1 INFO-VAX	Sat, 08 Apr 2000	Volume 2000 : Issue 196       Contents:# Compaq Song and Dance (the lack of) ' Re: Compaq Song and Dance (the lack of) 1 Re: create your own personal SYS$COMMON directory % Re: Critical/important bug on VMS ...   DECserver 90M booting problem...$ Re: DECserver 90M booting problem...$ Re: DECserver 90M booting problem...
 Freeware MIME  Re: Freeware MIME  Re: Freeware MIME  Re: Freeware MIME 2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory2 Re: Galaxy (or perhaps Wildfire) and shared memory< Re: Hobbyist VMS V7.2 TK50 DECnet-Plus installation failure?8 Hobbyist VMS V7.2 TK50 DECnet-Plus installation failure?* Re: Hoff's Email Address No Longer Valid ? Re: HSD05 DSSI Controller * Re: I can't run/detach in my own X session Re: In need for a console  Re: In need for a console  Re: In need for a console $ Re: Microsoft loses, we get punished$ Re: Microsoft loses, we get punished( Re: Microvax 3100-96 4000-106A psychosis multiple qmans in a cluster?  Re: multiple qmans in a cluster?  Re: multiple qmans in a cluster?# Network errors and CLUEXIT bugcheck  Re: Old vax * Re: old vs. new IRON (Was FS: Digital VAX)* Re: old vs. new IRON (Was FS: Digital VAX)7 Re: OVMS : Should I retrain back to the latest version?  Re: Re submit a batch job  Re: So who will buy VMS ? & Re: Sun considers VMS a "mainframe" OS& Re: Sun considers VMS a "mainframe" OS& Re: Sun considers VMS a "mainframe" OS0 Re: VMS today: % users x-widnows vs VT terminals0 Re: VMS today: % users x-widnows vs VT terminals Re: VRC16 Monitor Problem  Re: VRC16 Monitor Problem ( Re: Warning: circuitsurgeon is a ripoff!( Re: Warning: circuitsurgeon is a ripoff!6 RE: Why are there no French people in this newsgroup ?6 Re: Why are there no French people in this newsgroup ?6 Re: Why are there no French people in this newsgroup ?6 Re: Why are there no French people in this newsgroup ?% Re: Wierd Elsa Gloria Synergy Problem % Re: Wierd Elsa Gloria Synergy Problem   F ----------------------------------------------------------------------  % Date: Fri, 07 Apr 2000 11:15:24 -0700 5 From: "Larry D Bohan, Jr" <LBohan@dbc.spam_less..com> , Subject: Compaq Song and Dance (the lack of)2 Message-ID: <zyTuOHHxxQ3preRLAMOGAoUxTYaE@4ax.com>  . Mr Magee  hits the nail on the head (for once) to wit:   6  * Compaq does not seem to make such a song and dance 2    about chip announcements as Intel and the rest.      / http://www.theregister.co.uk/000407-000005.html   & Posted 07/04/2000 9:30am by Mike Magee    % Alpha chip powers Celera genome burst   ; Biotech firm Celera said yesterday that it has now finished 8 sequencing 99 per cent of the human genetic pattern, and? confirmed it will complete its corporate push during the course  of this year.   8 Sources close to the firm's plans said it has used "mass@ quantities" of an as yet unreleased 667MHz Alpha processor, each> of which includes 8Mb of cache*, incorporated in four way ES40 AlphaServers.   = Further, the next step of Celera's plan involves using faster A Compaq technology to put the sequences together to form the human : genome, thus giving it a massive lead over an alternative,6 government sponsored scheme. This is probably Wildfire technology.   @ The Human Genome Project wants Celera to share its data with it,@ and is backed by both Mr Tony Blair, the English prime minister,B and President Bill Clinton, who currently runs the United States.   = Informed sources said that Celera has chosen Alpha technology A because of its number crunching capabilities, and it is not alone ? amongst biotech firms in choosing the microprocessor over other  competing technologies.   A Sandia Labs has an installed base of nearly 3,000 Alpha CPUs, and 8 has recently added 1,300 AlphaServer DS10L Slates to its computing armory.   = Compaq is expected to formally announce its 667MHz-8Mb Alphas " during the course of next week. R   ? * Compaq does not seem to make such a song and dance about chip ? announcements as Intel and the rest. According to a contributor ? on our Bulletin Board, these have been available for some time. ? He adds that Alphas are particularly suitable for biotech apps. < But the chips that are available only have 4Mb of cache. Now= wouldn't it be great if you could still get Alpha for NT on a  666/8Mb cache system?    ------------------------------  " Date: Fri, 7 Apr 2000 19:56:54 GMT0 From: "Terry C. Shannon" <shannon@world.std.com>0 Subject: Re: Compaq Song and Dance (the lack of)& Message-ID: <FsnwBp.1GG@world.std.com>  L As the source of Magee's speeds-n-feeds info, methinks Compaq is waiting for. the >1GHz Alphae before it makes a big splash.   cheers,    terry s   @ "Larry D Bohan, Jr" <LBohan@dbc.spam_less..com> wrote in message, news:zyTuOHHxxQ3preRLAMOGAoUxTYaE@4ax.com...0 > Mr Magee  hits the nail on the head (for once)	 > to wit:  > 7 >  * Compaq does not seem to make such a song and dance 4 >    about chip announcements as Intel and the rest. >  >  > 1 > http://www.theregister.co.uk/000407-000005.html  > ( > Posted 07/04/2000 9:30am by Mike Magee >  > ' > Alpha chip powers Celera genome burst  > = > Biotech firm Celera said yesterday that it has now finished : > sequencing 99 per cent of the human genetic pattern, andA > confirmed it will complete its corporate push during the course  > of this year.  > : > Sources close to the firm's plans said it has used "massB > quantities" of an as yet unreleased 667MHz Alpha processor, each@ > of which includes 8Mb of cache*, incorporated in four way ES40 > AlphaServers.  > ? > Further, the next step of Celera's plan involves using faster C > Compaq technology to put the sequences together to form the human < > genome, thus giving it a massive lead over an alternative,8 > government sponsored scheme. This is probably Wildfire
 > technology.  > B > The Human Genome Project wants Celera to share its data with it,B > and is backed by both Mr Tony Blair, the English prime minister,C > and President Bill Clinton, who currently runs the United States.  > ? > Informed sources said that Celera has chosen Alpha technology C > because of its number crunching capabilities, and it is not alone A > amongst biotech firms in choosing the microprocessor over other  > competing technologies.  > C > Sandia Labs has an installed base of nearly 3,000 Alpha CPUs, and : > has recently added 1,300 AlphaServer DS10L Slates to its > computing armory.  > ? > Compaq is expected to formally announce its 667MHz-8Mb Alphas # > during the course of next week. R  > A > * Compaq does not seem to make such a song and dance about chip A > announcements as Intel and the rest. According to a contributor A > on our Bulletin Board, these have been available for some time. A > He adds that Alphas are particularly suitable for biotech apps. > > But the chips that are available only have 4Mb of cache. Now? > wouldn't it be great if you could still get Alpha for NT on a  > 666/8Mb cache system?  >  >    ------------------------------  % Date: Sat, 01 Apr 2000 07:48:20 -0400 % From: JM <vmswiz@geonospamcities.com> : Subject: Re: create your own personal SYS$COMMON directoryO Message-ID: <768797888A8B614F.9A7978CA78AF59E0.4B40157C8B000A55@lp.airnews.net>    > > > "John E. Malmberg" wrote:  > > > F > > > If it were _MY_ choice, I'd like to see the SYS$SYTEM path look  > > > like this: > > > + > > > "SYS$SYSTEM" = "SYS$SYSROOT:[SYSEXE]" & > > >         = "SYS$SYSROOT:[USREXE]" > >   B Here's a procedure I use in SYLOGICALS.COM to do just what you are@ asking except mine does ALL of the SYS$COMMON: and SYS$SPECIFIC:F directories. I use it to create a "cluster common" directory so that IF can maintain separate system disks for each system in the cluster, yetE share most of the same system command files. It also makes it easy to E separate my own applications from the generic VMS system directories. C USERCOMMON goes on one machine in the cluster and everybody "nests" $ SYS$SYSROOT to add it to their list.  ; It is good for isolating 3rd party products in SYS$SYSTEM:, C SYS$LIBRARY:, SYS$HELP:, SYS$MANAGER:, etc, even in a non-clustered  environment.  & $       DEFINE/SYS/EXEC SYS_SHAREDROOT' $1$DKA0:[USERCOMMON.]/TRANS=(CONC,TERM)  $ ( $       DEFINE/SYSTEM/EXEC SYS$SYSROOT -                 > 'F$TRNLNM("SYS$SPECIFIC","LNM$SYSTEM")'/TRANS=(CONCE,TERMI), -                 SYS$COMMON:, -                 SYS_SHAREDROOT:                                 *JM*   vmswiz@geonospamcities.com   ------------------------------  $ Date: Fri, 7 Apr 2000 14:43:52 -0400' From: "Bill Todd" <billtodd@foo.mv.com> . Subject: Re: Critical/important bug on VMS ...( Message-ID: <8cla69$156$1@pyrite.mv.net>  7 Bob Koehler <koehler@eisner.decus.org> wrote in message   news:2000Apr7.084820.1@eisner...L > In article <8cjb7j$s04$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:  > > ? > > JF Mezei <jfmezei.spamnot@vl.videotron.ca> wrote in message  > L > >> Even though I am well outside the earth, the cockpit still indicates anH > >> airspeed. Since airspeed is measured from an air pressure measuring pitot J > >> outside the plane, and since there is no air around me, shouldn't the > > airspeed& > >> indicator indicate a speed of 0 ? > > L > > One might assume that the mechanism adjusts for air density.  Perhaps it isI > > sufficiently sensitive to be able to compute 'air speed' based on the + > > occasional hydrogen atom you encounter.  > >  > G > Actually you'ld only be able to measure your velocity relative to the 
