0 INFO-VAX	Mon, 20 Jan 2003	Volume 2003 : Issue 40      Contents:5 Re: (OT) lots of spam in vmsnet.sdk.openvms.fieldtest 5 RE: (OT) lots of spam in vmsnet.sdk.openvms.fieldtest 2 Re: Alpha, Itanic and Opteron benchmark comparison2 Re: Alpha, Itanic and Opteron benchmark comparison" Re: ALphaServer 2100A Disk Options) Re: Altavista Personal Search substitute? * BACKUP/IMAGE - Estimating time to complete" Compiling Mozilla on OpenVMS 7.3-1. EV7 hits, blows away Sun, IBM, 2-3x VMS gains! Re: How to Backup OSX  Re: How to Backup OSX  Re: How to Backup OSX  Re: How to Backup OSX  Re: Is that possible Re: Is that possible+ RE: Itanium ahead of itself, not behind ... - Re: Itanium with alpha goodies on the way ... - Re: Itanium with alpha goodies on the way ... - Re: Itanium with alpha goodies on the way ...  Re: MAIL 11 protocol definition  Re: MAIL 11 protocol definition  Marvel Performance' Re: MicroVAX upgrade (to Alpha, maybe?) ' Re: MicroVAX upgrade (to Alpha, maybe?) ' Re: MicroVAX upgrade (to Alpha, maybe?) & Re: MX 5.3-2 ignoring vmsmail forwards. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. RE: Network Time Protocol (NTP) for OpenVMS ??. Re: Network Time Protocol (NTP) for OpenVMS ??. RE: Network Time Protocol (NTP) for OpenVMS ??  no wonder Intel wanted alpha ... OpenVMS Alpha V6.2 media wanted P OpenVMS.org: Marvel article and HP's press release for Marvel and Alpha Retain TP Re: OpenVMS.org: Marvel article and HP's press release for Marvel and Alpha RetaP Re: OpenVMS.org: Marvel article and HP's press release for Marvel and Alpha Reta+ Re: Oracle LMON and LMDO buffered I/O usage + Re: Oracle LMON and LMDO buffered I/O usage 4 Re: osu/http-server + gecko browser support for .csv3 Security threats seen big for 2003 ... not for VMS! 7 Re: Security threats seen big for 2003 ... not for VMS! L TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet cardP Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet cardP Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet cardP Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet cardP Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet card  F ----------------------------------------------------------------------  % Date: Mon, 20 Jan 2003 10:58:16 -0000 2 From: "Chris Sharman" <chris.sharman@sorry.nospam>> Subject: Re: (OT) lots of spam in vmsnet.sdk.openvms.fieldtest4 Message-ID: <b0gkno$9g5$1$830fa7b3@news.demon.co.uk>  1 "Shane Smith" <ssmith@icius.com> wrote in message * news:01C2BE46.85281560@sulfer.icius.com...C > Incidentally, William, taxidermy may not be involved, but there's C > already an electronic "severed heads" page for usenet abusers who B > tangled with the alt.gothic "special forces". Have a chuckle at:- > http://thingy.apana.org.au/~fun/agsf/heads/  > (much snipped)  L The latest 'incident' logged there is very early '98 - presumably long dead.   Chris    ------------------------------  % Date: Mon, 20 Jan 2003 10:45:37 -0800 $ From: Shane Smith <ssmith@icius.com>> Subject: RE: (OT) lots of spam in vmsnet.sdk.openvms.fieldtest0 Message-ID: <01C2C071.3681F840@sulfer.icius.com>  B It's an attractive thought, but unfortunately impractical. In thisH instance, what would you do, block all of Spain from the internet? We'reA not talking about a group you can easily throw a blockade around. D They're all over the place and mostly on DHCP addresses so you can'tE block them for long. And if you try, you block the next person to get 0 that address assigned, who hasn't done anything.   Shane    -----Original Message-----0 From: bill@cs.uofs.edu [mailto:bill@cs.uofs.edu]) Sent: Saturday, January 18, 2003 10:01 AM  To: Info-VAX@Mvb.Saic.Com > Subject: Re: (OT) lots of spam in vmsnet.sdk.openvms.fieldtest    * In article <b0c06e$2r$1@aquila.mdx.ac.uk>,. 	david20@alpha1.mdx.ac.uk (David Webb) writes: > G > If they have to do this then they should setup their own alt group !! H > They should definitely not be doing it in any of the other hierachies.  B And just how would they get News Admins worldwide to create and/orC carry a newsgroup that's sole purpose is to violate copyright law?? C More and more News Admins are refusing to carry the alt.* hierarchy B at all and without wide distribution it doesn't do what they want,D thus, they barge in on legitimate groups with good distributions and make them useless.  @ Personally, I think the civilized world should start cutting offC routing to those areas (and we all know where they are) that refuse B to play by the rules.  When they have no access at all, maybe theyC will consider cleaning up their act.  None of the areas in question B have anything they offer on the INTERNET that outweighs the damage they do.  F This is nothing new, the UDP (USENET Death Penalty) has been effectiveE in many cases and I think it should be extended to the INTERNET.  Not = to mention how well "shunning" has worked for the Amish.  :-)    bill   --  C Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three  wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  # Date: Mon, 20 Jan 2003 06:49:43 GMT - From: bdc@world.std.com (Brian 'Jarai' Chase) ; Subject: Re: Alpha, Itanic and Opteron benchmark comparison & Message-ID: <H902yv.6vL@world.std.com>  O In article <3E271256.1040803@vajhoej.dk>, Arne Vajhj  <arne@vajhoej.dk> wrote:  > Shane Smith wrote:  J > > Unfortunately (at least for me) in German, but the picture says enough > > to be interesting. > > 8 > > http://www.heise.de/newsticker/data/as-16.01.03-002/ > > & > > It's labeled SAP SD @ tier, 4 way: > > # > > HPServer rx5670 Itanium 2   470 # > > Alpha Server ES45           426 # > > Proliant DL590 Itanium      206 # > > IBM xSeries 440             330 # > > AMD Opteron 1.6ghz        ca600  >  > So:  > 1. Itanium
 > 2. Alpha > 3. PowerPC > 4. x86 > < > (Opteron is just a chip not a system I can go out and buy)  G There aren't any PowerPC systems being compared in the original list.   E The IBM xSeries systems are Intel based.  The IBM pSeries systems are  Power based.   -brian.  --  F --- Brian Chase | bdc@world.std.com | http://world.std.com/~bdc/ -----H This counter is [6,177,399,753] times as pointless as a real one.  -- K.   ------------------------------   Date: 20 Jan 2003 09:28:06 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren); Subject: Re: Alpha, Itanic and Opteron benchmark comparison 0 Message-ID: <b0gff6$ama$1@pegasus.csx.cam.ac.uk>  ) In article <b0facv0214c@drn.newsguy.com>, ' Alan Greig  <a.greig@virgin.net> wrote: I >In article <dYCTgppwaxSu@elias.decus.ch>, p_sture@elias.decus.ch says...  >>> E >>"Capellas: Mein style is now once very informal. I love skirt music 7 >>over all, it runs permanently with me in the office."  > O >It then went on to say "I am a cheerleader", IIRC, just to complete the image!   ? Well, yes.  That was precisely what his role was in the New HP!   C Admittedly, the thought of Capellas in a tutu isn't attractive ....      Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679    ------------------------------  # Date: Mon, 20 Jan 2003 14:53:23 GMT A From: "Colin Butcher" <colin_DOT.butcher_AT@xdelta_DOT.co_DOT.uk> + Subject: Re: ALphaServer 2100A Disk Options = Message-ID: <D3UW9.2030$AY4.13317346@news-text.cableinet.net>   K Assuming that you're using the SWXCR RAID controller then that will support J 9Gb discs, unfortunately no bigger. However, I'm a firm believer in having@ your storage subsystems outside the box and using external arrayF controllers, such as the HSZ family. I'd recommend shopping around for- second user HSZs and discs / storage shelves.    -- Hope this helps. Cheers, Colin. ' (colinDOT.butcherAT@xdeltaDOT.coDOT.uk)    ------------------------------  % Date: Mon, 20 Jan 2003 18:00:40 +0100 1 From: SAP Trainee <Didier.Morandi.nospam@Free.fr> 2 Subject: Re: Altavista Personal Search substitute?& Message-ID: <3E2C2B38.2080206@Free.fr>  F Thanks for the pointer. I tried it. Sounds good. However, it does not  use nor create an index.  B But I found that within XP there IS an indexed search function. I C activated it and asked for a daily update. As I have now more than  G 390'000 files on my PC (because of SAP and its doc) the system is more  # busy doing I/Os than CPU cycles :-(    D.   Michael Unger wrote:H > There is a PC based tool called "superior search" which claims to findG > text and files quickly and reliably, even in compressed formats (ZIP,  > PDF, ...): >  > http://www.superiorsearch.com  > / > (page available in English, German and Dutch)  > H > I had just heard of it but never used it. As far as I know it is about7 > 30 Euros and there seems to be a trial version of it.    ------------------------------    Date: 20 Jan 2003 08:39:22 -0800' From: ultrajoe@spamcop.net (Joe Sewell) 3 Subject: BACKUP/IMAGE - Estimating time to complete = Message-ID: <a55b951e.0301200839.17edda91@posting.google.com>   F I've been assigned to add to an existing tool a more accurate means of? estimating the percent complete of a disk-to-disk BACKUP that's F currently cooking in a batch job.  The backup isn't using tape, nor isE it reading from nor writing to a saveset.  I already have the code in - place to ferret out the PID of the batch job.   D What I've been doing is using the number of I/O operations (direct +A buffered) as a guesstimate, using numbers empirically obtained to D convert that to number of blocks transferred.  Since I know how many@ blocks need to be transferred, that has given me a fairly decent- estimate of how far the operation has gotten.   @ That is, until the powers-that-be decided the destination shouldB always be ODS-5, not ODS-2.  Experimentation has led me to believeB that the characteristics are dramatically different for ODS-5, andA that the number of I/O operations is no longer a good determining  factor.   3 Is there any better way to handle this requirement?    ------------------------------  + Date: Mon, 20 Jan 2003 12:45:31 +0000 (UTC) 2 From: setala@phys-staff7.kolumbus.fi (Saku Setala)+ Subject: Compiling Mozilla on OpenVMS 7.3-1 1 Message-ID: <b0gr1b$hcd$1@phys-news1.kolumbus.fi>    Hi all,   A some time ago I remember looking into a document which instructed , how to compile mozilla source under OpenVMS.  : Now I can't locate this document anymore, does anyone have	 pointers?    Thanks.      --   Saku Setl  System Planning Manager ) Network Services, Elisa Internet Oy (LTD)    ------------------------------    Date: 20 Jan 2003 10:10:54 -0800( From: bob@instantwhip.com (Bob Ceculski)7 Subject: EV7 hits, blows away Sun, IBM, 2-3x VMS gains! = Message-ID: <d7791aa1.0301201010.54c780ca@posting.google.com>   8 HP's AlphaServer GS1280 Launch: A MARVELous Development C Posted by Terry Shannon (Sunday January 19 2003 @ 03:51PM EST) [ ]     = It's been a long time coming, but the enterprise server first > discussed a decade ago at Digital Equipment Corporation is now@ reality. As predicted in SKHPCin which we've been writing aboutE Marvel for nearly two years nowHP today opened a barrel of whoop-ass E on competitive enterprise server vendors with the public debut of the E HP AlphaServer GS1280 "Marvel" system family. The first three members D of the Alpha EV78-inside follow-on to the AlphaServer GS320 WildfireB enterprise server have been up and running in nearly 10 field test@ sites worldwide for nearly a year, and two through 16-way MarvelB systems are now ready for prime time. Here's SKHPC's first take onD HP's evolutionaryand some would argue revolutionaryreplacement for the GS320 family.    