop forever !    Regards    FC  - --- Richard Brodie <R.Brodie@rl.ac.uk> wrote:  > = > "Fabio Cardoso" <fabiopenvms@yahoo.com.br> wrote in message < > news:20030916133043.54749.qmail@web20202.mail.yahoo.com... > C > > We need to generate some data under OVMS using the UTF-8 coding  > > for XML treatment. >  > Fine.  > B > > This product is being discontinued by HP ! Another product can: > > subsititute it?  From HP ? Attunity ? Other reseller ? > ? > I don't understand. Why do you need to buy something in to do  > a job as trivial as encoding?  >  >      =====  ========================== Fbio dos Santos Cardoso OpenVMS System Manager Rio de Janeiro - Brazil  fabiopenvms@yahoo.com.br ==========================  " __________________________________ Do you Yahoo!?? Yahoo! SiteBuilder - Free, easy-to-use web site design software  http://sitebuilder.yahoo.com   ------------------------------  % Date: Wed, 17 Sep 2003 10:10:24 +0500 & From: Valentin Likoum <me@privacy.net>  Subject: Re: " USP for OpenVMS "9 Message-ID: <bk8qc0$qmvhm$1@ID-184585.news.uni-berlin.de>    Fabio Cardoso wrote:	 > Richard  > 6 > Is there any other way to encode this data in OVMS?  [snip]  :    Yes, certanly. NCS! We use it to encode/decode UTF8/16 ; data. It's ability to decode UTF16 was a real surprise for  = me. Contact me directly (decode e-mail in the bottom) if you  6 need some help on it. Our conversion routines are for : russian code page only, but it's not a big deal to figure / out the principle and remake it for portuguese.    --  
 Best regards, 
   Valentin)   valentin.likoum at ncc dot volga dot ru    ------------------------------  # Date: Tue, 16 Sep 2003 19:26:23 GMT # From: "John Smith" <a@nonymous.com> 4 Subject: Re: "HP breakup on the way" - Merrill LynchF Message-ID: <ztJ9b.45958$DZ.8112@news04.bloor.is.net.cable.rogers.com>   Done.    David J Dachtera wrote: 0 > "John Smith" <a@nonymous.com> wrote in messageH > news:<WXr9b.469094$4UE.325375@news01.bloor.is.net.cable.rogers.com>...	 >> [snip]  >  > John,  > @ > Please e-mail me privately. Use my e-mail address from earlierD > messages, use dls in place of fsi and, of course, drop the obvious > bit. > H > I registered to post on Google as we're in process of moving right now> > and I don't have my usual e-mail/posting capability at home. >  > Thanx  >  > David J Dachtera > dba DJE Systems    ------------------------------  % Date: Tue, 16 Sep 2003 17:01:52 -0400 ( From: David Froble <davef@tsoft-inc.com>4 Subject: Re: "HP breakup on the way" - Merrill Lynch, Message-ID: <3F677A40.7020900@tsoft-inc.com>   Brian Tillman wrote:  , >>There is one thing that I do know 100% forI >>certain on this deal....before my customer will sign for something like  >> > this > K >>they want Carly on their doorstep giving them assurances in writing, with K >>penalty clauses, that certain actions will be undertaken by HP in respect  >> > of > 3 >>VMS, and one of those includes advertising VMS...  >> > J > You might as well kiss your deal goodbye right now, if this is what theyM > expect.  I haven't heard of many CEOs who would gop along with this.  Would . > Sun's CEO?  Would Oracle's CEO?  Fat chance. >   P Well, I'm not spending $25 million dollars.  But if I was, and upper management L at company 'A' wouldn't address any concerns I had, and upper management at M company 'B' took the time and made the effort to address any concerns I had,  ! guess who would get the business.   Q There's gotta be some number that will get the attention of a CEO.  I don't know  8 what that number is, but I feel strongly that it exists.  Q I'll allow that some of the above demands are a bit excessive.  Not sure where a   line should be drawn.    Dave   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Tue, 16 Sep 2003 17:21:21 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>4 Subject: Re: "HP breakup on the way" - Merrill Lynch) Message-ID: <3F677EBD.95E8026F@istop.com>    David Froble wrote: R > There's gotta be some number that will get the attention of a CEO.  I don't know: > what that number is, but I feel strongly that it exists.  J I agree with this when it comes to sale involving a core strategic productJ that the vendor really cares about. But in the case of a product which theI vendor does not consider to be core and/or strategic, then the vendor may C decide that it is best to lose that sale than to be encumbered with 7 commitments the vendor is trying so hard to get rid of.   N For instance, how big an MPE or Tru64 sale opportunity would it take for HP toJ agree to make the commitments Mr John Smith expects from HP ? Sure, if theE amount is high enough, HP might reconsider its decision to kill those 0 products, buit the amount would have to be HUGE.  G In the case of VMS, the problem is twofold. If the customer expects the J systems now, they would have to be Alpha based and the customer would thenJ require some hard commitments on the remaining life of the alpha platform.N Extending the life of Alpha systems goes against strategic directions by HP to, copncentrate on that IA64 chip nobody wants.  L Secondly, by all appearances, HP seems to have decided that VMS is to remainN available only to existing customers with no push to extend its market. **IF**K this is the actual strategy by HP, then Carly would balk at requirememts to M market and grow VMS at the expense of competing products, namely Unix (HP-UX) 5 and Windows, both of which are core, strategic to HP.    ------------------------------  % Date: Wed, 17 Sep 2003 12:39:36 +0800 , From: Paul Repacholi <prep@prep.synonet.com>4 Subject: Re: "HP breakup on the way" - Merrill Lynch- Message-ID: <87vfrs9gfr.fsf@prep.synonet.com>    Can I add a `me too' :)    And just a reply will work...    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Tue, 16 Sep 2003 17:21:14 -0500  From: PEALYKWBAY@med.VA.gov $ Subject: RE: Block size in show dev?A Message-ID: <124534A39403D31182C80000F8312F6A03BD213C@VHABAYEXC1>   ' DSM Blocks vs. VMS Blocks vs. DSM Maps   1 DSM Block = 1024 bytes 1 VMS Block =  512 bytes 1 DSM Map   =  400 DSM blocks * GigaBytes DSM Maps   DSM BLKS     VMS BLKS. ---------            --------     --------    * 1/4 Gb= 640 maps      256,000      512,000+ 1/2 Gb =1,280 maps    512,000    1,024,000  * 1   Gb =2,560 maps    1,024,000  2,048,000* 2   Gb =5,120 maps    2,048,000  4,096,000* 3   Gb =7,680 maps    3,072,000  6,144,000* 4   Gb =10,240 maps   4,096,000  8,192,000 -----Original Message-----? From: David McKenzie [mailto:david.mckenzie@paradigm-shift.biz] * Sent: Saturday, September 13, 2003 4:58 AM To: Info-VAX@Mvb.Saic.Com $ Subject: Re: Block size in show dev?    . "IBM" is not a bad word, it is an acryronm :-)   --   David McKenzie  7 David.Mckenzie@paradigm-shift.dot.biz  remove the "dot"  OpenVMS IT Privacy and Law   http://www.paradigm-shift.biz 5 "John Brandon" <brandon@dalsemi.com> wrote in message + news:03091018390986@dscis6-0.dalsemi.com... - > > John Brandon <brandon@dalsemi.com> wrote: 3 > > > > SHOW DEV DKA0 shows 17,773,524 total blocks @ > > > > are these 512 byte blocks?  What is the block size here? > > > Block size = 512 bytes. . > > > Roughly 17,773,524 / 2,000,000 = 8.88 GBL > > > The value of 2,000,000 is not exact but it is close enough for my use. > > > Isn't it like 209,000,00?  > >  > Tony Arnold wrote:J > > Isn't 2,000,000 actually exact? If a block is 512 bytes, then 1Gb, hasG > > to have 2,000,000 blocks exactly? Or has my maths gone rusty? (I am H > > assuming that 1Gb is 1024 megabytes, which in turn is 1024 kilobytes	 > > etc.) 	 > > Tony.  >  > Your math is fine. > J > Two definitions (check the definition by IBM - ooo was that a bad word?) > C > http://whatis.techtarget.com/definition/0,,sid9_gci212542,00.html  >  > C > If you assume that 1-GB is 1,024 mega-bytes then you are correct.  > ' >      17,773,524 * 512 = 9,100,044,288 + >      9,100,044,288 / 1,024,000,000 = 8.88  >           -or-$ >      17,773,524 / 2,000,000 = 8.88 >  > = > However if you assume that 1-GB is 1,000 mega-bytes then... 9 > mega means 1,000,000 bytes and giga means 1,000,000,000 5 > Therefore 1-GB is 1,000 mega-bytes or 1,000,000,000  > ' >      17,773,524 * 512 = 9,100,044,288 5 >      9,100,044,288 / 1,000,000.00 = 9,100 megabytes 5 >      9,100,044,288 / 1,000,000,000 = 9.10 gigabytes  > E > I have seen bytes represented both ways - and since the division by  2,000,00' > works for me... well it works for me.  > L > And I bet the disk is labeled completely differently... need to check that one  > out! > H > So that still does not explain where I came up with the 2,090,000 - oh	 bother...  >  > Then what Hoff said : C > http://h71000.www7.hp.com/wizard/faq/vmsfaq_018.html#index_x_1092  >  >  >  >  > J*o*h*n B*r*a*n*d*o*n  > VMS Systems Administrator , > firstname.lastname.spam.me.not@dalsemi.com   ------------------------------   Date: 17 Sep 2003 01:28:55 GMT From: healyzh@aracnet.com , Subject: Re: Booting hobbyist CD on PWS433au, Message-ID: <bk8dcn01i3p@enews2.newsguy.com>  % Jeff Campbell <n8wxs@arrl.net> wrote: N > The KZPBA needs to be install in one of the 3 upper PCI slots, not in either@ > of the 2 lowest PCI slots. It needs to be behind a PCI bridge.  L Bleep.  What about a narrow SCSI card?  This would explain the problem I wasL having trying to add a second UW-SCSI card to my system.  My top PCI slot isH dead, and I was trying to add the 2nd KZPBA to one of the two low slots.   			Zane    ------------------------------  + Date: Tue, 16 Sep 2003 20:33:35 +0000 (UTC) 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> * Subject: Choosing between ASTs and Threads/ Message-ID: <bk7s2t$k3t$1@titan.btinternet.com>   	 Hi Vince,   K Let's  cut  to  the  chase; IMHO threads on VMS are *only* ever useful when K porting  UNIX code based on a threaded architecture. The more software that K runs  on  VMS  the better, and I applaud VMS engineering's *very expensive* K efforts  to  re-engineer  the  entire  operating  system to schedule around K kernel   threads  so  that  VMS  can  not  only  run  these  UNIX-developed 2 applications, but also run them blisteringly fast.  K Having  said  that,  if you are developing software *on* VMS *for* VMS then . ASTs are unquestionably the right way to go!!!  G > If I read it right, a program designed around ASTs cannot make use of  > more than one CPU at a time   I Bollocks! This is the sort of lie (or at best half-truth) spun by the ADA G worshipping Larry Kilgallens of this world who know full well that in a K True Master/Slave scenario (for example), the AST driven Master process can G delegate work to any number of worker processes (1:1 Kernel Threads) to I achieve all of the benefits of parallelism and Symmetric multi-processing  that they espouse.  J I've waited and waited (until the cock crowed 3 times in fact) for Hoff toI come in and tell you how wonderful threads are and that ASTs are destined E for the obsolete features manual (along with Macro32 and User-Written K System Services) but spurred on by his silence, let me help you "learn some 3 of the ways in which programming is unique on VMS."   K You  do  not  tell  us  much about what your application/problem is, so for K argument's  sake  I  will assume that you want to have thousands of network K connections  to  a  VMS  server  application.  A common way of solving this K problem is to have a single "Communication Server" process that handles all K of  the  Network I/O and Authentication and a variable number of "Execution K Server" processes that perform the application specific functions (database K I/O  etc). This pool of "Execution Server" processes can grow and shrink as K user  demand  requires  in  accordance with parameters set up by the System  Manager.  K The  communication  server is obviously AST threaded and must perform *all* K network  I/O,  timeouts and server management asynchronously. The *absolute K beauty*  of  this  is  that  as  long  as  you  have  your alignment and/or K granularity  settings  correct  then  there  is  *no* need to lock or mutex K around memory accesses. As long as all processing is done at AST level with K the  mainline doing something equivalent to SYS$HIBER, VMS will ensure that K no AST (of similar access mode) will interrupt another AST, so let VMS just K queue  them  up on a first come first served basis. (The demarcation of AST @ functionality at different access modes isn't that much harder.)  K But  surely  there  must  come  a  saturation point where the Communication K Server  is  swamped  with  work  and is competing with the Execution Server K processes?  Well  maybe yes, but this is not unique to the AST paradigm and K is  usually  addressed  by  setting  the  VMS scheduling priority of the CS K higher than that of the ESs. If it's still a problem then look at something K link  sys$process_affinity  so that the CS can be put on a different CPU to K that  of  the ES pool. (See sys$getsyi sys$_active_cpu_mask and cap$m_cpuN) K If you're still not bloody happy then get a cluster and some load balancing , software from UCX et al and scale endlessly!  K Now,  just  try and have thousands of threads in a single process trying to K achieve  the  same functionality. Mutext here; Lock there; Just to add 1 to K active_link_count.  "But  we've only got 16 (or 32?) kernel threads/process K all  the  rest  are  just bullshit?" Well then why don't you try that other K UNIX joke and FORK off a separate process for each one of your thousands of  users and see how that scales!  K In  closing  Vince,  if  you want more information then please ask away and K provide  a bit more detail about what you actually want to do. You may also K want  to  have  a  look  at  Windows2000  and why Microsoft moved away from / threads and re-embraced Asynchronous Callbacks.    THREADS JUST WON'T SCALE!!!    Wake up to it!   Regards Richard Maher   , From: Vince (vince@the[NOSPAM]berrymans.com)* Subject: Choosing between ASTs and Threads   View this article only Newsgroups: comp.os.vms  Date: 2003-09-07 04:47:33 PST   D I am new to VMS and am currently trying to learn some of the ways inB which programming is unique on VMS.  One question I am looking forD clarification on is the difference between using ASTs and threads on systems with more than one CPU.   E If I read it right, a program designed around ASTs cannot make use of H more than one CPU at a time since only one AST can execute at a time and; the main program can't execute while an AST is in progress.   E With threads, however, it appears I can have as many sections of code  executing as there are CPUs.   Is my understanding correct?   ------------------------------  % Date: Tue, 16 Sep 2003 17:11:49 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>. Subject: Re: Choosing between ASTs and Threads) Message-ID: <3F677C82.1C4391DD@istop.com>    Richard Maher wrote:  M > True Master/Slave scenario (for example), the AST driven Master process can I > delegate work to any number of worker processes (1:1 Kernel Threads) to K > achieve all of the benefits of parallelism and Symmetric multi-processing  > that they espouse.  M Are you talking about a master process that has a number of threads under it, L or a master process that has a number of totally separate worker processes ?  G In the case of having totally separate worker processes, what about the J overhead of inter process communications ? If you use stuff like decnet orM ICC, there is overhead for the comms, and if you use shared memory, you still : have to deal with some form of lock manager and signaling.   ------------------------------  % Date: Tue, 16 Sep 2003 17:48:56 -0400 * From: "Bill Todd" <billtodd@metrocast.net>. Subject: Re: Choosing between ASTs and Threads2 Message-ID: <XZ-cnXvkKovWGPqiXTWJkw@metrocast.net>  > "Richard Maher" <maher_rj@hotspamnotmail.com> wrote in message) news:bk7s2t$k3t$1@titan.btinternet.com...  > Hi Vince,  > H > Let's  cut  to  the  chase; IMHO threads on VMS are *only* ever useful when6 > porting  UNIX code based on a threaded architecture.  K While I haven't used VMS since long before threads existed there, it sounds C here as if you're completely ignoring the issues of multi-processor C systems - even though you explicitly quote that concern just below.    ...   I > > If I read it right, a program designed around ASTs cannot make use of  > > more than one CPU at a time  > K > Bollocks! This is the sort of lie (or at best half-truth) spun by the ADA I > worshipping Larry Kilgallens of this world who know full well that in a I > True Master/Slave scenario (for example), the AST driven Master process  can I > delegate work to any number of worker processes (1:1 Kernel Threads) to K > achieve all of the benefits of parallelism and Symmetric multi-processing  > that they espouse.  