 > solar wind.   H I may be being exceptionally dense today, but that kind of sounds like a, pretty good analogue to 'air speed' to me...   - bill   ------------------------------  # Date: Fri, 07 Apr 2000 18:42:10 GMT 0 From: "Randy Knight" <REKnight@worldnet.att.net>) Subject: DECserver 90M booting problem... F Message-ID: <6_pH4.2361$8v5.140323@bgtnsc06-news.ops.worldnet.att.net>  L I have multiple 90M units, when any one of them are powered on it boots fromG the flashrom gives the message "Image load complete" to console. It the J proceeds to flash the top 4 port lights then the bottom 4 port lights, theE repeats the top/bottom cycle for 10 min, before stopping then working K normally. I have DHCP and BOOTP off, so I don't know what it is waiting on. H Has anyone else had this problem? how do I fix it? Otherwise the servers work fine after booting.   Randy Knight REKnight@worldnet.att.net    ------------------------------  % Date: Fri, 07 Apr 2000 21:44:03 -0500 + From: Shael Richmond <ksrich@bellsouth.net> - Subject: Re: DECserver 90M booting problem... - Message-ID: <38EE9CF3.DE4DEE64@bellsouth.net>    Randy Knight wrote:  > N > I have multiple 90M units, when any one of them are powered on it boots fromI > the flashrom gives the message "Image load complete" to console. It the L > proceeds to flash the top 4 port lights then the bottom 4 port lights, theG > repeats the top/bottom cycle for 10 min, before stopping then working M > normally. I have DHCP and BOOTP off, so I don't know what it is waiting on. J > Has anyone else had this problem? how do I fix it? Otherwise the servers > work fine after booting. >   C This is normal operation.  The flashing is the decompressing of the  image.B The fix is to load mneng3 instead of mneng2.  You can't save it to flash(unlessG you have the newer one w/more memory) but it can load over the network.    Shael    ------------------------------  % Date: Fri, 07 Apr 2000 22:31:17 -0400 * From: David A Froble <davef@tsoft-inc.com>- Subject: Re: DECserver 90M booting problem... - Message-ID: <38EE99F5.58C06541@tsoft-inc.com>   L I have noticed that the 90M takes a while to boot up.  Seem to remember thatL it's 3-5 minutes before it's working.  Don't remember 10 minutes, but then I' wasn't counting.  Might just be normal.    Dave     Randy Knight wrote:  > N > I have multiple 90M units, when any one of them are powered on it boots fromI > the flashrom gives the message "Image load complete" to console. It the L > proceeds to flash the top 4 port lights then the bottom 4 port lights, theG > repeats the top/bottom cycle for 10 min, before stopping then working M > normally. I have DHCP and BOOTP off, so I don't know what it is waiting on. J > Has anyone else had this problem? how do I fix it? Otherwise the servers > work fine after booting. >  > Randy Knight > REKnight@worldnet.att.net    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596; 170 Grimplin Road               E-Mail: davef@tsoft-inc.com  Vanderbilt, PA  15486    ------------------------------  $ Date: Fri, 7 Apr 2000 14:20:05 +02002 From: "Frits A.M. Storms" <frits@storms.tmfweb.nl> Subject: Freeware MIME* Message-ID: <8ckjuo$rij$1@cyan.nl.gxn.net>  H I am looking for a simple freeware tool to encode & decode MIME files on OpenVMS.@ Could someone point me to the appropriate download site please ?  
 much obliged,    Frits A.M. Storms    ------------------------------   Date: 7 Apr 2000 21:16:29 GMT 2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman) Subject: Re: Freeware MIME6 Message-ID: <8clj7d$g35$1@mailint03.im.hou.compaq.com>  _ In article <8ckjuo$rij$1@cyan.nl.gxn.net>, "Frits A.M. Storms" <frits@storms.tmfweb.nl> writes: I :I am looking for a simple freeware tool to encode & decode MIME files on 	 :OpenVMS.   A   This capability is part of OpenVMS V7.2 and later.  (An ECO kit A   that fixes some problems found in the base V7.2 MIME support is    also available.)  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------   Date: 7 Apr 2000 19:12:00 -0500 / From: jlauret@?.chem.sunysb.edu (Jerome LAURET)  Subject: Re: Freeware MIME. Message-ID: <38ee6b40_1@dilbert.ic.sunysb.edu>  _ In article <8ckjuo$rij$1@cyan.nl.gxn.net>, "Frits A.M. Storms" <frits@storms.tmfweb.nl> writes: J |>I am looking for a simple freeware tool to encode & decode MIME files on
 |>OpenVMS.B |>Could someone point me to the appropriate download site please ? |> |>much obliged,  |> |>Frits A.M. Storms  |> |>  9 	I don't know of such a universal tool but if you look at I http://nucwww.chem.sunysb.edu/htbin/software_list.cgi for a package named L mmencode, this tool will do the base64 encode/decode for you. You will still need to extract things by hand.   F 	Apart from that, a must is pine (simple, compiles fine etc ...). Does_ the attachement + encoding/decoding (check out http://www.agh.cc.kcl.ac.uk/files/vms/pine-vms/)       --  6                   Jerome LAURET S.U.N.Y. @ Stony Brook$        ,,,,,      Dept. of Chemistry+       ( o o )     Stony Brook NY 11794-3400 ;   ---m---U---m--------------------------------------------- &   E-mail: jlauret@mail.chem.sunysb.edu<   URL   : http://nucwww.chem.sunysb.edu/jlauret/jlauret.html   ------------------------------   Date: 7 Apr 2000 19:18:20 -0500 / From: jlauret@?.chem.sunysb.edu (Jerome LAURET)  Subject: Re: Freeware MIME. Message-ID: <38ee6cbc_1@dilbert.ic.sunysb.edu>  H 	Oups ! And I forgot to specify that indeed sys$system:mime.exe is there on OpenVMS7.2 and up ...   --  6                   Jerome LAURET S.U.N.Y. @ Stony Brook$        ,,,,,      Dept. of Chemistry+       ( o o )     Stony Brook NY 11794-3400 ;   ---m---U---m--------------------------------------------- &   E-mail: jlauret@mail.chem.sunysb.edu<   URL   : http://nucwww.chem.sunysb.edu/jlauret/jlauret.html   ------------------------------  $ Date: Fri, 7 Apr 2000 15:02:47 -0400' From: "Bill Todd" <billtodd@foo.mv.com> ; Subject: Re: Galaxy (or perhaps Wildfire) and shared memory ( Message-ID: <8clb9p$26s$1@pyrite.mv.net>  ; JF Mezei <jfmezei.spamnot@vl.videotron.ca> wrote in message ) news:38EE013D.9F08571A@vl.videotron.ca...  > Joshua Cope wrote:K > > You're right; a crash during a system service which is modifying shared  memoryE > > is interesting. I'm finding it hard to talk in the abstract here;g
 perhaps weG > > can construct a specific example of such a failure and use that fore discussion?t > K > If I know that an application has a block of data (either records/file orPI > memory) which is accessed concurrently by other processes, I would thenI takeJ > precautions to ensure that the data's integrety is assured (locking, and6 > probably verification of integrity before using it). >NG > If you are using multiple threads in a process, you also have to takeY steps to > maintain data integrity. >OG > So, unless the shared memory is there without the process explicitely K > requesting it isn't it fair to assume that the application is fully aware  thatE > a bunch of memory is shared not only between processes, but betweenS systemssH > and that as such, steps must be taken to synchronize access and verify
 > integrity ?e  J The point is that the steps necessary to verify integrity when memory thatF can survive failure of a system is shared between systems constitute aI superset of the steps necessary to verify integrity when memory is sharedaJ only between processes on a single system that fails atomically for all ofF them - i.e., there are new failure modes that your existing code won'tL handle unless augmented, and the complete required set of such augmentationsD may not be immediately obvious (though it certainly includes the newI possibility that system services operating on the shared memory that in atH single system have known completion semantics in terms of the state they= leave that memory in will no longer have those semantics withn inter-system-shared memory).  L The issue is similar for system use of such inter-system-shared memory, saveL for the fact that, unlike memory-sharing among processes, the system *never*J really did something very similar before (Memory Channel use excepted, but? we already know that's pretty bizarre), so the problem requiresjF significantly more analysis.  One can at least start by looking at theG issues involved in multi-initator disk access, or even just recovery of L single-system persistent disk data after a crash - or, if you take a greyingJ PDP-8 programmer to lunch, how core memory was treated on restart when one wished to continue processing.   - bill   >tJ > It's not as if all your global variables in your C program magically getK > allocated in a shared memory between all the systems in the galaxy, is ite ?s   ------------------------------  $ Date: Fri, 7 Apr 2000 15:33:01 -0400' From: "Bill Todd" <billtodd@foo.mv.com>d; Subject: Re: Galaxy (or perhaps Wildfire) and shared memory0( Message-ID: <8cld2e$3tk$1@pyrite.mv.net>  5 Joshua Cope <Joshua.Cope@Compaq.com> wrote in messageM$ news:38EDF8A8.D28B8FAE@Compaq.com...D > bt> ...the possibility that some system service (say, some kind of	 nominally G > bt> atomic conversion process not requiring any I/O) operating on theg sharedI > bt> memory may be interrupted by the instance crashing in the middle of  its I > bt> processing, whereas in a traditional SMP if that happened the wholec boxrC > bt> crashed and no one could ever see the partially-updated data.O > I > You're right; a crash during a system service which is modifying sharedo memoryK > is interesting. I'm finding it hard to talk in the abstract here; perhapsi weE > can construct a specific example of such a failure and use that for  discussion?d  G Unfortunately, I haven't used VMS for over 13 years, so anything beyondeL abstractions is difficult for me.  But (as I just replied to JF) new failureL modes would seem inevitable with just about any system service (other than aH library service, which presumably can fail at any point even on a singleF system) that purports to leave the memory upon which it operates in anJ enumerable set of final states (including error-return states) rather thanI just saying that the state of the memory within its scope of operation iscF undefined on errors - and even in that case, the application instancesD running on other system instances sharing the memory must be able toA differentiate between lock-release after operation completion and H lock-release after an instance failure (for all I know, the DLM providesC this information already in some manner - but it's necessary unlessrC lock-release notification can be guaranteed to occur only after any J instance-failure notification that the application may be able to registerH to receive) and for that matter the application instances likely must beF changed to use locks rather than (local system) mutexes to synchronizeB access to the shared memory (though that's a simple change, and ifK application-controlled mutexes are created in the shared memory itself they L may be usable for such synchronization in a manner less expensive than locksI as long as the applications receive instance-failure notification and canj' clean up the mutex state if necessary).i  J Sorry to have rambled above, but since as I said I don't know exactly whatH services VMS provides there are quite a few possibilities to cover.  