Evolution and Revolution    = In fact, the new HP AlphaServer GS1280 family represents both E evolution and revolution. Indeed while these new systems represent an @ evolutionary follow-on to the GS320 family, the AlphaServer 1280A incorporates three major technology revolutions which will impose D minimal customer disruption. First among these is "built-in SMP." InA an effort to build a platform that delivers industry-leading apps E performance and far more scalable performance than its predecessor or B competitive products, HP aggregated all the constituents of an SMPA system onto a single Alpha EV78 CPU. These components include the ? memory controller, the L2 cache controllers, and interprocessor   communications and coordination.  A Around the EV78 processor, HP developed an environment it calls a ? switch-less meshed architecture that dispenses with much of the ? infrastructure, complexity, and expense of incumbent enterprise A servers. By interconnecting dual-processor modules with dedicated = memory and I/O directly to one another, HP managed to deliver F virtually truly linear scalability because there's nothing between oneF processor module and another except a cable. In addition, HP engineersC did significant work on the switchless architecture to fault harden 9 the environment to make it even more robust and reliable.    Simplicity = Stability  F The "Lego Block" approach to GS1280 design provides a number of unique? benefits to customers. System reliability dramatically improved @ through simplicity: since a single EV78 module contains multiple@ components that existed as discrete entities on GS320 systems, aC GS1280 system contains fewer components than its predecessor. Fewer @ components translate to fewer opportunities for failure, hence a> GS1280 system provides a very significant increase in inherentF single-system reliability. What's more, key components including CPUs,F redundant power supplies, fans, and I/O modules are hot-swappable, andC an advanced diagnostics subsystem not only monitors virtually every F aspect of system performance, but provides prefailure information thatC allows customers to take components offline and replace them before  they actually fail.   = The GS1280 family also sports optional RAID memory, a concept @ pioneered by HP's ISS group in IA-32-based systems. Based on the@ architectural simplicity of the AlphaServer GS1280 alone, HP hasD experienced a 15 to 30 percent improvement in MTBF in the new system> when compared with the more complex AlphaServer GS320. The newD instrumentation, console, and RAS features further contribute to the> stability and reliability of the GS1280 family. Additional RAS? features will be exploited by future versions of Tru64 UNIX and  OpenVMS opating systems.   Fast and Frugal Scale-Up  B Finally, HP is basing the AlphaServer 1280 family on several basicF building block modules that provide a foundation for building scalable= systems with reusable components. Significantly, these shared D components mean that customers need not purchase more infrastructureE than they need, nor will they confront daunting upgrade costs as they A grow their GS1280 systems to meet growing demand. For instance, a B customer can start with the smallest GS1280 system (and its modestD entry price) and grow it to the largest supported configuration in aE modular way that protects the vast majority of initial investments by A carrying all of the core products and modules and building blocks ? forward. At the same time, the GS1280 is binary-compatible with E incumbent Alpha systems, rendering the system upgrade path smooth and  free of pitfalls.    Benchmarks Eclipse Sun, et al   ? The new HP AlphaServer GS1280 systems have established industry ? leadership in both the SPEC_rate2000 and the STREAM benchmarks, A demonstrating their outstanding "whole" systems performance and a F unique ability to scale linearly in performanceand nearly linearly inF price as additional processor modules are added. While cache-intensiveE apps will demonstrate little improvement over EV68-based systems (the B EV7 CPU uses the EV68 core and contains a 7-way associative 1.75MB< cache while the latest EV68 sports a 32MB cache), real-world= application performance benchmarks shine on the GS1280, as do B SPEC_rate2000 and STREAMS numbers. In the STREAMS benchmark, whichB measures sustainable memory bandwidth for computer systems, the HP? AlphaServer GS1280 demonstrated five to 10 times greater memory E bandwidth than comparable enterprise systems from IBM and Sun. Due to A the new system architecture, memory bandwidth increases as GS1280 ? systems get larger: when compared with a 4-CPU GS320, a GS 1280 C quadprocessor offers ~40 percent greater memory bandwidth. On 8-way @ systems, the GS1280 offers ~64 percent greater bandwidth, and onB 16-way configurations, the GS1280 delivers ~70 percent more memory( bandwidth than does an equivalent GS320.   Superlative SPEC_rate Results   B The GS1280's superlative SPEC_rate 2000 results go well beyond theF leadership performance of today's AlphaServer systems, particularly in@ technical computing where the benefits of the systems' technical@ innovations are well demonstrated with nearly 100 percent linear@ scalability from one to 32 processors. The vastly reduced memory> latency of the AlphaServer GS1280 is a key contributor to thisF substantial improvement. As further proof points, HP plans to announceB SAP and Oracle application benchmarks in the coming months. (SKHPC< expects to see 35-50 percent performance gains on Tru64 UNIXB applications.) Initial reports from field test sites indicate thatE early adopters are absolutely delighted with the GS1280 system. Tru64 A UNIX customers are extremely pleased. OpenVMS customers are blown A away: for example, Cerner Corporation saw a threefold increase in E OpenVMS performance during its Marvel field test; the Bank of Austria D saw VMS performance double on an HP AlphaServer GS1280 equipped withC eight CPUs, and a 16-way system showed a performance gain of 150 to 2 200 percent over an equivalently-configured GS320.   Customer Assurance Ensured  F The new HP AlphaServer GS1280 is one of the cornerstones of HP VP RichF Marcello's RetainTrust Program. While previously reported in detail inF SKHPC, the RetainTrust program was formally unveiled during today's HPF AlphaServer GS1280 launch. With the initial eight to 16-way members ofD the HP GS1280 family available today, smaller systems in March, thenE 32-way systems in midyear and 64-way behemoths by the end of the year F (with several per-processor memory capacity increases slated to appearF this year), not to mention new releases of OpenVMS and Tru64 UNIX thatE better exploit the GS1280, Alpha customers should feel confident that B HP will be with them for the long haul. In addition, HP is alreadyC sampling prototype EV79 parts. EV79-Inside GS1280 systems should be D available in late 2004, thus providing another performance boost forE OpenVMS and Tru64 customers, and increasing the length of their Alpha E to IPF transition window. As things stand today, Alpha GS1280 systems D will be sold through at least 2006 and supported for at least half a? decade thereafter. Based on past experience with PDP-11 and VAX @ systems, SKHPC wouldn't be surprised to see GS1280 systems stillE available well into the second half of the decade. Accordingly, while D OpenVMS and Tru64 users ultimately will migrate to OpenVMS-on-IPF orF to the Consolidated Enterprise UNIX on IPF, the AlphaServer GS1280 andB its EV79-based successor remain very safe bets, and should be moreE than capable of meeting customer needs until such time as they choose B to migrate to HP's next-generation IPF systems. So yet again, HP'sE "promises made are promises kept" strategy remains in play. As one HP D executive said, "If there's a problem with application compatibilityB or functionality on the GS1280, it's HP's problem, not the ISV's."    Order Quick, They're Going Fast!  A The high-end AlphaServer GS1280 is already shipping in eight- and B 16-way configurations sporting 1.15 GHz Alpha EV78 processor-basedD systems. U.S. list prices for this highly-expandable system start atD $117K for a two-CPU configuration. GS1280 systems with 32 processorsF should be available by the middle of the year and 64-processor systems; are slated to ship by the end of the year. AlphaServer ES80 E departmental systems with up to eight 1GHz Alpha EV78 processors will 8 start shipping in March with prices starting at $81K forE dual-processor configurations (by comparison, equivalently-configured F incumbent GS80 systems cost about 20 percent more.) The ES47 workgroupB systems also are available now and come in two and four 1GHz AlphaE processor configurations with U.S. list prices starting at $39.7K for F a two-processor tower system, just slightly higher than the price of aB dual-processor DS25 system. An ES47 rackmount system with two CPUs? goes for about $64.4K, similar to the price of a dual-processor F AlphaServer ES45. Quadprocessor ES47 rackmount systems cost from $133K	 to $137K.   & Pricing Pressure... on the Competition  @ The price differential is even more significant in larger systemF configurations, a marketing decision that should attract new customersB while causing major headaches at IBM and Sun Microsystems. Case inF point: when compared with an eight-processor GS320, an eight-processor@ GS1280 is priced 60 percent lower than its predecessor; a 16-CPUC GS1280 is yours for 28 percent less than a 16-processor AlphaServer  GS320.   An Enterprise Server Done Right   A SKHPC has been tracking high-end systems from DEC, Compaq, HP and F their rivals since well before the October 1984 debut of the VAX 8600.D In comparison with previous VAX and Alpha offerings ranging from theB VAX 8800 and 9000 through the TurboLaser and Wildfire systems, the< AlphaServer GS1280 represents a quantum leap forward in manyC dimensions. In SKCHPC's opinion, the AlphaServer GS1280 is the best D enterprise server to be fielded by any systems vendor, and we expectC customer adoption rates will reflect the superiority of HP's latest F Alpha-Inside product offering for quite some time to come. In summary,D HP's latest entry in the enterprise systems derby is sure to concern; competitors while delighting customers. And as usual, SKHPC < subscribers will be provided with more details on the new HPE AlphaServer GS1280 system in future issues of this subscription-based  newsletter.   B (c) 2003 by Terry C. Shannon, IT Consultant and Publisher, Shannon	 Knows HPC   F Press Inquiries through 22 January: 505-400-8513 (cellular) 23 January and beyond 505-792-2324 (voice)   C Want ALL the details on Marvel and future AlphaServer and Superdome 4 systems? Easy: subscribe to SKHPC. Drop me a line at3 tshannon3@comcast.net for subscription information.    ------------------------------  # Date: Mon, 20 Jan 2003 12:00:20 GMT % From: Tony Lawrence <tony@pcunix.com>  Subject: Re: How to Backup OSX, Message-ID: <oxRW9.61882$hl1.6733@sccrnsc04>   Howard S Shubs wrote: . > In article <lXDW9.56181$1q3.7145@sccrnsc01>,) >  Tony Lawrence <tony@pcunix.com> wrote:  >  > ! >>You remain convinced otherwise.  >  > A > Yup.  I'm mired in reality, rather than ivory tower theorising.    THAT I cannot ignore :-)  E I'm an independent, and have been since 1983.  I'm not a var pushing  I product, I'm not a contract admin who stays at the same place for months  I at a time: I'm a guy who is out in the trenches every day, flitting from  B customer to customer ('cept nowadays my flitting is more ssh than  automobile).  C My comments about backup are based on my experience with literally  H thousands of customers.  In the old days (not all that long ago) I'd be F involved in hard drive failures at least once a month.  Nowadays it's H once a year, if that, but I've seen the problems with differentials and G incrementals and I DO NOT RECOMMEND THEIR USE UNLESS IT IS UNAVOIDABLE.   > If that's "ivory tower", I'm a 15th century blacksmith having  hallucinatory visions.  " That's it.  Go and sin no more :-)   --  
