1 INFO-VAX	Wed, 30 May 2001	Volume 2001 : Issue 298       Contents:> Re: %SYSTEM-I-GSDFULL, global section descriptor table is full Re: 7.3 kits Re: 7.3 kits Re: Affordable VMS Workstations  Re: Affordable VMS Workstations  Re: Affordable VMS Workstations O Re: Affordable VMS Workstations... All of a Sudden Alpha Looks More Competitive O Re: Affordable VMS Workstations... All of a Sudden Alpha Looks More Competitive P Re: Affordable VMS Workstations... All of a Sudden Alpha Looks More Competitive N Re: Affordable VMS Workstations... All of a Sudden Alpha Looks MoreCompetitive- AIXL, AIX5L - binary compatibility with Linux & Re: Alphaserver 300 and OpenVMS V7.2-16 Andrew to vanish from c.o.v.? Sun bans Internet access
 anonymous ftp  Re: anonymous ftp  Re: anonymous ftp + Re: Antigen found ExceedinglyInfected virus   Anyone know what happened to OMI% Re: Apache WebServer - How do I query  Re: Backup to CD? ! Can't SET HOST to cluster member.  CD-RW on a VMS system.2 Re: COMPAQ Service Contracts (Compax x EDS x DELL)2 Re: COMPAQ Service Contracts (Compax x EDS x DELL)3 Re: Contents of 7.3 kits -- missing VAX binaries CD ' RE: DECEvent, WEBESs - what's going on? > Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lost> Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lost> Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lost> Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lost Re: disk benchmarking  Re: disk benchmarking  Re: disk benchmarking ; Re: Hobbyist Licenses - was Re: Affordable VMS Workstations ; Re: Hobbyist Licenses - was Re: Affordable VMS Workstations ; Re: Hobbyist Licenses - was Re: Affordable VMS Workstations 0 Re: How to store results of Command in a symbol?$ Re: HSDSA Scrip in a Batch Procedure$ Re: HSDSA Scrip in a Batch Procedure$ Re: HSDSA Scrip in a Batch Procedure$ Re: HSDSA Scrip in a Batch Procedure$ Re: HSDSA Scrip in a Batch Procedure$ Re: HSDSA Scrip in a Batch Procedure IBM opens their Linux Mainframe  Re: JRE.DMP  Re: JRE.DMP  Re: Login/Logout Problem Re: Login/Logout Problem Re: Login/Logout Problem Re: Login/Logout Problem. Re: OpenVMS locked up - How do I find out why?5 RE: Oracle Export 8.1.7 much slower than 7.3.2 on VMS  PCSI and Account creation.D Re: Presentation Collateral From VMS Technical Update Days in London Re: Rambus loses another one Re: The future of VMS  Re: The future of VMS ( Re: UCX 5.x equiv for Multinet SMTP opts Re: VAX 8650 boot failure  Re: VAX 8650 boot failure  Re: vfork/exec question  Re: vms Emulation on NT : Re: VMS Home Automation (was: Affordable VMS Workstations): Re: VMS Home Automation (was: Affordable VMS Workstations): Re: VMS Home Automation (was: Affordable VMS Workstations): Re: VMS Home Automation (was: Affordable VMS Workstations): Re: VMS Home Automation (was: Affordable VMS Workstations) Re: VMS mail problem Re: VMS password cracking  Re: VMS password cracking  Re: VMS password cracking  Re: VMS password cracking ' Re: VMS price increases, rumor or fact? ' Re: VMS price increases, rumor or fact? ' Re: VMS price increases, rumor or fact? ' Re: VMS price increases, rumor or fact? & Re: VXT2000 hardware/firmware question What does READ/NEW do ? % [OT] Current Microsoft v DOJ status ? ) Re: [OT] Current Microsoft v DOJ status ?   F ----------------------------------------------------------------------  # Date: Wed, 30 May 2001 15:11:08 GMT 1 From: "Mark D. Jilson" <jilly@clarityconnect.com> G Subject: Re: %SYSTEM-I-GSDFULL, global section descriptor table is full 2 Message-ID: <3B150EC2.250C3D37@clarityconnect.com>  
 This error' global_alloc: out of global space, file ! would appear to be for GBLPAGFIL.    Peter Weaver wrote:  > : > "Peter Weaver" <peter.weaver@stelco.ca> wrote in message0 > news:fCxQ6.242120$Z2.2712479@nnrp1.uunet.ca...9 > > "Hamlyn Mootoo" <univms@bigfoot.com> wrote in message ) > > news:3B129A19.A45CF755@bigfoot.com... A > > > Increase the number of global sections in sysgen or autogen  > > (preferably),  > > >... > > C > > I took GBLSECTIONS up to 2000, so I have 1,209 free. I am still  > > getting; > > 4 > >     Exception occurred during event dispatching:" > >     java.lang.OutOfMemoryError/ > >     global_alloc: out of global space, file = > >     USER2$:[JFEX.JFEX_VM.SRC.RUNTIME]GLOBALS.C;1, line 29  > > G > > I'm going to forward this off to one of the WEBES people. I promise  > toH > > never say anything bad about DECevent if they would just kill Compaq+ > > Analyze and this whole WEBES garbage...  > G > Sorry, I should have mentioned before the GSDFULL error was appearing D > in the Process Termination record in the accounting file, the JavaC > error above is what appeared in my terminal window after the CCAT  > Window died.   --  D Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  # Date: Wed, 30 May 2001 13:29:33 GMT 1 From: "Mark D. Jilson" <jilly@clarityconnect.com>  Subject: Re: 7.3 kits 2 Message-ID: <3B14F6F2.2A3EECA9@clarityconnect.com>  F 1) The art folks making up the manuals selected the photo from a stock
 of photos.E 2) Once the art work was set the 'error' was not considered to be any 7 greater than a nit and it was too expensive to pick it.   D No this was not a staged photo for the manual set.  It was one of anE environment that is obviously multivendor and reflected the desire of H upper management to get across that OpenVMS plays nice in a company's IT sandbox.   Or so I've been told.    Kenneth Randell wrote: >  > So, who are the folks who: >  > 1) Suggested this picture?! > 2) Let it go through unchecked?  > < > I suppose it's safe to say that this was a staged picture. >    --  D Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  % Date: Wed, 30 May 2001 10:27:57 -0400 . From: "Kenneth Randell" <kenr@datametrics.com> Subject: Re: 7.3 kits + Message-ID: <9f2vvn$4ma$1@bob.news.rcn.net>   K I'm not trying to start a "Let's figure out what was in the minds of 'upper G management'", but...how does this photo reflect 'an environment that is K obviously multivendor'?  Looking on the box, I find the following hardware:   ) 1 DEC (?) monitor with a text application  1 LK-style keyboard . 3 Obviously Sun monitors with GUI applications Numerous other keyboards Some type of fax and/or printer  A phone   H The photo on the box has what little of the DEC (?) monitor cropped off.2 It's also cropped off a little of the top as well.  4 In my opinion, this box cover qualifies for the UPG.   Ken Randell   : Mark D. Jilson <jilly@clarityconnect.com> wrote in message, news:3B14F6F2.2A3EECA9@clarityconnect.com...H > 1) The art folks making up the manuals selected the photo from a stock > of photos.G > 2) Once the art work was set the 'error' was not considered to be any 9 > greater than a nit and it was too expensive to pick it.  > F > No this was not a staged photo for the manual set.  It was one of anG > environment that is obviously multivendor and reflected the desire of J > upper management to get across that OpenVMS plays nice in a company's IT
 > sandbox. >  > Or so I've been told.  >    ------------------------------  % Date: Wed, 30 May 2001 09:11:08 +0100 % From: Alan Greig <a.greig@virgin.net> ( Subject: Re: Affordable VMS Workstations8 Message-ID: <6ia9ht8qqsc9aaplnnu4ufueemh3mmea2v@4ax.com>  7 On Tue, 29 May 2001 10:48:08 -0500, "David J. Dachtera"  <djesys.nospam@fsi.net> wrote:   >Tim Llewellyn wrote:   K >> Anyway, latest marketting in the UK includes "buy 2 servers and get free J >> VMS or Tru64 cluster licence" which isn't a bad offer IMHO, in fact oneS >> that people here and elsewhere have been pushing for, Actually, reading further, T >> you could have instead a free unlimited user licence for VMS. Shame you can't use >> the two offers together.   @ But when I called the toll free number on this brochure for moreC Alphaserver marketing material to pass to my Norwegian colleagues I > received around 10 assorted Compaq Intel/.Armada/Proliant/iPAQE brochures and no Alpha literature whatsoever. Various Compaq folk are  investigating.  
 ><pipe_dream>  > 0 >If only they would offer that here in the U.S.!  5 Hopefully they'd sort out the followup details first!      -- Alan   ------------------------------  % Date: Wed, 30 May 2001 09:17:26 +0100 % From: Alan Greig <a.greig@virgin.net> ( Subject: Re: Affordable VMS Workstations8 Message-ID: <ssa9htc7nm3oeng7c96rl8647p6v7roqmd@4ax.com>  1 On Tue, 29 May 2001 18:06:39 +0100, Tim Llewellyn   <tim.llewellyn@bbc.co.uk> wrote:     ] L >Terry, this is all very reasonable, and I am sure if you were CEO of CompaqJ >and made this statement then I would consider it OK to proceed like this.I >However, if serious about commercial development it would not be prudent O >IMHO to violate licencing terms that might potentially affect any earnings one J >made for one's hard work (likke Compaq sueing the crap outta my bizness).  F And I well recall how my former boss, Arrick Wilkinson, a former DECUSB chairman was sued by DEC for publishing a third party DEC softwareC guide. Verbally DEC said "OK". Then they sued him for a few million F dollars while saying "nothing personal". Reading between the lines DECF at the time did not want anyone promoting VMS software to new clients.E Disgusting, shameful behaviour on DEC's part and I personally did not , attend a DECUS meeting for years afterwards. -- Alan   ------------------------------  % Date: Wed, 30 May 2001 15:36:14 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk> ( Subject: Re: Affordable VMS Workstations) Message-ID: <3B15055E.825D001C@bbc.co.uk>    Alan Greig wrote:   3 > On Tue, 29 May 2001 18:06:39 +0100, Tim Llewellyn " > <tim.llewellyn@bbc.co.uk> wrote: >  > ] N > >Terry, this is all very reasonable, and I am sure if you were CEO of CompaqL > >and made this statement then I would consider it OK to proceed like this.K > >However, if serious about commercial development it would not be prudent Q > >IMHO to violate licencing terms that might potentially affect any earnings one L > >made for one's hard work (likke Compaq sueing the crap outta my bizness). > H > And I well recall how my former boss, Arrick Wilkinson, a former DECUSD > chairman was sued by DEC for publishing a third party DEC softwareE > guide. Verbally DEC said "OK". Then they sued him for a few million H > dollars while saying "nothing personal". Reading between the lines DECH > at the time did not want anyone promoting VMS software to new clients.G > Disgusting, shameful behaviour on DEC's part and I personally did not . > attend a DECUS meeting for years afterwards.  I hmmm., maybe we should be thinking about developing on linux instead :-(.   -- 6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those of  MedAS or the BBC.    ------------------------------    Date: 30 May 2001 22:53:05 +0800, From: Paul Repacholi <prep@prep.synonet.com>X Subject: Re: Affordable VMS Workstations... All of a Sudden Alpha Looks More Competitive- Message-ID: <878zje7usu.fsf@prep.synonet.com>   6 "Terry C. Shannon" <terryshannon@mediaone.net> writes:  B > Also, note the memory in this configuration... one would surmise> > that Dell would want to keep its entry-level price as low asD > possible, all the better to upsell. The 1GB memory makes we wonder > if Itanium is a memory hog.   A Fridays Euro spot market was $20US for 128M. So that is less than  $200 on current prices.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.H Spam-To: uce@ftc.gov,enforcement@sec.gov,sness@fcc.gov,hfurchtg@fcc.gov,#   mpowell@fcc.gov,gtristan@fcc.gov     ------------------------------  % Date: Wed, 30 May 2001 13:37:12 -0400 8 From: "Island Computers US Corp" <dbturner@islandco.com>X Subject: Re: Affordable VMS Workstations... All of a Sudden Alpha Looks More Competitive/ Message-ID: <thabocnn8r5p35@news.supernews.com>   H Or you could buy an Alpha 500Mhz from Island , with 1GB RAM, 18GB Ultra29 Disk Next day replacement parts for less than $2000   :0)    Hmmm -  I Granted theyre refurbished - but they look good and they work - what else  does one need ?    David T      -- Island Computers US Corporation  2700 Gregory Street 	 Suite 150  Savannah GA 31404  Tel: 912 447 6622  Fax: 912 201 0096  sales@islandco.com www.islandco.com' http://www.islandco.com/legal-email.htm   < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3B145D62.3F268DC5@fsi.net...  > Robert Deininger wrote:  > > H > > In article <slrn9h7v5q.jbj.strpic@lorien.elf.hr>, strpic@bofhlet.net (Vid > > Strpic) wrote: > > L > > > > Starting at $7,999(a), the Dell Precision Workstation 730 comes with a E > > > > single Intel Itanium processor at 733 MHz, 1 GB SDRAM, Matrox  Millenium G450D > > > > graphics card, 18 GB(b) SCSI hard disk drive, and a one-year
 next-business D > > > > day onsite service(c). The Dell Precision Workstation 730 is
 available for 3 > > > > order today from Dell's direct sales teams.  > > G > > Is this thing actually obtainium now?  Or is it just orderium, with ' > > deliverium some time in the future?  > * > Actually, I think it's more like Itanic. >  > -- > David J. Dachtera  > dba DJE Systems  > http://www.djesys.com/ > < > Unofficial Affordable OpenVMS Home Page and Message Board:! > http://www.djesys.com/vms/soho/  > H > This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings > is to be expected. > B > Feel free to exercise your rights of free speech and expression. > H > However, attacks against individual posters, or groups of posters, are > strongly discouraged.    ------------------------------  % Date: Wed, 30 May 2001 15:18:00 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk> Y Subject: Re: Affordable VMS Workstations... All of a Sudden Alpha Looks More Competitive  ) Message-ID: <3B150118.34DA29D4@bbc.co.uk>    "Terry C. Shannon" wrote:   G > > Is this thing actually obtainium now?  Or is it just orderium, with ' > > deliverium some time in the future?  > E > Seems to be a Daze of Future Passed sorta thing. If you want actual ! > unobtainium, Bailey Has It Now.  > . > http://www.claytonbailey.com/unobtainium.htm  A the noise that page makes is somewhat disturbing. Sounds like the  world is ending :-).   --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those of  MedAS or the BBC.    ------------------------------  % Date: Wed, 30 May 2001 10:33:39 +0100   From: steven.reece@quintiles.comW Subject: Re: Affordable VMS Workstations... All of a Sudden Alpha Looks MoreCompetitive H Message-ID: <OFCB2FDF4D.8BCFDBE6-ON80256A5C.0034292E@qedi.quintiles.com>  6 But then again how do you improve system performance : - increase memory; - increase CPU capacity; - increase I/O capacity * and others, not necessarily in that order.  I The "requirement" for 1GB of memory may be to give acceptable performance C rather than being a prerequisite.  i.e. the system has a bottleneck E somewhere in the I/O or CPU so increasing memory means it can do more + caching etc to overcome (hide) the problem.    Steve.   Jason Andrade wrote: >>> 6 "Terry C. Shannon" <terryshannon@mediaone.net> writes:  K >Also, note the memory in this configuration... one would surmise that Dell E >would want to keep its entry-level price as low as possible, all the  betterF >to upsell. The 1GB memory makes we wonder if Itanium is a memory hog.  H turn back the clock several years.  we bought all our DEC 3000/400s withH 32M of ram.  we very quickly had to upgrade to 64M because "the alpha isF an absolute memory hog with its 64bit stuff, compared to our sparc20s, hp 7xx, intel 486, etc".  G i don't doubt the itanium will need more memory than x86, but i suspectLI the 1G is based more on the sales pitch than any technical consideration.A  F marketing have probably realised that trying to pitch the leading edgeE at least to their technical customers makes them look stupid to offer.I the system with 128M of memory.  as it stands, it looks like a reasonablyiG rounded out system if you (for whatever reason) decided you had to haveh one. <<<.        A jason@azure.dstc.edu.au (jason andrade) on 30-05-2001 02:52:58 AMu  9 Please respond to jason@azure.dstc.edu.au (jason andrade)k   To:   Info-VAX@Mvb.Saic.Comp) cc:    (bcc: Steven Reece/QRED/Quintiles)p  H Subject:  Re: Affordable VMS Workstations... All of a Sudden Alpha Looks       More Competitive    6 "Terry C. Shannon" <terryshannon@mediaone.net> writes:  K >Also, note the memory in this configuration... one would surmise that DelltE >would want to keep its entry-level price as low as possible, all the  betterF >to upsell. The 1GB memory makes we wonder if Itanium is a memory hog.  H turn back the clock several years.  we bought all our DEC 3000/400s withH 32M of ram.  we very quickly had to upgrade to 64M because "the alpha isF an absolute memory hog with its 64bit stuff, compared to our sparc20s, hp 7xx, intel 486, etc".  G i don't doubt the itanium will need more memory than x86, but i suspectiI the 1G is based more on the sales pitch than any technical consideration.   F marketing have probably realised that trying to pitch the leading edgeE at least to their technical customers makes them look stupid to offerPI the system with 128M of memory.  as it stands, it looks like a reasonablyPG rounded out system if you (for whatever reason) decided you had to haven one.   -jason   ------------------------------   Date: 30 May 2001 15:33:39 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog)6 Subject: AIXL, AIX5L - binary compatibility with Linux, Message-ID: <9f33sj$g9u@gap.cco.caltech.edu>  I Many of us have suggested in this forum that it would be a useful featurec? if there was a binary compatibility mode common to Linux/Alpha,uH Tru64/Alpha, and OpenVMS/Alpha. Apparently the good folks at IBM agree -I but of course, for their own products, not for the good ship Alpha (whichcF at the moment Compaq is busy steering towards several huge icebergs.)    Here's a link to the AIX info: o  9    http://www-1.ibm.com/servers/aix/products/aixos/aix5l/e  D No, I didn't just stumble onto this information - unlike Compaq, IBMK actually advertises the enterprise products it wants to sell.  In this caseVG it was a big ad in Infoworld.  The AIXL documents use "affinity" a lot,MH but maybe that will restore the honor of this maligned word, because I'mK reasonably sure that IBM is serious about it, and that their affinity won'to# be the total fraud that DEC's was. m   Regards,   David Mathog mathog@caltech.edu? Manager, sequence analysis facility, biology division, Caltech WJ **************************************************************************J *                                RIP VMS                                 *J **************************************************************************   ------------------------------  % Date: Wed, 30 May 2001 08:08:30 -0400n, From: "J. Scott Greig" <jsgreig@geminaq.com>/ Subject: Re: Alphaserver 300 and OpenVMS V7.2-1l2 Message-ID: <em5R6.3168$Mb7.237876@brie.direct.ca>  @ "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message, news:EQQQ6.800$fi2.20419@news.cpqcorp.net...I > 1) I assume you have an ISA config setting with ISA--VGA as the handle.@         Correct. > @ > 2) I assume that this system booted V6.2 WITH THE IDENTICAL HW
 CONFIGURATIONfJ         Not quite. The original system had only 64 MB memory.  When one of the banks "broke",L         field service replaced the 32MB bank with a 128MB bank, hence 160MB. However,J         apart from the memory diffference, V6.2 and V7.1 (with all the Y2K patches)  ran on thisf          identical configuration. >[K > 3) The 1.6GB of memory seems quite a bit for this little system - is thatd7 > much supported on it?  Was it in the machine on V6.2?          See above. >d5 > 5) I assume you have **NO** ISA configuration file.l        Correct >WI > 6) Edit the file DECW$AUTOCONFIG.DAT and add something at the top like:r >  > DEBUG = -1 > C > (at least I think that's the syntax)  and boot the system from an 	 alternate  > console and see what you get.i  ?     I editted (created) DECW$USER_AUTOCONFIG.DAT with DEBUG=-1,HG     and experienced no differences - i.e. the DECWindows server processoJ     hangs (or more precisely, probably) loops (status= COM) at priority 6.  H     I believe that something is "missing" in the configuration step i.e. when II     boot minimum using the graphics console, I can log in at the graphicstF     console.  When I do SYSMAN> IO AUTO, the console "hangs" (i.e. see/     the original note about the top line &etc.)a  I     Certainly, if something is wrong during configuration, I would expectt1     problems with DECWindows attempting to start.h       Any other thoughts?e  	     Scott  > % > J. Scott Greig wrote in message ...t > >nB > >"Hoff Hoffman" <hoffman@xdelta.zko.dec.nospam> wrote in message/ > >news:w%xO6.609$fi2.15159@news.cpqcorp.net...tE > >> In article <OhvO6.75$Mb7.10326@brie.direct.ca>, "J. Scott Greig"s  > ><jsgreig@geminaq.com> writes: > >>I > >> :When attempting to install VMS V7.2-1 on an Alphaserver 300 (4/266) E > >> :the system "hangs" while configuring the graphics adapter.  The  adaptera) > >> :in question is an ATI MACH64 (ISA).n > >>E > >>   The ISA version?  AFAIK, that widget was comparatively rare onc OpenVMS.K > >>   AFAIK, you can have only one of these ISA widgets installed, and youw > >needaK > >>   to be very careful about ISA physical address assignment collisions.  > >1F > >I don't think that there is an addressing conflict- the same system worked
 > >fine under-; > >prior versions of VMS (6.2 and 7.1 with all Y2K patches)r > >L > >>D > >> :The 7.2-1 CD-ROM will boot when using a serial console, and soG > >> :VMS can get installed, but attempting to boot the newly installedLE > >> :software (as yet, unconfigured for DECnet, TCP/IP &etc.), using F > >> :the graphics console, and logging into the system using a serial- > >> :port, DECwindows server process is COM.s > >>J > >>   Boot off the serial console, install any available DECwindows Motif ECO J > >>   kit (availability is dependant on which version of DECwindows Motif youlK > >>   have), and install the OpenVMS ECO kits (all mandatory kits, and the K > >>   GRAPHICS kit), and check the DECW*.LOG files, and make sure you haveE) > >>   enough physical memory configured.l > >kI > >Done.  The graphics terminal hangs during the configuration portion of  > >startup.wI > >That is: after loading the bootstrap image, we get the startup herald,i then > >afterH > >the network device is configured (EWA0 in this case), when the screen > >normallyoI > >"changes modes" (?), instead of the normal "console" screen, the firsto	 > >three- L > >quarters of the top line on the monitor switch from white letters on blueE > >background to red letters on a black background.  After this, when I > >DECWindows tries to start up, the server process DECW$SERVER_0 "hangs"tF > >(state is COM), and the monitor, keyboard and mouse do not respond. > >uH > >Crashing the server process yields nothing but empty DECW*.LOG files.B > >The system has 160MB of memory (don't ask how that happened :)) > >>L > >Is it possible that this ISA device is no-longer supported, or, at least, > >wasL > >missed during the 7.2 release? (note: 7.1 CD boots, neither 7.2 nor 7.2-1 > >can)n > >e< > >Or, can someone give me a work-around (like something for > >SYS$USER_CONFIG.DAT)??e > >m > >Regards,e > >Scott > >a > >>+ > >>  ---------------------------- #includeu* > ><rtfaq.h> -----------------------------8 > >>       For additional, please see the OpenVMS FAQ -- > www.openvms.compaq.com/ > >>  --------------------------- pure personalr& > >opinion ---------------------------4 > >>    Hoff (Stephen) Hoffman   OpenVMS Engineering > >hoffman#xdelta.zko.dec.com  > >> > >y > >  >l >f   ------------------------------  % Date: Wed, 30 May 2001 15:04:50 +0100s% From: Alan Greig <a.greig@virgin.net>-? Subject: Andrew to vanish from c.o.v.? Sun bans Internet access 8 Message-ID: <8bv9ht0jhbo2mq1q39jfsu3buuvgnchoho@4ax.com>   According to The Register: >Sun Microsystems has told it employees NOT to use the Internet in an effort to shave a few dollars off its phone bill and thereby make its annual results look a bit better.  >a  5 See http://www.theregister.co.uk/content/7/19277.htmla   What will Andrew do all day? -- Alan   ------------------------------  % Date: Wed, 30 May 2001 07:17:35 -0700n! From: Tom Linden <tom@kednos.com>c Subject: anonymous ftp9 Message-ID: <CIEJLCMNHNNDLLOOGNJIMEHKCKAA.tom@kednos.com>w   running Multinet 3.3 on VMS 6.2u   How do I enable anonymous ftp?   User (norns:(none)): anonymous3 331 anonymous user ok. Send real ident as password.p	 Password:  530  ftp disallowed.