AndL while such additional failure modes are things application developers should9 be aware of, I'm more concerned about system-level use ofoE inter-system-shared memory, since system code hasn't had to deal with I *anything* at all similar before (save for things like on-disk persistento	 storage)..   >nG > I can see how such a failure would be a problem unique to partitionedn systems:F > bad data is now sitting in shared memory, and is still accessible to runninglD > processes on the other instances. I'm still not sure how CPU cache	 coherencyf! > comes into the picture, though.i  G It doesn't, as I've kept repeating:  people have just kept insisting onnK dragging my original question off into this other interesting but unrelated  topic.   - bill   >c> > ------------------------------------------------------------8 > The above opinions and information are not necessarily' > those of Compaq Computer Corporation.D> > ------------------------------------------------------------   ------------------------------  % Date: Fri, 07 Apr 2000 14:55:45 -0500>% From: Chris Scheers <asi@airmail.net> ; Subject: Re: Galaxy (or perhaps Wildfire) and shared memory O Message-ID: <21685555D1262E29.AA372B211D0D206C.88434F3685342C02@lp.airnews.net>r   Bill Todd wrote: > M > Apologies for replying to myself (again), but I had an interesting off-linehL > conversation that may help shed light on the kind of problem I'm trying to2 > describe (name omitted to protect the innocent): > L > Actually, my real worry was that the hardware interconnections required toN > join caches across partitions constituted a potential cross-partition singleN > failure point in and of themselves (of a kind that your option '2' would notJ > necessarily protect against, though hardware isolation on these paths inM > excess of that normally found on SMP systems could) - a failure point which-G > conventionally-separated cluster nodes do not have since they have noR) > inter-node shared memory to begin with./ > N > That may be kind of a nit (or a failure mode of sufficiently low probabilityJ > that no one is really going to care much), but if true it does mean thatL > fault-isolation truly equivalent to that of conventional clustering cannotL > be achieved if cachable shared memory is used between partitions - whereasL > uncached shared memory could maintain the same level of fault-isolation in > this area. > N > Good point about rebooting needing to flush rather than destroy cached data,' > and it brought to mind another issue:l > N > If write-back caching is used with such shared memory, modified data not yetK > requested by anyone else could be sitting in a processor cache long after0J > the interlocks protecting shared access to it had been released - and ifK > that partition then died in any manner such that the dirty cache contentsMN > were lost, the shared memory would be left in an undetectibly-corrupt state.J > On the other hand, if write-back caching is not used when modifying suchJ > shared memory, then I'm not sure that preserving cache contents across aN > reboot is necessary:  whatever interlocks are guarding the modifications canL > remain in place until whatever recovery mechanisms executing subsequent toL > the failure clean up the shared state (though I'm not quite sure that suchN > sequencing fits well with the existing VMS distributed lock manager recoveryJ > model - something I've observed w.r.t. distributed file-system issues asN > well), and they won't really care just how much of that state did or did notF > get modified before the failure.  On the *third* hand, however, thisK > highlights the necessity if write-back caching *is* used of ensuring thateL > the rebooting is somehow synchronized with the recovery (being executed byH > someone else) such that dirty but now stale data won't be written back& > *after* the recovery has executed... > L > All this may mean that one simply does not write-back-cache data in memoryM > shared across partitions, though if you ensure that a reboot never destroysdI > cache contents and believe that any hardware problem that would destroyiJ > cache contents or cache connectivity is low enough to ignore I guess youM > could write-back cache.  On the other hand (not sure which one we've got to M > by now, but I feel like an octopus), just prohibiting write-back caching tosG > such memory might be enough that no special additional steps would be F > necessary to protect surviving sharing members against virtually anyI > hardware failure - and preserving the ability to read-cache such shared)L > memory likely is far more important to its performance than the ability to > write-back-cache it.    = When I looked into this once, I found that there are hardwared? multiprocessor write back caching schemes which will ensure the 1 integrity of both local caches and global memory.e  G Disclaimer: I'm not a hardware guru, so I may have some of these points  wrong.  6 For this to work correctly, several things must occur.  F 1) The cache should maintain a dirty bit, i.e., whether or not a value' in the cache has been modified locally.h  G 2) The cache must snoop the memory bus to see what other processors aree doing.  G 3) If some other processor writes to a location in the local cache, theiH current value in the cache must either be invalidated or replaced by the written value.  E 4) If some other processor reads from a location that is dirty in the G local cache, the read must be pended while the location is written from.D the local cache to global memory, then the read can proceed with the updated value.  F I believe that a major part of the SMP support hardware is involved inB satisfying these requirements.  This mechanism is SMP specific andF Galaxy independent.  It does not care how a Galaxy operates.  The view@ of global memory should remain correct indpendently from what is currently cached.h  H Items (3) and (4) are interesting when considered together and the localD cache has a dirty copy.  This is the case when two processors try toF write the same location without an intervening read.  Which value endsG up in global memory depends on the particular cache implementation.  Ita> may not be deterministic which value ends up in global memory.  G It appears that the Alpha LDx_L/STx_C instructions are designed to dealSG with this problem.  It would appear that a STx_C instruction would need C to force the cache to write through mode (instead of write back) to  maintain coherency.t  ; I hope that this helps (and is reasonably accurate <grin>).m  G -----------------------------------------------------------------------s$ Chris Scheers, Applied Synergy, Inc.  G 817-237-3360 (Voice)    817-237-3074 (Fax)    Internet: asi@airmail.netn   ------------------------------  $ Date: Fri, 7 Apr 2000 17:14:16 -0400' From: "Bill Todd" <billtodd@foo.mv.com>l; Subject: Re: Galaxy (or perhaps Wildfire) and shared memory ( Message-ID: <8clj0c$9o1$1@pyrite.mv.net>  = Hoff Hoffman <hoffman@xdelta.zko.dec.nospam> wrote in messagea0 news:8ckv5r$8d3$1@mailint03.im.hou.compaq.com... > L > In article <8cjdd4$18m$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:  >   >   re: Galaxy and Shared Memory >cI >   Alpha provides specific instruction sequence requirements to maintaineI >   and to update the CPU caches appropriately -- the memory barriers...>lF >   Appropriate use of these barriers is required on any OpenVMS Alpha system.n >aL >   OpenVMS Galaxy partitions and instances (or even hard partitions) do notI >   specifically enter into the discssion of processor cache coherency --o thetG >   box has to work right as an SMP box.  Of interest here, however, istJ >   coherency of cluster resources, and the same coherency mechanisms usedK >   elsewhere within an OpenVMS Cluster are used across any cluster membersa' >   located within an OpenVMS Galaxy...B >GF >   The OpenVMS Ask The Wizard area has a general discussion of memoryI >   barriers and hardware memory synchronization in topic 2681, with more G >   details in the OpenVMS Programming Concepts Manual and in the Alphar$ >   Architecture Reference Manual...  B Thanks for the pointer(s):  6 or so years ago I actually thought IL understood memory barrier use, but I've never had reason to use an Alpha andL I'd forgotten whatever understanding I may have had - or even that MBs might$ be relevant to the question at hand.  K Ah, yes.  Use w.r.t. instruction reordering - likely unrelated.  Use w.r.t.eK flushing write-back-cached data - quite possibly relevant, depending on thenK details of the memory architecture.  The critical fact that I had forgotteneJ was that Alpha appears to eliminate a good deal of SMP hardware complexityJ by pushing responsibility for explicit management of shared memory updatesJ back onto the programer - by no means necessarily a bad trade-off, but oneI which I think is less common in other architectures (though may be fairlyeI standard for RISC architectures - the little hardware knowledge I have isg( for the most part horribly out of date).  I In particular, if an MB write-flush causes data to be written all the wayeL back to main memory, thus completely eliminating inter-partition connectionsG at any cache level (of the kind that exist in some SMP architectures tocE allow inter-processor data-sharing without the latency of main-memory J accesses), then my concern that inter-cache connections between processorsE executing in different partitions might compromise fault-isolation is K groundless:  the only inter-partition connection will be through the sharedsC memory itself, which will presumably pass along any necessary cache J invalidations as write activity occurs (though my understanding of exactlyF how this occurs is regrettably fuzzy), and the hardware connecting theG caches to main memory is presumably already sufficiently robust to keep B local faults from affecting access by others to any shared memory.  