K It didn't sound as if the original poster wished to delegate work like that I (and of course such delegation has significant overheads of its own):  he L appeared to want to make use of multiple processors within the activities of a single process.    ...   I > You  do  not  tell  us  much about what your application/problem is, so  for E > argument's  sake  I  will assume that you want to have thousands of  network / > connections  to  a  VMS  server  application.   E You assume a workload that may lend itself to multi-process execution E (because the overheads of delegation pale beside the overheads of the F activities themselves - and those activities can be prosecuted largelyA independently by the multiple server processes without noticeable 8 interaction between them or with the delegating parent).  C What if the workload *doesn't* meet those assumptions?  What if the D delegation cost *is* important and/or the multiple concurrent serverF activities *do* need to coordinate closely among themselves (and henceE desire the efficiency of doing so within a single process rather than G between processes)?  The certainly sounds like the kind of workload the L original poster was trying to deal with (and if not, it's certainly the kindI of workload *some* people have to deal with, hence is still interesting).      A common way of solving thisI > problem is to have a single "Communication Server" process that handles  all B > of  the  Network I/O and Authentication and a variable number of
 "ExecutionC > Server" processes that perform the application specific functions 	 (database J > I/O  etc). This pool of "Execution Server" processes can grow and shrink asF > user  demand  requires  in  accordance with parameters set up by the System
 > Manager. > G > The  communication  server is obviously AST threaded and must perform  *all* C > network  I/O,  timeouts and server management asynchronously. The 	 *absolute F > beauty*  of  this  is  that  as  long  as  you  have  your alignment and/orG > granularity  settings  correct  then  there  is  *no* need to lock or  mutex H > around memory accesses. As long as all processing is done at AST level withH > the  mainline doing something equivalent to SYS$HIBER, VMS will ensure thatH > no AST (of similar access mode) will interrupt another AST, so let VMS justI > queue  them  up on a first come first served basis. (The demarcation of  AST B > functionality at different access modes isn't that much harder.)  D This is very similar to the way spinlocks are typically handled in aH single-processor kernel:  the processor priority is simply raised on theK spinlock request to a level that won't allow another spinlock request to be J made, and lowered when the spinlock is released.  The problem, once again,E is that the mechanism doesn't allow one to capitalize upon additional E processors when they are present to perform the work in parallel:  by < definition, the process is executing only one AST at a time.   ...   J > Now,  just  try and have thousands of threads in a single process trying to# > achieve  the  same functionality.   L Some people do that, but (as I explained in my earlier answer) it's far fromC optimal (too much unnecessary contention).  So, by the way, is your F suggested solution above, unless your 'execution server' processes areC limited in number to no more than the number of CPUs expected to be I available.  But then they may *under*-utilize the available CPU resources K unless they themselves execute asynchronously to keep busy - at which point K any apparent advantage in simplicity over the model I advocated disappears.   *  Mutext here; Lock there; Just to add 1 to > active_link_count.  A The same objection applies to your suggested solution - and those I independent server processes are far less efficient in communicating with L each other than an equal number of threads within a single process would be.  9   "But  we've only got 16 (or 32?) kernel threads/process G > all  the  rest  are  just bullshit?" Well then why don't you try that  other J > UNIX joke and FORK off a separate process for each one of your thousands of  > users and see how that scales!  G As I noted earlier, the optimal solution is to have a number of threads D equal to the expected number of available processors, and to operateH asynchronously within each such thread to keep them all busy.  This doesB require some modest programming to ensure that they never wait forJ *anything* for any significant period - e.g., you need to implement simpleL lock and mutex mechanisms which return to the thread instead of quiescing itI (after setting up structures which will place an appropriate item on your : work queue when the requested resource becomes available).   > I > In  closing  Vince,  if  you want more information then please ask away  and H > provide  a bit more detail about what you actually want to do. You may alsoH > want  to  have  a  look  at  Windows2000  and why Microsoft moved away from1 > threads and re-embraced Asynchronous Callbacks.  >  > THREADS JUST WON'T SCALE!!!   L I think you may be confused about the reasons for (and benefits of) threads.I Not only can they be easier to work with when performance doesn't matter, H but they allow use of multiple processors within a single process (while6 avoiding the overheads of multi-process interactions).   - bill   ------------------------------    Date: 16 Sep 2003 17:45:03 -0500- From: Kilgallen@SpamCop.net (Larry Kilgallen) . Subject: Re: Choosing between ASTs and Threads3 Message-ID: <qVu19+erfXih@eisner.encompasserve.org>   e In article <bk7s2t$k3t$1@titan.btinternet.com>, "Richard Maher" <maher_rj@hotspamnotmail.com> writes:  > Hi Vince,   H >> If I read it right, a program designed around ASTs cannot make use of >> more than one CPU at a time > K > Bollocks! This is the sort of lie (or at best half-truth) spun by the ADA I > worshipping Larry Kilgallens of this world who know full well that in a M > True Master/Slave scenario (for example), the AST driven Master process can I > delegate work to any number of worker processes (1:1 Kernel Threads) to K > achieve all of the benefits of parallelism and Symmetric multi-processing  > that they espouse.  E That begins to fall apart when the control patterns get complex.  For D someone doing safety critical work (i.e., people will die if you getB it wrong), using the SPARK subset of Ada allows formal proofs thatG would just not be possible on a home-grown load-sharing implementation.   D And as recently posted here by Rod Chapman of Praxis, Spark ExaminerD _is_ supported on VMS (even on VAX, since he was looking for TK50s).  > For anyone who thinks that handling such simultaneity is easy,< I recommend a look at "Concurrency in Ada" by Alan Burns andD Andy Wellings, regardless of what programming language you will use.   ------------------------------  % Date: Tue, 16 Sep 2003 19:32:50 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>. Subject: Re: Choosing between ASTs and Threads) Message-ID: <3F679D86.FB6EF39D@istop.com>   K When you take VMS clustering into consideration, then having an application M architecture that is multi-process based instead of multi-threaded results in H greater capabilities and potential in terms of tolerance to failures and/ automatic failover when one node fails etc etc.    ------------------------------  # Date: Wed, 17 Sep 2003 02:59:01 GMT / From: "Richard L. Dyson" <rick-dyson@uiowa.edu> , Subject: CSwing and OpenVMS v7.3-1, Again.../ Message-ID: <V5Q9b.481056$o%2.214650@sccrnsc02>   F I believe that the problems that I had read about in the past and thenJ experienced first hand with the old CSwing (v3.7.6) program and the v7.3-1 filesystem have gone away.  I I recently installed from scratch an AlphaServer 2000 with v7.3-1 and all J the ECOs and TCPIP v5.3, etc.  I happened to accidentally copy over an oldM copy of CSwing from another AlphaServer and was using it to read and discover B ODS-5 volumes, etc. when I realized that was not supposed to work!  N I can only guess that the recent RMS ECO might be the difference?  Does anyoneI else know or if they are aware of the CSwing problems can they check too? J I seem to recall I couldn't do directory tree "walks" (the S function) forL ODS-2 or ODS-5.  There were some other problems too, but I can't recall them just right now.    Does anyone else remember?   Regards, Rick --  J Richard L. Dyson                                      rick-dyson@uiowa.eduK   _   _  _____                      http://www-pi.physics.uiowa.edu/~dyson/ J | | | ||_   _|  Senior Systems Analyst   --   INFORMM-Cerner Systems Group< | | | |  | |    The University of Iowa Hospitals and ClinicsJ | \_/ | _| |_   Information Systems Dept. BT1000 GH   Office: 319/384-7016K   \___/ |_____|  Iowa City, IA 52242-1052                 FAX: 319/384-7020 E                  (Consulting to the Physics and Astronomy Department)    ------------------------------  % Date: Tue, 16 Sep 2003 13:48:02 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>5 Subject: Re: FTP on AlphaServer running OpenVMS 7.2-1 ) Message-ID: <3F674CCD.AF2BE286@istop.com>     shawnm@cotepe.carenet.org wrote:M > 1 particular directory.  Basically the user will only connect to the server 