 Tony Lawrence , Free SCO, Mac OS X and  Linux Skills Tests: % http://aplawrence.com/skillstest.html    ------------------------------    Date: 20 Jan 2003 10:00:48 -0600+ From: young_r@encompasserve.org (Rob Young)  Subject: Re: How to Backup OSX3 Message-ID: <Hoh+qG8gl6mO@eisner.encompasserve.org>   T In article <oxRW9.61882$hl1.6733@sccrnsc04>, Tony Lawrence <tony@pcunix.com> writes:   > E > My comments about backup are based on my experience with literally  J > thousands of customers.  In the old days (not all that long ago) I'd be H > involved in hard drive failures at least once a month.  Nowadays it's J > once a year, if that, but I've seen the problems with differentials and I > incrementals and I DO NOT RECOMMEND THEIR USE UNLESS IT IS UNAVOIDABLE.  > @ > If that's "ivory tower", I'm a 15th century blacksmith having  > hallucinatory visions. >  	  	No, just very out of date.   C 	I do incrementals - always.  Folks are somewhat uncomfortable with A 	that.  The warm and fuzzy is to do (typical) weekly fulls, daily = 	incrementals back to the weeklies.  With TSM, you are always  	doing incrementals.  B 	The problem is conceptual.  Think of your backups taking place toA 	a fault-tolerant database.  Each day, an export goes off-site so E 	your backed up files reside in two places (this is typical for most  & 	of us regardless of backup solution).  ; 	What is occuring is files that are *not* changing are only + 	backed up initially!  incrementals forever   9 http://adsm.nerdc.ufl.edu/local/incrementals-forever.html   D 	Picking a file at random, one created in 1999, you can see when it  	was last backed up:  < $ abc show backup site$disk:[username.routines]TST_AA_IF.TXT; Archive Backup Client for ADSM on OpenVMS, Version V3.1.0.1 8 Copyright 1996-2001, Storage Solutions Specialists, Inc.N DISK$SITE_USERS:[USERS.USERNAME.ROUTINES]TST_AA_IF.TXT;1 (A) 6097 20-SEP-2001 	  10:40:43   E 	Last backed up over a year and a half ago.  I know I can restore it, : 	as the database/tapepool containing it is fault-tolerant.   	So, I disagree with this:  J > once a year, if that, but I've seen the problems with differentials and I > incrementals and I DO NOT RECOMMEND THEIR USE UNLESS IT IS UNAVOIDABLE.   C 	You have a poor backup solution in use, would be my first guess as , 	to why you so strongly oppose incrementals.   				Rob    ------------------------------  % Date: Mon, 20 Jan 2003 13:14:45 -0500 ' From: Howard S Shubs <howard@shubs.net>  Subject: Re: How to Backup OSX< Message-ID: <howard-92EBC4.13144520012003@enews.newsguy.com>  3 In article <Hoh+qG8gl6mO@eisner.encompasserve.org>, -  young_r@encompasserve.org (Rob Young) wrote:   > > $ abc show backup site$disk:[username.routines]TST_AA_IF.TXT= > Archive Backup Client for ADSM on OpenVMS, Version V3.1.0.1 : > Copyright 1996-2001, Storage Solutions Specialists, Inc.P > DISK$SITE_USERS:[USERS.USERNAME.ROUTINES]TST_AA_IF.TXT;1 (A) 6097 20-SEP-2001  >  10:40:43   E Note that this particular syntax is very system specific.  I've been  < trying to keep the discussion more general when I added the  cross-posting to COV.    --  4 Today, on Paper-view: The World Origami Championship   ------------------------------  % Date: Mon, 20 Jan 2003 13:13:03 -0500 ' From: Howard S Shubs <howard@shubs.net>  Subject: Re: How to Backup OSX< Message-ID: <howard-517C0F.13130020012003@enews.newsguy.com>  , In article <oxRW9.61882$hl1.6733@sccrnsc04>,'  Tony Lawrence <tony@pcunix.com> wrote:   E > My comments about backup are based on my experience with literally  J > thousands of customers.  In the old days (not all that long ago) I'd be H > involved in hard drive failures at least once a month.  Nowadays it's J > once a year, if that, but I've seen the problems with differentials and I > incrementals and I DO NOT RECOMMEND THEIR USE UNLESS IT IS UNAVOIDABLE.    It's almost always unavoidable.   H When you've got TB, or even a few hundred GB, to backup every night and I it has to go over a network, even GB ethernet isn't enough to get it all  G done in the wee hours.  You have to reduce the needed bandwidth to get  & it done.  Incrementals are sufficient.   --  4 Today, on Paper-view: The World Origami Championship   ------------------------------  + Date: Mon, 20 Jan 2003 11:00:39 +0000 (UTC) + From: david20@alpha2.mdx.ac.uk (David Webb)  Subject: Re: Is that possible + Message-ID: <b0gksn$fl5$1@aquila.mdx.ac.uk>   g In article <f948cf20.0301191442.3591d848@posting.google.com>, taupin974@hotmail.com (taupin974) writes:  >Hi = >I'd like to know if i can run open vms on a vax simulator as  >Charon-VAX, Ts10 or Simh ? > >If all programing tools of vms will work if it is possible !! >  >thx  D The emulator's emulate the VAX hardware. You then load VMS onto that "hardware". J Hence everything in VMS should work so long as the underlying hardware and emulator support it.H For instance networking obviously won't work if your PC doesn't have anyN networking or if (as I believe happened with the CHARON-VAX hobbyist version -L which is no longer available anyway) the emulator didn't support networking.    
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------    Date: 20 Jan 2003 10:22:58 -0800' From: taupin974@hotmail.com (taupin974)  Subject: Re: Is that possible = Message-ID: <f948cf20.0301201022.382b1d62@posting.google.com>   5 And what emulator are avable on my pc with windows ?!     ^ david20@alpha2.mdx.ac.uk (David Webb) wrote in message news:<b0gksn$fl5$1@aquila.mdx.ac.uk>...i > In article <f948cf20.0301191442.3591d848@posting.google.com>, taupin974@hotmail.com (taupin974) writes:  > >Hi ? > >I'd like to know if i can run open vms on a vax simulator as  > >Charon-VAX, Ts10 or Simh ? @ > >If all programing tools of vms will work if it is possible !! > >  > >thx > F > The emulator's emulate the VAX hardware. You then load VMS onto that
 > "hardware". L > Hence everything in VMS should work so long as the underlying hardware and > emulator support it.J > For instance networking obviously won't work if your PC doesn't have anyP > networking or if (as I believe happened with the CHARON-VAX hobbyist version -N > which is no longer available anyway) the emulator didn't support networking. >  >  > David Webb > VMS and Unix team leader > CCSS > Middlesex University   ------------------------------  % Date: Mon, 20 Jan 2003 10:47:59 -0800 $ From: Shane Smith <ssmith@icius.com>4 Subject: RE: Itanium ahead of itself, not behind ...0 Message-ID: <01C2C071.748B0320@sulfer.icius.com>  H But ahead compared to which generation of the roadmap? They've moved theE goalposts further out several times, so what if they're coming back a  little?    Shane    -----Original Message-----6 From: bob@instantwhip.com [mailto:bob@instantwhip.com]) Sent: Saturday, January 18, 2003 11:44 AM  To: Info-VAX@Mvb.Saic.Com 0 Subject: Itanium ahead of itself, not behind ...    7 if you are Bill Todd, then disregard this as pure b.s., 6 otherwise, I know I can read as others can and Itanium5 is actually ahead of schedule, not behind like others - (we know who) have asserted ... from cnet ...     # Intel accelerates Itanium schedule     By Michael Kanellos  Staff Writer, CNET News.com  January 15, 2003, 9:00 PM PT   Read more about chips     ? Intel has changed the release schedule of its Itanium chips for D servers, adding a new chip for 2004 and moving up the launch date of6 an Itanium with two processor cores to 2005 from 2007.@ The changes reflect Intel's confidence in its ability to releaseB high-end server chips faster than competitors and thereby gain theC performance high ground, said Jonathan Eunice, principal analyst at  Illuminata.   E Itanium 2 ranks with the best server chips in the market, but the new F release schedule will likely enhance the chip's attractiveness and putF pressure on competitors to step up their own schedules, something that they have been loath to do.        D Intel's "design teams and design resources are well stocked, so theyE can do a shrink early or do a dual-core (chip) early. They have a lot A of leeway that would stress out a Sparc development team," Eunice0D said, referring to shrinking the size of components on a chip and to' Sun Microsystems' UltraSparc processor.   D Under the new schedule, the Santa Clara, Calif.-based chipmaker this? summer will release Madison, a souped-up version of the currentp? Itanium 2 with 6MB of level 3 cache, according to Jason Waxman,lD marketing manager for enterprise processors at Intel. Increasing the@ cache, a reservoir of memory located on the processor, generally enhances performance.o  E The chip, which will contain around 500 million transistors, will runBE at 1.5GHz. Madison, like the entire Itanium family, is a 64-bit chip,aF meaning that it can digest data in 64-bit chunks (as opposed to 32-bitC Pentium chips). Typically, 64-bit chips fit into the most expensiveh and powerful servers.a  E Soon after, Intel will release Deerfield, an energy-efficient Itanium  2 for rack and blade servers.V  F Then, in 2004, the company will come out with a new version of MadisonC that will contain 9MB of level 3 cache. Most server chips currentlyi? come with 1MB of cache. This chip was previously not on Intel'se product road maps, Waxman said.n   Two cores better than oneoA In 2005, Intel will follow with Montecito, which will contain twoMD Itanium processor cores on the same piece of silicon. Dual-processorE chips are pretty much what they sound like: single chips that containnB two separate "brains" so they can best a single-core chip but cost less than two separate ones.  C It's one of the hot design ideas in the chip world. IBM has alreadyPB come out with the first dual-processor chip for the server market,D Power4. Sun's UltraSparc IV, debuting toward the end of the year, is( expected to contain two processor cores.  C Analysts have also noted that Advanced Micro Devices' Opteron chip,lB coming out in the first half, could be redesigned to accommodate a second processor core.  C Originally, Montecito, due in 2004, wasn't a dual-core chip, but itbF was morphed after engineering and manufacturing teams concurred that a@ dual-processor chip could be mass-manufactured at Intel by 2005.  > "Our dual-core (chip) was originally planned for the following" generation of chips," said Waxman.  C Besides enhancing performance, Intel may use its dual-core chips to1E undercut IBM, said Kevin Krewell, senior editor at the MicroprocessorrE Report. In larger servers, Oracle and other software vendors charge a:4 licensing fee for every processor in a given server.  F To date, IBM has said that the Power4, although a single chip, has twoB processors. Hence, software customers have to buy two licenses forC each Power4 chip. Intel is already indicating that it will considerSB Montecito a single processor, requiring only one software license,