 Login failed.a ftp>   ------------------------------  # Date: Wed, 30 May 2001 14:41:50 GMT"- From: goathunter@goatley.com (Hunter Goatley)C Subject: Re: anonymous ftp/ Message-ID: <3b150655.2098687@news.process.com>g  F On Wed, 30 May 2001 07:17:35 -0700, Tom Linden <tom@kednos.com> wrote:    >running Multinet 3.3 on VMS 6.2 >  >How do I enable anonymous ftp?- >- >User (norns:(none)): anonymous:4 >331 anonymous user ok. Send real ident as password.
 >Password: >530  ftp disallowed.  >Login failed. >ftp>D  A You have to create an ANONYMOUS account in the SYSUAF file.  This9E URL will take you to the V4.3A documentation for setting up anonymousr: FTP.  I'm pretty sure the procedure was the same for V3.3.  J http://www.multinet.process.com/multinet-docs/admin_guide/Ch11.htm#E15E115   Hunter ------9 Hunter Goatley, Process Software, http://www.process.com/e9 goathunter@goatley.com     http://www.goatley.com/hunter/n   ------------------------------  % Date: Wed, 30 May 2001 10:41:31 -0400', From: "Richard Whalen" <WhalenR@process.com> Subject: Re: anonymous ftp+ Message-ID: <9f30rq$r5t$1@news.process.com>s  G The most likely problem is that the ANONYMOUS username in the SYSUAF is @ either not present, has the wrong password (should be GUEST), is) expired/password expired, or is disabled."    . "Tom Linden" <tom@kednos.com> wrote in message3 news:CIEJLCMNHNNDLLOOGNJIMEHKCKAA.tom@kednos.com...e! > running Multinet 3.3 on VMS 6.2m >l  > How do I enable anonymous ftp? >a  > User (norns:(none)): anonymous5 > 331 anonymous user ok. Send real ident as password.  > Password:t > 530  ftp disallowed. > Login failed.  > ftp>   ------------------------------  % Date: Wed, 30 May 2001 10:31:28 +0100e% From: Alan Greig <a.greig@virgin.net>y4 Subject: Re: Antigen found ExceedinglyInfected virus8 Message-ID: <pse9ht01rsh182bvquckrreaqg4oqr66jp@4ax.com>  3 On Tue, 29 May 2001 12:51:25 -0400, ANTIGEN_NTS0146-, <ANTIGEN_NTS0146@nts0047.beehive.com> wrote:  K >Antigen for Exchange found hs.zip infected with ExceedinglyInfected virus.oI >The file is currently Removed.  The message, "RE: HSDSA Scrip in a Batch. >Procedure", wasI >sent from Kent, Philip  JW1811  and was discovered in IMC Queues\Inboundo" >located at Bernstein/NYC/NTS0146.  C What a load of bollocks!!! The zip file only contained DCL. I guessiF "Antigen for Exchange" thinks that everything that ends with .COM mustA be for Windows. What "Antigen" seems to be saying  is that it haskD found a ZIP file infected with VMS. At least I know never to go near Antigen now.  C And now I know now why somebody complained about posting viruses to- c.o.v/info-vax     -- Alan   ------------------------------  % Date: Wed, 30 May 2001 12:36:17 -050000 From: Patrick Spinler <spinler.patrick@mayo.edu>) Subject: Anyone know what happened to OMIs( Message-ID: <3B152F91.D653B65A@mayo.edu>   Folks:  H At one point there was a package "Oscar's Menu Interpreter" which was to% be had from http://www.oveas.net/omi.r  F Unfortunately, oveas.net no longer exists.  Does anyone have a current reference to this package ?o   Thanks,m -- Pat  ? Refs: http://www.goatley.com/wwwboard/fileserv/messages/56.htmlb and : r http://www.google.com/search?q=cache:F6G1BlqpJBY:software.oveas.net/omi/prod.html+oscar%27s+menu+interpreter&hl=en --  ?       This message does not represent the policies or positions 1 	     of the Mayo Foundation or its subsidiaries.s3   Patrick Spinler			email:	Spinler.Patrick@Mayo.EDU '   Mayo Foundation			phone:	507/284-94854   ------------------------------  # Date: Wed, 30 May 2001 05:53:50 GMT/ From: Dirk Munk <munk@home.nl>. Subject: Re: Apache WebServer - How do I query' Message-ID: <3B148AED.279A7CE2@home.nl>5  J It is quite simple. You have to edit the httpd.conf file, and add an aliasD for every directory tree outside the standard Apache directory tree.  
 For instance:r  ! Alias /vmsdocs/ "/dka200/000000/"   : <Directory "/dka200/000000">       << no end slash here !!     options bla bla bal .... </Directory>     Rob Buxton wrote:r  	 > Hi All,1 >B0 > Just installed the Apache WebServer under VMS.G > Cool, it all works, and after copying some of our documentation I can 
 > see it all.o > C > What I want to do is be able to serve the VMS Docs CD Internally.i >h+ > So, what does the href command look like?d >uE > I'm not quite sure how I'm going to be able to call a device that'sh- > external to the Apache directory Structure.. >cB > The CD will be mounted on zippy$dka400:, so I need to be able toH > traverse back up from the Apache Server root directory and back to the > CD.o > tried various things:O< > href = "../zippy$dka400:/000000/index.html      through to! > href = "zippy$dka400/index.htmlo > etc. etc.r >g1 > Clues or pointers to documentation appreciated.a >c > Many Thanks in Advance,r >o > Rob.   ------------------------------  % Date: Wed, 30 May 2001 16:11:16 +0100-- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>m Subject: Re: Backup to CD?) Message-ID: <3B150D94.C3F337EB@bbc.co.uk>    Alan Greig wrote:f  3 > On Tue, 29 May 2001 12:39:36 +0100, Tim Llewellyn3" > <tim.llewellyn@bbc.co.uk> wrote: >C > >nP > >Anyone know how one determine the actual compression factor for a TZ8* drive,
 > >EASILY? >iD > Don't recall how far back this goes but certainly on the later DLTH > libraries with front panel LCD display and menus one of the menu itemsE > is compression ratio IIRC. I'm not near a drive to check right now.p  > Might be under the maint menu. >n  N A DLT library would be a considerable luxury for me. Something that works withH a raw drive and doesn't involve me doing the calculations would be best.  --w6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk-  A I speak for myself only and my views in no way represent those oft MedAS or the BBC.R   ------------------------------  % Date: Wed, 30 May 2001 08:31:58 -0700v! From: Tom Linden <tom@kednos.com>6* Subject: Can't SET HOST to cluster member.9 Message-ID: <CIEJLCMNHNNDLLOOGNJIGEHOCKAA.tom@kednos.com>n  H I have three machines that form a cluster (6.2 and 7.1 on AXp and 7.1 on VAX)  3 All of a sudden I can't get to the VAX or out of itG  = %SYSTEM-F-UNREACHABLE, remote node is not currently reachablee  I if I do a SHOW CLUSTER it looks fine, so there must be connectivity.  Theo' activity lights on the hub look normal.o   Any thoughts are appreciated   TIAt Tomo   ------------------------------  % Date: Wed, 30 May 2001 17:53:43 +0100a; From: "Leigh G. Bowden" <LGBowden@bowdenfamily.fsnet.co.uk>c Subject: CD-RW on a VMS system. / Message-ID: <9f38ik$3q0$1@newsg2.svr.pol.co.uk>r  K This is an update of something I posted a couple of months ago. I shouldn'toG really complain about freeware I suppose but trying to find informatione? about writing to a CD-RW from VMS has basically been confusing,h8 contradictary and a litany of busted websites and links.  < System: VAXstation 4000/60 with OpenVMS 6.2 and Y2K patches.% CDRECORD: 1.8a9 (only doco is 1.8.1). H CD-RW: Plextor Plexwriter 12/4/32 SCSI-2 (HD50) in a tabletop enclosure.B This Plextor is the closest I could find that matched the CDRECORDL information. It works as I have attached it to an NT system and it was fine.  B The Plextor has some selector switches on the back for parity (on)H termination (on -  last in chain), test (off) and block. The manual saysK that turning this on makes it ready for Unix environments. There is nothing J more in the manual about this than that. Everything below has been done in  the Unix and "Windows" settings.  K * At the console prompt the VAXstation recognises the device and that it ise a Plextor device.p  J * VMS recognises the device and recognises that it is a Plextor. See below/ but you'll need courier font to see it aligned.e  @ Disk PHJ$DKA200:, device type PLEXTOR CD-R PX-W124TS, is online,
 file-oriented 0     device, shareable, error logging is enabled.<     Error count                    1    Operations completed 4p1     Owner process                 ""    Owner UICa [SYSTEM]0     Owner process ID        00000000    Dev Prot S:RWPL,O:RWPL,G:R,W0;     Reference count                0    Default buffer sizec 512"  L * I have not been able to mount anything with this drive though with variousH combinations of MOUNT/MEDIA=CDROM/UNDEFINED_FAT=STREAM:2048 etc etc etc.K Some of these commands are from the VMS FAQ. Any form of mount generates annC error on the device count. An analysis of the error is shown below.a    D ---------------------- Forwarded by Leigh G Bowden/IT/North East/RAC Motoring8 Services on 29/05/2001 15:04 ---------------------------  3 "PHJ::BOWDENL"@re1.rac.co.uk on 29/05/2001 14:59:14y8 To:   Leigh G Bowden/IT/North East/RAC Motoring Services Subject:    E V A X / V M S        SYSTEM ERROR REPORT         COMPILED 29-MAY-2001  13:57:47J                                                                       PAGE 1.. ******************************* ENTRY     944. ******************************* E ERROR SEQUENCE 26.                              LOGGED ON:        SIDs 12000003E DATE/TIME 29-MAY-2001 13:56:26.74                            SYS_TYPEp 04140002 SYSTEM UPTIME: 0 DAYS 03:50:39J SCS NODE: PHJ                                                 VAX/VMS V6.27 DEVICE ERROR  KA46  CPU FW REV# 3.  CONSOLE FW REV# 1.4 ( GENERIC DK SUB-SYSTEM, UNIT _PHJ$DKA200:        HW REVISION     34302E31.9                                        HW REVISION = 1.04         ERROR TYPE            06nG                                        INVALID MODE SENSE DATA RETURNED         SCSI ID               02g3                                        SCSI ID = 2.a        SCSI LUN              00r4                                        SCSI LUN = 0.        SCSI SUBLUN           00u7                                        SCSI SUBLUN = 0.t        PORT STATUS     00000001aJ                                        %SYSTEM-S-NORMAL, NORMAL SUCCESSFUL2                                         COMPLETION        SCSI CMD        003F001A*                            00FF 1                                        MODE SENSE*        SCSI STATUS           00*+                                        GOOD  MODE SENSE DATA-        MEDIUM TYPE           00 :                                        DEFAULT MEDIUM TYPE        FLAGS                 0034                                        WRITE ENABLED        BLOCK DESC 1.   00000000i                        00080000cA                                        _DEFAULT DENSITY OF MEDIUM =                                        _NUMBER OF BLOCKS = 0.h<                                        _BLOCK LENGTH = 2048.        PAGE DESC 1.    0A000601                         00000000u8                                        PAGE CODE = 01(X)@                                        ERROR RECOVERY PARAMETERS7                                        _RETRY CNT = 10.r<                                        _CORRECTION SPAN = 0.<                                        _HEAD OFFSET CNT = 0.C                                        _DATA STROBE OFFSET CNT = 0. E                                        _RECOVERY TIME LIMIT = 0. MSEC         PAGE DESC 2.    F0040E02                         00000000                         00000000E                        00000000a8                                        PAGE CODE = 02(X)E V A X / V M S        SYSTEM ERROR REPORT         COMPILED 29-MAY-2001l 13:57:47J                                                                       PAGE 2.J                                        DISCONNECT/RECONNECT CONTROL PARAMS>                                        _BUFFER FULL RATIO = 4.A                                        _BUFFER EMPTY RATIO = 240.mF                                        _BUS INACTIVITY LIMIT = 0. USECG                                        _DISCONNECT TIME LIMIT = 0. USECdD                                        _CONNECT TIME LIMIT = 0. USEC        PAGE DESC 3.    C4013205                         00000008g                        00000000p                        96000000s                        00000000                         00000000                         00000000g                        00000000h                        00000000t                        00000000i                        00000000>                        00000000r                        00000000 8                                        PAGE CODE = 05(X)A                                        FLEXIBLE DISK DRIVE PARAMS D                                        _452. K BIT/SEC TRANSFER RATE<                                        _NUMBER OF HEADS = 8.8                                        _0. SECTORS/TRACKA                                        _0. DATA BYTES/PHYS SECTORe@                                        _NUMBER OF CYLINDERS = 0.D                                        _START CYL WRITE PRECOMP = 0.H                                        _START CYL REDUCE WRITE CURR = 0.>                                        _DRIVE STEP RATE = 150.C                                        _DRIVE STEP PULSE WIDTH = 0.n>                                        _HEAD SETTLE DELAY = 0.;                                        _MOTOR ON DELAY = 0.f<                                        _MOTOR OFF DELAY = 0.<                                        _HEAD LOAD DELAY = 0.D                                        _STARTING SECTOR, SIDE 0 = 0.D                                        _STARTING SECTOR, SIDE 1 = 0.        PAGE DESC 4.    0A000607o                        00000000h8                                        PAGE CODE = 07(X)/                                        RESERVEDE        PAGE DESC 5.    00000A08h                        00000000r                        00000000h8                                        PAGE CODE = 08(X)/                                        RESERVED         PAGE DESC 6.    0B00060Du                        4B003C00e8                                        PAGE CODE = 0D(X)/                                        RESERVEDn        PAGE DESC 7.    00040E0Ee                        00000000aE V A X / V M S        SYSTEM ERROR REPORT         COMPILED 29-MAY-2001  13:57:47J                                                                       PAGE 3.                        FF02FF01                         00000000 8                                        PAGE CODE = 0E(X)/                                        RESERVEDo        PAGE DESC 8.    0707142A                         23296F73                         00010D16                         0D160010                         45080000,                            4508 8                                        PAGE CODE = 2A(X)A                                        VENDOR UNIQUE PAGE FORMATS(        UCB$B_ERTCNT          04O;                                        4. RETRIES REMAINING         UCB$B_ERTMAX          00l;                                        0. RETRIES ALLOWABLEF        ORB$L_OWNER     00010004n:                                        OWNER UIC [001,004]        UCB$L_CHAR      1EC54008i;                                        DIRECTORY STRUCTUREDm4                                        FILE ORIENTED/                                        SHARABLE-0                                        AVAILABLE4                                        ERROR LOGGING0                                        ALLOCATED<                                        SOFTWARE "WRITE LOCK"7                                        CAPABLE OF INPUTo8                                        CAPABLE OF OUTPUT4                                        RANDOM ACCESS        UCB$W_STS           0000t        UCB$L_OPCNT     00000000i9                                        0. QIO'S THIS UNIT         UCB$W_ERRCNT        0001-:                                        1. ERRORS THIS UNIT        IRP$W_BCNT          0000 ?                                        TRANSFER SIZE 0. BYTE(S).        IRP$W_BOFF          0000n<                                        TRANSFER PAGE ALIGNED        IRP$L_PID       00010012n6                                        REQUESTOR "PID"        IRP$Q_IOSB      00000001lC                        000F0000        IOSB, 0. BYTE(S) TRANSFERRED1? ANAL/ERROR/SINCE=29-MAY-2001 13:45:00.00/INCLUDE=DKA200/OUT=T.Tu  J * I haven't really been able to try CDRECORD. I did do a CDRECORD -SCANBUSD DEV=0,2,0 out of interest and it reported that unable to find device DKA002!! Which was a surprise.  K * I've tried this on a Alphastation 255 and the results are pretty much the:; same except an error is reported as "invalid media format".w   ------------------------------  % Date: Wed, 30 May 2001 07:51:15 -0300a) From: fabio_compaq@ep-bc.petrobras.com.bri; Subject: Re: COMPAQ Service Contracts (Compax x EDS x DELL)uL Message-ID: <OF08BDB080.88A09027-ON03256A5C.003B8B1E@ep-bc.petrobras.com.br>  J How is the situation of Compaq X EDS if a few  days ago, DELL announced an agreement with EDS ...     Regards    FC        < "Andy Proctor" <aproctor@hotmail.com> em 29/05/2001 18:44:51  7 Favor responder a "Andy Proctor" <aproctor@hotmail.com>e             Info-VAX@Mvb.Saic.Com       7 Assunto: Re: COMPAQ Service Contracts for Alpha Systemsy    H The blame in my opinion lies with COMPAQ. If they employ EDS (Do they in thesB US for example?) then it is for them to ensure that EDS do the job	 properly,nK as they are clearly not. If the worldwide contracting services are with EDS K then the top guys at COMPAQ should hold back one of these big cheques untilMC the service provided by EDS is brought to an acceptable level. IMHO    Andy Proctor  7 "Richard Maher" <maher_rj@hotmail.c*m> wrote in message - news:9f0v5l$jbh$1@neptunium.btinternet.com...  > Hi,s >d, > So why are you blaming Compaq and not EDS? >tH > Maybe Compaq could take this ancillary function (making money) back on > board. >lI > Can anyone guess at the size of the check that Compaq gives EDS *every*t > year.  >s > Regards Richard Maher. >r@ > > "Andy Proctor" <aproctor@hotmail.com> em 28/05/2001 19:48:54 > >T; > > Favor responder a "Andy Proctor" <aproctor@hotmail.com>g > >r > >a > >o > >       Info-VAX@Mvb.Saic.Comg > >e > >v > >o7 > > Assunto: COMPAQ Service Contracts for Alpha SystemsA > >p > >  > > All, > >uH > > Maybe someone from COMPAQ who follows this group can assist? (I have > talkedF > > with Kerry Main on this subject before, but I am of the opinion it needs  > > further action) G > > I currently am responsible for our customer base worldwide customer  > > support,G > > and in this position one of the things I do is place contracts withi	 > various J > > COMPAQ service organisations around the world, for many types of AlphaK > > workstation. For eg. US alone is approx $50k per annum. Not the biggest  > you J > > may say, but worldwide it ends up a lot of money. Why then are certainF > > COMPAQ countries singularly incapable of performing the most basic salesnK > > functions for service contracts? I make things easier by only requiringt > NBDwD > > 9x5 hardware only cover. I sometimes approach the COMPAQ service centrese@ > > directly (eg Austria - excellent service/Germany - ExcellentB > > Service/France - has been poor but now improving slightly) and	 sometimes  > > usewK > > 3rd party providers to effect my requirements.