K Furthermore, my fear that a cache might hold dirty data for lengthy periodsAD before writing it back to shared memory, thus drastically increasingK exposure to any failure which could keep that write-back from occurring, is L groundless if in fact MB flushes will *always* be explicitly issued on *all*J processors in a partition that have participated in any interlocked updateI activity to an inter-partition-shared piece of memory prior to release of H whatever synchronization mechanism is used - which it sounds as if is noA stronger a requirement than exists today for memory-sharing in an8H unpartitioned Alpha SMP system.  It does, however, sound as if processorK cache content needs to be flushed or invalidated on instance failure (which I may already be standard procedure) to ensure that stale updates won't get=+ applied to already-recovered shared memory.S  J So considerate of Fred to have asked such a comprehensive question, and ofI course of the Wizard to have provided such an informative answer.  Again,EG thanks for the pointer, and apologies for raising what I did originallyhG state might be a dumb question (though it did raise some interesting ify& unrelated explorations along the way).   - bill   >oK >   In addition to the discussion of cache coherency, OpenVMS Galaxy bringst7 >   in (at least) two additional design considerations:  >pE >     o the potential that shared memory will survive a system crash. C >       This isn't specifically a CPU cache issue, this is an issuee2 >       around recovering from interrupted access. > E >       Options include use of the OpenVMS Galaxy event services, the G >       lock manager, as well as timers and aging and keepalive schemesbE >       in the shared memory, as well as "careful" updates -- updates E >       performed in a sequence that closes or reduces the likelyhood G >       of problems from crashes.  (RMS uses "careful" updates, and foru >       just this reason...) > G >     o Differing memory access speeds.  This is obviously not an issue H >       with currently available systems, but will be a consideration onF >       certain AlphaServer GS series platforms -- the current OpenVMSD >       spinlock design effectively has an assumption of access timeD >       uniformity, and has been modified to provide better fairnessH >       to the protected resource when non-uniform memory configurations >       are in use.p >e >s, >  --------------------------- pure personal# opinion ---------------------------i1 >    Hoff (Stephen) Hoffman   OpenVMS Engineeringl hoffman#xdelta.zko.dec.com >    ------------------------------  $ Date: Fri, 7 Apr 2000 17:29:48 -0400' From: "Bill Todd" <billtodd@foo.mv.com>t; Subject: Re: Galaxy (or perhaps Wildfire) and shared memory ( Message-ID: <8cljtd$c28$1@pyrite.mv.net>  0 Chris Scheers <asi@airmail.net> wrote in messageI news:21685555D1262E29.AA372B211D0D206C.88434F3685342C02@lp.airnews.net....   ...   ? > When I looked into this once, I found that there are hardware A > multiprocessor write back caching schemes which will ensure the 3 > integrity of both local caches and global memory.o >(I > Disclaimer: I'm not a hardware guru, so I may have some of these pointso > wrong. >r8 > For this to work correctly, several things must occur. >tH > 1) The cache should maintain a dirty bit, i.e., whether or not a value) > in the cache has been modified locally.t >sI > 2) The cache must snoop the memory bus to see what other processors are, > doing. >oI > 3) If some other processor writes to a location in the local cache, thelJ > current value in the cache must either be invalidated or replaced by the > written value. >nG > 4) If some other processor reads from a location that is dirty in the-I > local cache, the read must be pended while the location is written fromBF > the local cache to global memory, then the read can proceed with the > updated value. >0H > I believe that a major part of the SMP support hardware is involved inD > satisfying these requirements.  This mechanism is SMP specific andH > Galaxy independent.  It does not care how a Galaxy operates.  The viewB > of global memory should remain correct indpendently from what is > currently cached.t  F The difference between normal SMP operation and Galaxy (or perhaps SunK 'domain' or S/390 'hardware-partitioned') operation is that when a hardwarec; fault occurs in a 'normal' SMP, the entire system stops and-2 cache/main-memory coherence ceases to be an issue.  K In a Galaxy, if a fault disables a cache that is holding dirty data that nonJ other processor has yet expressed an interest in, then that data is lost -A but subsequent use of the related but unupdated memory can occur.n  J Hoff just reminded me that the required explicit use of memory barriers inJ Alpha environments may make this a non-issue (at least in any sense beyondI the issue that already must be addressed explicitly in 'normal' Alpha SMPiK configurations - where processor and cache hardware failures bring down themJ entire system - to keep shared memory consistent).  But there are some SMPH environments that do not require such explicit handling of shared-memoryF consistency - which may well make those environments considerably moreD difficult to adapt to inter-partition memory sharing with good faultF isolation (though I suspect that simply disabling caching or even justJ write-back caching of inter-partition shared memory may be sufficiently to address such difficulties).o   - bill   >hJ > Items (3) and (4) are interesting when considered together and the localF > cache has a dirty copy.  This is the case when two processors try toH > write the same location without an intervening read.  Which value endsI > up in global memory depends on the particular cache implementation.  Itm@ > may not be deterministic which value ends up in global memory. >nI > It appears that the Alpha LDx_L/STx_C instructions are designed to dealfI > with this problem.  It would appear that a STx_C instruction would needeE > to force the cache to write through mode (instead of write back) toh > maintain coherency.e >p= > I hope that this helps (and is reasonably accurate <grin>).  >tI > ----------------------------------------------------------------------- & > Chris Scheers, Applied Synergy, Inc. >VI > 817-237-3360 (Voice)    817-237-3074 (Fax)    Internet: asi@airmail.neta   ------------------------------   Date: 7 Apr 2000 21:33:29 GMTe2 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman); Subject: Re: Galaxy (or perhaps Wildfire) and shared memoryd6 Message-ID: <8clk79$g35$3@mailint03.im.hou.compaq.com>  R In article <8clj0c$9o1$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:* :...The critical fact that I had forgottenK :was that Alpha appears to eliminate a good deal of SMP hardware complexityoK :by pushing responsibility for explicit management of shared memory updatesc :back onto the programer...t  H   VAX SMP has similar underlying constraints -- you have to use the REI F   to drain or invalidate the caches as appropriate, and shared memory G   operations require the use of the interlocked instructions to ensure -I   that the data is visible across all processors in an SMP configuration.i  J :In particular, if an MB write-flush causes data to be written all the wayM :back to main memory, thus completely eliminating inter-partition connections-H :at any cache level (of the kind that exist in some SMP architectures toF :allow inter-processor data-sharing without the latency of main-memoryK :accesses), then my concern that inter-cache connections between processorssF :executing in different partitions might compromise fault-isolation is :groundless...  E   And that is effectively what the memory barrier does -- it providesnC   a barrier to further execution, pending the completion of memory n%   operations currently "in flight"...   L :Furthermore, my fear that a cache might hold dirty data for lengthy periodsE :before writing it back to shared memory, thus drastically increasingaL :exposure to any failure which could keep that write-back from occurring, isM :groundless if in fact MB flushes will *always* be explicitly issued on *all*aK :processors in a partition that have participated in any interlocked updatepJ :activity to an inter-partition-shared piece of memory prior to release of. :whatever synchronization mechanism is used...  G   That is part of the basic synchronization involved in interlocking...rG   This synchronization is a descendent of the interlocking capabilitiesoH   used on VAX -- a differing implementation, but the same basic purpose     and same basic capabilities...  E   BTW, the URL pointer (currently) referenced in Ask The Wizard topich*   2681 has effectively been superceded by:  3     ftp://ftp.compaq.com/pub/products/alphaCPUdocs/   G   Also please see the OpenVMS FAQ for some other pointers to low-level  E   Alpha microprocessor and architectural documentation.  (Updates to aC   the ATW reply to topic 2681 and to the OpenVMS FAQ are underway.)r  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  " Date: Fri, 7 Apr 2000 22:47:33 GMT2 From: kilgallen@eisner.decus.org (Larry Kilgallen); Subject: Re: Galaxy (or perhaps Wildfire) and shared memory,& Message-ID: <2000Apr7.184733.1@eisner>  R In article <8cjdd4$18m$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:  L > Actually, my real worry was that the hardware interconnections required toN > join caches across partitions constituted a potential cross-partition singleN > failure point in and of themselves (of a kind that your option '2' would notJ > necessarily protect against, though hardware isolation on these paths inM > excess of that normally found on SMP systems could) - a failure point whichrG > conventionally-separated cluster nodes do not have since they have nor) > inter-node shared memory to begin with.