 Krewell said.o  F "I think they are going to sell it that way to make it cost effective"E to switch to Itanium from other servers with different chips, Krewell5 said.e  C Montecito, Krewell added, will be made on the 90-nanometer process, ; which means the average feature inside the chip measures 90rC nanometers. Madison, Madison II and Deerfield will contain features : measuring 130 nanometers. The current Itanium 2 comes withA 180-nanometer features. Reducing feature size allows companies toe# squeeze more transistors on a chip.    Not as sweet by any other nameF Continuity is another theme. Madison, Deerfield and Madison II will be sold under the Itanium 2 name.  E All of the forthcoming chips, including Montecito, will also fit into6F the same motherboard sockets and be capable of using the same chipsetsD currently used in Itanium 2 servers, said Waxman. In turn, this willE reduce the need for server makers to redesign their servers with eachr4 new chip release, smoothing the commercial adoption.  D Despite its long and often controversial history, the Itanium familyE appears to be gaining momentum in the market, according to Eunice. In'C the mid-1990s, analysts speculated that the chip, designed by IntelcF and Hewlett-Packard, would become one of the most popular for high-endD servers. However, the first version, formerly code-named Merced, was? delayed several times and offered only middling performance. Ita+ finally debuted in 2001; sales were dismal.d  ; Itanium 2, formerly McKinley, came out in July 2002 and wassF substantially different from the first Itanium. Analysts have given it fairly positive reviews.  E Itanium 2 "competes or outperforms the fastest Alpha and Power chips.rF It is right up there at the elite country club of performance," Eunice said.d  ? Still, despite strong benchmark scores, sales started slow. ThelB economy was in a slump, which discouraged interest from customers,( software developers and hardware makers.  F Interest, though, appears to be growing, Eunice said. Laboratories and> other scientific customers are increasingly offering Itanium 2D servers. More software tools are also coming onto the market. If the" tide changes, Intel could benefit.  C "2002 was a terrible time to launch a new technology," said Eunice.    ------------------------------    Date: 20 Jan 2003 01:22:31 -0800' From: icerq4a@spray.se (David Svensson)(6 Subject: Re: Itanium with alpha goodies on the way ...= Message-ID: <734da31c.0301200122.6ac60da1@posting.google.com>n  M > Itanic looked better exactly once:  after McKinley appeared and proved thatnL > the architecture wasn't a complete dud (just a pig in terms of power, chipH > area, and development effort).  It is indeed all down-hill from there,G > because the McKinley core isn't going to change in Madison/Deerfield,eL > Madison II, or Montecito, but only get shrunk and acquire more cache (plusL > dual cores in 2005) - while its competition will be getting actual *designM > enhancements* (like SMT) as well as simple shrinks, dual cores, more cache,  > etc. >  > - bill  C I haven't seen anything that says that Montecito will have the samet core as McKinley?o   ------------------------------   Date: 20 Jan 2003 09:32:49 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)6 Subject: Re: Itanium with alpha goodies on the way ...0 Message-ID: <b0gfo1$aq9$1@pegasus.csx.cam.ac.uk>  = In article <734da31c.0301200122.6ac60da1@posting.google.com>, ( David Svensson <icerq4a@spray.se> wrote:N >> Itanic looked better exactly once:  after McKinley appeared and proved thatM >> the architecture wasn't a complete dud (just a pig in terms of power, chipaI >> area, and development effort).  It is indeed all down-hill from there,sH >> because the McKinley core isn't going to change in Madison/Deerfield,M >> Madison II, or Montecito, but only get shrunk and acquire more cache (plusaM >> dual cores in 2005) - while its competition will be getting actual *design N >> enhancements* (like SMT) as well as simple shrinks, dual cores, more cache, >> etc.- > D >I haven't seen anything that says that Montecito will have the same >core as McKinley?  E Intel have said that it will be 100% compatible and deliver the extraeB performance without recompilation (unlike McKinley), and used someB term like "only minor enhancements" which is generally reckoned to3 be marketese for "bug fixes and bigger tables etc."   F That was before the dual-core plan was announced, but I doubt that has changed anything.      Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679h   ------------------------------    Date: 20 Jan 2003 10:34:14 -0800( From: bob@instantwhip.com (Bob Ceculski)6 Subject: Re: Itanium with alpha goodies on the way ...= Message-ID: <d7791aa1.0301201034.73f40861@posting.google.com>3  d "Bill Todd" <billtodd@metrocast.net> wrote in message news:<Jh6dnWjpFfjiobajXTWcqw@metrocast.net>...7 > "Bob Ceculski" <bob@instantwhip.com> wrote in messagec9 > news:d7791aa1.0301190634.72f65155@posting.google.com...e >  > ...D > E > > and who in the heck cares?  as long as they get the thing runningi4 > > well (and they will with the help of alpha team) > L > Bob, you're ineducable.  The Alpha team has nothing - NOTHING - to do with+ > Montecito, whether single- or dual-cored.x >  >  and it runs vmsD > > as well as it runs now on alpha with improvements in the future, > M > It won't.  EV7 has performance characteristics in large systems that ItaniclI > won't begin to approach until such time (assuming that occurs) as thereaM > really *is* an Alpha-derived Itanic.  I've said for the past 18 months thatFL > wouldn't be before 2006 - 2007, and every indication still points to thoseI > dates.  (At one time I thought Itanic might at least get some EV7-styleHL > on-chip glue in 2005, but the dual-core Montecito announcement proves that > it won't.) >  > ...  >  >  by the looks of things thedG > > alpha team is slowly using EV7/EV8 alpha designs and slowly puttingeC > > them into itanium and the result is that itanium is starting toC > > look better and better > J > You utter idiot:  *NOTHING* from Alpha is making it into any Itanic thatK > Intel has said anything about in public, except possibly Chivano (and the B > 'Chivano' name is all that Intel has mentioned about that chip). >   B that is what I am saying idiot, that chivano and post chivano will> have alpha influence ... I have been saying it for over a year? now, and I will continue to ... EV7 is now in place to hold the-? fort for vms until it gets here (2005?) ... nothing has changeda in the game plan ...   ------------------------------  % Date: Mon, 20 Jan 2003 04:43:45 -040010 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>( Subject: Re: MAIL 11 protocol definition/ Message-ID: <3E2BB6BB.3DB7AD00@vl.videotron.ca>c   Antonio Carlini wrote:; > how up to date that document was). I don't think I have a 8 > copy of it anymore but you might be able to find it on3 > the Freeware CDs or just floating around the net.e  L NMAIL, available on the freeware, contains a few more details on the foreignK mail protocol. I was able to figure out how to access the TLD structure (mylM original doc was wrong, NMAIL sees the TLD as a pointer to a descriptor whose  contents are a tld).  L I can see 17 fields (when a foreign message is being sent, (numbered 1 to 17J oddly :-), but have no idea what they mean. It is supposedly documented inH some "appendix C" which wasn't released. (this is the function code 19).   Also, when I do :r  ? MAIL/FOREIGN NAKEDLADY.JPG  jfmezei/subject="Interesting image"f  N How does the IMAP server figure out what type of file it was ? It doesn't seem> like the original file name is stored anywhere in the message.  N For a non-decnet email, I can use the supplied RAB to find the FAB to find theN NAM and get a full file name. But when a  mail message is sent via decnet, the name is unavailable.    N Also, since an SMTP header depends on the message contents (whether the quotedM printable kicks in, whether the message contains attachements etc), one needsa1 to scan the message prior to building the header.s  K Does the MAIL protocol allow the use of SYS$REWIND on the file being sent ?c (even for decnet operations)    K For instance, I could do $GET for all records in a message to decide how tonN treat said message, then $REWIND , then build the header, then do $GET (again)* and apply the proper encoding to the data.  M If I can't use $REWIND, it means I need to copy the file to another temporarymM file while scanning it, and then open that temporary file for processing into 1 another temporary file which finally gets sent...r   ------------------------------    Date: 20 Jan 2003 04:58:14 -0800' From: mdeblis@hotmail.com (Mike Deblis)h( Subject: Re: MAIL 11 protocol definition= Message-ID: <bc45f679.0301200458.205319ac@posting.google.com>c  g JF Mezei <jfmezei.spamnot@vl.videotron.ca> wrote in message news:<3E2A1DD3.2C563C28@vl.videotron.ca>...aK > Years ago, I remember having a command procedure one could invoke to sendhF > email via DECNET and which could impersonate anyone. The goal was toO > document/demonstrate the MAIL-11 protocol between nodes (the procedure didn'tvG > invoke MAIL, it just wrote to a decnet link to a remote mail server).P > M > Unfortunatly, with lack of hindsight, I delete the file... Now, I'd like toyP > have it or an equivalent so I can better understand the protocol and hopefully$ > see some of the flags definitions. > O > Does anyone know of that command procedure or good definitions of the MAIL 112  > protocol at the DECNET level ?    ' Nick de Smith did this a while back...    : http://vms.process.com/scripts/fileserv/fileserv.com?PROTO   HTH    Mike   ------------------------------  # Date: Mon, 20 Jan 2003 13:50:17 GMTa# From: "John Smith" <a@nonymous.com>l Subject: Marvel PerformanceeH Message-ID: <t8TW9.51384$ej1.48671@news02.bloor.is.net.cable.rogers.com>  4 http://h18003.www1.hp.com/hps/announce/jan_perf.html   Found this on HP's web site.  