(for the US) However wheno= > > there is a delay or problem, it always traces back to thes sales/contractB > > admin dept of the relevant COMPAQ country. For example, the US > organisationC > > takes approx 3 weeks to get me a quotation, then once I issue an purchaseB > > order (normally same day) it can be 4 further weeks to get the	 contract,- > orJ > > even confirmation that the order is being processed! NOT GOOD SERVICE. InH > > have the same (to a lesser extent) with Canada (Kerry, sorry but the > > problemCI > > still exists), Australia, Holland to shame but a few. (I can go on) In feelH > > that maybe I am not alone and I should bring this problem to generalJ > > attention. If COMPAQ want to keep customers then they should damn well get0G > > the staff able to process quotations and orders within a reasonable/ time.r > > AsI > > I'm in the UK, I would expect to wait 4 days for a quote from the US,  as aMK > > maximum. Am I being unfair? After all I have money I want to spend with F > > COMPAQ and keep using Alpha's and OpenVMS. The inability of COMPAQF > > sales/contract admin is driving me away from the fine products! If thereo > isI > > someone here who can at least bypass organisations like "Global Bid",  andeH > > bring this to the attention of more senior COMPAQ staff, please feel free > > toB > > pass this on. The situation is starting to become intolerable. > >  > >e > > Andy Proctor' > > (reply address is personal address)0B > > Work:(remove the obvious) Andy.Proctor@deleteme.spirentcom.com > >  > >. > >r > >  > >e > >p > >r > >d > >n > >n >p >i   ------------------------------  % Date: Wed, 30 May 2001 15:38:25 +0100-- From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>0; Subject: Re: COMPAQ Service Contracts (Compax x EDS x DELL)R) Message-ID: <3B1505E1.4DDBBF26@bbc.co.uk>3  * fabio_compaq@ep-bc.petrobras.com.br wrote:  L > How is the situation of Compaq X EDS if a few  days ago, DELL announced an > agreement with EDS ... >w   EDS will take anybody's money.     --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those ofk MedAS or the BBC.h   ------------------------------  % Date: Wed, 30 May 2001 09:34:18 -0400u. From: "Kenneth Randell" <kenr@datametrics.com>< Subject: Re: Contents of 7.3 kits -- missing VAX binaries CD+ Message-ID: <9f2sr3$jbs$1@bob.news.rcn.net>,   OK -- one more time.  L I got my second kit today...part #QA-MT3AA-H8, OVMS VAX+ALPHA CDROM KIT.  It! has the VAX and ALPHA binary CDs.   > The first kit I received is part #QA-MT1AA-H8, OVMS ALPHA UPD.  8 I suppose I should pay attention to the shipping labels.  I This explains that mystery, but not why the 'Guide to OpenVMS Version 7.3cI CD-ROMs' #AA-QSBRE-TE does not differentiate between the kits; it clearlyCL states (to me anyway) that both the VAX and ALPHA binary CDs are part of the kit.  H Oh well.  At least I have some nice pictures of SUN monitors to look at.   Ken RandellE  8 Tim Llewellyn <tim.llewellyn@bbc.co.uk> wrote in message# news:3B13B008.5715E3A0@bbc.co.uk...e >s >e > Kenneth Randell wrote: >yK > > Upon further examination, my kit is apparently missing the VAX Binarieso CD,-J > > which is described in Section 1.1 of the 'Guide to OpenVMS Version 7.3 > > CD-ROMS' as: > >B5 > > * OpenVMS VAX Version 7.3 Operating System CD-ROMM > >t# > > Anybody else have this problem?2 >r; > yup, but it is possible mine is a service contract issue,p >o > --8 > Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project2 > MedAS at the BBC, Whiteladies Road, Bristol, UK.C > Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk  >iC > I speak for myself only and my views in no way represent those of  > MedAS or the BBC.a >e >    ------------------------------  % Date: Wed, 30 May 2001 08:07:05 -0500C+ From: "Main, Kerry" <Kerry.Main@compaq.com>e0 Subject: RE: DECEvent, WEBESs - what's going on?R Message-ID: <DC4745D1A85CA04180C83CDC706A9D180D964D@cthexc02.americas.cpqcorp.net>   Rob,  . re: pointers .. here is a good place to start:L http://www.support.compaq.com/svctools/webes/webes_docs.html (includes WEBES V4) 7 http://www.support.compaq.com/svctools/webes/index.htmln   Regards,    
 Kerry Main Senior Consultanto Compaq Canada Inc. Professional Servicesn Voice: 613-592-4660d Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----< From: rob.buxton@wcc.govt.nz [mailto:rob.buxton@wcc.govt.nz] Sent: May 29, 2001 8:46 PM To: Info-VAX@Mvb.Saic.Comu, Subject: DECEvent, WEBESs - what's going on?    
 G'day all,  A Various News items have stated the demise of DECevent, many while  discussing WEBES.l  * So, much ignorance here, what is going on.F We're running DECEvent 3.3, it's particularly flaky on the VAXes and IE have to keep replacing one of the KNL files with the Factory version.0   Has DECEvent come to an end?  
 What's WEBES?e  " Pointers to info much appreciated.   Rob.   ------------------------------  % Date: Wed, 30 May 2001 10:14:13 +0100g% From: Alan Greig <a.greig@virgin.net>EG Subject: Re: Decnet Phase IV - Circuit EWA-1, Line synchronization losta8 Message-ID: <gae9ht49cbe14fude2b2dittba608rj78p@4ax.com>  1 On Tue, 29 May 2001 11:32:28 -0400, "Todd Nelson"o  <toddnelson_@hotmail.com> wrote:   >Hello All,b >eK >I have a Compaq DS10 Alphaserver running VMS 7.2-1, Decnet Phase IV, TCPIP- >v5.0a.- >-3 >I keep getting the following sequence of events...1 >9 >                NCP>dI >                %%%%%%%%%%%  OPCOM  29-MAY-2001 11:21:52.19  %%%%%%%%%%%S4 >                Message from user DECNET on MYALPHA> >                DECnet event 4.7, circuit down, circuit faultB >                From node 1.13 (MYALPHA), 29-MAY-2001 11:21:52.189 >                Circuit EWA-1, Line synchronization lost  >  >                NCP>GI >                %%%%%%%%%%%  OPCOM  29-MAY-2001 11:21:52.19  %%%%%%%%%%% 4 >                Message from user DECNET on MYALPHA. >                DECnet event 4.10, circuit upB >                From node 1.13 (MYALPHA), 29-MAY-2001 11:21:52.18 >                Circuit EWA-1  > Could you please post the output of the following NCP commands NCP> SHOW KNOWN LINE COUNT NCP> SHOW KNOWN CIRC COUNT  8 I suspect a hardware/config problem with your ethernet..    J >The network interconnect is Ethernet.. and I am using port A (the onboard >card has A and B ports).  >   >TCPIP finctionality seems fine. > L >I am wondering if I need to be using EWA-0 instead of EWA-1.  Problem is -- >no EWA-0 is present.U >N! >Output of NCP circuits follows..  > ' >            MYALPHA$ncp sho known circPF >            Known Circuit Volatile Summary as of 29-MAY-2001 11:22:58 >NA >               Circuit          State                   LoopbackO	 >Adjacent G >                                                                  Name 
 >Routing Node  > # >              EWA-1             onT >S, >            MYALPHA$ncp sho known circ char >  > E >            Known Circuit Volatile Characteristics as of 29-MAY-2001 	 >11:23:01  >  >            Circuit = EWA-1 > * >            State                    = on0 >            Service                  = disabled* >            Cost                     = 10* >            Maximum routers allowed  = 33* >            Router priority          = 64* >            Hello timer              = 150 >            Type                     = Ethernet > 8 >If any other information is needed, please let me know. > ! >Thanks in advance for your help.  >  >Todd. >  >  >    -- Alan   ------------------------------  % Date: Wed, 30 May 2001 08:33:34 -0400E- From: "Todd Nelson" <toddnelson_@hotmail.com>EG Subject: Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lost / Message-ID: <th9q3vt4mquua8@corp.supernews.com>O  0 I did the following... and things seem good now.   $NCP CLEAR CIRC EWA-1T $NCP PURGE CIRC EWA-1    $NCP SET CIRC EWA-0 STATE OFFT $NCP DEF CIRC EWA-0 STATE OFFG $NCP SET CIRC EWA-0 STATE ON $NCP DEF CIRC EWA-0 STATE ON  I Now EWA-0 seems to be working... as it corresponds to the 'A' port on the @ dual ethernet controller. (We only use one port - the 'A' port).  K The crazy thing now is that in TCPIP (we have Compaq / Digital TCPIP v5.0a) 2 I cannot seem to rid the machine of interface WE1.  J TCPIP> SET NO INTERFACE WE1 - seems to do the trick - until TCPIP services= are restarted.. then both WE0 and WE1 are present once again.I  L WE1 does not appear to interfere with the operation of TCPIP in any way, butC I really would like to see it gone permanently.  Maybe I am missing F something when I am issuing the command.  Perhaps a /perm qualifier or
 something.   Thanks    5 "Richard Brodie" <R.Brodie@rl.ac.uk> wrote in message % news:9f0i5u$e3m@newton.cc.rl.ac.uk...  > : > "Todd Nelson" <toddnelson_@hotmail.com> wrote in message) news:th7g7fc77pgf15@corp.supernews.com...0 >0@ > > I am wondering if I need to be using EWA-0 instead of EWA-1. >0; > Yeah. Decnet line/circuit names are rather non-intuitive.0 > EWA-1 would be port B. > & > > Problem is -- no EWA-0 is present. >0A > IIRC, you can bring only one circuit up on nonrouting phase IV.07 > Check what the permanent databases say with NCP LIST.(9 > Possibly you are starting UCX before DECNET; that mightI2 > prevent DECNET binding to the primary interface. >. >B >S   ------------------------------  % Date: Wed, 30 May 2001 08:37:09 -0400 - From: "Todd Nelson" <toddnelson_@hotmail.com>SG Subject: Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lostS/ Message-ID: <th9qam62n87g42@corp.supernews.com>U  % As requested, here are the outputs of  > NCP> SHOW KNOWN LINE COUNT > NCP> SHOW KNOWN CIRC COUNT  K I suspect that EWA-1 counters are still there until I reboot the machine... A the EWA-1 circuit has been cleared and purged... and EWA-0 added.   D Things seem okay, but here are the counters anyway.. if you have any additional input._   Thanks   Todd.    (Output Follows...)     MYALPHA$ncp sho known line count  . Known Line Counters as of 30-MAY-2001 08:26:37   Line = EWA-0  '       >65534  Seconds since last zeroedE"       596205  Data blocks received'       590711  Multicast blocks receivedS            0  Receive failure      93049290  Bytes received&     92789009  Multicast bytes received            0  Data overrun         8788  Data blocks sent#         4730  Multicast blocks sent=.            3  Blocks sent, multiple collisions+            1  Blocks sent, single collision -           16  Blocks sent, initially deferred        488210  Bytes sent"       199376  Multicast bytes sent            0  Send failure,            0  Collision detect check failure,           32  Unrecognized frame destination'            0  System buffer unavailable %            0  User buffer unavailable    Line = EWA-1  '       >65534  Seconds since last zeroed "            0  Data blocks received'            0  Multicast blocks receivedO            0  Receive failureY            0  Bytes received&            0  Multicast bytes received            0  Data overrun            0  Data blocks sent#            0  Multicast blocks sent .            0  Blocks sent, multiple collisions+            0  Blocks sent, single collision -            0  Blocks sent, initially deferred             0  Bytes sent"            0  Multicast bytes sent&          453  Send failure, including:$                 Carrier check failed,            0  Collision detect check failure,            0  Unrecognized frame destination'            0  System buffer unavailableN%            0  User buffer unavailable       MYALPHA$ncp sho known circ count  1 Known Circuit Counters as of 30-MAY-2001 08:26:43T   Circuit = EWA-0   '       >65534  Seconds since last zeroedR*           45  Terminating packets received&           43  Originating packets sent)            0  Terminating congestion loss &            0  Transit packets received"            0  Transit packets sent%            0  Transit congestion loss             0  Circuit down$            0  Initialization failure            0  Adjacency down            0  Peak adjacencies         4476  Data blocks sent       225283  Bytes sent"           45  Data blocks received         3050  Bytes received,            0  Unrecognized frame destination%            0  User buffer unavailableo     MYALPHA$2 "Alan Greig" <a.greig@virgin.net> wrote in message2 news:gae9ht49cbe14fude2b2dittba608rj78p@4ax.com...3 > On Tue, 29 May 2001 11:32:28 -0400, "Todd Nelson"0" > <toddnelson_@hotmail.com> wrote: > 
 > >Hello All,' > >SG > >I have a Compaq DS10 Alphaserver running VMS 7.2-1, Decnet Phase IV,  TCPIP1	 > >v5.0a.I > >T5 > >I keep getting the following sequence of events...  > >  > >                NCP>EK > >                %%%%%%%%%%%  OPCOM  29-MAY-2001 11:21:52.19  %%%%%%%%%%% 6 > >                Message from user DECNET on MYALPHA@ > >                DECnet event 4.7, circuit down, circuit faultD > >                From node 1.13 (MYALPHA), 29-MAY-2001 11:21:52.18; > >                Circuit EWA-1, Line synchronization lost1 > >5 > >                NCP>=K > >                %%%%%%%%%%%  OPCOM  29-MAY-2001 11:21:52.19  %%%%%%%%%%%A6 > >                Message from user DECNET on MYALPHA0 > >                DECnet event 4.10, circuit upD > >                From node 1.13 (MYALPHA), 29-MAY-2001 11:21:52.18  > >                Circuit EWA-1 >e@ > Could you please post the output of the following NCP commands > NCP> SHOW KNOWN LINE COUNT > NCP> SHOW KNOWN CIRC COUNT >q: > I suspect a hardware/config problem with your ethernet.. >  >mL > >The network interconnect is Ethernet.. and I am using port A (the onboard > >card has A and B ports).t > >t" > >TCPIP finctionality seems fine. > >,H > >I am wondering if I need to be using EWA-0 instead of EWA-1.  Problem is --c > >no EWA-0 is present.m > >2# > >Output of NCP circuits follows..  > >d) > >            MYALPHA$ncp sho known circ H > >            Known Circuit Volatile Summary as of 29-MAY-2001 11:22:58 > >hC > >               Circuit          State                   Loopbackl > >AdjacenteI > >                                                                  Namej > >Routing Nodes > > % > >              EWA-1             ona > > . > >            MYALPHA$ncp sho known circ char > >h > >hG > >            Known Circuit Volatile Characteristics as of 29-MAY-2001b > >11:23:01a > >a > >            Circuit = EWA-1 > >R, > >            State                    = on2 > >            Service                  = disabled, > >            Cost                     = 10, > >            Maximum routers allowed  = 33, > >            Router priority          = 64, > >            Hello timer              = 152 > >            Type                     = Ethernet > > : > >If any other information is needed, please let me know. > >t# > >Thanks in advance for your help.T > >  > >Todd. > >  > >d > >c >" > -- > Alan   ------------------------------  % Date: Wed, 30 May 2001 09:00:27 -0400 - From: "Todd Nelson" <toddnelson_@hotmail.com> G Subject: Re: Decnet Phase IV - Circuit EWA-1, Line synchronization lost / Message-ID: <th9rmc64l9784d@corp.supernews.com>k   I fixed TCPIP problem with   TCPIP> SET NOINTERFACE WE1! TCPIP> SET CONFIG NOINTERFACE WE1   E Once I rebooted, the WE1 does not show up now and everything is good!c   Thanks everyone!!,   Todd.     8 "Todd Nelson" <toddnelson_@hotmail.com> wrote in message) news:th9q3vt4mquua8@corp.supernews.com...t2 > I did the following... and things seem good now. >k > $NCP CLEAR CIRC EWA-1  > $NCP PURGE CIRC EWA-1  >  > $NCP SET CIRC EWA-0 STATE OFF, > $NCP DEF CIRC EWA-0 STATE OFFf > $NCP SET CIRC EWA-0 STATE ON > $NCP DEF CIRC EWA-0 STATE ON >aK > Now EWA-0 seems to be working... as it corresponds to the 'A' port on the?B > dual ethernet controller. (We only use one port - the 'A' port). > F > The crazy thing now is that in TCPIP (we have Compaq / Digital TCPIP v5.0a)4 > I cannot seem to rid the machine of interface WE1. >rL > TCPIP> SET NO INTERFACE WE1 - seems to do the trick - until TCPIP services? > are restarted.. then both WE0 and WE1 are present once again.v >wJ > WE1 does not appear to interfere with the operation of TCPIP in any way, butaE > I really would like to see it gone permanently.  Maybe I am missing H > something when I am issuing the command.  Perhaps a /perm qualifier or > something. >a > Thanks >  >r7 > "Richard Brodie" <R.Brodie@rl.ac.uk> wrote in messaget' > news:9f0i5u$e3m@newton.cc.rl.ac.uk...t > >d< > > "Todd Nelson" <toddnelson_@hotmail.com> wrote in message+ > news:th7g7fc77pgf15@corp.supernews.com...r > >tB > > > I am wondering if I need to be using EWA-0 instead of EWA-1. > >f= > > Yeah. Decnet line/circuit names are rather non-intuitive.  > > EWA-1 would be port B. > >r( > > > Problem is -- no EWA-0 is present. > >mC > > IIRC, you can bring only one circuit up on nonrouting phase IV.u9 > > Check what the permanent databases say with NCP LIST. ; > > Possibly you are starting UCX before DECNET; that mighto4 > > prevent DECNET binding to the primary interface. > >l > >a > >n >I >n   ------------------------------  % Date: Wed, 30 May 2001 02:06:48 -0400i' From: "Bill Todd" <billtodd@foo.mv.com>  Subject: Re: disk benchmarking' Message-ID: <9f22hk$qk$1@pyrite.mv.net>e  5 "Hamlyn Mootoo" <univms@bigfoot.com> wrote in message % news:3B1469EA.536EE9F3@bigfoot.com... I > It is entirley possible that you did not read through the entire threadh leadingo > up to the discussion,n  E That is correct - but I *did* read the initial post, which explicitlyaK referred to an application mix, and a couple of others which were similarly 	 eclectic.   5  so let me reiterate: We were originally defining our K > domain to a single disk unit being optimized for LARGE sequential writes,m so theL > "mix of access patterns and sizes" that do refer to does not exist for the > purposes of the discussion.   H Incorrect:  the size mix may not exist (though as far as I can tell onlyG because you defined it away by stating that sizes *shouldn't* be mixed,nH which itself is a debatable proposition), but unless you've narrowed theG discussion down to a single instance of a purely serial application theeG *access* mix still exists - and disk utilization will be compromised by H using small per-disk chunks (since the array will perform a larger totalE number of seeks for a given amount of data transferred) even when all-G individual transfers are large, as long as parallel activity is presentl (i.e., arm contention exists).  5   Second, since there is a sufficient amount of cacheeI > (usually designated to BOTH writeback, and read) VMS receives notice of  its I/O H > write request completion in the hundreds of microseconds range (if not less) in= > most cases, not the "3 to 5 milliseconds" as you indicated.O  H If so, the underlying disks aren't being driven anywhere near saturationL levels, in which case performance will be largely insensitive to chunk size.     This lets the RAIDJ > controller take the I/O request and spread it (per its cache degradationC > algorithm) at its leisure to as many disks as it can (still being.