 > via ftp. > L > Is there a way I can set permissions or something up in his login.com file; > to keep him from moving up a level from where I drop him.   N This was discussed before. No, there is no way to 100% garantee he won't strayB to other directories. However, you can make things harder for him:  : created a rooted logical name which acts as a device name.  J so instead of having his sys$login point to dka200:[users.jdoe] you defineN USRDIR to point to dka200:[users.]. In authorize, his default device is USRDIR$ and his default directory is [JDOE].N Then, you secure the dka200:[000000]users.dir file so that jdoe cannot see its directory contents.   N This means that the use will basically see only his own directory and won't be able to navigate up.  J However, if the user knows of another directory, he can jump to a specificL directory assuming he has permission to do so, just as a normal regular userC can do a DIR of many system directories (such as SYS$HELP etc etc).    ------------------------------    Date: 16 Sep 2003 13:28:48 -0700% From: tadamsmar@yahoo.com (Tom Adams) 2 Subject: How to use shareable image to share data?= Message-ID: <ea44f5a1.0309161228.713de53f@posting.google.com>   ? I am trying to set up a shareable image for sharing a data area  between F processes.  I have an old app that was developed on the Vax and is notE working on the Alpha (7.2-1). It uses shareable images to communicate B data.  I set up a simple test case and it is also not working, not
 sharing data.    Here is my test case:   * I am trying to have a shared common /test/  
 I have C.FOR:  	integer*4 x         byte b(9000)         common /test/ x,b  	volatile /test/         end    I compile and link this with: 
 	FOR C.FOR 	LINK/SHAR C.FOR   Then I install it:( 	install := $sys$system:install /command3 	install create DSA0:[ADAMS]C.EXE /OPEN/SHARE/WRITE      I define clognam: ! 	define clognam dsa0:[adams]c.exe   
 I have X.FOR:          integer*4 x          byte b(9000)         common /test/ x,b  	volatile /test/         type *,x
         x = 5          type *,x         end   
 and X.OPT:    clognam/share   I compile and link X:     for X    link X,X/OPT   5 Here are the fortran and link symbols that were used:   4  FOR*TRAN == "FORTRAN /NOLIST /DEBUG /NOOPT /NOALIGN$       /CHECK=(BOUN,FP_E,OVER,NOUNDE)&       /WARN=(NOALIGN,NOUNINIT)/EXTEND"  '  LINK == "LINK /DEBUG /NOMAP /NONATIVE"   B I have demonstrated in various ways that the executable X does notB share the common /TEST/ when I run the executable in two different
 processes.   What am I doing wrong?   ------------------------------  % Date: Tue, 16 Sep 2003 17:06:11 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>6 Subject: Re: How to use shareable image to share data?( Message-ID: <3F677B31.AA20425@istop.com>   Tom Adams wrote:D > I have demonstrated in various ways that the executable X does notD > share the common /TEST/ when I run the executable in two different > processes. >  > What am I doing wrong?  L Isn't that how shareable images are supposed to behave ? the code is common, but each has its own data.  E Shouldn't you be using global sections instead of a shareable image ?    ------------------------------  + Date: Wed, 17 Sep 2003 05:47:54 +0000 (UTC) 3 From: "Richard Maher" <maher_rj@hotspamnotmail.com> 6 Subject: Re: How to use shareable image to share data?2 Message-ID: <bk8sia$23h$1@hercules.btinternet.com>   Hi,   3 On Alpha in your linker options file something liker( Symbol_table=(blah=procedure,test=psect)   Regards Richard Maherr  0 Tom Adams <tadamsmar@yahoo.com> wrote in message7 news:ea44f5a1.0309161228.713de53f@posting.google.com... A > I am trying to set up a shareable image for sharing a data areah	 > between H > processes.  I have an old app that was developed on the Vax and is notG > working on the Alpha (7.2-1). It uses shareable images to communicatehD > data.  I set up a simple test case and it is also not working, not > sharing data.2 >  > Here is my test case:e > , > I am trying to have a shared common /test/ >h > I have C.FOR:o
 > integer*4 xb >         byte b(9000) >         common /test/ x,br > volatile /test/ 
 >         endl >  > I compile and link this with:A > FOR C.FOR  > LINK/SHAR C.FORA >t > Then I install it:) > install := $sys$system:install /command44 > install create DSA0:[ADAMS]C.EXE /OPEN/SHARE/WRITE >5 >n > I define clognam:s" > define clognam dsa0:[adams]c.exe >u > I have X.FOR:  >         integer*4 xo >         byte b(9000) >         common /test/ x,bn > volatile /test/T >         type *,x >         x = 5r >         type *,x
 >         endN >o > and X.OPT: >    clognam/share >i > I compile and link X: 
 >    for X >    link X,X/OPTo >e7 > Here are the fortran and link symbols that were used:e >T6 >  FOR*TRAN == "FORTRAN /NOLIST /DEBUG /NOOPT /NOALIGN& >       /CHECK=(BOUN,FP_E,OVER,NOUNDE)( >       /WARN=(NOALIGN,NOUNINIT)/EXTEND" >:) >  LINK == "LINK /DEBUG /NOMAP /NONATIVE"m >tD > I have demonstrated in various ways that the executable X does notD > share the common /TEST/ when I run the executable in two different > processes. >. > What am I doing wrong?   ------------------------------    Date: 16 Sep 2003 15:56:13 -07001 From: keithparris_NOSPAM@yahoo.com (Keith Parris)-6 Subject: HP extends lead in open storage area networks= Message-ID: <cf15391e.0309161456.154aac76@posting.google.com>c  C "HP holds firm no. 1 position in worldwide disk storage systems and3 external disk markets.  B HP (NYSE:HPQ) demonstrated strong results with its HP StorageWorksE systems worldwide in the second calendar quarter of 2003, maintaining B a clear No. 1 position in the total worldwide disk storage systemsC market by factory revenues and posting year-over-year gains in manydB market segments, according to IDC's Worldwide Disk Storage Systems" Quarterly Tracker, September 2003.  ? In networked disk storage systems, HP led the open storage area2D networks (SANs) market for the fourth consecutive quarter and gained8 three percentage points of market share year-over-year." ...o9 http://www.hp.com/hpinfo/newsroom/press/2003/030905a.htmln   ------------------------------  % Date: Tue, 16 Sep 2003 19:34:44 -0400A* From: JF Mezei <jfmezei.spamnot@istop.com>: Subject: Re: HP extends lead in open storage area networks) Message-ID: <3F679DF8.55672CA9@istop.com>=  M And Steve Jobs, in a BBC interview, said that there were only 2 profitable PC G businesses: Dell because they are high volume and efficient producer of=F commodity stuff, and Apple because of their innovation. All others areT struggling and must subsidize their PC operations with profits from other divisions.   ------------------------------  % Date: Tue, 16 Sep 2003 17:49:13 -0400r* From: JF Mezei <jfmezei.spamnot@istop.com>% Subject: IA64 to be dual core by 2005 ) Message-ID: <3F678544.CB974C58@istop.com>e  n > http://story.news.yahoo.com/news?tmpl=story&cid=569&ncid=738&e=2&u=/nm/20030916/tc_nm/tech_intel_otellini_dc  * (sorry if it wraps, essential text below:)M SAN JOSE, Calif. (Reuters) - Intel Corp.  the world's largest chip maker, on tN Tuesday outlined plans for two new chips that will have two or more processorsJ on a single piece of silicon, boosting the performance of corporate server computers.    N The new chips are a 32-bit Xeon server processor MP, code-named "Tulsa," whichG will be its first dual-core chip, and a new 64-bit Itanium server chip,-J code-named "Tanglewood," Intel President Paul Otellini said in his keynote& address at the Intel Developer Forum.   