D No news announcement on Marvel from the HP home page news link. They? make more fuss about a new printer cable than they do about thee/ fastest general purpose computer in the market.h   ------------------------------  % Date: Mon, 20 Jan 2003 13:04:51 -0500i# From: Tom Rymes <tomnews@rymes.net>e0 Subject: Re: MicroVAX upgrade (to Alpha, maybe?)E Message-ID: <tomnews-FDD395.13045120012003@news.comcast.giganews.com>a  E In article <tomnews-39F4DC.18560317012003@news.comcast.giganews.com>,o%  Tom Rymes <tomnews@rymes.net> wrote:    <snip>  eE > 2.) What is the likelihood that we could move to an Alpha platform hG > instead? Licensing issues? Would our software run on Alpha, or would pE > there be recompilation, tweaking, and who knows what else involved?e   <snip>  H Many thanks to everyone who posted feedback on my questions. Assuming I G can get a straight answer out of my SW vendor, I am going to look into nF an Alpha version of our software. If not, a 3100-98 will provide some F more breathing room, though I'm not certain how much. (256-512 MB RAM E instead of 80MB max in our 3100-95, and slightly faster processor, I d  believe) Now two more questions:  A 1.) What is the minimum required VMS version to run on a 3100-98?u  F 2.) We currently have a load of terminals, printers, etc connected to F our VAX via terminal servers and MMJ jacks. Alpha Servers do not have I MMJ jacks, as far as I can tell. Where does this leave us for connecting hI our equipment? Multiple serial port cards? Just adapt the Alpha's Serial h port to MMJ and go?   G Thanks again for the great answers thus far, and hopefully for more in y response to this post!   Toms   ------------------------------  + Date: Mon, 20 Jan 2003 19:19:28 +0100 (MET)P9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> 0 Subject: Re: MicroVAX upgrade (to Alpha, maybe?); Message-ID: <01KRGW4R8FKO96VR7Q@sysdev.deutsche-boerse.com>A  J > Many thanks to everyone who posted feedback on my questions. Assuming I I > can get a straight answer out of my SW vendor, I am going to look into u$ > an Alpha version of our software.   + In many cases, a recompile is all you need.b  D The fastest VAX machines were about 40--50 VUPs I think.  The first H ALPHA machines were around 60, IIRC.  An ALPHA 255/233 I bought in 1997 F would be about 200 or so.  Modern ALPHA machines are probably another  factor 30 or so faster.n  H > 2.) We currently have a load of terminals, printers, etc connected to H > our VAX via terminal servers and MMJ jacks. Alpha Servers do not have # > MMJ jacks, as far as I can tell. e  G But they do have serial connectors.  You can even get adapters between u1 the two, or move to the pinned serial connectors.t  H > Where does this leave us for connecting our equipment? Multiple serial@ > port cards? Just adapt the Alpha's Serial port to MMJ and go?   F What about leaving the stuff connected to the VAX and build a cluster  with the ALPHA?d   ------------------------------  % Date: Mon, 20 Jan 2003 18:34:02 +0000 + From: John Laird <john@laird-towers.org.uk>t0 Subject: Re: MicroVAX upgrade (to Alpha, maybe?)8 Message-ID: <7afo2vkiobm1figt83sn70ekaim26lssqq@4ax.com>  H On Mon, 20 Jan 2003 13:04:51 -0500, Tom Rymes <tomnews@rymes.net> wrote:  F >In article <tomnews-39F4DC.18560317012003@news.comcast.giganews.com>,& > Tom Rymes <tomnews@rymes.net> wrote: >s ><snip>o > F >> 2.) What is the likelihood that we could move to an Alpha platform H >> instead? Licensing issues? Would our software run on Alpha, or would F >> there be recompilation, tweaking, and who knows what else involved? >i ><snip>  >2I >Many thanks to everyone who posted feedback on my questions. Assuming I iH >can get a straight answer out of my SW vendor, I am going to look into G >an Alpha version of our software. If not, a 3100-98 will provide some  G >more breathing room, though I'm not certain how much. (256-512 MB RAM tF >instead of 80MB max in our 3100-95, and slightly faster processor, I ! >believe) Now two more questions:0 > B >1.) What is the minimum required VMS version to run on a 3100-98?  K I look after a couple of hybrid machines (/98 motherboard with the original6L /96 CPU).  These run 6.2 just fine - I don't recall off-hand whether the /98G was supported by this release as these odd machines identify themselves2K exactly the same as other genuine /96s (but with more memory).  For us, themI 128-512Mb upgrade was worth far more than the increment in CPU power.  WehI still find the SCSI performance a bit dismal, though.  A third-party firmeI (Nemonix?) made an UW upgrade card, but we could not justify the expense.e YMMV.   G >2.) We currently have a load of terminals, printers, etc connected to hG >our VAX via terminal servers and MMJ jacks. Alpha Servers do not have eJ >MMJ jacks, as far as I can tell. Where does this leave us for connecting J >our equipment? Multiple serial port cards? Just adapt the Alpha's Serial  >port to MMJ and go?   Adapt the serial ports and go.     	Johna   ------------------------------  % Date: Mon, 20 Jan 2003 10:49:03 -0000e2 From: "Chris Sharman" <chris.sharman@sorry.nospam>/ Subject: Re: MX 5.3-2 ignoring vmsmail forwardsr4 Message-ID: <b0gk6f$poa$1$8302bc10@news.demon.co.uk>  % "Peter LANGSTOEGER"  wrote in messager, news:TcwW9.67405$TY.611542@news.chello.at...F > In article <b05t6i$cc0$1$8302bc10@news.demon.co.uk>, "Chris Sharman"$ <chris.sharman@sorry.nospam> writes:0 > >Upgraded MX from 5.0 to 5.3 eco 2 last night.2 > >It appears it's now ignoring VMS mail forwards: >t1 > Only on your site. Mine is working properly ;-)   A Traced to name_conversion (ie removing name_conversion fixes it).mK Can't see anything wrong with name_conversion - it's been running unchanged  since June '98.1E All it does is change usernames to dotted names, for consistency (and J security - better if vms usernames aren't visible to the outside world) onI outbound messages. Seems fireproof, and still works, but (by some unknown D mechanism) prevents the recognition of some mail forwards on inboundK messages, causing them to bounce. I'm stuck - nothing else to try, really -a# and living without name_conversion. F It's written in Pascal, and uses VAR dsc1$type & str$ routines for the& strings, which I think should be safe. [GLOBAL] FUNCTION convert(  VAR p: pcontext;n  code: convtype;!  VAR inadr: [READONLY] dsc1$type;s  VAR outadr: dsc1$type
  ): UNSIGNED;t     VAR   pn: pnode VALUE ZERO;  in_s: VARYING [255] OF CHAR;i*     BEGIN ! convert names on outgoing mail     ESTABLISH(lib$sig_to_ret);G     str$upcase(destination_string := %descr in_s, source_string := %refw inadr);eG     IF debug THEN WRITELN(log, pid:8:16, ' Convert', code, ': "', in_s,l '"');eL     IF (code<>from_user) OR_ELSE (LENGTH(in_s)>12) OR_ELSE NOT lookup(p, pn,
 in_s) THEN  BEGIN=  IF debug THEN WRITELN(log, pid:8:16, ' No conversion done');S  convert := 0;  END     ELSE  BEGIN/  str$copy_dx(%ref outadr, %descr pn^.fullname);eI  IF debug THEN WRITELN(log, pid:8:16,' converted to "',pn^.fullname,'"');P  convert := 1;  END;s     END;    I The lookup function uses lib$insert_tree to search/maintain a username tosK dotted name mapping list, and looks up new ones from a file. I'll post morey if anyone wants to see it.   > But I've another hint: >e< > MAIL> SET FORWARD/USER=JOHNNY.CASH SMTP%"CASH@domain.name"  L Ours are all simple, local forwards, used purely to create aliases for a vms	 username.HJ We might have john.cash, johnny.cash, jonathan.cash, all pointing to cash.   ------------------------------    Date: 20 Jan 2003 10:11:53 +0100' From: huber@mppmu.mpg.de (Joseph Huber)l7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ?? + Message-ID: <W4Z5tkEBqLFa@vms.mppmu.mpg.de>e  b In article <3E284E33.ACCA3FBB@vl.videotron.ca>, JF Mezei <jfmezei.spamnot@vl.videotron.ca> writes: > Dan Allen wrote:7 >> http://www.boulder.nist.gov/timefreq/service/its.htmp > D > many thanks. Aren't there time servers in other countries though ?  = In Germany the central and official time service is from the d( "Physikalisch-Technische Bundesanstalt":      ptbtime1.ptb.de .    ptbtime2.ptb.de    -- 9N Joseph "Sepp" Huber   mailto:joseph.huber@web.de   http://www.huber-joseph.de/   ------------------------------    Date: 20 Jan 2003 09:17:50 -0600+ From: young_r@encompasserve.org (Rob Young)f7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??r3 Message-ID: <d2IMJJjFLYj+@eisner.encompasserve.org>n  U In article <W4Z5tkEBqLFa@vms.mppmu.mpg.de>, huber@mppmu.mpg.de (Joseph Huber) writes:rd > In article <3E284E33.ACCA3FBB@vl.videotron.ca>, JF Mezei <jfmezei.spamnot@vl.videotron.ca> writes: >> Dan Allen wrote:d8 >>> http://www.boulder.nist.gov/timefreq/service/its.htm >> pE >> many thanks. Aren't there time servers in other countries though ?h > ? > In Germany the central and official time service is from the n* > "Physikalisch-Technische Bundesanstalt": >  >    ptbtime1.ptb.de   >    ptbtime2.ptb.de >     > 	The concept of "time server" is old school.  Forget about it.  < 	For less than a $500 U.S. you can get a time source that isB 	accurate to a microsecond level - or less.  Most of us don't need= 	that accuracy.  Since there are a number of these across ther= 	Internet setup as Stratum 1 servers , we (collective we) cansA 	point to Stratum 2 servers that are pointed to GPS Stratum 1 andi 	and obtain very good accuracy.s  > 	Here is how you can check the accuracy of your NTP source(s):  " http://www.gpsclock.com/check.html  : 	Off that page is a link to a gps source they are selling:  " http://www.gpsclock.com/specs.html  : PPS pulses accurate to plus or minus 1 microsecond of UTC . Tunable to plus/minus 250 nanosecond accuracy 5 Can maintain timing accuracy with only one satellite S@ Includes web page builder, NTP patches and DOS support software  Low cost ($380)   : 	There are similar products out and about.  Here is a much" 	more robust list with accuracies:  , http://www.eecis.udel.edu/~ntp/hardware.html  8 	A number of good GPS links hanging off that, including:   http://www.gpsy.com/gpsinfo/   	From there:  + http://www.edu-observatory.org/gps/gps.htmla  ! 	How accurate are the GPS clocks?u  C http://www.globalsecurity.org/space/library/news/1997/970505-2f.htm,  K "The GPS IIF spacecraft will have four frequency standards composed of both:L cesium and rubidium technologies. These extremely accurate GPS atomic clocks- can keep time to within 8 nanoseconds a day."5  ? 	Bottom Line:  The days of a "time server" (national, military)>8 	are past/unnecessary.  Nameless, faceless, GPS is best.   				Roby  : And the wind shall say:  "Here were decent godless people:>                           Their only monument the asphalt road:                           And a thousand lost golf balls."