 restricted byv$ > chunk size of each physical disk).  H Again, if the disks are so lightly loaded that there's plenty of time toI write back the data, this discussion is irrelevant.  But if the disks are H working their tails off, then while the write-back cache in concert withK per-disk queue optimization can help them, they'll *still* benefit from theOG reduction in seek-count that larger transfers provide:  caching doesn't E repeal the need to get the data back to disk (save to the degree thatrL overwrites to the same data can be coalesced before being destaged, but thatJ certainly doesn't seem to be the kind of access under discussion here), it just helps smooth it out..  .   Also, VMS is not CLAIRVOYANT! An applicationI > writing beyond the initial allocation on an RMS file only allocates ther amountJ > given by the RMS extend size (which I beleive has a default of 5 on mostH > sysytems, so it will allocate a least one cluster or group of clusters only upmG > to the size of the write request).  Any subsequent writes allocate anD
 additional	 > extent.1  H Perhaps I misunderstood what you meant by 'large' writes:  to me, that'sI well into the MB range, and thus well beyond the size that tuning clustercK sizes can help much.  So I read your previous statements as indicating thatdJ you did not realize that RMS would perform all allocation required for anyH *single* write in a single contiguous (if possible) allocation operation@ (and on re-reading them they still suggest that interpretation).  K Fortunately, RMS (unlike, e.g., Unix) provides a per-file settable 'defaultMK extension quantity' which specifies a lower bound on how much to extend the F file by when extension is required.  This is the way to tune extensionF behavior with a much finer-grained brush than tinkering with the basic2 volume structure (to the detriment of other uses).   - bill  G   This is not the same as a copy operation where, VMS has the benfit ofyK > knowing (without calling a psychic mind you) what the output size will ben in7 > advance, and allocate on a contiguous-best-try basis.- >  > HM   ------------------------------  % Date: Wed, 30 May 2001 04:07:15 -0400u( From: Hamlyn Mootoo <univms@bigfoot.com> Subject: Re: disk benchmarking+ Message-ID: <3B14AA33.A392D674@bigfoot.com>@  & --------------8DA0808920BD35B09A2F52E6* Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bitI  P Any analysis of tuning an I/O subsystem first starts with the establishment of aM model based on a single characteristic I/O stream, then proceeds to expand on P the model by adding multiple streams, all the while keeping in mind that, in ANYJ saturation case, everything goes to hell anyway.  If there are two clearlyN identifiable patterns of application I/O behavior, they should be separated asP much as possible for better overall performance of each.  This is why there is aM ten items or fewer checkout lane in the supermarket (or at least that was theAN original intent of this lane).  You seem to keep wanting to stand in the 10 orK fewer lane with your full shopping cart (metaphorically speaking).  For the N purposes of discussion, lets say that we have an application that has 10 usersL generally and 5 simultaneous users at any one time, and writes records of  2N megabytes at a time - not unreasonable I would think.  We also have a RAID 0+1P disk unit consisting of eight striped mirrors.  The chunk size is set to 256K onP the unit (remember chunk size is the size of the chunk  on a single disk device,M giving a 2MB stripe size).  Lets also say that the cluster size is 512 blocksiI (256K).  So, when one instance of the application gets ready to write, it L allocates 8 clusters (essentially "forewarning" the controller algorithm andI cache that its about to write 2MB).  Before the data is actually sent theeO controller gets a little bit of time to flush its cache if need be and/or ready M some disks for an actual write. We will assume for the purposes of discussion M that this is the first write and that the cache is relatively empty and it is O configured for writeback. How many disks? Eight of the mirrors will participate M in this operation based on the allocation "hint".  So the cache gets the data O then sends I/O completion status to the host.  It may or may not write the data K yet, I do not pretend to know sofisticated caching algorithms.  Along comesyM another write, so I'm guessing the cache probably figures it's a good idea to O write out the first 2 MB of data, whilst getting ready to receive another 2 MB. E Then, of course, along comes another 2MB from another instance of the J application - so far the cache is buying us a lot of headroom to wait, butO eventually, if it get enough simultaneous data, it will become overwhelmed.  It N is at this point that the disks will be in full operation (but temporarily) toL degrade the cache.  At the point of cache saturatioen, when the last requestO into the (I assume) FIFO queue clears this queue and get to be written, it gets J written AS FAST AS POSSIBLE because all disks are participating at once toP satify the write; therefore from the standpoint of the application ( and user ofM said app) the I/O happened in less time vis-a-vis if the chunk size were, say P 512K and only 4 mirrors were participating in the write.  In an uncached case, aO single application would  run faster with the smaller chunk size.  Why penalizeeN applications running on the system when there is little I/O contention, for anL almost identical performance under system load when the chunk size is largerL (because the controller will cause all disks to be in operation anyway underM full cache saturation -AND will handle the data efficiently by removing those $ pernicious "extra seeks" you claim)?  P This is analogous to single threaded processes (and those which are specificallyK not written to use multiple processors) running on a multiprocessor system. O Adding extra cpu's don't help the actual elapsed real world time of these apps,eM assuming it is the only app running on the system at the time - it only helpsrN alleviate cpu taken away by other apps in contention for the cpu when multipleO apps are running.  In short, for single threaded apps, multiple cpu's allow you P more load (more users,apps,etc.) not more speed (execution time of a single app)P - it still takes 9 months to make a baby, regardless if you put two women on the project or not.   J And yes, extend size should be set correctly once the behavior of app/file? extending habits have been established - at least we agree hereh   HM   .r Bill Todd wrote:  7 > "Hamlyn Mootoo" <univms@bigfoot.com> wrote in messager' > news:3B1469EA.536EE9F3@bigfoot.com...oK > > It is entirley possible that you did not read through the entire threadu	 > leadingE > > up to the discussion,r >nG > That is correct - but I *did* read the initial post, which explicitlyeM > referred to an application mix, and a couple of others which were similarlyi > eclectic.y >d7 >  so let me reiterate: We were originally defining our.M > > domain to a single disk unit being optimized for LARGE sequential writes,  > so theN > > "mix of access patterns and sizes" that do refer to does not exist for the > > purposes of the discussion.t >sJ > Incorrect:  the size mix may not exist (though as far as I can tell onlyI > because you defined it away by stating that sizes *shouldn't* be mixed, J > which itself is a debatable proposition), but unless you've narrowed theI > discussion down to a single instance of a purely serial application theoI > *access* mix still exists - and disk utilization will be compromised byeJ > using small per-disk chunks (since the array will perform a larger totalG > number of seeks for a given amount of data transferred) even when alleI > individual transfers are large, as long as parallel activity is presentb  > (i.e., arm contention exists). > 7 >   Second, since there is a sufficient amount of cacheaK > > (usually designated to BOTH writeback, and read) VMS receives notice ofe	 > its I/OrJ > > write request completion in the hundreds of microseconds range (if not
 > less) in? > > most cases, not the "3 to 5 milliseconds" as you indicated.c > J > If so, the underlying disks aren't being driven anywhere near saturationN > levels, in which case performance will be largely insensitive to chunk size. >  >   This lets the RAIDL > > controller take the I/O request and spread it (per its cache degradationE > > algorithm) at its leisure to as many disks as it can (still beingf > restricted by & > > chunk size of each physical disk). > J > Again, if the disks are so lightly loaded that there's plenty of time toK > write back the data, this discussion is irrelevant.  But if the disks are1J > working their tails off, then while the write-back cache in concert withM > per-disk queue optimization can help them, they'll *still* benefit from theeI > reduction in seek-count that larger transfers provide:  caching doesn'ttG > repeal the need to get the data back to disk (save to the degree thatgN > overwrites to the same data can be coalesced before being destaged, but thatL > certainly doesn't seem to be the kind of access under discussion here), it > just helps smooth it out.  >e0 >   Also, VMS is not CLAIRVOYANT! An applicationK > > writing beyond the initial allocation on an RMS file only allocates the  > amountL > > given by the RMS extend size (which I beleive has a default of 5 on mostJ > > sysytems, so it will allocate a least one cluster or group of clusters	 > only up3I > > to the size of the write request).  Any subsequent writes allocate ano > additional > > extent.o >qJ > Perhaps I misunderstood what you meant by 'large' writes:  to me, that'sK > well into the MB range, and thus well beyond the size that tuning cluster M > sizes can help much.  So I read your previous statements as indicating that L > you did not realize that RMS would perform all allocation required for anyJ > *single* write in a single contiguous (if possible) allocation operationB > (and on re-reading them they still suggest that interpretation). >CM > Fortunately, RMS (unlike, e.g., Unix) provides a per-file settable 'default M > extension quantity' which specifies a lower bound on how much to extend the H > file by when extension is required.  This is the way to tune extensionH > behavior with a much finer-grained brush than tinkering with the basic4 > volume structure (to the detriment of other uses). >  > - bill >lI >   This is not the same as a copy operation where, VMS has the benfit ofCM > > knowing (without calling a psychic mind you) what the output size will bei > in9 > > advance, and allocate on a contiguous-best-try basis.e > >n > > HM  & --------------8DA0808920BD35B09A2F52E6) Content-Type: text/html; charset=us-asciis Content-Transfer-Encoding: 7bitn  > <!doctype html public "-//w3c//dtd html 4.0 transitional//en"> <html>K Any analysis of tuning an I/O subsystem first starts with the establishment H of a model based on a single characteristic I/O stream, then proceeds toH expand on the model by adding multiple streams, all the while keeping inH mind that, in ANY saturation case, everything goes to hell anyway.&nbsp;K If there are two clearly identifiable patterns of application I/O behavior, K they should be separated as much as possible for better overall performanceoF of each.&nbsp; This is why there is a ten items or fewer checkout laneQ in the supermarket (or at least that was the original intent of this lane).&nbsp;RH You seem to keep wanting to stand in the 10 or fewer lane with your fullN shopping cart (metaphorically speaking).&nbsp; For the purposes of discussion,F lets say that we have an application that has 10 users generally and 5K simultaneous users at any one time, and writes records of&nbsp; 2 megabytes I at a time - not unreasonable I would think.&nbsp; We also have a RAID 0+1,F disk unit consisting of eight striped mirrors.&nbsp; The chunk size isK set to 256K on the unit (remember chunk size is the size of the chunk&nbsp;-G on a single disk device, giving a 2MB stripe size).&nbsp; Lets also sayoG that the cluster size is 512 blocks (256K).&nbsp; So, when one instancehL of the application gets ready to write, it allocates 8 clusters (essentiallyH "forewarning" the controller algorithm and cache that its about to writeI 2MB).&nbsp; Before the data is actually sent the controller gets a littleiH bit of time to flush its cache if need be and/or ready some disks for anH actual write. We will assume for the purposes of discussion that this isK the first write and that the cache is relatively empty and it is configured-G for writeback. How many disks? Eight of the mirrors will participate inPF this operation based on the allocation "hint".&nbsp; So the cache getsF the data then sends I/O completion status to the host.&nbsp; It may orI may not write the data yet, I do not pretend to know sofisticated cachingpF algorithms.&nbsp; Along comes another write, so I'm guessing the cacheF probably figures it's a good idea to write out the first 2 MB of data,J whilst getting ready to receive another 2 MB.&nbsp; Then, of course, alongG comes another 2MB from another instance of the application - so far the G cache is buying us a lot of headroom to wait, but eventually, if it getfI enough simultaneous data, it will become overwhelmed.&nbsp; It is at this K point that the disks will be in full operation (but temporarily) to degradepI the cache.&nbsp; At the point of cache saturatioen, when the last request.G into the (I assume) FIFO queue clears this queue and get to be written,.J it gets written AS FAST AS POSSIBLE because <i>all disks are participatingQ at once</i> to satify the write; therefore from the standpoint of the applicationeF ( and user of said app) the I/O happened in less time vis-a-vis if theF chunk size were, say 512K and only 4 mirrors were participating in theF write.&nbsp; In an uncached case, a single application would&nbsp; runK faster with the smaller chunk size.&nbsp; Why penalize applications runningsJ on the system when there is little I/O contention, for an almost identicalH performance under system load when the chunk size is larger (because theJ controller will cause all disks to be in operation anyway under full cacheM saturation -AND will handle the data efficiently by removing those perniciousr "extra seeks" you claim)?kF <p>This is analogous to single threaded processes (and those which areP specifically not written to use multiple processors) running on a multiprocessorI system.&nbsp; Adding extra cpu's don't help the actual elapsed real world H time of these apps, assuming it is the only app running on the system atM the time - it only helps alleviate cpu taken away by other apps in contentioncF for the cpu when multiple apps are running.&nbsp; In short, for singleH threaded apps, multiple cpu's allow you more load (more users,apps,etc.)I not more speed (execution time of a single app) - it still takes 9 monthstF to make a baby, regardless if you put two women on the project or not.M <p>And yes, extend size should be set correctly once the behavior of app/filea? extending habits have been established - at least we agree herel <p>HMs <p>. <br>Bill Todd wrote:F <blockquote TYPE=CITE>"Hamlyn Mootoo" &lt;univms@bigfoot.com> wrote in messagehZ <br><a href="news:3B1469EA.536EE9F3@bigfoot.com">news:3B1469EA.536EE9F3@bigfoot.com</a>...F <br>> It is entirley possible that you did not read through the entire thread <br>leadingt <br>> up to the discussion,eH <p>That is correct - but I *did* read the initial post, which explicitlyO <br>referred to an application mix, and a couple of others which were similarlym
 <br>eclectic. = <p>&nbsp;so let me reiterate: We were originally defining our G <br>> domain to a single disk unit being optimized for LARGE sequentials writes,n
 <br>so theH <br>> "mix of access patterns and sizes" that do refer to does not exist for thet! <br>> purposes of the discussion.iF <p>Incorrect:&nbsp; the size mix may not exist (though as far as I can	 tell onlyeK <br>because you defined it away by stating that sizes *shouldn't* be mixed,uH <br>which itself is a debatable proposition), but unless you've narrowed the G <br>discussion down to a single instance of a purely serial applicationp theaH <br>*access* mix still exists - and disk utilization will be compromised byF <br>using small per-disk chunks (since the array will perform a larger totalrI <br>number of seeks for a given amount of data transferred) even when allhK <br>individual transfers are large, as long as parallel activity is presentO" <br>(i.e., arm contention exists).= <p>&nbsp; Second, since there is a sufficient amount of cachetJ <br>> (usually designated to BOTH writeback, and read) VMS receives notice of <br>its I/ObH <br>> write request completion in the hundreds of microseconds range (if not  <br>less) inA <br>> most cases, not the "3 to 5 milliseconds" as you indicated.RK <p>If so, the underlying disks aren't being driven anywhere near saturationeJ <br>levels, in which case performance will be largely insensitive to chunk size.  <p>&nbsp; This lets the RAIDN <br>> controller take the I/O request and spread it (per its cache degradationG <br>> algorithm) at its leisure to as many disks as it can (still beingi <br>restricted by ( <br>> chunk size of each physical disk).H <p>Again, if the disks are so lightly loaded that there's plenty of time toH <br>write back the data, this discussion is irrelevant.&nbsp; But if the	 disks arenG <br>working their tails off, then while the write-back cache in concertl withF <br>per-disk queue optimization can help them, they'll *still* benefit from theH <br>reduction in seek-count that larger transfers provide:&nbsp; caching doesn'ttI <br>repeal the need to get the data back to disk (save to the degree thatxG <br>overwrites to the same data can be coalesced before being destaged,  but thatK <br>certainly doesn't seem to be the kind of access under discussion here),  it <br>just helps smooth it out.k6 <p>&nbsp; Also, VMS is not CLAIRVOYANT! An applicationI <br>> writing beyond the initial allocation on an RMS file only allocatese theM
 <br>amountF <br>> given by the RMS extend size (which I beleive has a default of 5 on mostlL <br>> sysytems, so it will allocate a least one cluster or group of clusters <br>only up-M <br>> to the size of the write request).&nbsp; Any subsequent writes allocateo an <br>additional