-  > Well, of course, sharing a memory bus (or switch) introduces a? degree of commonality, and an arrangement with shared resourcesl@ is not so immune to propagation of failures as is an arrangement with no shared resources.u   I thought of that as obvious.h   ------------------------------  $ Date: Fri, 7 Apr 2000 20:31:39 -0400' From: "Bill Todd" <billtodd@foo.mv.com>o; Subject: Re: Galaxy (or perhaps Wildfire) and shared memoryd( Message-ID: <8cluid$or8$1@pyrite.mv.net>  = Larry Kilgallen <kilgallen@eisner.decus.org> wrote in messageu  news:2000Apr7.184733.1@eisner...L > In article <8cjdd4$18m$1@pyrite.mv.net>, "Bill Todd" <billtodd@foo.mv.com> writes:j > K > > Actually, my real worry was that the hardware interconnections requiredc toI > > join caches across partitions constituted a potential cross-partitions singleL > > failure point in and of themselves (of a kind that your option '2' would notsL > > necessarily protect against, though hardware isolation on these paths inI > > excess of that normally found on SMP systems could) - a failure pointh whicheI > > conventionally-separated cluster nodes do not have since they have no + > > inter-node shared memory to begin with.a >e@ > Well, of course, sharing a memory bus (or switch) introduces aA > degree of commonality, and an arrangement with shared resourcesnB > is not so immune to propagation of failures as is an arrangement > with no shared resources.c >e > I thought of that as obvious.)  F Of course it's obvious.  What was not obvious to me was that isolationC mechanisms had been added over and beyond those present in a normalcJ (unpartitionable) SMP that would allow the claim that a Wildfire or MarvelF box offered fault isolation comparable to that of a cluster comprisingL independent nodes to be true even when memory was shared between partitions.  L I did assume that the shared memory itself was reasonably protected from theC kind of undetectible corruption that I described as possible when aPK partition suffered arbitrary hardware failure - making the shared memory asiK recoverable after a failure as distributed shared resources coordinated viaeG message-passing mechanisms are - though of course the memory itself canmL fail, just as communication paths can fail, and if you don't want the memoryK itself to form a single point of failure (though the cooperating partitionsoI could use it in a way that was robust across such failure), you mirror ith? just as you duplicate communications paths for the same reason.h  D But inter-partition cache communication seemed to open an additionalG indirect avenue to such corruption that it was not clear to me had beendL covered, unless such communication paths had been made more failure-tolerantH than is necessary in a normal SMP.  The substance of other responses wasF that cache coherence works exactly the same way it works in a 'normal'I [Alpha] SMP - which is indeed the case, but did not reassure me because I-F did not remember that Alpha handles cache coherence by not handling itL (except as a by-product of actual accesses to the related memory itself) andL leaving it up to the programer to:  my error, but it took Hoff to figure out( what I was talking about and explain it.   - bill   ------------------------------  * Date: Fri, 7 Apr 2000 15:42:24 -0400 (EDT)2 From: "Douglas S. Meade" <vaxboy@inforum2.umd.edu>E Subject: Re: Hobbyist VMS V7.2 TK50 DECnet-Plus installation failure?aF Message-ID: <Pine.LNX.4.20.0004071540520.4242-100000@inforum2.umd.edu>   Steven,   C Actually, I would be interested in getting this procedure debugged tB too.  I cut some TK50 tapes of my own for an Old MicroVax II, but > I could only get the base installation done.  I ended up doingA some of the other save sets and .pcsi files one by one, which wasn very slow.    
 Doug Meade= P.S. - Actually, I think you were the one who pointed out to c) me how to make the tapes.  Thanks again! s  * On Thu, 6 Apr 2000 sms@antinode.org wrote:  J >    I've recently been playing with putting the Hobbyist VMS V7.2 CD-ROM H > installation kit onto TK50 tape.  After extracting the data concerningI > which save sets go onto which tapes, I thought that all was well.  ThenhC > I (foolishly, it seems) tried an installation which failed at the  > DECnet-Plus stage: >  > [...]  > : > * Enter name of drive holding the DECnet-Plus kit: mua0:* > %SYSTEM-F-IVLOGNAM, invalid logical name > R > ******************************************************************************** > 2 >     Installation of OpenVMS VAX V7.2 incomplete. > < >     Error - exiting from Mandatory Update phase procedure.1 >     Restart this phase by rebooting the system.I >  > [...]  > H >    When I attempted to "[r]estart this phase by rebooting the system", > the error recurred.o > H >    The cause turned out to be a typo in [VMS$COMMON.SYSEXE]STARTUP.MANG > (in VMS072.B), proving conclusively that "net_def" is not the same asuA > "net_dev".  It affects only installations from tape, naturally.s > D >     Has anyone else run into this one, or am I the only one looney > enough to try this?g > J > ------------------------------------------------------------------------ > E >    Steven M. Schweda               (+1) 651-699-9818  (voice, home)uE >    382 South Warwick Street        (+1) 763-781-0308  (voice, work)eI >    Saint Paul  MN  55105-2547      (+1) 763-781-0309  (facsimile, work) ; >    sms@antinode.org                sms@provis.com  (work)  >    ------------------------------  $ Date: Sat, 8 Apr 2000 00:31:04 -04002 From: "Richard B. Gilbert" <DRAGON@compuserve.com>A Subject: Hobbyist VMS V7.2 TK50 DECnet-Plus installation failure?i7 Message-ID: <200004080031_MC2-A05C-3C22@compuserve.com>y  F         I experienced exactly this behavior a couple of months ago and reported it here.i  J         I didn't have software support for VAX/VMS so I reported it here.=  =  H I never went so far as to pick apart the kit and locate the cause of theE error.   I was simply taking advantage of the availability of two VAXhB 6000/410 to install and configure a CI cluster for the first time!  J         My workaround was to copy the savesets to disk and install from t= he copy.  That did work.   1 Message text written by INTERNET:sms@antinode.org J >   I've recently been playing with putting the Hobbyist VMS V7.2 CD-ROM =  F installation kit onto TK50 tape.  After extracting the data concerningG which save sets go onto which tapes, I thought that all was well.  ThenfA I (foolishly, it seems) tried an installation which failed at theb DECnet-Plus stage:   [...]   8 * Enter name of drive holding the DECnet-Plus kit: mua0:( %SYSTEM-F-IVLOGNAM, invalid logical name  J *************************************************************************= ** *****s  0     Installation of OpenVMS VAX V7.2 incomplete.  :     Error - exiting from Mandatory Update phase procedure./     Restart this phase by rebooting the system.f   [...]r  F    When I attempted to "[r]estart this phase by rebooting the system", the error recurred.s  F    The cause turned out to be a typo in [VMS$COMMON.SYSEXE]STARTUP.MANE (in VMS072.B), proving conclusively that "net_def" is not the same ass? "net_dev".  It affects only installations from tape, naturally.e  B     Has anyone else run into this one, or am I the only one looney enough to try this?a <n   ------------------------------  # Date: Sat, 08 Apr 2000 02:59:51 GMTa From: gordon@waylay.zko.dec.come3 Subject: Re: Hoff's Email Address No Longer Valid ?n+ Message-ID: <38eea07d.12074382@news.mv.net>y  4 On Tue, 4 Apr 2000 16:27:50 -0400, "Fred Kleinsorge"$ <kleinsorge@star.zko.dec.com> wrote:  
 >Hi Jerry. >h9 >Everyone in Compaq can be found by First.Last@compaq.coms >nK >the routers that handled the old internal names are gone, but you can sendn; >to the IP address if you know it - probably something likef >hoffman@xdelta.zko.dec.comv >s >_Fred  > While that's true for all the Fred Kleinsorges at Compaq, the 7 two Doug Gordons, at a minimum, have generated a lot of"4 confusion.  He was Doug.Gordon@compaq.com first.  HeC gets a lot of my external mail and I get a lot of his internal mailh% (because I'm listed first in the GAL)o  	 			--Dougu  			 (not Doug.Gordon@compaq.com)   ------------------------------  # Date: Fri, 31 Mar 2000 21:07:30 GMTi* From: cajunmikemike@netscape.net (Mike M.)" Subject: Re: HSD05 DSSI Controller1 Message-ID: <38e512c2.88455051@enews.newsguy.com>V  4 On Fri, 31 Mar 2000 07:28:28 -0800, "Wolf, Gerald J"# <Gerald.Wolf@F22.Boeing.com> wrote:   F Are you saying that you are *currently* using a 4 and 9 gigabyte drive  in a BA350 driven by an HSD05?    ? Because I was under the possibly mistaken impression that HSD05-; controllers would not support anything larger than 2.1gb.     M >There is a VAX4000-500a here with a BA350 shelf and HSD05 DSSI Controller in  >it. >-K >The disk shelf currently has (1) 2GB, (1) 4GB and (2) 9GB narrow drives in  >it. >lL >The 2GB narrow drive keeps filling up with important development data, so I >wants- >to replace the 2GB narrow with a 9GB narrow.c >rJ >Do I have to configure the slot the 2GB disk is in to accept a larger 9GB >disk?; >Will this shelf accept 9GB , 18GB and 32GB VW type drives?t   ------------------------------  % Date: Fri, 07 Apr 2000 21:20:53 +0200h+ From: Arne Bergseth <Arne.Bergseth@dnv.com>t3 Subject: Re: I can't run/detach in my own X sessionc' Message-ID: <38EE3515.7666E7EB@dnv.com>    Ildefonso Junquero wrote:    > Hello! >oI >     I'm having an extrange problem... I can't run/detach an x client ono > my own xserver.y >tI >     I've a new CDE and a configured security manager with node * user *  > and transport *m > J >     If I run : run decw$utils:xmag (for example), all works ok, but if i > run :h >p- >         run/detach  decw$utils:xmag/out=a.ae >h> >     the client says in a.a : xmag: unable to open display "" >a  >     can anyone help me please? >)
 >     Thanks!p >e >     Ilde.o >  > --H > ______________________________________________________________________ >c  A I have not tried this out on my  own Alpha with CDE, but in otherr