N  Tanglewood will come some time after 2005 and Tulsa in two to three years, he said.   L  In an interview, Otellini declined to say what plans Intel has for bringingM 64-bit chips to PCs. Currently, its chips for PCs crunch 32 bits of data at as time.   H  "The production operating systems are not there yet" for 64-bit desktopO computers, he said. "The mainstream applications won't exist until next year." e  J  Apple Computer Inc.  already offers a personal computer based on a 64-bitI chip, while Intel rival Advanced Micro Devices Inc.  is set to unveil itsh$ Athlon 64 for PCs later this month.   K  A future Itanium processor, code-named "Montecito," will be the first chip G with one billion transistors, according to Otellini. It is targeted forR production in 2005, he said. r  E  Otellini also gave a timeline for pending chip manufacturing process G technologies that allow Intel to shrink the circuitry and space between L transistors to fit more of them onto a chip, thus increasing their computing( power with no increase in cost or size.   M  He showed for the first time a wafer containing chips made with 65 nanometern7 process, which will be in production in 2005, he said. o  G  Those chips will be followed by 45 nanometer process chips in 2007; 32 J nanometer process chips in 2009; and, in 2011, 22 nanometer process chips,H which will have transistor space that is smaller than the width of a DNA molecule, he said.   ------------------------------  % Date: Tue, 16 Sep 2003 17:52:19 -0400p* From: JF Mezei <jfmezei.spamnot@istop.com>) Subject: Re: IA64 to be dual core by 2005t) Message-ID: <3F6785FD.D135B863@istop.com>e  	 question:d  K what does dual core mean from the point of view of the operating system and  scheduling ?  M Does it appear as a single (faster) chip to the OS, or does the OS see 2 cpusI which it can use for SMP etc ?   ------------------------------  # Date: Wed, 17 Sep 2003 00:26:25 GMTe# From: "John Smith" <a@nonymous.com> ) Subject: Re: IA64 to be dual core by 2005 J Message-ID: <RSN9b.476449$4UE.283772@news01.bloor.is.net.cable.rogers.com>  ( Then there's this article and quote.....  3 http://www.eweek.com/article2/0,4149,1269660,00.aspA  0       Intel CTO: Desktop Isn't Ready for 64 Bits       By Mark Hachman2       September 15, 2003    J       SAN JOSE, Calif.-Sixty-four bit computing won't be needed on desktopH PCs for several more years, according to Pat Gelsinger, chief technology officer at Intel Corp. ......   F       ie. no volume in the 64-bit market from Intel for years.....read- higher prices....read lower adoption rates...h   ------------------------------  # Date: Wed, 17 Sep 2003 02:09:42 GMTv+ From: Ryan Moore <rmoore@rmoore.dyndns.org>a) Subject: Re: IA64 to be dual core by 2005r5 Message-ID: <Pine.LNX.4.56.0309161906170.6107@jaipur>i  $ On Tue, 16 Sep 2003, JF Mezei wrote: > question:b >dM > what does dual core mean from the point of view of the operating system ande > scheduling ? > O > Does it appear as a single (faster) chip to the OS, or does the OS see 2 cpusJ  > which it can use for SMP etc ?  I Assuming it's the same sort of technology as the new P4 "hyperthreading", J then it would show up as two CPUs on one piece of silicon.  In the case ofJ P4 "hyperthreading", there are two sets of execution units, but some otherI components are shared.  Thus, the CPU isn't as fast as two complete CPUs.t  I The more componenets that are shared between the cores, the less speed-uph that can be accomplished.t  I VMS would see two CPUs and would do SMP scheduling.  There would probablyo. need to be support in the OS to do it, though.   -RyanM   ------------------------------   Date: 17 Sep 2003 02:33:25 GMT From: healyzh@aracnet.com ) Subject: Re: IA64 to be dual core by 2005l, Message-ID: <bk8h5l01onj@enews2.newsguy.com>  , Ryan Moore <rmoore@rmoore.dyndns.org> wrote:& > On Tue, 16 Sep 2003, JF Mezei wrote:
 > > question:, > >tO > > what does dual core mean from the point of view of the operating system and  > > scheduling ? > >cQ > > Does it appear as a single (faster) chip to the OS, or does the OS see 2 cpus " > > which it can use for SMP etc ?  K > Assuming it's the same sort of technology as the new P4 "hyperthreading", L > then it would show up as two CPUs on one piece of silicon.  In the case ofL > P4 "hyperthreading", there are two sets of execution units, but some otherK > components are shared.  Thus, the CPU isn't as fast as two complete CPUs.a  K Dual Core refers to two seperate CPU's on the same chip.  I believe IBM hasuJ been doing this for at least a year or two with POWER4.  Basically it lets9 you have a true dual processor system with only one chip.V   			Zanej   ------------------------------  % Date: Wed, 17 Sep 2003 00:49:35 -00002! From: Z  <zarlenga@conan.ids.net>c> Subject: init/queue/start/processor=ucx$telnetsym/on=... hangs/ Message-ID: <vmfbsv4bbc09ac@corp.supernews.com>k  + Queue shows as "starting" but never starts.   7 The $init never completes (^Y/^C must be used to abort)   B The node has telnet enabled ... must I also start telnet symbiont?   If so, how?=   ------------------------------   Date: 17 Sep 2003 01:41:38 GMT From: healyzh@aracnet.coma Subject: KZPBA Question0, Message-ID: <bk8e4i11i3p@enews2.newsguy.com>  H On a DEC PWS 433au, can you have disks attached to both the Internal andI External Wide connections?  I seem to recall that there is a problem withM doing this.   I Is this simply something that is unsupported like using both the InternalcM Narrow and Wide connectors, or is this something that just plain won't work. m  @ I can see the disk on the internal connector, but can't INIT it:   $ init dka0: testo! %INIT-F-MEDOFL, medium is offlinee $    			Zane0   ------------------------------  % Date: Tue, 16 Sep 2003 23:26:28 -0400>* From: "Stanley F. Quayle" <stan@stanq.com> Subject: Re: KZPBA Questionr- Message-ID: <3F679C24.23916.AE07AD@localhost>e  J > On a DEC PWS 433au, can you have disks attached to both the Internal andK > External Wide connections?  I seem to recall that there is a problem withr
 > doing this.   A You might not have properly terminated SCSI bus.  If the card is 0B terminating, for example, but is in the middle of the bus, you'll  have problems.  F On my 500au, I finally gave up and bought a second interface just for  external connections...i  
 --Stan Quaylel Quayle Consulting Inc.  
 ----------C Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363  Fax: +1 614 868-1671M1 8572 North Spring Ct. NW, Pickerington, OH  43147e= Preferred address:  stan@stanq.com       http://www.stanq.comf   ------------------------------    Date: 16 Sep 2003 16:06:00 -0700  From: rich@gocmt.com (gocmt.com)+ Subject: Re: Looking for TK50s in the UK...== Message-ID: <d926a291.0309161506.75ad446c@posting.google.com>   l Tim Llewellyn <tim.llewellyn@blueyonder.co.uk> wrote in message news:<3F47BA80.3E30FC8A@blueyonder.co.uk>... > Rod Chapman wrote: > > 5 > > Does anyone know where I can buy blank TK50 tapes 2 > > (also known as DLT type I) in the UK?  Believe: > > it or not, we still have to supply our biggest VAX/VMS > > user using TK50 media... >  > C We at CMT can sell you TK 50's(we are sitting on a lot of them)at an@ great price if your interested shoot me an email rich@gocmt.com  Thanks Richi    9 You may be lucky and find one of the office suppliers who " > carries them, but I'm not sure.  > ; > I am wondering just what the maintenance charges for your < > customer's TK50 drives  are, and how many CDrom drives one; > could buy  out of that. OK, there are bootability issues, < > but there are ways round that like booting VMS on a WintelA > running a VAX emulator and booting enough of a cluster to copy a > data from CD to a local disk.m >  mz   >  e > >  > >  Thanks,: > >   Rod Chapman, Praxis (yes..that's _the_ Praxis! :-) )   ------------------------------  # Date: Tue, 16 Sep 2003 20:33:03 GMT 4 From: brad@.gateway.2wire.net (Bradford J. Hamilton) Subject: Re: Montagar website?/ Message-ID: <3sK9b.367458$Oz4.144372@rwcrnsc54>   Z In article <1063711284.32205.0@lotis.uk.clara.net>, issinoho <issinoho@slayme.com> writes: !Bradford J. Hamilton wrote: !s] !> In article <1063703678.29910.0@lotis.uk.clara.net>, issinoho <issinoho@slayme.com> writes:l! !> !Anyone know what's happening>-3 !> !Have been unable to connect for a few days now.0 !> ! !> rJ !