 <br>> extent.:I <p>Perhaps I misunderstood what you meant by 'large' writes:&nbsp; to me,- that'sM <br>well into the MB range, and thus well beyond the size that tuning clusternO <br>sizes can help much.&nbsp; So I read your previous statements as indicating  thatF <br>you did not realize that RMS would perform all allocation required for anysL <br>*single* write in a single contiguous (if possible) allocation operationD <br>(and on re-reading them they still suggest that interpretation).N <p>Fortunately, RMS (unlike, e.g., Unix) provides a per-file settable 'defaultK <br>extension quantity' which specifies a lower bound on how much to extende the O <br>file by when extension is required.&nbsp; This is the way to tune extensiontJ <br>behavior with a much finer-grained brush than tinkering with the basic6 <br>volume structure (to the detriment of other uses).	 <p>- billmL <p>&nbsp; This is not the same as a copy operation where, VMS has the benfit ofG <br>> knowing (without calling a psychic mind you) what the output sizen will bee <br>in; <br>> advance, and allocate on a contiguous-best-try basis.w <br>>i <br>> HM</blockquote>D </html>k  ( --------------8DA0808920BD35B09A2F52E6--   ------------------------------  % Date: Wed, 30 May 2001 13:14:28 -0400 ' From: "Bill Todd" <billtodd@foo.mv.com>i Subject: Re: disk benchmarking( Message-ID: <9f39lf$ima$1@pyrite.mv.net>  L I think you'll find that many Usenet users prefer plain-text to HTML message format.   5 "Hamlyn Mootoo" <univms@bigfoot.com> wrote in messager% news:3B14AA33.A392D674@bigfoot.com... ? > Any analysis of tuning an I/O subsystem first starts with thei establishment of aL > model based on a single characteristic I/O stream, then proceeds to expand onK > the model by adding multiple streams, all the while keeping in mind that,  in ANY2 > saturation case, everything goes to hell anyway.  F That doesn't work all that well when there *is* no 'characteristic I/OG stream', which is the situation typical systems have to deal with.  AndnL unless the load is sufficient to approach device saturation (i.e., result inD non-null device queues a fair amount of the time), at least half the7 optimization challenge (and potential gain) evaporates.      If there are two clearlyC > identifiable patterns of application I/O behavior, they should bet separated as: > much as possible for better overall performance of each.  L There are clearly common cases in which this is not true.  For example, whenH the resulting disk load differs dramatically (even on a moment-by-momentJ basis), separating the disk resources ensures that overall throughput willK suffer relative to a comparable system in which all resources are available J to each stream.  Or, to put it in the terms you use below, it's equivalentK to having all the 10-or-fewer checkout lanes standing idle because shopperst* happen to be buying in bulk at the moment.     This is why there is aK > ten items or fewer checkout lane in the supermarket (or at least that was- thefJ > original intent of this lane).  You seem to keep wanting to stand in the 10 or.D > fewer lane with your full shopping cart (metaphorically speaking).  H Yup:  that's the way to optimize throughput.  And request-prioritizationK (something harder to implement without creating customer rancor in a singlesG checkout line, but perfectly reasonable for a disk queue) is the way touE ensure that, e.g., real-time or interactive requests still get promptD service.  	   For thefJ > purposes of discussion, lets say that we have an application that has 10 usersgK > generally and 5 simultaneous users at any one time, and writes records ofe 2oL > megabytes at a time - not unreasonable I would think.  We also have a RAID 0+1dJ > disk unit consisting of eight striped mirrors.  The chunk size is set to 256K on J > the unit (remember chunk size is the size of the chunk  on a single disk device,tH > giving a 2MB stripe size).  Lets also say that the cluster size is 512 blocksK > (256K).  So, when one instance of the application gets ready to write, iteJ > allocates 8 clusters (essentially "forewarning" the controller algorithm ando% > cache that its about to write 2MB).   I Last I knew, that's not the way applications typically worked:  they tendyI just to issue write requests and let the system allocate space as needed. C And it would surprise me if RMS interacted in any way with the RAIDoI controller (and its cache) when allocating space (even if the applicationcE *did* do so explicitly), so I suspect your analysis in this regard isn flawed.a  &   Before the data is actually sent theK > controller gets a little bit of time to flush its cache if need be and/ore readyeD > some disks for an actual write. We will assume for the purposes of
 discussionE > that this is the first write and that the cache is relatively empty   B No, we won't, since in any typical (i.e., relatively steady-state)J real-world situation the cache is always full - and in any write-dominatedL situation that's driving the underlying disks anywhere near their saturationK points it's full of dirty data.  This is why the main benefit of write-back A caching for write-saturated activity is allowing disk-level queuefJ optimization, and why the resulting average write times are still measuredE in milliseconds rather than microseconds (where read activity is more*K prevalent, write performance is more like what you state:  subsequent ReadslG that *would have* been cached pay a price for this, but with reasonableo% victim choices it's often a net win).s  
  and it isE > configured for writeback. How many disks? Eight of the mirrors willy participate 3 > in this operation based on the allocation "hint".c  J I assert that there is no such hint, but would be happy to be corrected by someone in the VMS group.e     So the cache gets the dataL > then sends I/O completion status to the host.  It may or may not write the dataG > yet, I do not pretend to know sofisticated caching algorithms.  Alongs comesuL > another write, so I'm guessing the cache probably figures it's a good idea toK > write out the first 2 MB of data, whilst getting ready to receive another  2 MB.DG > Then, of course, along comes another 2MB from another instance of theaL > application - so far the cache is buying us a lot of headroom to wait, but@ > eventually, if it get enough simultaneous data, it will become overwhelmed.  It@ > is at this point that the disks will be in full operation (but temporarily) toi > degrade the cache.  L This situation is not temporary:  it's typical, steady-state behavior in any4 situation where the steady state is write-dominated.  :   At the point of cache saturatioen, when the last requestL > into the (I assume) FIFO queue clears this queue and get to be written, it getsL > written AS FAST AS POSSIBLE because all disks are participating at once toJ > satify the write; therefore from the standpoint of the application ( and user oftK > said app) the I/O happened in less time vis-a-vis if the chunk size were,e say : > 512K and only 4 mirrors were participating in the write.  L Notice that you have just (tacitly) accepted my earlier assertion that chunkD size should usually be *at least* 256 KB (compared with your earlier: assertion that "128K (or even smaller)" was the way to go.  D But you're still focusing on response time in the absence of overallE throughput considerations.  While with smaller chunks each individualhF request does indeed complete 'as fast as possible' *when the disks getK around to servicing it*, the *wait time* for service increases, because the L disks are satisfying requests at a slower *rate* due to the additional seeks? (plus rotational latencies) required by finer-grained requests.r  L In a lightly-loaded system, no wait-time penalty exists (i.e., the disks sitL idle waiting for each request, and therefore the write-back cache allows allH writes to appear to complete at memory rather than disk speeds) - but in@ such a system it doesn't matter *what* chunk size you use.  In aK heavily-loaded system, large chunks not only improve overall throughput buttJ frequently response-time as well (by reducing queuing delay more than theyI increase transfer time).  And in a typical system (where loads vary), you J compromise by using the largest chunk size that won't significantly worsenH light-load response time (at least 256 KB with today's disks, as the fewD milliseconds it may sometimes add is at least partly balanced by theC reduction in probability that more disks - and hence larger average F seek/rotational latency times - will be involved in a given transfer).     In an uncached case, a > single application  K Make that 'a single instance of a single application'.  And the observationrD applies to the cached case as well if the writes saturate the cache.  =  would  run faster with the smaller chunk size.  Why penalizetI > applications running on the system when there is little I/O contention,e  L That's why I suggested a chunk size minimum of 256 KB (rather than somethingH even larger) in responding to your original suggestion of "128K (or even
 smaller)".    for anlG > almost identical performance under system load when the chunk size isc largerH > (because the controller will cause all disks to be in operation anyway under&I > full cache saturation -AND will handle the data efficiently by removing  thoseh& > pernicious "extra seeks" you claim)?  K 'Claim' implies that there is some doubt, which is not the case.  The extra C seeks *will* occur unless the write-back cache is large enough that D subsequent writes to an individual disk *by a single instance of theG application* arrive before the previous writes to the same disk by that J instance have been destaged (in which case they can be concatenated into aH single larger write that executes just about the same way a larger chunkH size would have - except that the larger chunk size guarantees that thisF optimization will *always* occur, not just with write-back caches thatJ happen to be large enough compared with the application mix to accommodate it).  D Otherwise, the write-back cache simply allows queue-optimization forG whatever individual writes exist.  And queue-optimization eliminates nocE seeks unless requests just happen to target the same disk track (evenuL head-switching on the same cylinder incurs approximately the same latency as a single-cylinder seek).   > E > This is analogous to single threaded processes (and those which arel specificallyE > not written to use multiple processors) running on a multiprocessore system.a  B No, it's not, unless you limit the storage to a single disk (whereJ discussions of chunk size become uninteresting):  even with only a single,H purely sequential process using the RAID, there's *still* some trade-offE between the increase in request-completion time caused by spreading atI request over more disks (unless they're spindle-synchronized, as is often F true with RAID-3) and the decrease attributable to parallel transfers.  G (And of course even for such a process an SMP system can provide *some*hH parallelism within the system in satisfying the application's requests.)  K > Adding extra cpu's don't help the actual elapsed real world time of thesea apps,sI > assuming it is the only app running on the system at the time - it onlys helpsIG > alleviate cpu taken away by other apps in contention for the cpu whenS multipleG > apps are running.  In short, for single threaded apps, multiple cpu'st	 allow youeF > more load (more users,apps,etc.) not more speed (execution time of a single app) K > - it still takes 9 months to make a baby, regardless if you put two women  on the > project or not..  L Your analogy is flawed in multiple respects, as noted above.  Another one isI that when you include a write-back cache, using larger chunks can improveeK throughput even for a single serial process, since it still makes destagingnJ dirty data more efficient - as long as the cache is large enough to acceptI new dirty data while it is writing back existing dirty data to the array.    - bill   > L > And yes, extend size should be set correctly once the behavior of app/fileA > extending habits have been established - at least we agree heree >p > HM   ------------------------------    Date: 30 May 2001 06:26:45 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)eD Subject: Re: Hobbyist Licenses - was Re: Affordable VMS Workstations3 Message-ID: <x8q936z87Fqg@eisner.encompasserve.org>v  [ In article <3B145BCF.46990A80@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:b  G > Even today, affordable OpenVMS-capable hardware remains as elusive as J > affordable end-user licensing for OpenVMS, with the exception that older6 > Alphas are now somewhat common on the auction sites.  D According to conventional beancounter wisdom, it is against Compaq'sG economic interest to encourage use of anything other than new machines.n  ? You can argue all you want that things should be different, but % that is their premise for operations.o  @ The Hobbyist program does not come from that end of the company.   ------------------------------  % Date: Wed, 30 May 2001 07:41:23 -0500b1 From: "David J. Dachtera" <djesys.nospam@fsi.net>hD Subject: Re: Hobbyist Licenses - was Re: Affordable VMS Workstations' Message-ID: <3B14EA72.AD01AA7F@fsi.net>    Larry Kilgallen wrote: > ] > In article <3B145BCF.46990A80@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:c > I > > Even today, affordable OpenVMS-capable hardware remains as elusive asnL > > affordable end-user licensing for OpenVMS, with the exception that older8 > > Alphas are now somewhat common on the auction sites. > F > According to conventional beancounter wisdom, it is against Compaq'sI > economic interest to encourage use of anything other than new machines.  > A > You can argue all you want that things should be different, butx' > that is their premise for operations.p > B > The Hobbyist program does not come from that end of the company.  < The left hand does not know what he right hand is doing, eh?   -- i David J. Dachterak dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/   F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------    Date: 30 May 2001 11:44:44 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)sD Subject: Re: Hobbyist Licenses - was Re: Affordable VMS Workstations3 Message-ID: <WwmU0DEZtWXc@eisner.encompasserve.org>g  [ In article <3B14EA72.AD01AA7F@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:l > Larry Kilgallen wrote: >> k^ >> In article <3B145BCF.46990A80@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes: >> cJ >> > Even today, affordable OpenVMS-capable hardware remains as elusive asM >> > affordable end-user licensing for OpenVMS, with the exception that olderp9 >> > Alphas are now somewhat common on the auction sites.n >>  G >> According to conventional beancounter wisdom, it is against Compaq'suJ >> economic interest to encourage use of anything other than new machines. >> eB >> You can argue all you want that things should be different, but( >> that is their premise for operations. >>  C >> The Hobbyist program does not come from that end of the company.  > > > The left hand does not know what he right hand is doing, eh?  $ I think, rather, that they disagree.   ------------------------------    Date: 30 May 2001 06:17:14 -0700/ From: pjohanse@eoc.ecs.nasa.gov (Pete Johansen)n9 Subject: Re: How to store results of Command in a symbol?o< Message-ID: <8242cdd.0105300517.50dd8e45@posting.google.com>  / Thanks to everyone.  The info was very helpful.e    \ carl@gerg.tamu.edu (Carl Perkins) wrote in message news:<24MAY200119071275@gerg.tamu.edu>.... > Hamlyn Mootoo <univms@bigfoot.com> writes...= > }No such construct exists in VMS.  The closest VMS gets is:s8 > }$ PIPE SHOW SYSTEM  |  SEARCH SYS$INPUT "SWAPPER" | -; > }   (READ SYS$INPUT LINEVAR  ;  WRITE SYS$OUTPUT LINEVAR)n > }  > }HM  >  > Try  > 7 > $ PIPE SHOW SYSTEM  |  SEARCH SYS$INPUT "SWAPPER" | -r= >    (READ SYS$INPUT LINEVAR  ;  DEFINE/JOB LINEVAR &LINEVAR)i > ? > which will leave a logical name in the job logical name tablet? > called LINEVAR holding the information. (Via the magic of thel > ampersand operator.) > D > Using data from logical names is only slightly more difficult thanC > using data from symbols. You can get at the contents of the abovea
 > logical viau > ' > $ foo = F$TrnLnm("linevar","LNM$JOB")e > E > (The second argument to the F$TRNLNM lexical makes sure you get theoB > right logical name, in case there is one of the same name in the > process logical name table.) > 
 > --- Carl   ------------------------------  % Date: Wed, 30 May 2001 09:57:00 +0100 % From: Alan Greig <a.greig@virgin.net>o- Subject: Re: HSDSA Scrip in a Batch Procedureo8 Message-ID: <u9d9htokj02rmq8kgd0d7tf7juvocq3p5h@4ax.com>  @ On Tue, 29 May 2001 23:20:04 GMT, Scott Vieth <svieth@wi.rr.com> wrote:  C >Since when is it *okay* to post binary attachments to comp.os.vms?y  @ In general it certainly isn't but when it's only 5K and contains/ useful VMS stuff I personally don't mind. YMMV.-  B >Also, thanks for shipping a virus to everyone who reads the list.  F I have never heard of a virus which can be launched by unzipping. It;s not a self extracting zip file.n   -- Alan   ------------------------------    Date: 30 May 2001 06:23:27 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)e- Subject: Re: HSDSA Scrip in a Batch Procedurel3 Message-ID: <pmnwjjEPKJ8H@eisner.encompasserve.org>i  P In article <3B142F13.AB443FBF@wi.rr.com>, Scott Vieth <svieth@wi.rr.com> writes:D > Since when is it *okay* to post binary attachments to comp.os.vms?   Never.  C > Also, thanks for shipping a virus to everyone who reads the list.y  & Not to those of us who read it on VMS.   ------------------------------  % Date: Wed, 30 May 2001 11:59:28 +0100a* From: "Richard Brodie" <R.Brodie@rl.ac.uk>- Subject: Re: HSDSA Scrip in a Batch Proceduret+ Message-ID: <9f2jqi$gas@newton.cc.rl.ac.uk>a  U "Scott Vieth" <svieth@wi.rr.com> wrote in message news:3B142F13.AB443FBF@wi.rr.com... D > Since when is it *okay* to post binary attachments to comp.os.vms?  @ I don't know. How long has spamming mailing lists and newsgroups with virus warnings been OK?   ------------------------------  % Date: Wed, 30 May 2001 12:23:01 +0100r% From: Alan Greig <a.greig@virgin.net>i- Subject: Re: HSDSA Scrip in a Batch Procedureb8 Message-ID: <3ql9ht4t9u94s42u1mv5ns7t3alcsevfik@4ax.com>  @ On 30 May 2001 06:23:27 -0500, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) wrote:  Q >In article <3B142F13.AB443FBF@wi.rr.com>, Scott Vieth <svieth@wi.rr.com> writes:tE >> Since when is it *okay* to post binary attachments to comp.os.vms?o >r >Never.n > D >> Also, thanks for shipping a virus to everyone who reads the list. > ' >Not to those of us who read it on VMS.   B It isn't even a virus on Windoze unless you define VMS DCL to be a? virus. Some crap Windoze virus scanning software is incorrectlyrE declaring that a few lines of DCL with a .COM extension is actually at- Microsoft executable .COM containing a virus.i  C There is no binary code whatsoever in the .ZIP file. It's all ASCIIa. text just zipped up to prevent line wraps etc.   -- Alan   ------------------------------    Date: 30 May 2001 08:23:26 -05009 From: Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)s- Subject: Re: HSDSA Scrip in a Batch Procedurea3 Message-ID: <B3exTxjaOFsn@eisner.encompasserve.org>o  ` In article <3ql9ht4t9u94s42u1mv5ns7t3alcsevfik@4ax.com>, Alan Greig <a.greig@virgin.net> writes:  E > There is no binary code whatsoever in the .ZIP file. It's all ASCIIi0 > text just zipped up to prevent line wraps etc.  : That is still wrong for a text newsgroup like comp.os.vms.   ------------------------------  % Date: Wed, 30 May 2001 14:43:20 +0100e% From: Alan Greig <a.greig@virgin.net> - Subject: Re: HSDSA Scrip in a Batch Procedurer8 Message-ID: <84t9htg28f923dlm0j585njeg4js3u6455@4ax.com>  @ On 30 May 2001 08:23:26 -0500, Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) wrote:  a >In article <3ql9ht4t9u94s42u1mv5ns7t3alcsevfik@4ax.com>, Alan Greig <a.greig@virgin.net> writes:l >tF >> There is no binary code whatsoever in the .ZIP file. It's all ASCII1 >> text just zipped up to prevent line wraps etc.  > ; >That is still wrong for a text newsgroup like comp.os.vms.   D Is vmsnet.sources still operational? Does it have a valid moderator?D If the answer to both is yes then fine but last time I tried to post! anything there it never appeared.f  A I agree a text newsgroup such as comp.os.vms is not the place forsC binary postings but I personally don't mind if the odd *very small*aF zip file of VMS DCL gets posted. I can see why you and others disagree though..     -- Alan   ------------------------------  % Date: Wed, 30 May 2001 13:07:41 -0300i) From: fabio_compaq@ep-bc.petrobras.com.bru( Subject: IBM opens their Linux MainframeL Message-ID: <OFB6DAB346.21923CA8-ON032569D7.003D03CA@ep-bc.petrobras.com.br>   Clickh  : http://www-1.ibm.com/servers/eserver/zseries/os/linux/lcds   To see what is marketing by IBMn   Regardsr   FC   ------------------------------  % Date: Wed, 30 May 2001 19:17:54 +0010d% From: paddy.o'brien@zzz.tg.nsw.gov.aur Subject: Re: JRE.DMP5 Message-ID: <01K46PEIQIWY000TLO@tgmail.tg.nsw.gov.au>e  $ Again, thanks to all the responders.   I've given up.  H The WEBES stuff was loaded by my "accomplice" whilst was on leave.  But M nothing (CA -wise) was being done at the time that the .DMP file was created.   K I've searched through all the SYS$COMMON:[COMPAQ.SVCTOOLS] stuff and found uN many mentions of JRE but not of JRE.DMP.  I've taken a snapshot of the access L violation from show/process on this file (as was suggested by some to try), ! and will be sending it to COMPAQ.r  N I've been intrigued by Peter Weaver's post in this thread and his comments in ( others regarding this new devil of a CA.  N Having deleted this file, if I get a new file created or other problems which : seem to be associated, I shall contact COMPAQ immediately.   Regards, Paddy   Paddy O'Brien, Transmission Development,t