 circumstances H the "unable to open display" message is related to the (virtual) display
 device WSAxx:   not beeing established or wrong.L Since the /detach option will create an independent process some symbols and logicalsK may not be available. Try  to examine the $SHOW DISPLAY command and look uptG Help for the $SET DISPLAY /CREATE ... , then the question is what thesen commandsC would do if they were issued in the newly created detached process.a  
 Arne Bergsetht   ------------------------------  % Date: Fri, 31 Mar 2000 12:22:50 -0800s< From: Chris Doran <chrisj.doranNOchSPAM@physics.org.invalid>" Subject: Re: In need for a console9 Message-ID: <147000d0.cea24c86@usw-ex0104-031.remarq.com>c  = In article <38E4E794.5774AED3@calstatela.edu>, miguel ramirez   <mramir18@calstatela.edu> wrote:? >    I'm running a  microvax 3100 model 20 computer and I woulde like to ( >know what type of console I should get.  9 Just about any RS-232 terminal will work (with some head- = scratching over cable connections and line setup) if you just'< want to see line-by-line console messages or are prepared to= write SMG definition files for cursor control. If you want tor; use VMS screen editors, you really need a DECpaq VTxxx (anyt model -- even a VT52!).x  = The easiest solution is probably an emulator running on a PC.k> CRT from http://www.vandyke.com/ (shareware, time-limited demo' free) works well, but there are others.u   Christ    L * Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *G The fastest and easiest way to search and participate in Usenet - Free!d   ------------------------------  # Date: Fri, 31 Mar 2000 19:20:30 GMTc. From: Michael Austin <maustin@nc.prestige.net>" Subject: Re: In need for a console/ Message-ID: <38E4F9CA.D8312146@nc.prestige.net>i  I If you have a PC laying around you can use something like reflections offn9 of the serial port.  VT320/420/5xx all work equally well.t   Michael Austin DBA Consultant   miguel ramirez wrote:a  H >     I'm running a  microvax 3100 model 20 computer and I would like toI > know what type of console I should get.  I've been sugested to look forfI > a VT-320.  If so, where could I find one.  I am also looking for a 4206n > or 4207 tektronix terminal. B >     If anyone has any information, i would really appreciate it. >-G >                                                                 thank, > you,H >                                                                 miguel   ------------------------------  % Date: Fri, 31 Mar 2000 17:32:25 -05002* From: David A Froble <davef@tsoft-inc.com>" Subject: Re: In need for a console- Message-ID: <38E52779.8D6F20FC@tsoft-inc.com>    miguel ramirez wrote:@ > H >     I'm running a  microvax 3100 model 20 computer and I would like toI > know what type of console I should get.  I've been sugested to look foroI > a VT-320.  If so, where could I find one.  I am also looking for a 4206l > or 4207 tektronix terminal.iB >     If anyone has any information, i would really appreciate it. > G >                                                                 thank  > you,H >                                                                 miguel  8 Normally, the newer the terminal, the better.  I'd want:   VT-525 with color monitord VT-520 or VT-510 Vt-420  = Only if I couldn't get the above, then a VT-320, VT-220, etc.w  P The VT-420 and above have a cut and paste feature, and can support more lines inP memory than is on the screen, thus allowing you to scrool back to see lines that% have moved off the top of the screen.a  H A decent PC with a terminal emulator works, but there are more points ofC failure.  For just a console, the VT terminals are my first choice.s   Dave   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596; 170 Grimplin Road               E-Mail: davef@tsoft-inc.comr Vanderbilt, PA  15486    ------------------------------  $ Date: Fri, 7 Apr 2000 14:40:59 -0400' From: "Bill Todd" <billtodd@foo.mv.com>t- Subject: Re: Microsoft loses, we get punishedb( Message-ID: <8cla0s$14u$1@pyrite.mv.net>  K Without in the least making excuses for Microsoft's execrable behavior overiJ the years, if Compaq can tout its 'one-touch Internet access' as the primeJ example of its own innovation Microsoft's claims to being innovative place% at best a distant second in chutzpah.t  K Of course, the beauty contest between Gates and an iPAQ is hardly a contesthI at all.  And one could say the same in the area of personality.  Too bad, I because Gates clearly does have qualities that could had they been put togI better use have been a positive force in both the industry and the world.a   - bill  = David Mathog <mathog@seqaxp.bio.caltech.edu> wrote in messageu& news:8cl09j$4e6@gap.cco.caltech.edu...K > Having just lost in the legal arena MS has now decided to take their casepA > to the court of public opinion.  Last night NBC ran the same MS  commercial,iK > featuring Bill Gates' lecturing us on innovation (sic), at least 3 times. D > Apparently they're running it hourly in prime time.  Somebody onceI > suggested to me that the proper punishment for a monopolist was summary-L > execution.   At the time I thought that punishment was a bit extreme.  ButI > after suffering through 3 exposures in one night to Mr. Gates' smirking2> > visage I'm just about ready to sign up for the firing squad. >)
 > Regards, >G > David Mathog > mathog@seqaxp.bio.caltech.edue@ > Manager, sequence analysis facility, biology division, Caltech   ------------------------------  # Date: Fri, 07 Apr 2000 21:56:51 GMTy From: Dirk Munk <munk@home.nl>- Subject: Re: Microsoft loses, we get punishedc' Message-ID: <38EE59A8.BFC8FC96@home.nl>r  & "Brian Schenkenberger, VAXman-" wrote: > c > In article <8cl09j$4e6@gap.cco.caltech.edu>, mathog@seqaxp.bio.caltech.edu (David Mathog) writes:eL > >Having just lost in the legal arena MS has now decided to take their caseN > >to the court of public opinion.  Last night NBC ran the same MS commercial,L > >featuring Bill Gates' lecturing us on innovation (sic), at least 3 times.E > >Apparently they're running it hourly in prime time.  Somebody oncemJ > >suggested to me that the proper punishment for a monopolist was summaryM > >execution.   At the time I thought that punishment was a bit extreme.  ButhJ > >after suffering through 3 exposures in one night to Mr. Gates' smirking? > >visage I'm just about ready to sign up for the firing squad.u > H > Could this be the same sound bite I heard this AM on CNN -- Bill GatesH > smirking as he speaks of "freedom"?  Funny how he can speak to freedomH > when court's evidence shows an extreme lack thereof in the restrictive% > covenant he made with PC suppliers.d  F What do you mean ? After all we are all free to buy Micro$oft products aren't we ??   >  > --P > VAXman- OpenVMS APE certification number: AAA-0001           VAXman@TMESIS.COM   ------------------------------  % Date: Fri, 31 Mar 2000 17:36:57 -0500i* From: David A Froble <davef@tsoft-inc.com>1 Subject: Re: Microvax 3100-96 4000-106A psychosisn- Message-ID: <38E52889.4DE5C83A@tsoft-inc.com>e   "Jay T. McCanta" wrote:h > G > Recently we purchased a used MicroVax 3100-96.  That's what the orderaG > says, that's what the model number on the back of the unit describes,sF > but the HW_NAME says it's a VAX 4000-106A.  So what gives?  I talkedB > to a Compaq tech about it and he says there is a listing for theB > 3100-96 in the VMS hardware tables so he didn't know why the box > didn't identify correctly. >  > I am running OpenVMS 7.1 > HW_NAME  VAX 4000-106A > HW_MODEL 487 > ) > Has anyone seen this?  Is it a problem?a > 
 > Jay Mccantas	 > ImmunexrG > -===================================================================- ; > Jay McCanta              |  My opinions are barely my ownl= > System Administrator     |  My employer doesn't necessarily ) > Immunex Corp.            |  share them.uG > -===================================================================-o  L Not a problem at all.  They are essentially the same processor and I believeN motherboard.  The major differences are packaging and the availability of DSSIC in the VAX 4000 model 100 systems.  Not sure how the DSSI attaches.a  O What you are seeing is the same for all MicroVAX 3100 model 96, model 98, and In/ think model 95 systems I've been involved with.    Dave   -- t4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596; 170 Grimplin Road               E-Mail: davef@tsoft-inc.como Vanderbilt, PA  15486t   ------------------------------  $ Date: Fri, 7 Apr 2000 11:53:49 -0700* From: "Jack Peacock" <peacock@simconv.com>% Subject: multiple qmans in a cluster?b? Message-ID: <J5qH4.17540$k7.2821544@news-west.usenetserver.com>   F Is it possible to have a separate queue manager run on each node in anG alpha vms 7.2 cluster?  I want drives and other resources to be shared,sH but each node has specific batch and print queues which are never sharedH with other nodes.  I would like the queues on each node to be completelyA independent.  I looked through the cluster manual but it seems toeG address only configurations where shared queues are used.  I don't wantuE queues on one node visible on other nodes.  One node has no queues at 1 all and would not have the queue manager running.   F Each node has a separate system disk.  Will SCS have any problems if I2 assign the QMAN logicals to the local system disk?   ------------------------------   Date: 7 Apr 2000 21:21:30 GMT02 From: hoffman@xdelta.zko.dec.nospam (Hoff Hoffman)) Subject: Re: multiple qmans in a cluster?g6 Message-ID: <8cljgq$g35$2@mailint03.im.hou.compaq.com>  l In article <J5qH4.17540$k7.2821544@news-west.usenetserver.com>, "Jack Peacock" <peacock@simconv.com> writes:  I   The more you have stuff configured as node-specific, the more problems o   you can make for yourself...  G :Is it possible to have a separate queue manager run on each node in an H :alpha vms 7.2 cluster?  I want drives and other resources to be shared,I :but each node has specific batch and print queues which are never sharedkI :with other nodes.  I would like the queues on each node to be completely B :independent.  