> If you are trying to connect for a hobbyist-related issue, try this URL !> instead:r !>  " !> http://www.openvmshobbyist.org/ !> t$ !> Same folks, just a new address... !>  M !> __________________________________________________________________________sD !> Bradford J. Hamilton                    "All opinions are my own"N !> bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-' 3 !>                                          with @"a !nJ !Thanks Brad, that gets me further, although it still attempts to contact = !Montagar (still failing!) when I try and Register a license.i !   E I just connected to montagar.com, both directly, and via the links ats openvmshobbyist.org.  J __________________________________________________________________________A Bradford J. Hamilton                    "All opinions are my own"sK bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-' r0                                          with @"   ------------------------------    Date: 16 Sep 2003 15:46:06 -0700( From: bob@instantwhip.com (Bob Ceculski) Subject: Re: OpenVMS Securitye= Message-ID: <d7791aa1.0309161446.153fa2a4@posting.google.com>S   Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<bk6o0m$mus$1@new-usenet.uk.sun.com>...e > Bob Koehler wrote: > > In article <bk4rhg$3j0$1@new-usenet.uk.sun.com>, Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> writes: > > = > >>What should however be entirely obvious to you and anyonew; > >>with sense is that CERT is not the place to go tallyings: > >>up OE tables of vunerabilites if OpenVMS is one of the7 > >>OE's in question. It would be an entirely dishonestn > >>excercise. > >  > > 3 > >    So what alternative do you want to point to?l > >  > 5 > OpenVMS vunerabilites are "reported" in a number ofo5 > places CERT sometimes, HP security updates, OpenVMSt7 > patch reports, BugTraq reports. There is no one placec= > and often advice differs on the same vunerability depending  > on where you look. > # > So you come up with a suggestion.  > ? > I have never suggested that this kind of excercise is usefulls9 > in the first place though obviously having some sort ofv" > definitive source would be good. > 	 > Regards  > Andrew Harrisonm  @ the trouble is you will not find that many ... and you will growB bored very fast waiting around for the 14th which may take another 5 years ... :)   ------------------------------  + Date: Tue, 16 Sep 2003 17:48:08 +0000 (UTC)e From: david20@alpha2.mdx.ac.uk6 Subject: Re: OT: Verisign messes with DSN root servers) Message-ID: <bk7ico$j1d$1@news.mdx.ac.uk>t  o In article <goH9b.4847$2r7.98@news.cpqcorp.net>, "John E. Malmberg" <Malmberg@dskwld.zko.dec.compaq.hp> writes:  >JF Mezei wrote:J >> added information: seems that Verisign have temporarily restorer properQ >> behaviour for the .com tld, but .net still has a wildcard response pointing too >> verisign's own IP address.e >wK My understanding is that the .com tld hasn't been corrected it's just that bN caching and propogation delays mean that you are currently not always getting  the wildcard response. e   ><snip>r > P >> Consider the whole RBL issue. RBL have become moot since any and all .com and0 >> .net domain names now translate to something. >nB >The RBLs seem to be unaffected.  The one below has a simple test 7 >address.  (Use spamcop.net as an RBL at your own risk)  > + >$ tcpip show host 1.0.0.127.bl.spamcop.net ) >%TCPIP-W-NORECORD, information not foundl >-RMS-E-RNF, record not found + >$ tcpip show host 2.0.0.127.bl.spamcop.netc >   K These shouldn't be affected since the net servers know about the subsiduaryaF domain spamcop and will pass DNS resolutions to that domain's servers.  O I think the only things which will be affected by this are where the typo is ina( the level immediately below .net or .com   ie  <     xxxcvbbb.spamcop.net  will receive the correct rejectionK     xxxcvbbb.spamcxop.net  will be return the ip address of verisigns site.a  ' Alpha1:ucx sh host xxxcvbbb.spamcop.netw( %TCPIP-W-NORECORD, information not found -RMS-E-RNF, record not found( Alpha1:ucx sh host xxxcvbbb.spamcxop.net        BIND database  ! Server:   158.94.0.14      alpha2    Host address    Host name   % 64.94.110.11    XXXCVBBB.SPAMCXOP.NET     O Of course the logical followup to this is for Verisign to put in a wildcard forTK in-addr.arpa addresses - that would really affect all unassigned addresses.0       >      BIND database > " >Server:   16.30.0.20       zkons1 >f >Host address    Host name > ) >127.0.0.2       2.0.0.127.BL.SPAMCOP.NETg >r >lG >If you are getting a bad response from an RBL, it probably means that m5 >there was a typo, or something specific to that RBL.g > H >The only thing in a spam filter that this should affect is if you were J >deleteing e-mails that referenced non-existant domains in their content, D >and that means that you were throwing out mail when someone made a  >typographical error.f >d  
 David Webb VMS and Unix team leader CCSS Middlesex University         >-John" >malmberg@dskwld.zko.dec.compaq.hp >Personal Opinion Only >    ------------------------------  % Date: Tue, 16 Sep 2003 13:42:56 -0400d* From: JF Mezei <jfmezei.spamnot@istop.com>6 Subject: Re: OT: Verisign messes with DSN root servers) Message-ID: <3F674B9B.A858C16F@istop.com>e   "John E. Malmberg" wrote:r, > $ tcpip show host 1.0.0.127.bl.spamcop.net* > %TCPIP-W-NORECORD, information not found > -RMS-E-RNF, record not found, > $ tcpip show host 2.0.0.127.bl.spamcop.net  K That is because spamcop.net exists and points to their own DNS server whichaV then responds with an NXDOMAIN when you query for a specific host inside their domain.  L When you query a .com or .net, your DNS "client" will first inquire with theJ root servers to see who is responsible for .net (you will be told to checkG with verisign). You then check with verisign for who is responsible forlG spamcop.net and verisign will give you the dns server that contains the-L resolution for spamcom.net, you then query that server for the specific host and get its IP address.R  = (caching means that you don't do those queries all the time).q  H The problem is that if you query www.wdjhfkjsh.net, verisign, instead ofL telling you that wdjhfkjsh.net is not a valid domain, it points you to a DNSK server which always responds with the same IP address no matter what domain 	 you want.U   ------------------------------  % Date: Tue, 16 Sep 2003 11:34:01 -0700e+ From: "Barry Treahy, Jr." <Treahy@MMaz.com> 6 Subject: Re: OT: Verisign messes with DSN root servers' Message-ID: <3F675799.7090603@MMaz.com>    JF Mezei wrote:a   >"Barry Treahy, Jr." wrote:  >  >  I >aF >>www.asdadasd.com, and it in fact redirected to this ad-site.  I, forH >>one, will create a bogus DNS entry for sitefinder.verisign.com to stop >>this without our networks. >>     >> >nN >This won't work properly. When you ask for www.blablabla.com or .net, you getO >a succesful DNS query with the ip address of 64.94.110.11. You host then tries L >to connect to that IP address. You'd have to block that specific IP address >from your routers.e >mN >Furthermore, if you use your own DNS to have sitefinder.verisign.com to point> >to something non-existent, then your browser will act fully:  >r >  c >eD I see what you mean.  Am I the only one that finds this an abuse of I Verisign's delegated role on the net?  Who in the hell would you contact eG to complain about something like this?  My Senator and Rep would never n understand this...   Barryu   --    > Barry Treahy, Jr                       E-mail: Treahy@MMaz.com> Midwest Microwave                          Phone: 480/314-1320> Vice President & CIO                         FAX: 480/661-7028                        e   ------------------------------  + Date: Tue, 16 Sep 2003 18:29:29 +0000 (UTC)i From: david20@alpha1.mdx.ac.uk6 Subject: Re: OT: Verisign messes with DSN root servers) Message-ID: <bk7kq9$jma$1@news.mdx.ac.uk>i  V In article <3F674B9B.A858C16F@istop.com>, JF Mezei <jfmezei.spamnot@istop.com> writes: >"John E. Malmberg" wrote:- >> $ tcpip show host 1.0.0.127.bl.spamcop.netv+ >> %TCPIP-W-NORECORD, information not foundy >> -RMS-E-RNF, record not foundi- >> $ tcpip show host 2.0.0.127.bl.spamcop.netr >sL >That is because spamcop.net exists and points to their own DNS server whichW >then responds with an NXDOMAIN when you query for a specific host inside their domain.  >wM >When you query a .com or .net, your DNS "client" will first inquire with theoK >root servers to see who is responsible for .net (you will be told to check"H >with verisign). You then check with verisign for who is responsible forH >spamcop.net and verisign will give you the dns server that contains theM >resolution for spamcom.net, you then query that server for the specific hostn >and get its IP address. >n> >(caching means that you don't do those queries all the time). >eI >The problem is that if you query www.wdjhfkjsh.net, verisign, instead ofbM >telling you that wdjhfkjsh.net is not a valid domain, it points you to a DNSeL >server which always responds with the same IP address no matter what domain