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australial   Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people, ; but that little whizz-bang apostrophe gives me little spam.a   ------------------------------  % Date: Wed, 30 May 2001 19:31:32 +0010e% From: paddy.o'brien@zzz.tg.nsw.gov.aue Subject: Re: JRE.DMP5 Message-ID: <01K46PVFYFSY000TGD@tgmail.tg.nsw.gov.au>a   P.S.  O And I'm going to PRODUCT REMOVE this until I hear that someone like Peter *is* p happy.   Regards, Paddy   Paddy O'Brien, Transmission Development,t
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australia    Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people,s; but that little whizz-bang apostrophe gives me little spam.(   ------------------------------  % Date: Wed, 30 May 2001 10:18:05 +0100 % From: Alan Greig <a.greig@virgin.net> ! Subject: Re: Login/Logout Problemo8 Message-ID: <dge9htsqp7ab8mtj7as0bjo19k5t7qsj0v@4ax.com>  5 On Tue, 29 May 2001 12:11:33 -0400, John Eisenschmidts <jeisensc@aaas.org> wrote:  >All our users are accessing the Alpha through Telnet (we have no terminals here). We're running into a problem with our financial system which tracks login/logout for concurrent processing. Users are logging in and then exiting their terminal emulator without logging out, which makes the financial system think they are still logged it (it increments and decrements a counter in a file called postctl.dat). e >a >The problem is, despite auditing login/logout I can't tell who is exiting normally and who isn't. I dumped the audit trail to a file and imported it into an Access database (I know I know, forgive me) to compare how many times each user logged in and out, and the counts looked the same. Does anyone have any suggestions on how I might catch who it is? We don't have the source to the financial software so I haven't seen the code that tracks how many people are in the system. They are all using a common lgicm> d, at one point I had thought about doing something with that.  C Typically commercial apps which record user status in a file have aeF way to reset this file to match reality. For example ManMan has SYU199C (SYstem Utility 199 - reset current user count). I'd check with thegB software suppliers for functionality such as this as a first step.     >Thanks, >Johns   -- Alan   ------------------------------  % Date: Wed, 30 May 2001 06:13:24 -0700 , From: "James Gessling" <jgessling@yahoo.com>! Subject: Re: Login/Logout Problems4 Message-ID: <9f2rlq$1tiva$1@ID-46415.news.dfncis.de>  G You may be able to do something with the LGI routines.  Specifically by 	 providingvC a LGI$ICR_LOGOUT routine.  It depends on how the process is gettingeI run down, it it's the equivalent of a stop/ide you're out of luck, but ite youdF are getting process termination records in your accouting file you can probablyJ use this.  I've personally only done the login side of the house, but see:  L http://www.openvms.compaq.com:8000/73final/4493/4493pro_034.html#installatio nlogoutk   Jim*   ------------------------------  # Date: Wed, 30 May 2001 16:54:23 GMTeA From: Simon Clubley<simon_clubley@remove_me.excite.com-Earth.UFP>a! Subject: Re: Login/Logout Problem 6 Message-ID: <3B9R6.5503$rn5.271045@www.newsranger.com>  L On Tue, 29 May 2001 12:11:33 -0400, in article <sb139200.025@aaas.org>, John Eisenschmidt wrote:  >aM >All our users are accessing the Alpha through Telnet (we have no terminals = M >here). We're running into a problem with our financial system which tracks =aH >login/logout for concurrent processing. Users are logging in and then =G >exiting their terminal emulator without logging out, which makes the =fE >financial system think they are still logged it (it increments and =s7 >decrements a counter in a file called postctl.dat).=20e >sL >The problem is, despite auditing login/logout I can't tell who is exiting =M >normally and who isn't. I dumped the audit trail to a file and imported it =hJ >into an Access database (I know I know, forgive me) to compare how many =J >times each user logged in and out, and the counts looked the same. Does =L >anyone have any suggestions on how I might catch who it is? We don't have =G >the source to the financial software so I haven't seen the code that =cH >tracks how many people are in the system. They are all using a common =D >lgicmd, at one point I had thought about doing something with that. >  >Thanks, >Johnp >l  > Have you checked the final status code in the accounting log ?  F I just tried a quick experiment on a Alpha with UCX 4.2 and a Win98 PC? running Kea 420. In each case, the process was at a DCL prompt.e  J When shutting down a session by exiting Kea, I got a %RMS-F-RER, file readI error in the accounting log. When shutting down a session by logging off,e? I got the standard %CLI-S-NORMAL, normal successful completion.    Simon.   -- e; Simon Clubley, simon_clubley@remove_me.excite.com-Earth.UFPeK Worrying idea #101: What if Microsoft goes into the Ada compiler business ?l   ------------------------------    Date: 30 May 2001 13:06:39 -0500+ From: kuhrt@encompasserve.org (Marty Kuhrt)9! Subject: Re: Login/Logout Problemi3 Message-ID: <jVq0bUC7LA04@eisner.encompasserve.org>o  Q In article <sb139200.025@aaas.org>, John Eisenschmidt <jeisensc@aaas.org> writes: N > All our users are accessing the Alpha through Telnet (we have no terminals =N > here). We're running into a problem with our financial system which tracks =I > login/logout for concurrent processing. Users are logging in and then =wH > exiting their terminal emulator without logging out, which makes the =F > financial system think they are still logged it (it increments and =8 > decrements a counter in a file called postctl.dat).=20 > M > The problem is, despite auditing login/logout I can't tell who is exiting =6N > normally and who isn't. I dumped the audit trail to a file and imported it =K > into an Access database (I know I know, forgive me) to compare how many =xK > times each user logged in and out, and the counts looked the same. Does =eM > anyone have any suggestions on how I might catch who it is? We don't have =dH > the source to the financial software so I haven't seen the code that =I > tracks how many people are in the system. They are all using a common =mE > lgicmd, at one point I had thought about doing something with that.n  A If I understand the problem, you have users that are still loggedeA in, but not really doing anything since they closed their telnet t? session without logging out.  My suggestion would be to get an  A idle process killer like Watcher to nuke sessions that have been   dormant for over xx minutes.   ------------------------------  # Date: Wed, 30 May 2001 16:13:47 GMTJ1 From: "Mark D. Jilson" <jilly@clarityconnect.com>H7 Subject: Re: OpenVMS locked up - How do I find out why?e2 Message-ID: <3B151D70.E09FB541@clarityconnect.com>  H If the system did not respond to the console or the hardware halt buttonG such that you got a usable >>> prompt then there is really nothing thatVH could have been done software wise.  Check the system's firmware versionF and update to whatever is the latest.  Have the errorlog analyzed justF to make sure there is nothing interesting in there and you may want toA rename the current one so that a fresh one gets started.  If thisaF happens again such that the console or halt button can't get you a >>>G prompt, leave the system the way it is an log a hardware call so that a @ field tech can examine all the visual indicators the system has.   Rich Seibel wrote: > : > We are not VMS experts, so a simple answer is preferred. > M > Our OpenVMS system locked up.  It was unresponsive to open windows on otherfL > machines and did not respond to its console.  The led numbers on the frontB > showed FF.  The box responded to a power cycle and came up fine. > J > The question is, where/how do I determine what might have happened?  How > do I view the system log?t > 
 > Rich Seibeld   -- tD Jilly	- Working from Home in the Chemung River Valley - Lockwood, NY0 	- jilly@clarityconnect.com			- Brett Bodine fan. 	- Mark.Jilson@Compaq.com			- since 1975 or so, 	- http://www.jilly.baka.com               -   ------------------------------  % Date: Wed, 30 May 2001 07:58:01 -0500e+ From: "Main, Kerry" <Kerry.Main@compaq.com>0> Subject: RE: Oracle Export 8.1.7 much slower than 7.3.2 on VMSR Message-ID: <DC4745D1A85CA04180C83CDC706A9D1860774A@cthexc02.americas.cpqcorp.net>  L >>> I can't find any reference to this bug number in the Metallink database.@ Can you give me some info on what it addresses? Does it apply toA Oracle 8.1.7? ( I can't see any patches that apply to 8.1.7 ) <<<>  J I would suggest contacting Oracle Support directly for a full explanation,L but my understanding is that 1416910 applies to both Oracle 8.1.6 and 8.1.7.   Regardsr  
 Kerry Main Senior Consultantw Compaq Canada Inc. Professional Servicesw Voice: 613-592-4660d Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com     -----Original Message-----B From: nothome@spammers.are.scum [mailto:nothome@spammers.are.scum] Sent: May 30, 2001 12:12 AMi To: Info-VAX@Mvb.Saic.Com > Subject: RE: Oracle Export 8.1.7 much slower than 7.3.2 on VMS    
 In articleG <DC4745D1A85CA04180C83CDC706A9D180D964A@cthexc02.americas.cpqcorp.net>,n- "Main, Kerry" <Kerry.Main@COMPAQ.com> writes:iE > A few misc notes ..assuming VMS V7.2-1 .. the following may help ..o > 2 > Ensure the following VMS patches are installed:  > - VMS721_SYS-V0800 (or later)i    SYS-V0900  ! > - VMS721_ACRTL-V0200 (or later)a    CHECK  . > - TCPIP V5.0A eco 2 (or TCPIP V5.1 eco 1) -     CHECK (5.1 eco 1) > I > Above VMS patches must be installed before the following Oracle patch -I > I > Ensure the Oracle patch that addresses bug report 1416910 in the Oraclen BUG  > database is installed.  J   I can't find any reference to this bug number in the Metallink database.@ Can you give me some info on what it addresses? Does it apply to> Oracle 8.1.7? ( I can't see any patches that apply to 8.1.7 )    ------------------------------  % Date: Wed, 30 May 2001 13:27:39 +0100   From: steven.reece@quintiles.com# Subject: PCSI and Account creation. H Message-ID: <OF811ED2C3.3297362D-ON80256A5C.0043D50C@qedi.quintiles.com>  G From various discussions that I've been part of and heard, I understand J that the recommended way of using PCSI for product installation kits is toK use PCSI and the PRODUCT command to place the product files onto the systemeI and then use a configuration tool (like NET$CONFIGURE or TCPIP$CONFIGURE)0> to carry out the configuration of the product being installed.  K How then does this work with the creation of accounts?  I can create a userfI account using the "account" statement in the PCSI$DESCRIPTION file.  ThistE will "create a system account".  I do realize that it is created as aeD managed object so is this the reason that it's available rather than1 leaving such creation to the configuration phase?o   Steve.   ------------------------------    Date: 30 May 2001 22:56:04 +0800, From: Paul Repacholi <prep@prep.synonet.com>M Subject: Re: Presentation Collateral From VMS Technical Update Days in Londonn- Message-ID: <874ru27unv.fsf@prep.synonet.com>M  6 "Terry C. Shannon" <terryshannon@mediaone.net> writes:  + > :> http://www.compaq.co.uk/campaigns/vms/r >  > Even wackier is: > , > Compaq PCs use genuine Microsoft Windows$ > www.microsoft.com/piracy/howtotell >  > F > Even better is the International Debt to Bill Gates calculator. ThisN > exciting cash-o-meter tells you, IN REAL TIME, just how much software piracy > costs! >  > ; > http://www.microsoft.com/piracy/howtotell/why/default.asp  >  > N > Seriously.... does anyone doing VMS and looking to see what went down at the= > Techical Update day care /what/ is installed on Compaq PCs?i  F Given what is there, I guess EVERYONE who reads what is there has too. They may not want to...o   -- e< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.H Spam-To: uce@ftc.gov,enforcement@sec.gov,sness@fcc.gov,hfurchtg@fcc.gov,#   mpowell@fcc.gov,gtristan@fcc.gov :   ------------------------------    Date: 30 May 2001 10:47:48 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>.% Subject: Re: Rambus loses another onetH Message-ID: <y43d9nnryj.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  # Shane.F.Smith@Healthnet.com writes:a  G > I hope so, but there is an alternative scenario. If Rambus goes bust,tH > wouldn't the liquidators sell off the assets? Isn't their intellectualL > property a sellable asset? If so (and remember, I'm not a lawyer so I have5 > no idea), Intel could buy it, and charge royalties.s  K In principle, yes. However, the point is that the patents might be declared ; unenforceable - you then have no property left to sell off.g  ? > You /know/ AMD would have an interesting time licensing it...I  I Not selling to a competitor (in another sector), while being the dominant G market player, will get the anti-trust guys on their back very quickly.    	Jan   ------------------------------  # Date: Wed, 30 May 2001 13:27:09 GMT B From: Simon Clubley <simon_clubley@remove_me.excite.com-Earth.UFP> Subject: Re: The future of VMS6 Message-ID: <Ny6R6.5174$rn5.256620@www.newsranger.com>  , On Tue, 29 May 2001 20:07:47 GMT, in articleN <q7v7htoh1gjsrutf1b7kd9d549lq1na0h3@4ax.com>, LBohan@dbc.spam_less..com wrote: >r= >On 29 May 2001 14:04:41 GMT, bill@triangle.cs.uofs.edu (Billo >Gunshannon) wrote:- >-
 >>Background: < >>   We have a required course for all of our seniors called; >>   "Senior projects". :-)  I am considering proposing thed< >>   porting of one (or maybe two) of the currently popular - >>   X toolkits to VMS as potential projects.s >> >u8 >if X toolkits look to be too hairy/too long a project,  >or hard to test:7 >l6 >the most common initial stumbling block in a vms port4 >of some unix tool, is often  the autoconf/configure* >steps, ie getting a usable build script.  > ' >Maybe look at a VMS port of autoconf ?s >i2 >perhaps not doable without a working unix shell;  >GNU bash might work for this, a/ >but, iirc, it assumes/uses v7.3 rtl libraries.  >b >other thoughts: >b: >  mySQL  (Dan O'Rielly, process.com had started on this?) >f# >or, perhaps one of the DBM api's  c8 >(My recollection is that autoconf was a stumbling block > here also) >n6 >SleepyCat (berkeley) had done a VMS port of at least ? >one of the DBM api's.  at least according to their web site.  n >   E A related idea: has anybody evaluated what would be involved in doing E a Cygwin style environment for VMS and would this be suitable for then above project ?s  . Some background for those not aware of Cygwin:  C Cygwin is the project to put the Unix tools on Windows. It works byr@ having a base set of code that emulates the Unix environment andJ works well enough that a wide selection of the Unix tools have been ported to Windows.r  H Given previous discussions in the newsgroup about filesystem differencesM as a cause of porting problems, you may be interested in the Cygwin approach, I as it could be used for VMS as well. If you look at a Cygwin installationsM from a MSDOS prompt, you see \CYGWIN\USR\* as been the top level of /usr, butiL from within BASH running under CYGWIN, /usr appears as a top level directoryN just as it does under Unix. Your normal Windows drives appear as, for example, /cygdrive/c/ for C:.   Simon.   -- /; Simon Clubley, simon_clubley@remove_me.excite.com-Earth.UFP K Worrying idea #101: What if Microsoft goes into the Ada compiler business ?i   ------------------------------   Date: 30 May 2001 16:01:48 GMT2 From: mathog@seqaxp.bio.caltech.edu (David Mathog) Subject: Re: The future of VMS, Message-ID: <9f35hc$g9u@gap.cco.caltech.edu>  ` In article <9f0a9p$162l$1@info.cs.uofs.edu>, bill@triangle.cs.uofs.edu (Bill Gunshannon) writes:@ >In an attempt to aid in the continuing existence of VMS, I have? >decided to try and take positive action.  However, I will need. >a bit of information first. >  >Background:; >   We have a required course for all of our seniors calledz: >   "Senior projects". :-)  I am considering proposing the; >   porting of one (or maybe two) of the currently popular m, >   X toolkits to VMS as potential projects. >J  I That could easily exceed the time the students have available to work on  I such a project, unless they work in teams, and maybe not even then.  X11 pA programs can be particularly difficult to port, see for instance:e  3    http://saf.bio.caltech.edu/www/X11_VMS_NOTES.TXT0  J As a somewhat more tractable alternative I'd like to suggest (and this is B just the embryo of the idea) a class in code portability where theE semester's project consists of developing a cross platform C specifictH "make" environment that must run on Unix, VMS, and WNT (W2K) and has the following requirements:   0   1.  must implement full dependency based buildQ   2.  it may not use external programs of any sort (ie, it can't awk, sed, SORT, oK       SEARCH, find, copy, cp, mv, ln -s etc) with the exception that it mayo=       emit "compile, library, and link" related instructions.lJ   3.  it must be written entirely in ANSI C.  Grade inversely proportionalG       to the number of #ifdefs required to run on a different platform.eI   4.  it must be able to discover platform specific properties, but only h@       through creating, compiling, and running small C programs.   extra credit for:e  %   5.  automatic dependency discovery.t  G A semester is a relatively short time, and this would be a good thing, sJ since it would force the students to keep the projects simple - hopefully > keeping them from evolving into monstrosities like "autoconf".  K Students could be divided into groups of 4, each such group being assigned mJ the same "gnu" program to test their cross platform method on.  The group J of 4 could collaborate on any porting within the program itself needed to J get it to run on the different platforms.  However, each student would be / responsible for his/her own "build" software.  n  F The whole class could agree upon a set of standard symbols for classesK of #ifdef usage.  (for instance, HAS_STRNCAT).  They could also agree upon Y% a common format for dependency files.n  L Part of the grade would include testing each student's build program on all E 4 platforms versus several of the gnu packages, using the agreed upon- dependency and #ifdef symbols.  J Anyway, something along those lines.  And something useful would come out H of it - full build environments for many of the gnu progrms on multiple 
 platforms.   Regards,   David Mathog mathog@caltech.edu? Manager, sequence analysis facility, biology division, Caltech e   ------------------------------  # Date: Wed, 30 May 2001 14:08:23 GMTa' From: Bob Bowles <rbowles@mediaone.net>t1 Subject: Re: UCX 5.x equiv for Multinet SMTP opts , Message-ID: <3B14FEB3.9209520F@mediaone.net>   Simon Clubley wrote:K > I am assuming that the messages are actually been routed to your machine,o > but are been rejected by UCX.t > " > If that is the case, then I use: > 6 > UCX> set host <original_node_name>/alias=foo.bar.net  7 Thanks, I bet that will do the trick.  With Multinet ifc7 I didn't have all the aliases set correctly and the DNSg5 MX record pointed to the box, I believe that the SMTPT4 messages would loop for a bit as the box would never2 be convinced that IT was the last stop in the SMTP path.s   ------------------------------  % Date: Wed, 30 May 2001 11:51:50 +0200e" From: "Hans Vlems" <hvlems@iae.nl>" Subject: Re: VAX 8650 boot failure( Message-ID: <9f2fpn$h6q$1@news.IAEhv.nl>  H You're probably right that the battery backup unit is not the reason for this> fault. IIRC the 8650 boots from an RL02 pack. Is that unit OK?   Hans, <gleason@encompasserve.org> wrote in message- news:xb1n4GeAr1TN@eisner.encompasserve.org...l >i >dF >   I have a VAX 8650, that refuses to boot - fails with "UNKNOWN MACHI > CHK" when I try. One theory being bruted about here is that the batterynB > back up is dead, and this is preventing the boot from succeding. >mC >   While I'm skeptical about that theory, I'd like to eliminate it=E > from consideration here. I recall dimly that there is a jumper that F > can be placed on the backplane that disables use of the battery backI > up circuitry, yet allows normal operation - anyone remember the detailso > of this trick? >n >o > Lee K. Gleason N5ZMR > Control-G Consultants. > lgleason@houston.rr.com    ------------------------------  # Date: Wed, 30 May 2001 13:11:35 GMTt/ From: "Alexandre SOUF" <alexandre.souf@free.fr> " Subject: Re: VAX 8650 boot failure5 Message-ID: <bk6R6.2648$hQ4.3810371@nnrp4.proxad.net>u  L in my opinion a battery problem would prevent the hardware from seeing disks3 I had this several times and it was always the samei  K the solution in case of battery failure is to configure the disk controlleri$ so that the strategy allows the boot  B something like CACHE_POLICY (for HSD10?) to set to 2 (boot anyway)   Regardsi     --   --F Alexandre SOUF - Consultant Indpendant en Informatique - Freelance IT