-                                 -- T.S. Eliota   ------------------------------  % Date: Mon, 20 Jan 2003 16:58:07 +0100tE From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>c7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??r+ Message-ID: <3E2C1C8F.EF9BB85A@mediasec.de>a  G >         The concept of "time server" is old school.  Forget about it.t  N And from where, do you think, do the GPS satellites get their reference time!?K (No, not via NTP, but it's the same clocks as, for instance, the PTB serverI uses.)   	Jan   ------------------------------    Date: 20 Jan 2003 10:28:44 -0600+ From: young_r@encompasserve.org (Rob Young)s7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??X3 Message-ID: <ZqUvLBd$Xtpr@eisner.encompasserve.org>l  s In article <3E2C1C8F.EF9BB85A@mediasec.de>, Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de> writes:bH >>         The concept of "time server" is old school.  Forget about it. > P > And from where, do you think, do the GPS satellites get their reference time!?M > (No, not via NTP, but it's the same clocks as, for instance, the PTB servere > uses.) >   E 	Well yes.  There has to be an ultimate source of time somewhere.  I )E 	think in context, I'm trying to press home - don't point to them (asaA 	mentioned elsewhere their public interfaces are often overloaded G 	to the point of being worthless).  Guess I could have worded that partS 	a bit better.  9 	The GPS source of course is much better in many aspects.s   				Rob    ------------------------------   Date: 20 Jan 2003 16:54:34 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ?? 0 Message-ID: <b0h9ka$5jv$1@pegasus.csx.cam.ac.uk>  3 In article <d2IMJJjFLYj+@eisner.encompasserve.org>,i- young_r@encompasserve.org (Rob Young) writes:  |> k= |> PPS pulses accurate to plus or minus 1 microsecond of UTC E1 |> Tunable to plus/minus 250 nanosecond accuracy u8 |> Can maintain timing accuracy with only one satellite C |> Includes web page builder, NTP patches and DOS support software t |> Low cost ($380) g  < Does it include a copy of Every Idiot's Guide to Relativity?     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679T   ------------------------------  % Date: Mon, 20 Jan 2003 18:12:42 +01009E From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>E7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??:+ Message-ID: <3E2C2E0A.D76BB2BB@mediasec.de>a  > > Does it include a copy of Every Idiot's Guide to Relativity?  G Not required - already implemented in the GPS receiver's software. (The,E TAI clocks and you typically share an inertial frame, or the relativenH velocities will be generally so small you can still usefully synchronise to the nanosecond or so.)i   	Jan   ------------------------------  % Date: Mon, 20 Jan 2003 18:13:52 +0100 E From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>:7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??u+ Message-ID: <3E2C2E50.DD91A7D3@mediasec.de>w  F Of course, the GPS clock is by nature a broadcast medium, so it's wellE suited to this sort of thing. OTOH, the PTB internet time servers aree4 very responsive. Guess not too many people use them.   	Jan   ------------------------------   Date: 20 Jan 2003 17:30:22 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??O0 Message-ID: <b0hbne$79c$1@pegasus.csx.cam.ac.uk>  s In article <3E2C2E0A.D76BB2BB@mediasec.de>, Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de> writes:gA |> > Does it include a copy of Every Idiot's Guide to Relativity?  |> -J |> Not required - already implemented in the GPS receiver's software. (TheH |> TAI clocks and you typically share an inertial frame, or the relativeK |> velocities will be generally so small you can still usefully synchronise0 |> to the nanosecond or so.)  ? Software developers still need to know that their code will not.? be safe if it is itself used on board a satellite, nor if it iseB used to synchronise local atomic clocks for use on board aircraft!  ? Two aircraft flying in different directions have a relativisticl8 time difference of c. 70 nanoseconds a day, for example.     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679B   ------------------------------  % Date: Mon, 20 Jan 2003 09:33:45 -0800 # From: "Tom Linden" <tom@kednos.com> 7 Subject: RE: Network Time Protocol (NTP) for OpenVMS ??o9 Message-ID: <CIEJLCMNHNNDLLOOGNJIAECCGHAA.tom@kednos.com>a   >-----Original Message-----g0 >From: Nick Maclaren [mailto:nmm1@cus.cam.ac.uk]' >Sent: Monday, January 20, 2003 9:30 AM  >To: Info-VAX@Mvb.Saic.Com8 >Subject: Re: Network Time Protocol (NTP) for OpenVMS ?? >s >- > 3 >In article <3E2C2E0A.D76BB2BB@mediasec.de>, Jan C.fA >=?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de> writes:tB >|> > Does it include a copy of Every Idiot's Guide to Relativity? >|>xK >|> Not required - already implemented in the GPS receiver's software. (The I >|> TAI clocks and you typically share an inertial frame, or the relativehL >|> velocities will be generally so small you can still usefully synchronise >|> to the nanosecond or so.)y > @ >Software developers still need to know that their code will not@ >be safe if it is itself used on board a satellite, nor if it isC >used to synchronise local atomic clocks for use on board aircraft!  >g@ >Two aircraft flying in different directions have a relativistic9 >time difference of c. 70 nanoseconds a day, for example.r  A During which time the aircraft will have moved about a third of a-$ micron if travelling at sonic speed.   >4 > 	 >Regards,: >Nick Maclaren, + >University of Cambridge Computing Service,o? >New Museums Site, Pembroke Street, Cambridge CB2 3QH, England.w >Email:  nmm1@cam.ac.ukl0 >Tel.:  +44 1223 334761    Fax:  +44 1223 334679 >g >---' >Incoming mail is certified Virus Free.9; >Checked by AVG anti-virus system (http://www.grisoft.com).mB >Version: 6.0.435 / Virus Database: 244 - Release Date: 12/30/2002 >o ---m& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.435 / Virus Database: 244 - Release Date: 12/30/2002t   ------------------------------    Date: 20 Jan 2003 12:33:16 -0600+ From: young_r@encompasserve.org (Rob Young)u7 Subject: Re: Network Time Protocol (NTP) for OpenVMS ??t3 Message-ID: <FP96lJE$gQ$f@eisner.encompasserve.org>'  s In article <3E2C2E50.DD91A7D3@mediasec.de>, Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de> writes:lH > Of course, the GPS clock is by nature a broadcast medium, so it's wellG > suited to this sort of thing. OTOH, the PTB internet time servers are-6 > very responsive. Guess not too many people use them. >   6 	They are.  Maybe something improved over the last few& 	years or I was doing something wrong.   				Rob0   ------------------------------   Date: 20 Jan 2003 18:29:58 GMT( From: nmm1@cus.cam.ac.uk (Nick Maclaren)7 Subject: RE: Network Time Protocol (NTP) for OpenVMS ?? 0 Message-ID: <b0hf76$a3s$1@pegasus.csx.cam.ac.uk>  9 In article <CIEJLCMNHNNDLLOOGNJIAECCGHAA.tom@kednos.com>,w% "Tom Linden" <tom@kednos.com> writes:0 |> :E |> >|> > Does it include a copy of Every Idiot's Guide to Relativity?o |> >|>N |> >|> Not required - already implemented in the GPS receiver's software. (TheL |> >|> TAI clocks and you typically share an inertial frame, or the relativeO |> >|> velocities will be generally so small you can still usefully synchronisec  |> >|> to the nanosecond or so.) |> >C |> >Software developers still need to know that their code will not C |> >be safe if it is itself used on board a satellite, nor if it isuF |> >used to synchronise local atomic clocks for use on board aircraft! |> >C |> >Two aircraft flying in different directions have a relativisticm< |> >time difference of c. 70 nanoseconds a day, for example. |> lD |> During which time the aircraft will have moved about a third of a' |> micron if travelling at sonic speed.n  ? A third of a micron a day?  Boy!  Talk about supercharging ....h     Regards, Nick Maclaren,* University of Cambridge Computing Service,> New Museums Site, Pembroke Street, Cambridge CB2 3QH, England. Email:  nmm1@cam.ac.uk/ Tel.:  +44 1223 334761    Fax:  +44 1223 334679l   ------------------------------    Date: 20 Jan 2003 05:51:36 -0800( From: bob@instantwhip.com (Bob Ceculski)) Subject: no wonder Intel wanted alpha ... = Message-ID: <d7791aa1.0301200551.1e5355e1@posting.google.com>m    HP's new Alpha Marvels revealed   ( Good stuff. No wonder Intel wanted Alpha  , By Eva Glass: Monday 20 January 2003, 10:03   F DIETRICH TELLS me he wants to be alone, so stresses not to mention hisB name in connection with the three AlphaServers HP will launch this week.6C On no account should we involve Dietrich in details of the GS1280 oE the high end HP server that will be available in eight way and 16 way F systems, and which apparently will allow two processor partitions that> can run different applications and separate operating systems.  D Like, say, you combined OpenVMS and Tru64 clusters for downtime thatD may never come. Later this year, said Dietrich, we are likely to see@ 32-64 way GS1280s. Right now the systems will support 8 to 64GB.  : SPECint_2000 marks of 875 or so, that's not bad, and 1,500E SPECfp_2000s. And some pretty sizzling reliability features includingrE auto server recovery, RAID memory, hot swapping server management LANfB components and redundo hot swap power fans. Sheesh! What a Marvel.  C These ES47 AlphaServers which use the EV7 Alphas aren't bad either,eE with HP likely to offer 2 CPU towers and 2-4 way racks. A system on an@ chip (SOC) feature gives built in symmetric multiprocessing, and@ they'll support Linux, Tru64 and OpenVMS too. The 1GHz EV7s will: support up to eight/16GB memory, and seem to be exhibiting9 SPECint_2000 marks of 763 and SPECfp_2000 marks of 1,290.n  E And as for the cunning ES80 systems, well Dietrich reckons they'll beoD out in March, again using 1GHz EV7s and will offer two, four, six orC eight CPU support, with bencmakrs similar to the above, for obviousS reasons.  C No wonder Mike Capellas ceded this Alpha technology to Intel  it'so surely good stuff.  D * MEANWHILE, some Alpha SAP benchmarks that got HP's knicks in a bitB of a twist have mysteriously disappeared off its web site with theF puzzling message: "we're very sorry! The page you requested can not be found".i   ------------------------------    Date: 20 Jan 2003 07:20:02 -0800  From: otto@programmer.net (Otto)( Subject: OpenVMS Alpha V6.2 media wanted= Message-ID: <c0b935a5.0301200720.26b3ffe9@posting.google.com>e   Hello!  5 We need a copy of the "OpenVMS Alpha V6.2" CD-ROM(s).I Good price will be offered.i   Otto   ------------------------------    Date: 20 Jan 2003 09:08:07 -0800& From: kfarmer@openvms.org (Ken Farmer)Y Subject: OpenVMS.org: Marvel article and HP's press release for Marvel and Alpha Retain To= Message-ID: <80eeab56.0301200908.37846b8c@posting.google.com>s   http://www.openvms.org  C Check out the articles.  I'm currently at the Partner Roundhouse in C Nashua now and will return home tomorrow.  Email responses may be ae little slow.   