 >you want.  O Seems some RBL setups may be affected by a secondary effect according to a postr on vmsnet.mail.pmdf      " H I think this is also breaking gateways that use dorkslayers.com as theirG RBL.  Since dorkslayers went off the air, lookups are being referred toS
 64.94.110.11.h  K If you go to samspade.org and enter any address in their "Blackhole" searchu box you'll get something like:  @ 192.42.95.16 listed in DORKS(64.94.110.11) DORKZTL(64.94.110.11)  I We're seeing mail returned to us saying that we're blackholed (we're not)t# and I think this may be the reason.d   "n  
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  # Date: Tue, 16 Sep 2003 19:00:17 GMTe& From: Rick Jones <foo@bar.baz.invalid>6 Subject: Re: OT: Verisign messes with DSN root servers1 Message-ID: <55J9b.4870$yD7.237@news.cpqcorp.net>n  * Barry Treahy, Jr. <Treahy@mmaz.com> wrote:E > I see what you mean.  Am I the only one that finds this an abuse ofuB > Verisign's delegated role on the net?  Who in the hell would youD > contact to complain about something like this?  My Senator and Rep  > would never understand this...  F Do the top-level registries get their charter's from the IANA or is it another of the Internet TLAs?n  
 rick jones --  G oxymoron n, Hummer H2 with California Save Our Coasts and Oceans platessF these opinions are mine, all mine; HP might not want them anyway... :)A feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...P   ------------------------------  % Date: Tue, 16 Sep 2003 15:04:29 -0400a* From: JF Mezei <jfmezei.spamnot@istop.com>6 Subject: Re: OT: Verisign messes with DSN root servers( Message-ID: <3F675EB3.7841E24@istop.com>   "Barry Treahy, Jr." wrote:E > I see what you mean.  Am I the only one that finds this an abuse of ) > Verisign's delegated role on the net?  c    K Nop. there are plenty of people who are very angry at this change. But mostoF are network managers and you don't really see the anger very publicly.    " >Who in the hell would you contactH > to complain about something like this?  My Senator and Rep would never > understand this...  ( Comments@icann.org is the primary place.  J it was mlarson@verisign.com who proudly made the announcement in a network! manager's mailing list yesterday.p  B However, I think that pressure on the media to point to verisign'sJ incompetence and backlash that has resulted would put a lot of pressure onS verisign since its stock might thumble at the news that it has ruined the internet.o   ------------------------------  % Date: Tue, 16 Sep 2003 17:12:15 -0400u* From: JF Mezei <jfmezei.spamnot@istop.com>6 Subject: Re: OT: Verisign messes with DSN root servers) Message-ID: <3F677C9C.9EC3F2D0@istop.com>s   Rick Jones wrote:aH > Do the top-level registries get their charter's from the IANA or is it > another of the Internet TLAs?d   http://www.icann.org/tlds/  M Seems that ICANN is the general authority that delegates to the organisations0B listed on that page. Verisign has responsability for .com and .net   ------------------------------  # Date: Wed, 17 Sep 2003 01:17:29 GMT 0 From: "Matt Muggeridge" <Matt.Muggeridge@hp.com># Subject: Re: personal name in TO: ?e@ Message-ID: <JCO9b.107321$bo1.105387@news-server.bigpond.net.au>   JF,o  K It worked for me.  I tried it on a V5.3 ECO-2 and pre-released V5.4 system. G I used the TCPIP$SMTP_FROM logical to define my username with the extra  info.   E When I sent to my Exchange account, the extra info had been stripped. F However, when I looked at the RFC header, I noticed the info field wasL in-tact.  I'm not sure where the filtering was occuring, but VMS was sending it.R  J I then sent to my yahoo account and I saw the information was displayed as	 expected.h   Matt.r   -- g= -------------------------------------------------------------i OpenVMS TCP/IP Engineering Enterprise Computing Group Hewlett-Packard CompanyD Gold Coast, AUSTRALIA = -------------------------------------------------------------M    7 "JF Mezei" <jfmezei.spamnot@istop.com> wrote in messagec# news:3F66B0CD.777FFAF1@istop.com...e? > On an SMTP mail client, one can enter an internet address as:t >b# > Chef Jacques <chef@chocolate.com>m >i; > is there a way to enter a similar address from vms mail ?r >h
 > If I try. > TO: smtp%"Chef Jacques <chef@chocolate.com>" >yI > then the "Chef Jacques" seems to get removed and the received message'sKD > headers only contain the actual email address (chef@chocolate.com) > F > Is there another syntax one can use so that the VMS SMTP sender will preserveJ > the personal name for the RFC headers while using only the email for the SMTP > commands such as RCPT TO: ?e   ------------------------------  % Date: Tue, 16 Sep 2003 21:47:46 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com># Subject: Re: personal name in TO: ?m( Message-ID: <3F67BD1C.C162BBF@istop.com>   Matt Muggeridge wrote:M > It worked for me.  I tried it on a V5.3 ECO-2 and pre-released V5.4 system. I > I used the TCPIP$SMTP_FROM logical to define my username with the extra3 > info.c  M Actually, I was talking about specifying a personal name in the TO field. ForoF the From field, it picks it up from the MAIL profile and then uses the9 TCPIP$SMTP_FROM internet address so that part works fine.t   ------------------------------  % Date: Tue, 16 Sep 2003 14:31:24 -0400 0 From: "Brian Tillman" <Tillman@sparkingwire.com>' Subject: Re: Reimport VMS mail problems $ Message-ID: <3f675700$1@news.si.com>  * >Why do you think he said that? He didn't.   Because I read his sentence:    F >On OpenVMS V7.3-1, I extracted my mails ( EXTRACT/ALL myfile.txt ) to >save disk space.e -- sI Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot com 5 Smiths Aerospace  Addresses modified to prevent SPAM.aD 3290 Patterson Ave. SE, MS 1B3 Replace "at" with "@", "dot" with "." Grand Rapids, MI 49512-1991e8        This opinion doesn't represent that of my company   ------------------------------   Date: 16 Sep 2003 18:17 CDTe' From: carl@gerg.tamu.edu (Carl Perkins) ' Subject: Re: Reimport VMS mail problemsb- Message-ID: <16SEP200318175699@gerg.tamu.edu>   Y In article <3f675700$1@news.si.com>, "Brian Tillman" <Tillman@sparkingwire.com> writes...r+ }>Why do you think he said that? He didn't.- }  }Because I read his sentence:  o } G }>On OpenVMS V7.3-1, I extracted my mails ( EXTRACT/ALL myfile.txt ) tom }>save disk space. }-- J }Brian Tillman         Internet: Brian.Tillman at smiths-aerospace dot com   To which you replied:   L }Why do you think that extracting the mail and reimporting it will save disk }space?a   Which is not the same thing.  D It is not what you said: extracting + reimporting = save disk space.< It is instead: extracting and compressing = save disk space.  D He extracted the mail and compressed the resulting text file to saveD space. And it does save space. Probably more than 50%. There is onlyI one assumption: that after extracting them to the text file the originals E were deleted an the space reclaimed (the COMPRESS command in mail cancG be very nice to use after you have deleted a bunch of messages when youFH are short on disk space, or disk quota). This is a safe assumption sinceG if they were not deleted he wouldn't need to import them to get at them " since they would already be there.   --- Carl   ------------------------------  # Date: Wed, 17 Sep 2003 02:45:17 GMTo0 From: Randy Park <rjpark@mindspring.nospaam.com>, Subject: Remove satellite node from cluster?8 Message-ID: <3vhfmvk39dfk3n31thqiecekka66a9q26h@4ax.com>  > I'm getting DCL error while using CLUSTER_CONFIG.COM to remove a satellite node.i  - Boot node:  VaxStation 3100/30 VMS 6.1 (16MB)s$ Satellite node:  MicroVax 2000 (4MB)  E Ok, their old, but they work.  