I looked through the cluster manual but it seems toH :address only configurations where shared queues are used.  I don't wantF :queues on one node visible on other nodes.  One node has no queues at2 :all and would not have the queue manager running.  J   In addition to executing jobs only on a particular node, you can easily H   use ACLs to block access to the queues from other nodes, and to block D   the visibility of the jobs in the queues...  Alternatively (or in H   conjunction), you can set up a logical name on each node that -- when F   used as the queue name -- causes submitted jobs to be queued to the    node-specific queue.  G   If you want the queues to be truely and completely disconnected, you  >   can use DQS or similar tool to remotely queue the entries...  G :Each node has a separate system disk.  Will SCS have any problems if I 3 :assign the QMAN logicals to the local system disk?o  G   No, it won't, as long as they (files, etc) are completely separate...-  N  --------------------------- pure personal opinion ---------------------------L    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   ------------------------------  $ Date: Fri, 7 Apr 2000 15:57:49 -0700* From: "Jack Peacock" <peacock@simconv.com>) Subject: Re: multiple qmans in a cluster?(? Message-ID: <tGtH4.18491$k7.3081672@news-west.usenetserver.com>2  = Hoff Hoffman <hoffman@xdelta.zko.dec.nospam> wrote in message.0 news:8cljgq$g35$2@mailint03.im.hou.compaq.com... >tG > In article <J5qH4.17540$k7.2821544@news-west.usenetserver.com>, "Jackm& Peacock" <peacock@simconv.com> writes: >hA >   The more you have stuff configured as node-specific, the morei problems  >   you can make for yourself... >oH I have three nodes in the cluster, each with its own set of printers and@ batch jobs, no failover (special customer requirement due to howC companies are structured).  There is no HSC or quorum disk, clusteruE should remain up if any one node goes down (that's the level of faulteF tolerance/graceful degradation customer requested).  For the most partB all nodes operate independently but there are some users and tasksG suitable for the cluster as a whole (i.e. backup for one), so it is not F configured as three independent units. If I go with a clusterwide QMANC file it becomes locked to a particular node, with no benefit to thesG customer.  With separate QMAN files I only lose a portion of the queues.E if a node goes down, no more than one company is affected until theirE2 application is manually restarted on another node.     Jack Peacock   ------------------------------  $ Date: Thu, 6 Apr 2000 10:33:30 -0400( From: Etienne Vogt <vogt@alice.obspm.fr>, Subject: Network errors and CLUEXIT bugcheckJ Message-ID: <CB1575D4D198D311A3F800600803947E019013AA@scopent3.mar.hp.com>  J This message is in MIME format. Since your mail reader does not understand< this format, some or all of this message may not be legible.  ' ------_=_NextPart_000_01BFA09D.A193A742g Content-Type: text/plain;" 	charset="iso-8859-1"n  	 Hi there,a  G   Yesterday, my VAXStation 3200 has started to give me serious trouble.  It isiF running OpenVMS VAX 6.2 and is the slave in a two node VAXCluster (the master node is a VAXStation 3100/76).G   The problem is that the VS3200 suddenly lost connection to the VS3100f and was G thrown out of the cluster. As I'm currently away from where the machine  isH located, I sent E-Mail to a collegue there, asking her to check what wasF going on. She found the VS3200 in "quorum hang" state, but without any
 indicationG about why the connection went down. The two machines are interconnectedb viau; thin Ethernet wires, with a DELQA controller on the VS3200.tE   She tried changing the Ethernet thinwire transceiver and the VS3200u regained? connectivity and proceeded with a CLUEXIT bugcheck as expected.m  E  So, my first conclusion was that the old transceiver had died. But ai	 few hourseC later, the same thing happened again, so it must be something else.bF  Doing an analyse/error after the first crash, I found that before the clusteroG connection went down, the PEA0 device logged some 'Fatal error detectedM byA datalink' events. I have attached the output of the analyse/error  commandrE to this message. If someone here has seen this kind of problem before2 and77 can point to a possible cause, it would be appreciated.oG Note that this machine is no longer under any kind of support contract,a so& calling field service is not possible.  
 Best regards,    -- h# 		Etienne Vogt (vogt @ na.astro.it)S    ' ------_=_NextPart_000_01BFA09D.A193A742l Content-Type: text/plain;e 	name="err.lis"0+ Content-Transfer-Encoding: quoted-printablet  Content-Disposition: attachment; 	filename="err.lis"   1  ******************************* ENTRY    5419. =  *******************************nH  ERROR SEQUENCE 7911.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 00:49:46.83                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 07:25:47eH  SCS NODE: MESOPK                                              VAX/VMS = V6.2  =  DEVICE ATTENTION  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2A     NI-SCS SUB-SYSTEM, MESOPK$PEA0:  '        FATAL ERROR DETECTED BY DATALINKe          STATUS          0000045Cm                        00001201         DATALINK UNIT       0001a        DATALINK NAME   41515803/                        00000000x                        00000000E                        00000000r>                                        DATALINK NAME =3D XQA1:        REMOTE NODE     00000000                         00000000                         00000000i                        00000000l        REMOTE ADDR     00000000o                            0000n        LOCAL ADDR      000400AAh                            461C-:                                        ETHERNET ADDR =3D = AA-00-04-00-1C-46         ERROR CNT           0001 F                                        1. ERROR OCCURRENCES THIS ENTRY        UCB$W_ERRCNT        0003e:                                        3. ERRORS THIS UNIT =0CnH  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:16G                                                                       =n	 PAGE   2.   1  ******************************* ENTRY    5420. =t *******************************lH  ERROR SEQUENCE 7912.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 00:49:48.10                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 07:25:48hH  SCS NODE: MESOPK                                              VAX/VMS = V6.2  =  DEVICE ATTENTION  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2D     NI-SCS SUB-SYSTEM, MESOPK$PEA0:  '        FATAL ERROR DETECTED BY DATALINKr          STATUS          00000007m                        00001200o        DATALINK UNIT       0001e        DATALINK NAME   41515803n                        00000000c                        00000000n                        00000000t>                                        DATALINK NAME =3D XQA1:        REMOTE NODE     00000000t                        00000000e                        00000000p                        00000000T        REMOTE ADDR     00000000                             0000         LOCAL ADDR      000400AA.                            461Co:                                        ETHERNET ADDR =3D = AA-00-04-00-1C-46t        ERROR CNT           0001 F                                        1. ERROR OCCURRENCES THIS ENTRY        UCB$W_ERRCNT        0004-:                                        4. ERRORS THIS UNIT1  ******************************* ENTRY    5421. =  *******************************eH  ERROR SEQUENCE 7916.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 01:26:55.69                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 08:02:56uH  SCS NODE: MESOPK                                              VAX/VMS = V6.2  7  TIME STAMP  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2e =0CoH  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:16G                                                                       =l	 PAGE   3.f  1  ******************************* ENTRY    5422. =e ******************************* H  ERROR SEQUENCE 7917.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 01:32:50.07                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 08:08:500H  SCS NODE: MESOPK                                              VAX/VMS = V6.2  ;  ERL$LOGMESSAGE  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2r  !  NI-SCS SUB-SYSTEM, _MESOPK$PEA0:   &        PORT HAS CLOSED VIRTUAL CIRCUIT  -        LOCAL STATION ADDRESS, FFFFFFFFFF00(X)d'        LOCAL SYSTEM ID, 00000000461C(X)   .        REMOTE STATION ADDRESS, 0000000000DE(X)(        REMOTE SYSTEM ID, 000000004542(X)          UCB$B_ERTCNT          32n<                                        50. RETRIES REMAINING        UCB$B_ERTMAX          32 <                                        50. RETRIES ALLOWABLE        UCB$W_ERRCNT        0005 :                                        5. ERRORS THIS UNIT        PPD$B_PORT            00 7                                        REMOTE NODE # 0.g        PPD$B_STATUS          002        PPD$B_OPC             00h5                                        UNKNOWN OPCODE         PPD$B_FLAGS           00  =0CaH  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:18G                                                                       = 	 PAGE   4..  1  ******************************* ENTRY    5423. =h *******************************bH  ERROR SEQUENCE 7918.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 15:36:14.51                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 22:12:14-H  SCS NODE: MESOPK                                              VAX/VMS = V6.2  =  DEVICE ATTENTION  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2      NI-SCS SUB-SYSTEM, MESOPK$PEA0:  '        FATAL ERROR DETECTED BY DATALINK           STATUS          0000045Co                        00001201u        DATALINK UNIT       0001         DATALINK NAME   41515803i                        00000000e                        00000000i                        00000000c>                                        DATALINK NAME =3D XQA1:        REMOTE NODE     00000000s                        00000000a                        00000000l                        00000000u        REMOTE ADDR     00000000u                            0000b        LOCAL ADDR      000400AA                             461Cs:                                        ETHERNET ADDR =3D = AA-00-04-00-1C-46.        