Keno   -- Kenneth Farmer <>< http://www.openvms.org   ------------------------------  # Date: Mon, 20 Jan 2003 17:23:09 GMTo# From: "John Smith" <a@nonymous.com>eY Subject: Re: OpenVMS.org: Marvel article and HP's press release for Marvel and Alpha RetapJ Message-ID: <1gWW9.400777$F2h1.97261@news01.bloor.is.net.cable.rogers.com>  3 "Ken Farmer" <kfarmer@openvms.org> wrote in messagef7 news:80eeab56.0301200908.37846b8c@posting.google.com...  > http://www.openvms.org >tE > Check out the articles.  I'm currently at the Partner Roundhouse inoE > Nashua now and will return home tomorrow.  Email responses may be a  > little slow.    9 http://www.openvms.org/stories.php?story=03/01/20/1480031.  B Sorry to say it Ken, but this isn't an 'article' - it's a HP press release.  E What HP pointedly leaves out are performance specs, and the fact thatnF today's Marvel is a much better system than the Itanic II systems that# will be released in 2-3 years time.    ------------------------------  # Date: Mon, 20 Jan 2003 17:34:17 GMTt# From: "John Smith" <a@nonymous.com>nY Subject: Re: OpenVMS.org: Marvel article and HP's press release for Marvel and Alpha RetadK Message-ID: <tqWW9.400839$F2h1.167628@news01.bloor.is.net.cable.rogers.com>d  3 "Ken Farmer" <kfarmer@openvms.org> wrote in messagee7 news:80eeab56.0301200908.37846b8c@posting.google.com...e > http://www.openvms.org >rE > Check out the articles.  I'm currently at the Partner Roundhouse inI > Nashua now  D Are they saying anything about serious VMS advertising and marketing1 efforts?. If not, I'd start taking Linux courses.    ------------------------------  % Date: Mon, 20 Jan 2003 09:07:33 +0100 * From: "Sven Tieste" <sven.tieste@sw-hb.de>4 Subject: Re: Oracle LMON and LMDO buffered I/O usage2 Message-ID: <b0gao5$gqa$1@f40-3.zfn.uni-bremen.de>   Hello Bill,iA we are using a AlphaServer 1000A 5/400 and had the same problems.nK First, these two processes are processes of Oracle Parallel Server (AFAIK). F We solved this problem by upgrade to OpenVMS 7.2-2 and Oracle 8.1.7.3.K After this all work much better than before (quite good but not perfect; IOi5 from >500 to about 75; CPU from >90% to mean of 15%). L I think this problem is well known at Oracle and HP but both of them did not resolved it in older versions. Hope this will help.	 greetings- Sven    A Bill McLaughlin <mcbill20@hotmail.com> schrieb in im Newsbeitrag:(2 e9cbc4f2.0301170918.296fa424@posting.google.com...F > I am running Oracle Enterprise Server version 8.1.6.0.0 on Alpha VMS> > 7.2. This is on a home system (AlphaStation 500 with 256MB). >yF > Oracle works OK but there are two processes that eat up huge amountsG > of buffered I/O's, even on a totally idle database. The processes are H > LMON and LMDO. When I do a MONITOR SYSTEM, the buffered I/O total sitsF > at about 1030 and the top users alternate between the LMDO and LMON,A > at 512 or 513 each. After three days of uptime (with no processlC > accessing Oracle for the first two), the LMON has done 6133499390kF > I/O's and the LMDO has done 6133587962 I/O's. Here's the output form > SHOW SYSTEM: >'H > 2020012E ORA_V8160000000 LEF      6      194   0 00:00:00.22       277 >    263H > 20200130 ORA_CALLS1_PMON HIB      6      117   0 00:00:04.60       895 >   1079H > 20200131 ORA_CALLS1_LMON HIB      6133499390   0 00:00:06.45      1551 >   1340H > 20200132 ORA_CALLS1_LMD0 HIB      6133587962   0 00:00:06.13       908 >    909H > 20200133 ORA_CALLS1_DBW0 HIB      6      168   0 00:00:00.41       939 >   1010H > 20200134 ORA_CALLS1_LGWR HIB      6      250   0 00:00:00.39       906 >   1056H > 20200135 ORA_CALLS1_CKPT HIB      5   169118   0 00:01:45.14       770 >   1115H > 20200136 ORA_CALLS1_SMON HIB      4      200   0 00:00:09.84      1634 >   1267H > 20200137 ORA_CALLS1_RECO HIB      4      103   0 00:00:00.96      1103 >   1381H > 20200138 ORA_LISTENER907 HIB      5   260834   0 00:00:10.93       793 >    596H > 20200199 ORA_CALLS1B4824 LEF      6      191   0 00:00:00.35      2835 >   1022 >k >d >  > Any ideas? >, > Thanks in advance. >t > Bill McLaughlin    ------------------------------  % Date: Mon, 20 Jan 2003 09:15:28 +0100i* From: "Sven Tieste" <sven.tieste@sw-hb.de>4 Subject: Re: Oracle LMON and LMDO buffered I/O usage2 Message-ID: <b0gb70$keg$1@f40-3.zfn.uni-bremen.de>   It is me again!sI I forget, the first thing we did were disabling Oracle Parallel Server in  init.ora and restart the DB.< After that there will be no LMD0-process and LMON needs less CPU and IO.d	 greetingsc Sven= Sven Tieste <sven.tieste@sw-hb.de> schrieb in im Newsbeitrag: ' b0gao5$gqa$1@f40-3.zfn.uni-bremen.de...n
 > Hello Bill,1C > we are using a AlphaServer 1000A 5/400 and had the same problems.-D > First, these two processes are processes of Oracle Parallel Server (AFAIK).H > We solved this problem by upgrade to OpenVMS 7.2-2 and Oracle 8.1.7.3.J > After this all work much better than before (quite good but not perfect; IO7 > from >500 to about 75; CPU from >90% to mean of 15%).WJ > I think this problem is well known at Oracle and HP but both of them did noti  > resolved it in older versions. > Hope this will help. > greetings  > Sven >  > C > Bill McLaughlin <mcbill20@hotmail.com> schrieb in im Newsbeitrag:i4 > e9cbc4f2.0301170918.296fa424@posting.google.com...H > > I am running Oracle Enterprise Server version 8.1.6.0.0 on Alpha VMS@ > > 7.2. This is on a home system (AlphaStation 500 with 256MB). > >uH > > Oracle works OK but there are two processes that eat up huge amountsI > > of buffered I/O's, even on a totally idle database. The processes arehJ > > LMON and LMDO. When I do a MONITOR SYSTEM, the buffered I/O total sitsH > > at about 1030 and the top users alternate between the LMDO and LMON,C > > at 512 or 513 each. After three days of uptime (with no processiE > > accessing Oracle for the first two), the LMON has done 6133499390 H > > I/O's and the LMDO has done 6133587962 I/O's. Here's the output form > > SHOW SYSTEM: > >oJ > > 2020012E ORA_V8160000000 LEF      6      194   0 00:00:00.22       277
 > >    263J > > 20200130 ORA_CALLS1_PMON HIB      6      117   0 00:00:04.60       895
 > >   1079J > > 20200131 ORA_CALLS1_LMON HIB      6133499390   0 00:00:06.45      1551
 > >   1340J > > 20200132 ORA_CALLS1_LMD0 HIB      6133587962   0 00:00:06.13       908
 > >    909J > > 20200133 ORA_CALLS1_DBW0 HIB      6      168   0 00:00:00.41       939
 > >   1010J > > 20200134 ORA_CALLS1_LGWR HIB      6      250   0 00:00:00.39       906
 > >   1056J > > 20200135 ORA_CALLS1_CKPT HIB      5   169118   0 00:01:45.14       770
 > >   1115J > > 20200136 ORA_CALLS1_SMON HIB      4      200   0 00:00:09.84      1634
 > >   1267J > > 20200137 ORA_CALLS1_RECO HIB      4      103   0 00:00:00.96      1103
 > >   1381J > > 20200138 ORA_LISTENER907 HIB      5   260834   0 00:00:10.93       793
 > >    596J > > 20200199 ORA_CALLS1B4824 LEF      6      191   0 00:00:00.35      2835
 > >   1022 > >. > >  > >o > > Any ideas? > >  > > Thanks in advance. > >w > > Bill McLaughlinh >  >    ------------------------------  % Date: Mon, 20 Jan 2003 10:27:19 -0000s2 From: "Chris Sharman" <chris.sharman@sorry.nospam>= Subject: Re: osu/http-server + gecko browser support for .csv 4 Message-ID: <b0gitm$nb2$1$8302bc10@news.demon.co.uk>  = "JF Mezei" <jfmezei.spamnot@vl.videotron.ca> wrote in messageD) news:3E283863.B694B06C@vl.videotron.ca...nF > Since those csv files are essentually text files, have youconsidered sendingb > them as text/plain ?G > This would make it more likely that the data will be saved in te OS'sp nativeG > text file type (but not garanteed) and then the user can import it ind$ > his/her/its favourite application.  9 Would certainly 'work' everywhere, but would require morer effort/understanding froml# Microsoft users - not a good thing.o  J The mis-typing by mozilla is a known bug, scheduled for fixing in the nextG release (1.3beta), whereby on windows a mimetype translates to only oneu	 filetype.e2 http://bugzilla.mozilla.org/show_bug.cgi?id=189434  H > Here is an intertesting one: this is a microsoft submitted document at iana to(D > officially define excell's mime type:  (the part about security is interesting !) >n' > >Media Type: application/vnd.ms-exceli5 > >--------------------------------------------------u( > >Name             Microsoft Excel (tm)& > >Required parameters:           None& > >Optional parameters:           name2 > >Encoding considerations:       base64 preferredL > >Security considerations:         As with most application types this data isH > >intended for interpretation by a program that understands the data onI > >the recipient's system. Recipients need to understand that they are attJ > >the "mercy" of the sender, when receiving this type of data, since dataG > >will be executed on their system, and the security of their machines4 > >can be violated.@  K Oh dear. Typical microshaft stuff. I'm uncertain how much risk there reallyD is.4K I've got Outlook running on 'restricted' (untrusted) security - ought to bel the default, but isn't.eI csv files can't contain macros, afaik, but they can contain formulae. The A csvs on our intranet are safe enough, anyway - just numbers/text.E   Chrisw   ------------------------------    Date: 20 Jan 2003 08:18:38 -0800( From: bob@instantwhip.com (Bob Ceculski)< Subject: Security threats seen big for 2003 ... not for VMS!= Message-ID: <d7791aa1.0301200818.50fde73c@posting.google.com>h   not for vms users ...l    	 INFOWORLDi   By Brian Fonseca a January 16, 2003 1:13 pm PTu  @  LESS THAN HALFWAY through the first calendar month of 2003, andB already end-users have had to stave off computer threats such as aF DHCP multiple buffer overflow vulnerability, the Sobig worm, the LirvaA worm, and remnants of the Yaha virus. Blended and longer lifespanfC malicious code attacks are expected to run rampant in the new year,-E but they are far from the only dangers angling to make life miserableo for security administrators.  B According to security experts, the rash of forecasted software andD product vulnerabilities will force customers to acquire workflow andE remediation tools for their OS and networks. In addition, integrationo@ and security event management's role within an existing securityF infrastructure will help define realistic business and ROSI (return onD security investment) expectations against inflated security hype andC vendor non-accountability, said security analyst Pete Lindstrom, of:! the Spire Group in Malvern, Penn.)  A "People are going to start realizing 2003 is a year that securitymB folks get their act together and ducks in line and understand that; individual point products just don't cut it from a securitye? architecture perspective and we'll need to start evaluating andtC justifying security products without slipping into hype mode," said 