I'm getting short on disk space on theeA boot disk of my VaxStation and since I don't use or need the 2000r@ any more, I figured I'll just remove that node from the cluster.  : So I started to run CLUSTER_CONFIG.COM to remove the node.A It prompts me for the node name to remove, and I enter it.  A fewoC seconds later it appears to enter an infinite loop with a DCL error ; of "DCL-W-SYMDEL".  Note that the satellite node is not up.o  B I can't just delete the files in [sys10...] and [sys010...] can I?= Aren't there a lot of aliases?  Should I do a SET FILE/REMOVE  on all those files?.  9 Or should I try to copy the CLUSTER.COM from my 6.2 AlphaeB system and run it on the VaxStation?  It's about 80 blocks longer.  > I would really like to recover the disk space.  Any help would+ be appreciated.  Thanks.  No e-mail please.    ------------------------------  % Date: Tue, 16 Sep 2003 23:53:24 -0400u* From: JF Mezei <jfmezei.spamnot@istop.com>0 Subject: Re: Remove satellite node from cluster?) Message-ID: <3F67DA86.C1314C46@istop.com>s   Randy Park wrote:nD > I can't just delete the files in [sys10...] and [sys010...] can I?? > Aren't there a lot of aliases?  Should I do a SET FILE/REMOVE  > on all those files?   I Inside the root directory , there is directory [SYSCOMMON], syscommon.dirsN isn't a file, it is an alias, so you need to use set file/remove for that one.3 But the rest seems to be quite normal directories. s  I There shouldn't be that many big files. (except for any page/swap files).)  H However, removing the root directory will still leave some lint here andH there. For instance, the MOP load (either in NCP or LANCP), and the nodeC definition in the cluster authorize.dat, as well as decnet etc etc.r   ------------------------------  % Date: Tue, 16 Sep 2003 22:23:23 +0300G" From: Guy Peleg <guy.peleg@hp.com>E Subject: Re: Strange DCL Problem:  DCL Can't Find My GOSUB Labels...."& Message-ID: <3F67632B.ED3A4E9D@hp.com>   GOSUB LABEL only.o   Guy    Michael Unger wrote:  ) > On 2003-09-16 07:09, "Guy Peleg" wrote:e >eL > > It is possible that DCL won't recognize a label although it is there andK > > it is legal. It will take about an hour trying to explain why, but  try< > > looking forsK > > a gosub statement which is jumping to a very remote label. Let's assume = >                                           ^^^^^^^^^^^^^^^^^n > > thatJ > > the first line of a command file is GOSUB LABEL and the label LABEL is > > definedeM > > after 2000 lines. While DCL is scanning the procedure for the label LABELh > > it will defineL > > any label it encountered in the way. Some labels might be defined in the > > wrong resultingiM > > in the USGOSUB error. If you have a valid support contract you may file a-
 > > call witheK > > your local support center and have the call escalated to engineering. In > > will provide you withH+ > > a new DCL.EXE that might behave better.s >-H > Is that true only for "GOSUB label" or for "CALL subroutine" too? YourM > explanation seems to imply it is independent of how the "big jump" is done.e >d	 > Michaela >r > --= > Real names enhance the probability of getting real answers.3B > Please do *not* send "Security Patch Notifications" or "SecurityC > Updates"; this system isn't running a Micro$oft operating system. ? > And don't annoy me <mailto:postmaster@[127.0.0.1]> please ;-)n   ------------------------------  % Date: Tue, 16 Sep 2003 12:38:59 -0700I( From: Alan Frisbie <Abuse@NelsonUSA.com>0 Subject: Re: Summary: KVMs, LK461 and DECwindows, Message-ID: <3F6766D3.7070902@NelsonUSA.com>   Ken Fairfield wrote:E >     Some time ago, I started a thread titled, "KVM between PWS600aum@ > and IBM T20 Laptop".  I now have an update and perhaps another > question or two. >  > Executive Summary: > ? >     The Rose Mini-Vista KVM "works" well between a PC runninglD > Windows XP and an Alpha running VMS and DECwindows.  However, whenE > using an LK461-A2 keyboard, the 6-key editing keypad, function keysaG > F18-F20, and the numeric keypad minus keys are mapped incorrectly on mI > the Alpha, and the top-row functions keys from F13 through F17, and the ( > Compose key, are not available at all. > = >     Most KVM vendors support only a standard PC 101/102-keyrD > keyboard, but in there higher-end KVM series, Rose say they fully  > support the LK461 keyboard.   I I am using the Raritan MX48 (Four consoles connecting to eight computers)UF and am happy as a clam (ignoring the issue of why or when a clam mightF be particularly happy).   Both here at the client's site, and at home,G I am using it with a mix of Alpha/VMS, PC/Windows and PC/Linux systems.a) All the consoles have LK461-A2 keyboards.h  I The only "gotcha" is similar to the one Ken described above.   The fix is5G to have each Alpha system use a Raritan APKME module, connected in-line:D with the keyboard only (not the mouse).   I have found these modulesC on ebay for $10-15 each.   I got my personal MX48 on ebay for $450,t including eight cables.r  G According to an engineer at Raritan, they have a microcode fix for somenD of their KVM switches that avoids the need for the APKME module, butD that they have not yet inserted that fix into the code for the MX48.  C They also make a MX416 and MX432, which I see occasionally on ebay.i   Alan   ------------------------------  % Date: Tue, 16 Sep 2003 16:21:39 -0700k, From: Ken Fairfield <My.Full.Name@intel.com>0 Subject: Re: Summary: KVMs, LK461 and DECwindows) Message-ID: <3F679B03.3FB7072E@intel.com>a   Alan Frisbie wrote:s >  > Ken Fairfield wrote:G > >     Some time ago, I started a thread titled, "KVM between PWS600auwB > > and IBM T20 Laptop".  I now have an update and perhaps another > > question or two. > >, > > Executive Summary: > > A > >     The Rose Mini-Vista KVM "works" well between a PC running F > > Windows XP and an Alpha running VMS and DECwindows.  However, whenG > > using an LK461-A2 keyboard, the 6-key editing keypad, function keys H > > F18-F20, and the numeric keypad minus keys are mapped incorrectly onK > > the Alpha, and the top-row functions keys from F13 through F17, and thet* > > Compose key, are not available at all. > >o? > >     Most KVM vendors support only a standard PC 101/102-key-E > > keyboard, but in there higher-end KVM series, Rose say they fully< > > support the LK461 keyboard.c > K > I am using the Raritan MX48 (Four consoles connecting to eight computers) H > and am happy as a clam (ignoring the issue of why or when a clam mightH > be particularly happy).   Both here at the client's site, and at home,I > I am using it with a mix of Alpha/VMS, PC/Windows and PC/Linux systems.m+ > All the consoles have LK461-A2 keyboards.w  A    And you use all the top-row function keys in EVE?  That's goodl6 to hear, Alan (but I thought you were an EDT-er :-).    A    It's unfortunate that the tech support person I talked with at-; Raritan told me flatly that only 101/102-key keyboards were @ supported.  I would have rather he said, as the Rose fellow did,B that he would check with Raritan engineering and get back to me...* Another lost sale due to misinformation...   	-Kene --6 I don't speak for Intel, Intel doesn't speak for me...  
 Ken Fairfieldi" D1C Automation VMS System Support  kenneth.h.fairfield#intel.comt   ------------------------------  % Date: Tue, 16 Sep 2003 17:07:18 -0400 ( From: David Froble <davef@tsoft-inc.com> Subject: Re: VMS and MP3st, Message-ID: <3F677B86.2020900@tsoft-inc.com>   David J Dachtera wrote:x  v > pmoreau@ath.cena.fr (Patrick MOREAU, CENA Athis, Tel: 01.69.57.68.40) wrote in message news:<Czm8H9nux8$E@sinead>... > 6 >>In article <wf5yHbu5AkrR@eisner.encompasserve.org>, 8 >>kuhrt@nospammy.encompasserve.org (Marty Kuhrt) writes: >>[...]  >>7 >>>Did you rip your CD's on the VMS box?  If so, how?  s >>>oO >>I don't know if CD rippers works on VMS, but once you have WAV files, you cannO >>compress them with lame on VMS (I'll need to post a more recent version than  @ >>the one on the DECW archive, the 3.93 works well on VMS ...).  >> > 4 > Has anyone posted a "How-To" page for lame on VMS? >  > D.J.D. >   N Has anyone posted a set of function definitions for these (I'm guessing) Unix * functions with the o-so-descriptive names?   -- e4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Roadc Vanderbilt, PA  15486s   ------------------------------   End of INFO-VAX 2003.515 ************************