 Consultant= mail: alexandre.souf@free.fr    http://alexandre.souf.free.fr:L Tl. +33 3 59 12 79 80 /  Mobile +33 6 62 56 79 80     /     Fax +33 3 20 07 22 52t  5 <gleason@encompasserve.org> a crit dans le message :@( xb1n4GeAr1TN@eisner.encompasserve.org... >s > F >   I have a VAX 8650, that refuses to boot - fails with "UNKNOWN MACHI > CHK" when I try. One theory being bruted about here is that the batteryrB > back up is dead, and this is preventing the boot from succeding. >iC >   While I'm skeptical about that theory, I'd like to eliminate itaE > from consideration here. I recall dimly that there is a jumper thatnF > can be placed on the backplane that disables use of the battery backI > up circuitry, yet allows normal operation - anyone remember the detailsw > of this trick? >e >d > Lee K. Gleason N5ZMR > Control-G Consultantsc > lgleason@houston.rr.comw >o   ------------------------------    Date: 30 May 2001 09:44:45 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de>n  Subject: Re: vfork/exec questionH Message-ID: <y466ejnuvm.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>  ' "Paul Dembry" <pade@trifox.com> writes:i  M > The other problem is that I do not think it is possible to pass down a filetG > descriptor, e.g. the accept()'d socket, to a $CREPRC'd process.  WithvN > vfork()/exec() I do a dup() of the socket to a known file descriptor, we useL > 8, and then the exec()'d process does a dup() on fd 8 when it comes up.  I: > do not think that I can do a similar thing with $CREPRC.  M The usual way to do this is to define one of the three special logical values F that are $CREPRC arguments (input, output and error, which will be theN assigned values of SYS$INPUT, SYS$OUTPUT and SYS$ERROR in the process created)M such that it transports the requried information. This is done by NETACP, for N instance, when it creates a process to handle a DECnet (Phase IV) connection -3 IIRC, SYS$INPUT translates to the value of the NCB.   N > > A good suggestion is to go with an adaptive number of processes in a pool.- > > Most web servers, for instance, use this. L > In most cases, this would be ok however the incoming connection determinesN > which driver will be used.  For example, one connection might be for Oracle,N > the next for Rdb, a third for ADABAS and so I would need a process ready for- > each of these.  It gets messy after awhile.   D But computers are good at doing the accounting for such scenarios...   	Jan   ------------------------------  # Date: Wed, 30 May 2001 07:45:35 GMT  From:  <remy@younes.nom.fr>   Subject: Re: vms Emulation on NT$ Message-ID: <1103_991208673@seattle>  W On Tue, 29 May 2001 22:28:19 GMT, "Terry C. Shannon" <terryshannon@mediaone.net> wrote:o > N > > is there any Open VMS Emulation for AXP available that runs under Windowns9 > > NT ? Is there any SAMBA for VMS under AXP available ?  > >  > 4 > closest thing I can think of is www.charon-vax.com >   This stands for VMS emulation. I can confirm the port of Samba on OpenVMS. Works fine, but NT integration tools by CPQ also do the job well.i >  >  > .    ------------------------------  # Date: Wed, 30 May 2001 13:04:03 GMT1. From: "Alphaman" <alphaman64@nixspam-home.com>C Subject: Re: VMS Home Automation (was: Affordable VMS Workstations).< Message-ID: <7d6R6.87499$I5.20704185@news1.rdc1.tn.home.com>  L Yuh, I've got the FireCracker module interfaced to my Apache web server withI a bit of DCL coding around it.  I've modified the original FireCracker tosE allow me to schedule events with it and to dim to a set amount (e.g.,yI firecracker a03 dim 5 "tomorrow+06:00").  Plus I can access it via my webaI interface from anyplace in the world (e.g., make sure I turned the coffeeaJ pot off after I get to work, or schedule the Halloweenie lights to come on before the kids get home.)  L Fun, but by no means the level others are talking about herein.  There is anF update to the X10 std which will allow for fully bidirectional signalsJ (command and query).  Not too many modules available for it, and the price( is typically 4-5x a command-only module.   Aaronl --> Aaron Sakovich  http://members.home.net/sakovich/alphaman.html> Make April 15 just another day:        http://www.fairtax.org/I "vi troff ps su fsck grep rm du - - - And they call this a 'language'???"h  8 JF Mezei <jfmezei.spamnot@videotron.ca> wrote in message& news:3B14344B.D4BCFCC4@videotron.ca...$ > Shane.F.Smith@HEALTHNET.COM wrote:K > > Sounds like you want an X10 controller for VMS. I believe there are X101H > > units for all those things except maybe the thermostat. There again, justJ > > because I haven't seen them, doesn't mean they don't exist. A VMS port > > would probably be possible,l > J > Been there done that back in 1989. Haven't looked at the source for thatL > program, but I do use it everyday. I never bothered with a GUI since it is a 9 > lot faster to just say "X10 OPEN C 2"  at the $ prompt.h   ------------------------------  # Date: Wed, 30 May 2001 13:07:29 GMTa3 From: "Tom Wade" <t.wade@vms.eurokom.ie.removespam>cC Subject: Re: VMS Home Automation (was: Affordable VMS Workstations)7, Message-ID: <lg6R6.227$ND2.3281@news.iol.ie>  K I have followed with interest this thread on home automation.  What I wouldtI like to do is to have the Alpha turn on and off individual room radiators J (oil fired central heating) at different times.  This would need some kindI of valve that could be opened and closed electrically, and I haven't seenoJ mention of any of these on various home automation web sites.  Anyone seen something like this ?e   Tom Wade   ------------------------------  % Date: Wed, 30 May 2001 08:19:05 -0500P1 From: "David J. Dachtera" <djesys.nospam@fsi.net> C Subject: Re: VMS Home Automation (was: Affordable VMS Workstations)s' Message-ID: <3B14F349.C931E018@fsi.net>e   Tom Wade wrote:h > M > I have followed with interest this thread on home automation.  What I would K > like to do is to have the Alpha turn on and off individual room radiatorsoL > (oil fired central heating) at different times.  This would need some kindK > of valve that could be opened and closed electrically, and I haven't seeneL > mention of any of these on various home automation web sites.  Anyone seen > something like this ?:  @ Simple switch. Could be done with an X10 lamp module (switch, noG dimming). May require a voltage reduction transformer for each valve ifBD line-voltage solenoid valves make you nervous. The number of controlH "zones" may stretch the limits of how many X10 devices you can have in a given setup.   -- t David J. Dachtera  dba DJE Systemsu http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/o  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------    Date: 30 May 2001 10:01:42 -05003 From: malmberg@encompasserve.org (John E. Malmberg)rC Subject: Re: VMS Home Automation (was: Affordable VMS Workstations)i3 Message-ID: <quYns2Epnzz6@eisner.encompasserve.org>a  [ In article <3B146571.7E958E91@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:c > "John E. Malmberg" wrote: 	 >> [snip]e. >> The cost obstacles for home automation are: >>M >> 1. Network infrastructure.  Unless you have expensive custom wiring, or ane7 >>    expensive wireless network, X-10 is what is left.  >tJ > Well, not really. Done right, simple "telephone wire" (commonly known asI > "class 2 wiring" or "low voltage wire") should be usable to connect thea2 > individual sensors to the computer interface(s).  J And then you need either line drivers to eliminate the noise for dedicatedJ wire, and the same on the other end of the wire.  While the costs for each) component is small, together they add up.m  E Add to that local building codes that may restrict what wire that youi are using to plenum rated.  I There is also the asthetics factor.  Many people do not want any of these I wires to be visible.  The labor costs to hide this additional wire can ber high.o  B The costs on the controller end, and the cost of wiring can be cutF considerably by using a shared network bus, similar to ethernet.  ThisF currently increases the cost of the sensors or actuators considerably.  G The prices have gotten better in the last few years, but they are still2C much higher for either method for most people to see any payback ine8 energy savings as compared to more conventional methods.  K >> 2. Regulatory passed devices that will operate the motors needed to opensO >>    or close windows / blinds, that also have the required network interface.oL >>    Have you priced these?  Not cheap.  Of course in many homes the is not) >>    a power outlet close enough to use.7 >lJ > That wheel has already been invented. X10 is little more than a toy, butD > its what we've got. Let's stick to monitoring and warnings/alarms. >iL >> 3. Sensors that can detect when the environment has changed.  The sensorsL >>    are actually cheap, until you need to connect them to either a networkM >>    or dedicated wire.  This part can actually double as a security system.R >lI > Maybe, once the product matures and has gained a modicum of acceptance.  > K >> 4. Incandesant light controllers are low cost, but ones that can dim canSK >>    not be used with flourescent lights.  Flourescent controllers are not  >>    cheap. >gJ > Again, let's not re-invent those wheels. We'll live with X10 toys, until > something better comes along.n  L The problem exists even with the "X10 toys".  If you put an incandescent X10J controller in a light switch, you must then use only incandescent bulbs inI the fixture.  If this dimmer is controlling an outlet, woe be to you whenh6 a helpful person unplugs the lamp to vacuum the floor.  J >>    Energy efficient light bulbs may not work well with dimmers that are! >>    not designed just for them.r > J > ??? Not sure what that might mean, unless "Energy efficient light bulbs"C > is used to deliberately obscure the line between incandescent and.G > flourescent lights. Here again, we'll leave the dimming task (DIMMIT,W > DAMNIT!) to the X10 toys.l  J If you want an X10 or other device to dim a florescent or other some otherD energy efficient light, expect to pay a small ransom.  This includes1 manual dimmers.  This stuff is not plug and play.i  3 But try to convince the average consumer otherwise.c  	 >> [snip].J >> Simple mechanical timers and thermostats are right now provide a betterK >> return on investment than any computer controlled system for home energyr; >> control, if you need to put this into an existing house.: >eH > ...but do not provide the flexibility of being sensitive to changes inD > the outdoor environment, so that the cooling can activated when anG > outdoor air temperature rise is followed by an indoor air temperatureS > rise.   G Why should I be concerned about activating cooling indoors based on the ; outdoor temperature?  A thermostat indoors will do the job.cD A time based thermostat that adjusts the indoors temperature so thatI cooling and heating energy is reduced when the home is not expected to betG activated have been available.  Having a computer control one so that auK complex schedule can be maintained is available now, but in most cases, theo* unit can not communicate over the network.  I The basic problem is you have to initiate the temperature adjustment from C standby to occupied a bit in advance of when it is actually needed.   I In order for this to work, the computer in your personal cell phone woulduC need to detect when you have left for work so it could initiate theeH switch to the standby, and then when it detected that you were returning5 home, and within the correct distance, leave standby.a  F Systems are available where you can manually make the call, or use theG internet.  That technology has been available for years.  The consumerstE have voted with their pocketbook that they do not want to bother withw doing this.t  F And none of this is cost effective on a home that is normally occupied during the day.m  G Basically while it can be shown that a general purpose computer runningtG OpenVMS or another system can do these jobs, they can also be done withm simple thermostats.e  D > Nor do such things provide for warnings or alarms when things suchJ > as furnace filters, air conditioning coils, etc. have accumulated, shall* > we say, too much "deferred maintenance".  E Filters can be checked with simple sensors, but it takes considerablefF deferment of maintenance before noticable impairment will be detected.  6 A simple timing circuit will do the job more reliably.  I >> For a new house, paying attention to energy efficient designs and HVAC + >> systems will also have the best payback.r >eJ > Energy efficient designs, oddly enough, are part of what makes this typeG > of system all that much more important: the "tighter" a house is, the G > greater the danger from CO due to flue blockage, or even use of a gas F > stove. Some new homes here in sub'n Chgo. are prohibited to have gasG > stoves because they were built to be so air-tight. One developer evenpJ > uses dual hot water tanks to provide both hot potable water and heat forG > the house (Bigelow Homes) - NO BOILER! ...and guarantees no more thanrH > $200/yr in heating costs (well, that's shot to heck now with the price > of heating fuels!).n  G You seem to have some confusion.  There are tankless hotwater on demandiI systems offered with some new housing, and there is hot water zoned heat.a  J Neither system uses a boiler, and the hot water zoned heat method has been around for a very long time.  I That homebuilder was making that claim (or similar) when I was in Chicagom( over 12 years ago.  Read the fine print.  C Proper placement of windows, eaves, and awnings can do quite a bit.2H Passive solar heat costs almost nothing to operate or maintain, and doesI not add much to the cost of the house.  This is where most of the savingsy comes from.   F From what I have seen, most homebuilders do not pay attention to theseJ as it requires that the length eaves be calculated based on your longitudeF and that the house be oriented a specific way on the lot.  They do notD think that it will sell, and it prevents using the same houseplan in all states.b  H > This problem is exacerbated by some efforts to reduce indoor pollutionH > using electrostatic air cleaners. There is currently, to my knowledge,J > no such thing as an "ozono-ostat" to limit the EAC and prevent excessiveC > levels of ozone indoors. The system proposed should be capable oftH > limiting the operation of an EAC to keep the indoor ozone level within > predetermined limits.s  I I would suspect that you would need to have gone extremely overboard with1F installing electrostatic air cleaners to get the indoor ozone level upF to a noticeable or hazardous level.  Of course those same people mightC purchase an over-ozone sensor.  They are probably just as likely to,# purchase anything sold by spammers.     G Remember there is still a group in the U.S. that is trying to get a lawaG passed to require all buckets to leak their contents out in less than 5t0 minutes to prevent babies from drowning in them.  Q >> In the case of systems to assist the disabled, multiple autonomous controllersmJ >> are still preferred instead of a central system.  These have to surviveN >> either a power failure, or a network failure.  I would expect these to have: >> even stricter regulatory overview than home automation. ><E > The idea here is to provide assistance (voice actuated?) for when aDI > "physically challenged" person is alone, for whatever reason, and needsbG > assistance. Rather depends on the assist, device, etc. Even so, thesesD > autonomous systems can be monitored and their "health" tracked for > warning/alarm purposes.c  F At what point does the cost of the backup system exceed it's benefits?  C Put in a $1000.00 home monitoring system to do the work of a common 4 cell phone that the person is likely to have anyway?  B And that $1000.00 is only if you are not charging for the labor of installing it.  C > Still, I appreciate what you're saying. Eventually, some level ofx9 > regulation will need to be dealt with. This is a given.  >sG > Regulations are there to protect, not to serve as obstacles. I preferaH > rather than view regulations as obstacles, to instead incorporate such& > contingencies into my business plan.  I What I am trying to point out, that right now, even if you cut all of thefI costs to the bone, and even avoiding some of the regulatory requirements,eF using a central computer for monitoring these things is a luxury item.  J A luxury item that has shown in market tests to be unwanted by most of theB buying public, and in the cases where it was wanted, many failuresG occured.  Lawsuits were filed in one case by the buyers of a "showcase"o home.s  J Unless you want a home automation system that is only operated by computerG experts that know it will fail if they abuse it, you must use expensiver components.u  E And unless you use the premium expensive components, you will not getnF 100% uptime except through pure luck.  (And this is just for the stuffB to do the automation, and gather data, not the central computer to control and monitor it.)    E If you are seriously looking at doing this as a business, you need toe@ get much more familiar with the current work that is being done.  I Home automation is coming, but it is still a long way from being anythingfD other than a luxury toy.  And some of those people that want it will pay the price for it.s    I The interoperability standards are deadlocked because only the X-10 stuffaI has sold in sufficient quantity to establish a standard.  And X-10 is wayvG too expensive and way to primative for job.  So this means that a smarthK home that is not based on X-10 will have a spare parts maintenance problem.n  G Also the people who know how to make this stuff work can make much more-7 in industrial automation than in the small home market.    -Johnd wb8tyw@qsl.network Personal Opinion Only.   ------------------------------    Date: 30 May 2001 10:14:02 -05003 From: malmberg@encompasserve.org (John E. Malmberg)oC Subject: Re: VMS Home Automation (was: Affordable VMS Workstations),3 Message-ID: <HeVKmzBUovlg@eisner.encompasserve.org>R  , In article <lg6R6.227$ND2.3281@news.iol.ie>,5 "Tom Wade" <t.wade@vms.eurokom.ie.removespam> writes:wM > I have followed with interest this thread on home automation.  What I would K > like to do is to have the Alpha turn on and off individual room radiatorsML > (oil fired central heating) at different times.  This would need some kindK > of valve that could be opened and closed electrically, and I haven't seennL > mention of any of these on various home automation web sites.  Anyone seen > something like this ?t  F You should contract heating supply stores, and a heating contractor onK upgrading your system to zoned heat.  That is where you will find the partsT6 and the people that can advise you on the feasability.  I The solinoid valves are just one part.  The other is to have a thermostat H for each valve to control it.  There may be pitfalls that are obvious to= a heating specialist that are not apparent to the rest of us..  D For example, it is unknown if you have a central water tank for thisA configuration, or not.  Where are the current master thermostats?   H Once you have upgraded to zoned heat, then the next step would be remote controlled thermostats.u   -John' wb8tyw@qsl.network Personal Opinion Onlyl   ------------------------------  % Date: Wed, 30 May 2001 13:07:21 -0400r; From: "Brian Tillman" <tillman_brian@notnoone.notnohow.com>m Subject: Re: VMS mail problem $ Message-ID: <3b1528e5$1@news.si.com>  & Guy Morris (guy@dateline.co.uk) wrote:   >The reason I needL >to get it right is that I access the mails from a PC using Outlook Express,K >and it needs to see the count in order to access the mails that are there. 4 >Anybody got an idea how to reset the count upwards?  F Demonstrably untrue.  Suppose I have no messages in the NEWMAIL folderL (hence, no NEWMAIL folder at all) and 50 messages in the MAIL folder.  I canH enter the commands SELECT MAIL and COPY/ALL NEWMAIL and I'll then have aK NEWMAIL folder containing 50 messages, none of which VMS Mail will claim tocG be new.  SHOW NEW_MAIL_COUNT will show zero.  Outlook Express, however,tI will, when I choose SEND/RECEIVE download them all as new mail.  How do IH know?  I do it all the time. --A Brian Tillman                   Internet: tillman_brian at si.comtA Smiths Aerospace                          tillman at swdev.si.comu= 3290 Patterson Ave. SE, MS      Addresses modified to preventi< Grand Rapids, MI 49512-1991     SPAM.  Replace "at" with "@"8        This opinion doesn't represent that of my company   ------------------------------  % Date: Wed, 30 May 2001 11:07:27 +0200s2 From: Mike Rechtman <michael.rechtman@digital.com>" Subject: Re: VMS password cracking+ Message-ID: <3B14D46F.5D546A54@digital.com>1  B I suggest you look for a program (IIRC) called UAF, which used to G be on the WKU (now Process software) Freeware site. There was a programsF called GUESS_PASSWORD in the kit which checked each password against a dictionary, * against the username, and the owner field.  : try on ftp://ftp.process.com/vms-freeware/fileserv/uaf.zip   Mike.l   Kari Keronen wrote:  >  > Hi,t > K > Our computer security people asked me if there exists a program, that cantH > check if people are using  as passwords the words that are listed in a: > file. That is if I have a file that includes for example > 