 Lindstrom.  D For instance, Lindstrom said that Homeland Security initiatives have? led cyberterrorism awareness to skyrocket as a focal point whensA organizations should be more concerned with the machinations of aI> "basic thief trying to break into a virtual bank" or corporate network.  D IT research firm Aberdeen Group predicts that security incidents andA financial losses due to identity theft will significantly rise ing@ 2003. Among Boston-based Aberdeen's other top 10 predictions forC security and privacy include "old-guard" security technologies that.E will fade into obscurity; new government programs that will fail; anduB a security tech-spending rush fueled by automation and pragmatism.  ? Led by a new breed of suppliers such as Patchlink, Big Fix, St.nD Bernard Software, Citadel, Harris Corp., and Shavlik Technologies toE name just a few, the necessity of automated vulnerability remediatione? is gaining steam, notes Eric Hemmendinger, research director of ! information security at Aberdeen.-  E The security analyst said customers do not want to be at the mercy ofgE existing holes which linger while they have the capability to preventt? damage or liability that could occur. Therefore, solutions that-@ detect, prioritize, assess, remediate, and track vulnerabilitiesB should be present within an infrastructure to stay one step ahead.  E "If you can't eliminate vulnerabilities in terms of what causes them, C what can you do to remediate them?" asked Hemmendinger. "You reallynC have to start looking at solutions that focus on remediation ratherr5 than waiting for any OS provider to fix the problem."l  C Having the appropriate security product in place, however, does nottE necessarily result in a successful defense position or plan of actioni; against an incoming computer assault if tuning, monitoring,M6 maintenance, and response procedures are not in place.  F According to a report released last week by Stanford, Conn.-based MetaC Group, for instance, organizations running IDS (intrusion detectionsF system) products as a technical solution encounter difficulty with theA "operational aspects of a process-intensive solution." The report F notes these obstacles can range from inability to compensate for false? positive overloads to inadequate break-ins or threat reactions.e   ------------------------------  % Date: Mon, 20 Jan 2003 17:44:38 +0000.( From: Nic Clews <sendspamhere@127.0.0.1>@ Subject: Re: Security threats seen big for 2003 ... not for VMS!) Message-ID: <3E2C3586.3A8DCF0A@127.0.0.1>    Bob Ceculski wrote:s > 	 (Article) D > According to security experts, the rash of forecasted software and ...o  G Those 'security experts' just happen to work for companies that providekC anti virus and firewall software for the systems under threat, theynH don't however work in OpenVMS marketing. This sort of 'story' borders onE being advertorial. They are just working on keeping themselves on theMB gravy train, not specifically pointing out the weaknesses of thoseF systems. Sorry Bob, this 'story' just doesn't say anything. But we all( know what it's trying to sell, don't we?   -- ,? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciencesn nclews at csc dot com5   ------------------------------  + Date: Mon, 20 Jan 2003 17:55:57 +0100 (MET)r9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>tU Subject: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet card ; Message-ID: <01KRGSW81W0696VR7Q@sysdev.deutsche-boerse.com>e  @ I broke my rule not to buy hardware that I can't plug a VT into.  G Problem: On my hobbyist cluster, I want to move from ISDN (with several ? "real", static, public, routable IP addresses) to DSL (with onehH externally visible, non-static public IP address and, behind the router,E static, private, non-routable IP addresses).  This is purely for costrF reasons since DSL flat rates are much cheaper than ISDN, even though IC have both dial-out AND dial-in on demand.  (The phone costs are notrB negligible, and the ISP charges a high basic rate for this serviceD (which includes port filtering at the ISP, domain management etc).)   G Thus, I bought a standard DSL router which does NAT (actually probably tI PAT as well, which is really what I need).  The only problem is that one eF has to be able to access it over the network in order to configure it = (preferably via HTTP, though it might be possible to FTP the sE configuration file).  It has 192.168.1.1 as its IP address, which is rC obviously not on the network I have now.  However, I don't want to SG change the IP addresses on the VMS machines, since I want both to work  F for a transitional period (I assume I just have to change the default F gateway to determine where outgoing connections go and can connect to G the (dynamic) address of the router until I get the DNS to point to it  ' rather than to my static IP addresses).l  C Is SET (CONFIGURATION) INTERFACE the way to go?  After that, can I iC automatically access anything on my old network and on that of the n router?e  G Presumably the default route only kicks in for things not on the local c	 networks.(  G Fortunately, I have another cluster where I can practice first to make dI sure I have the router configuration correct, so I'll then know that any n, problems come only from the pseudointerface.   ------------------------------  + Date: Mon, 20 Jan 2003 18:37:51 +0100 (MET)n9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>nY Subject: Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet cardI; Message-ID: <01KRGUSFFXZW96VR7Q@sysdev.deutsche-boerse.com>c  J > I think you should be able to define a route just to that IP address youG > need...in any case, if you enter a numeric address, and have only oneGA > interface to which the default route is set, that should work. S  I That was my first thought.  However, nothing I tried worked.  (Sometimes oI it would be immediately inaccessible, sometimes it would try and hang.)  CI Of course, I don't want to change the default route until I am sure that GF everything works!  Is there a way to accomplish what I want without a 7 pseudointerface and without changing the default route?e  I I thought I could set a route directly to this address, but that doesn't m appear to be possible.   ------------------------------  % Date: Mon, 20 Jan 2003 18:16:20 +0000E+ From: John Laird <john@laird-towers.org.uk>lY Subject: Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet cardw8 Message-ID: <endo2vcfjeq2ue4dnhoo83hs82dgprcqn1@4ax.com>  8 On Mon, 20 Jan 2003 18:37:51 +0100 (MET), Phillip Helbig+ <HELBPHI@sysdev.deutsche-boerse.com> wrote:s  K >> I think you should be able to define a route just to that IP address youaH >> need...in any case, if you enter a numeric address, and have only oneB >> interface to which the default route is set, that should work.  >+J >That was my first thought.  However, nothing I tried worked.  (Sometimes J >it would be immediately inaccessible, sometimes it would try and hang.)  J >Of course, I don't want to change the default route until I am sure that G >everything works!  Is there a way to accomplish what I want without a g8 >pseudointerface and without changing the default route? > J >I thought I could set a route directly to this address, but that doesn't  >appear to be possible.t  G The problem is probably at the other end - your router thinks that onlypH 192.168.1.0 is "inside".  It will be ignoring connect requests from yourL existing IP address range until you tell it *that* is the network on its LANI side (either ignoring them or trying to send the acknowledgement out overcK the Internet !).  So, you'll need to configure your VMS machine temporarily@J in the same 192.168.1.0 network.  My ADSL router allows me to add internalG networks either singly as the network to be routed outside (directly or@L NAT), or multiply as static routes.  I know the management tools work on theG routed network, and assume they would likewise work on the static route>H side.  I have a final option which is to allow external management, fromK known networks or addresses - this may be useful for you if it is availabley on your hardware.e  I TCP/IP is not like Decnet - two machines on the same lan but in differentTG networks can talk without an intervening router provided they both have E routing entries for one another.  They will learn where each other isoF physically, using the Ethernet MAC addresses.  It's all handled in theL ARP/RARP stuff.  (Maybe a little bit more like LAT than DECnet if you like.)     	Johnt   ------------------------------  + Date: Mon, 20 Jan 2003 19:50:13 +0100 (MET) 9 From: Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com>eY Subject: Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet carda; Message-ID: <01KRGX8IRK7296VR7Q@sysdev.deutsche-boerse.com>g  I > The problem is probably at the other end - your router thinks that onlyoJ > 192.168.1.0 is "inside".  It will be ignoring connect requests from yourJ > existing IP address range until you tell it *that* is the network on itsJ > LAN side (either ignoring them or trying to send the acknowledgement out > over the Internet !).  h  D OK, makes sense, but in order to tell it that, I have to be able to F access it over the network!  As I said, one should never buy hardware 5 one can't configure via a VT and a serial connection!p  G > So, you'll need to configure your VMS machine temporarily in the sames > 192.168.1.0 network. c  C Not wanting to get rid of the real IP address, could an additional r$ (pseudo)interface do the trick here?  H > My ADSL router allows me to add internal networks either singly as theG > network to be routed outside (directly or NAT), or multiply as staticBF > routes.  I know the management tools work on the routed network, andE > assume they would likewise work on the static route side.  I have a I > final option which is to allow external management, from known networks F > or addresses - this may be useful for you if it is available on your > hardware.   D I actually have two different routers which appear to have the same B functionality.  I can't look at them in detail though until I can  connect to them!  A > TCP/IP is not like Decnet - two machines on the same lan but inoI > different networks can talk without an intervening router provided theyaH > both have routing entries for one another.  They will learn where eachJ > other is physically, using the Ethernet MAC addresses.  It's all handledJ > in the ARP/RARP stuff.  (Maybe a little bit more like LAT than DECnet if
 > you like.) e  G This is what I thought, and my original motivation for trying to set a 0E route.  Assume for the moment that it's not a DSL router which might :F have other ideas but, say, another VMS machine with an address on the C other network.  What would be the SET ROUTE command to allow me to dH access it?  (Presumably, I would have to do SET ROUTE on both machines, ' so that each knows where the other is.)r   ------------------------------  % Date: Mon, 20 Jan 2003 18:17:57 +0100tE From: Jan C. =?iso-8859-1?Q?Vorbr=FCggen?= <jvorbrueggen@mediasec.de>tY Subject: Re: TCPIP: pseudo-interface: 2 IP addresses on 1 VMS system with 1 ethernet card + Message-ID: <3E2C2F45.5F093134@mediasec.de>I  I I think you should be able to define a route just to that IP address you mE need...in any case, if you enter a numeric address, and have only oneh	 interfaceb4 to which the default route is set, that should work.   	Jan   ------------------------------   End of INFO-VAX 2003.040 ************************