 > password
 > 12345678 > etc... > J > the program checks from SYSUAF.DAT how many users are using listed (bad)I > passwords and makes some kind of statistics out of it. I have been toldkL > that this is easy to do in NT environment . I think this can be done usingE > $GETUAI etc ,but if someone already has this kind of program I'd be  > grateful.e > ) > OS. 7.2-1H1 Alpha and we have DEC C 6.0i > 	 > Thanks,i >  > -Kari- >  >  >    -- iE ---------------------------------------------------------------------lE Usual disclaimer: All opinions are mine alone, perhaps not even that.a? Mike Rechtman                            *rechtman@tzora.co.il* F Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  B   "20% of a job takes 80% of the time, the rest takes another 80%"E ---------------------------------------------------------------------. -----BEGIN GEEK CODE BLOCK-----a Version: 3.1: GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$6 PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@ ------END GEEK CODE BLOCK------f   ------------------------------   Date: 30 May 01 05:17:23 PDT From: mckinneyj@cpva.saic.come" Subject: Re: VMS password cracking( Message-ID: <hjXu$uTq9qFk@cpva.saic.com>  2 In article <01c0e85a$f6211dc0$120ba8c0@rlhkikker>,4  "Kari Keronen" <kari.keronen@radiolinja.fi> writes: > Hi,  > K > Our computer security people asked me if there exists a program, that can H > check if people are using  as passwords the words that are listed in a: > file. That is if I have a file that includes for example > 
 > password
 > 12345678 > etc... > J > the program checks from SYSUAF.DAT how many users are using listed (bad)I > passwords and makes some kind of statistics out of it. I have been toldrL > that this is easy to do in NT environment . I think this can be done usingE > $GETUAI etc ,but if someone already has this kind of program I'd ber > grateful.x > ) > OS. 7.2-1H1 Alpha and we have DEC C 6.0  > 	 > Thanks,o >  > -Kari- >  >  i >  "  8 	You might search the usual repositories for the old UAF6 	package. It include a GUESS_PASSWORD program that did6 	what you describe. Following is it's AAAREADME.TXT...   UAF and GUESS_PASSWORD  B         UAF is a general purpose utility for searching through the= authorization file based on any information stored within thesA authorization file, including privileges (specific privileges, ora) privilege classes), last login time, etc.n  D         GUESS_PASSWORD is a simple password guesser. You can feed itE a dictionary (see the source code, you'll need to change one #define) F of words to use as passwords, otherwise the default list is used (thisI was simply the passwords tried by the Internet Worm, I assumed Mr. MorrisbH had a reason for selecting the ones he did, maybe he didn't, who knows?)F You can also pass it passwords to guess (if you define it as a foreign	 command).,   ------------------------------  # Date: Wed, 30 May 2001 13:08:43 GMTe1 From: "Kari Keronen" <kari.keronen@radiolinja.fi>e" Subject: Re: VMS password cracking2 Message-ID: <01c0e909$a6dc5950$120ba8c0@rlhkikker>   Thanks Mike,   just what I was looking for.   -Kari-  = Mike Rechtman <michael.rechtman@digital.com> wrote in articles" <3B14D46F.5D546A54@digital.com>...D > I suggest you look for a program (IIRC) called UAF, which used to I > be on the WKU (now Process software) Freeware site. There was a program H > called GUESS_PASSWORD in the kit which checked each password against a
 > dictionary,n, > against the username, and the owner field. > < > try on ftp://ftp.process.com/vms-freeware/fileserv/uaf.zip >  > Mike.n >  > Kari Keronen wrote:w > >  > > Hi,  > > I > > Our computer security people asked me if there exists a program, that  canaJ > > check if people are using  as passwords the words that are listed in a< > > file. That is if I have a file that includes for example > >  > > password > > 12345678
 > > etc... > > F > > the program checks from SYSUAF.DAT how many users are using listed (bad)iK > > passwords and makes some kind of statistics out of it. I have been toldeH > > that this is easy to do in NT environment . I think this can be done usingtG > > $GETUAI etc ,but if someone already has this kind of program I'd bei
 > > grateful.l > > + > > OS. 7.2-1H1 Alpha and we have DEC C 6.0n > >  > > Thanks,o > > 
 > > -Kari- > >  > >  > >  >  > -- gG > --------------------------------------------------------------------- G > Usual disclaimer: All opinions are mine alone, perhaps not even that.iA > Mike Rechtman                            *rechtman@tzora.co.il*nH > Kibbutz Tzor'a.                          Voice (home): 972-2-9908337  D >   "20% of a job takes 80% of the time, the rest takes another 80%"G > ---------------------------------------------------------------------t! > -----BEGIN GEEK CODE BLOCK-----m > Version: 3.1< > GCM/CS d(-)pu s:+>:- a++ C++ U-- L-- W++ N++ K? w--- V+++$8 > PS+ PE-- t 5? X- tv-- b+ DI+ D-- G e++ h--- r+++ y+++@! > ------END GEEK CODE BLOCK------  >    ------------------------------  + Date: Wed, 30 May 2001 14:26:17 +0000 (UTC) ' From: david20@alpha1.mdx.ac.uk (D.Webb)n" Subject: Re: VMS password cracking+ Message-ID: <9f2vu9$np3$1@aquila.mdx.ac.uk>d  [ In article <3B13D70C.E5955652@fsi.net>, "David J. Dachtera" <djesys.nospam@fsi.net> writes:i >Kari Keronen wrote: >>   >> Hi, >> eL >> Our computer security people asked me if there exists a program, that canI >> check if people are using  as passwords the words that are listed in ae; >> file. That is if I have a file that includes for examples >> e >> passworda >> 12345678d	 >> etc...a >> sK >> the program checks from SYSUAF.DAT how many users are using listed (bad)oJ >> passwords and makes some kind of statistics out of it. I have been toldM >> that this is easy to do in NT environment . I think this can be done usingoF >> $GETUAI etc ,but if someone already has this kind of program I'd be >> grateful. >> h* >> OS. 7.2-1H1 Alpha and we have DEC C 6.0 >pE >GETUAI will return the hashed password. It cannot be "cracked". EveniF >OVMS must hash the password given at login time and compare it to the? >stored hashed password to see if the right password was given.t >hH >To do the kind of checking you're talking about will require re-hashingE >every file entry for every username, and comparing the stored hashedrG >pasword to the resulting hashed file entry. I believe there's a systemh= >call to do this for each entry. CPU intensive, but possible.h > 	 >AFAIK...t >   / Before doing this search the system dictionary nC sys$library:vms$password_dictionary.data for the words in the list.    eg  N Alpha2:search SYS$COMMON:[SYSLIB]VMS$PASSWORD_DICTIONARY.DATA  abstention     P abstention<NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL>P <NUL><NUL><NUL><NUL><NUL><NUL><NUL><NUL>                                        P Alpha2:                                                                           N By default VMS will check passwords against this list when user's change their
 passwords.  O Hence you may be able to reduce the time taken to check all the users passwordsl> by weeding out words which are in the system dictionary first.    1 You can add your own passwords to this list by :-r  . 1) Create a file containing the new passwords 5    Each password on a separate line and in lowercase.g    d   eg  ' $ CREATE LOCAL_PASSWORD_DICTIONARY.DATAf
 somefamous localheroese [Ctrl/Z]    ! 2) Merge in your local additions a   $ SET PROCESS/PRIVILEGE=SYSPRV4 $ CONVERT/MERGE/PAD LOCAL_PASSWORD_DICTIONARY.DATA -- _$ SYS$LIBRARY:VMS$PASSWORD_DICTIONARY.DATA  g    = The above is taken from the OpenVMS systems manager's manual.p    N I believe the system also performs other checks for weak passwords when users M change their passwords (eg not allowing the password to be the same as their 2I userid) but I don't know where a full list of these checks is documented.h        
 David Webb VMS and Unix team leader CCSS Middlesex University   ------------------------------  % Date: Wed, 30 May 2001 07:58:46 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> 0 Subject: Re: VMS price increases, rumor or fact?' Message-ID: <3B14EE86.19356191@fsi.net>d   Rich Jordan wrote: > M > My boss tells me  he heard that VMS is getting a 15% across the board price K > hike on licenses.  He gets all the admin and pricing info, I get the tech K > stuff, so I don't know what he got that says that, from Compaq or anotherhN > source, or what specific products or licenses are supposedly being affected. > Is this in fact the case?o  , I couldn't say, but I wouldn't be surprised.  4 In public transit, this is called "terminal spiral":  "  1. Revenue and ridership are low.1  2. Services are reduced and fares are increased.n<  3. Ridership falls due to reduced service and higher fares.  4. Go to 1.  B This can end in only one final result: public transit will die. In7 OpenVMS, the tune is the same - only the lyrics change:r  %  1. Revenue and market share are low. 2  2. Services are reduced and prices are increased.@  3. Market share falls due to reduced service and higher prices.  4. Go to 1.  8 This can end in only one final result: OpenVMS will die.  G If true, the already dead OpenVMS market in many parts of the U.S. will 2 be virtually assured of never reviving ever again.   -- t David J. Dachterao dba DJE Systemsz http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/h  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.o   ------------------------------  # Date: Wed, 30 May 2001 13:39:05 GMTp4 From: "Terry C. Shannon" <terryshannon@mediaone.net>0 Subject: Re: VMS price increases, rumor or fact?= Message-ID: <ZJ6R6.11703$zl5.4217925@typhoon.ne.mediaone.net>t  < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3B14EE86.19356191@fsi.net...t > Rich Jordan wrote: > > I > > My boss tells me  he heard that VMS is getting a 15% across the boarde pricesH > > hike on licenses.  He gets all the admin and pricing info, I get the techE > > stuff, so I don't know what he got that says that, from Compaq ork another F > > source, or what specific products or licenses are supposedly being	 affected.  > > Is this in fact the case?0 > . > I couldn't say, but I wouldn't be surprised. >   H Best to check with Compaq before propagating an unverified rumour... theJ most recent US Alpha Systems Price Book (mid-May) does not reflect a price hike.i   ------------------------------  % Date: Wed, 30 May 2001 10:03:54 -0300t) From: fabio_compaq@ep-bc.petrobras.com.bre0 Subject: Re: VMS price increases, rumor or fact?L Message-ID: <OF91B7B268.58908C21-ON03256A5C.0047A425@ep-bc.petrobras.com.br>   There is a way of revenge...  & Dont migrate to Compaq solutions ! ! !  C Instead of migrate from OpenVMS to Unix/WNT under the Compaq Brand,eA jump to HP, Sun. IBM, Dell and forget the Q .... if you can ! ! !n     Regardso   FC        B "David J. Dachtera" <djesys.nospam@fsi.net> em 30/05/2001 09:58:46  = Favor responder a "David J. Dachtera" <djesys.nospam@fsi.net>a             Info-VAX@Mvb.Saic.Com,      0 Assunto: Re: VMS price increases, rumor or fact?     Rich Jordan wrote: >aG > My boss tells me  he heard that VMS is getting a 15% across the boardr price K > hike on licenses.  He gets all the admin and pricing info, I get the techsK > stuff, so I don't know what he got that says that, from Compaq or anotheruD > source, or what specific products or licenses are supposedly being	 affected.s > Is this in fact the case?n  , I couldn't say, but I wouldn't be surprised.  4 In public transit, this is called "terminal spiral":  "  1. Revenue and ridership are low.1  2. Services are reduced and fares are increased.a<  3. Ridership falls due to reduced service and higher fares.  4. Go to 1.  B This can end in only one final result: public transit will die. In7 OpenVMS, the tune is the same - only the lyrics change:o  %  1. Revenue and market share are low.s2  2. Services are reduced and prices are increased.@  3. Market share falls due to reduced service and higher prices.  4. Go to 1.  8 This can end in only one final result: OpenVMS will die.  G If true, the already dead OpenVMS market in many parts of the U.S. willl2 be virtually assured of never reviving ever again.   -- David J. Dachterao dba DJE Systemse http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/0  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.d   ------------------------------  # Date: Wed, 30 May 2001 16:24:09 GMTt& From: "john nixon" <jnixon@cfl.rr.com>0 Subject: Re: VMS price increases, rumor or fact?? Message-ID: <J89R6.347306$fs3.56082500@typhoon.tampabay.rr.com>a  E I believe that trying to to verify the truth is what he was trying tolF accomplish.  "CHECK WITH COMPAQ"???   Please tell me how that is done?  ? "Terry C. Shannon" <terryshannon@mediaone.net> wrote in messageo7 news:ZJ6R6.11703$zl5.4217925@typhoon.ne.mediaone.net...y > >e >aJ > Best to check with Compaq before propagating an unverified rumour... theL > most recent US Alpha Systems Price Book (mid-May) does not reflect a price > hike.  >  >i >d >l   ------------------------------    Date: 30 May 2001 15:45:35 +0200G From: Jan Vorbrueggen <jan@mailhost.neuroinformatik.ruhr-uni-bochum.de> / Subject: Re: VXT2000 hardware/firmware questionsH Message-ID: <y466ejrlvk.fsf@mailhost.neuroinformatik.ruhr-uni-bochum.de>   dgordon@compaq.com writes:  C > Nope.  Removal of the floating point bits and changes to make theeD > InfoServer 150 use a standard rather than a custom cab (as the 100H > required) were done to reduce the cost of the hardware.  The 150 had a, > faster (and cheaper) network chip as well.  D You sold so many of them a re-spin of the CVAX chip was worth-while? I am astounded.m   	Jan   ------------------------------  % Date: Wed, 30 May 2001 04:24:53 -0400k- From: JF Mezei <jfmezei.spamnot@videotron.ca>t  Subject: What does READ/NEW do ?, Message-ID: <3B14AE4F.53EAB8EF@videotron.ca>  P Often, the "cure all" soution to MAIL problems is to issue the READ/NEW command.  N What exactly does this command do ?  Does it scan MAIL.MAI for any record thatJ has the unread bit set and then write back the resulting cound in the mail profile data file ?n   ------------------------------  # Date: Wed, 30 May 2001 12:37:51 GMTaB From: Simon Clubley <simon_clubley@remove_me.excite.com-Earth.UFP>. Subject: [OT] Current Microsoft v DOJ status ?6 Message-ID: <zQ5R6.5115$rn5.255132@www.newsranger.com>  K On Tue, 29 May 2001 14:15:11 -0500, in article <3B13F53F.4C55D9C9@fsi.net>,e David J. Dachtera wrote: > H >Yes, certain benefits have been gained by playing ball with the RedmomdE >bullies. Of late, however, their monopolistic tactics have become so G >intolerable that even the Justice Department has seen fit to intervene & >(much too little and years too late). >   , What is the current Microsoft v DOJ status ?  L I'm British and live in the UK, so I don't really know which way the currentK US administration is likely to go with this. I do know that prior to the USsF election, the feeling in this newsgroup was that a Bush administrationC would be less likely to continue with the action against Microsoft.f  J Is this likely to still be the case or has the DOJ viewpoint not changed ?  D BTW, is there a general move in the US to all things Microsoft or doF people still regard VMS/Unix/whatever as the correct tools for the jobL when it comes to critical systems ? Here in the UK, we have to deal with the
 following:  1 http://www.theregister.co.uk/content/4/19239.html    Simon.   -- r; Simon Clubley, simon_clubley@remove_me.excite.com-Earth.UFP K Worrying idea #101: What if Microsoft goes into the Ada compiler business ?8   ------------------------------  % Date: Wed, 30 May 2001 08:13:45 -0500n1 From: "David J. Dachtera" <djesys.nospam@fsi.net>s2 Subject: Re: [OT] Current Microsoft v DOJ status ?' Message-ID: <3B14F209.C0A7D90C@fsi.net>    Simon Clubley wrote: > M > On Tue, 29 May 2001 14:15:11 -0500, in article <3B13F53F.4C55D9C9@fsi.net>,e > David J. Dachtera wrote: > >hJ > >Yes, certain benefits have been gained by playing ball with the RedmomdG > >bullies. Of late, however, their monopolistic tactics have become soaI > >intolerable that even the Justice Department has seen fit to intervene ( > >(much too little and years too late). > >  > . > What is the current Microsoft v DOJ status ? > N > I'm British and live in the UK, so I don't really know which way the currentM > US administration is likely to go with this. I do know that prior to the US-H > election, the feeling in this newsgroup was that a Bush administrationE > would be less likely to continue with the action against Microsoft.h  A Experience with the "Dubbya" admin. tends to bear out the group's-D initial take on the matter. He's a corporate lackey - period, end of
 statement.  :L > Is this likely to still be the case or has the DOJ viewpoint not changed ?  E The DOJ tends to be strongly influenced by the feelings of the gov't.e leaders.  fF > BTW, is there a general move in the US to all things Microsoft or doH > people still regard VMS/Unix/whatever as the correct tools for the job% > when it comes to critical systems ?h  G With the release of the news that even the U.S. military is buying into8F Redmond's bulls--t, the perception that Micro$hit is all things to allE {sheep,lemmings,whatever-works-for-you} is likely to be proliferated.f  * > Here in the UK, we have to deal with the > following: > 3 > http://www.theregister.co.uk/content/4/19239.htmlb  D This is yet another example of Redmond's bully tactics: lock out theG competition, rather than ensuring the greatest possible user base. They-F may have achieved their greedy goals in the UK - so be it. My feeling,B however, is that other such entities will be inclined to learn theH lessons of this debacle (no guarantees, naturally) and ban M$ from theirD gov't "portal"s for such counter-competitive, monopolistic nonsense.  B Here's a bit of "innovation" I'll bet you'll never see come out ofH Redmond: play nice with the rest of the world, instead of trying to rule it.e   -- t David J. Dachtera  dba DJE SystemsA http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/i  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.t   ------------------------------   End of INFO-VAX 2001.298 ************************