ERROR CNT           0001 F                                        1. ERROR OCCURRENCES THIS ENTRY        UCB$W_ERRCNT        0006a:                                        6. ERRORS THIS UNIT =0CoH  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:18G                                                                       =n	 PAGE   5.   1  ******************************* ENTRY    5424. =u ******************************* H  ERROR SEQUENCE 7919.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 15:36:20.70                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 22:12:20cH  SCS NODE: MESOPK                                              VAX/VMS = V6.2  =  DEVICE ATTENTION  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2y     NI-SCS SUB-SYSTEM, MESOPK$PEA0:  '        FATAL ERROR DETECTED BY DATALINKc          STATUS          8133B6FCm                        00001200V        DATALINK UNIT       0001<        DATALINK NAME   41515803m                        00000000i                        00000000t                        00000000o>                                        DATALINK NAME =3D XQA1:        REMOTE NODE     00000000c                        00000000e                        00000000c                        00000000l        REMOTE ADDR     00000000i                            0000e        LOCAL ADDR      000400AAm                            461C :                                        ETHERNET ADDR =3D = AA-00-04-00-1C-46         ERROR CNT           0001sF                                        1. ERROR OCCURRENCES THIS ENTRY        UCB$W_ERRCNT        0007u:                                        7. ERRORS THIS UNIT =0C-H  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:18G                                                                       =i	 PAGE   6.a  1  ******************************* ENTRY    5425. =A *******************************iH  ERROR SEQUENCE 7920.                            LOGGED ON:        SID = 0A000004H  DATE/TIME  6-APR-2000 15:36:22.77                            SYS_TYPE = 01120102  SYSTEM UPTIME: 0 DAYS 22:12:22 H  SCS NODE: MESOPK                                              VAX/VMS = V6.2  ;  FATAL BUGCHECK  KA650  CPU FW REV# 5.  CONSOLE FW REV# 1.2   -  CLUEXIT, Node voluntarily exiting VAXclusterc          PROCESS NAME    NUL=00h          PROCESS ID      00010000s          ERROR PC        8152FB21o        ERROR PSL       04080004 ,                                        Z-BITG                                        INTERRUPT PRIORITY LEVEL =3D 08.g?                                        PREVIOUS MODE =3D KERNELw>                                        CURRENT MODE =3D KERNEL6                                        INTERRUPT STACK<                                        FIRST PART DONE CLEAR    STACK POINTERS=    =20E  KSP 7FFE7800  ESP 7FFE9800  SSP 7FFECA44  USP 00163944  ISP 81C574DCt    GENERAL REGISTERS    =20E  R0  00000001  R1  813AC340  R2  8141B284  R3  8134B600  R4  81391790eE  R5  813AC300  R6  00000000  R7  00000034  R8  81394040  R9  00077B64=E  R10 00096ADC  R11 00096C7C  AP  00163968  FP  00163944  SP  81C57574e    SYSTEM REGISTERSs    =20        P0BR            82219800e?                                        P0 PTE BASE (VIRT ADDRS)a        P0LR            000018C3o5                                        TOTAL P0 PAGESh        P1BR            81A51600 ?                                        P1 PTE BASE (VIRT ADDRS)e        P1LR            001FF6BFaB                                        TOTAL NON-EXISTENT P1 PAGES        SBR             017A4200FC                                        SYSTEM PTE BASE (PHYS ADDRS)-        SLR             00014980 D                                        TOTAL PAGES "SYSTEM" VIRT MEM        PCBB            00ACAC20@<                                        PCB BASE (PHYS ADDRS)        SCBB            0179DA00w<                                        SCB BASE (PHYS ADDRS)        ASTLVL          00000004n7                                        NO AST'S PENDINGe        SISR            00000000aF                                        INTERRUPT REQUEST ACTIVE =3D 0.        ICCS            00000040a =0CdH  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:18G                                                                       =n	 PAGE   7.o  7                                        INTERRUPT ENABLEa        TODR            41C5F934e        CADR            000000FC >                                        1ST LEVEL CACHE STATUS:8                                        _D STREAM ENABLED8                                        _I STREAM ENABLED5                                        _SET 1 ENABLEDu5                                        _SET 2 ENABLED8        MSER            00000080a;                                        1ST LEVEL CACHE MISSs        DSER            00000000c        QBEAR           0000000Ao        DEAR            00000000o        IPCR0           00000020 E                                        LOCAL MEMORY EXTERNAL ACCESS =e ENABLEDr          MEMCSR16        00000044sF                                        PHYSICAL PAGE ADDR =3D 00000(X)C                                        ECC ERROR SYNDROME =3D 44(X)w        MEMCSR17        00001045e2                                        ECC ENABLED<                                        CRD INTERRUPT ENABLEDD                                        MAIN MEM CYCLE SELECT =3D 5/3        MEMCSR0         80000016rH                                        MEMORY MODULE TYPE =3D MS650 (8 = MB)u:                                        SYSTEM BANK =3D 00.        MEMCON          002A0333C<                                        MEMORY CONFIGURATION:;                                        _BANKS ENABLED =3D =e 0000001100110011F                                        _MEMORY MODULE #1. - MS650 (8 = MB) F                                        _MEMORY MODULE #2. - MS650 (8 = MB) F                                        _MEMORY MODULE #3. - MS650 (8 = MB)y;                                        MEMORY ERROR STATUS:r9                                        _MEMORY MODULE #1.h        CACR            FFB97C90h>                                        2ND LEVEL CACHE STATUS:/                                        _ENABLEDe>                                        CYCLE SPEED CODE =3D 2.        CBTCR           C0000004SF                                        CDAL BUS TIMEOUT INTERVAL =3D = 0.4 MSECD                                        TIMEOUT DURING CPU READ/WRITE7                                        CDAL BUS TIMEOUTm =0CnH  V A X / V M S        SYSTEM ERROR REPORT         COMPILED  6-APR-2000 = 15:57:18G                                                                       =$	 PAGE   8.n  0 ANAL/ERROR/EXCLUDE=3DVOLUME/SINCE=3D6-APR-2000 =# 00:00:00.00/OUT=3DPOUBELLE3:ERR.LIS.  ) ------_=_NextPart_000_01BFA09D.A193A742--w   ------------------------------  $ Date: Sat, 1 Apr 2000 07:54:46 +12008 From: "Antony Wardle" <antony.wardle@nospamme.met.co.nz> Subject: Re: Old vax3 Message-ID: <Kj7F4.60077$3b6.236074@ozemail.com.au>   ! Thanks to everyone for the reply.e    2 I might grab a couple and setup my own vaxcluster.2 trouble is sneaking in the house past the wife ;-)  7 You can use the scsi disks in your pc, well I have usede5 rz26 's anyway. Plus I have a storage works box, so Ie- could at least put an operating system on it.   ; the vlc's are too skinny to put an rz26 in, fingers crossedh; that I might be able to take the boss into throwing out the = model 90. Then I'd be in. I guess I will have to try and findo an external cd rom.c  ? The linux question was more a what else can I run, if I want toi try something else and not vms.q   Thanks   Antony   ------------------------------  % Date: Sat, 01 Apr 2000 00:39:24 -0500m* From: David A Froble <davef@tsoft-inc.com>3 Subject: Re: old vs. new IRON (Was FS: Digital VAX)h- Message-ID: <38E58B8C.A252B225@tsoft-inc.com>t   Ian Parker wrote:> > F > Assuming a DEC-3000-400 is a DEC 3400s and a VAX-4000-705a is 4700A,# > then the VAX shows as 60% faster!r  N What I wonder, is why anyone would take more specific naming, and turn it into more ambiguous naming?   Dave   -- m4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596; 170 Grimplin Road               E-Mail: davef@tsoft-inc.com0 Vanderbilt, PA  15486e   ------------------------------  # Date: Sat, 01 Apr 2000 13:31:49 GMT ' From: benmyers@ultranet.com (Ben Myers) 3 Subject: Re: old vs. new IRON (Was FS: Digital VAX)23 Message-ID: <38e5f9fd.1016661@news.ma.ultranet.com>   E Obfuscation!  Keep your customers in the dark about your products.  A D trademark of name-brand companies selling proprietary iron.  And theE Wintel industry is not immune to this phenomenon, either... Ben Myerso  + David A Froble <davef@tsoft-inc.com> wrote:n   >Ian Parker wrote: >> eG >> Assuming a DEC-3000-400 is a DEC 3400s and a VAX-4000-705a is 4700A,h$ >> then the VAX shows as 60% faster! >rO >What I wonder, is why anyone would take more specific naming, and turn it intot >more ambiguous naming?x >t >Dave  >, >-- 5 >David Froble                       Tel: 724-529-0450u5 >Dave Froble Enterprises, Inc.      Fax: 724-529-0596 < >170 Grimplin Road               E-Mail: davef@tsoft-inc.com >Vanderbilt, PA  15486   ------------------------------  # Date: Fri, 07 Apr 2000 21:53:31 GMTs1 From: Part Time VMS person <hcarmcar@my-deja.com> @ Subject: Re: OVMS : Should I retrain back to the latest version?) Message-ID: <8cllck$kg9$1@nnrp1.deja.com>   ) In article <38E88A5A.108DAA8A@bbc.co.uk>,i    tim.llewellyn@bbc.co.uk wrote: >  >e > bfb wrote: >e > >o > >eB > > To get back on topic, there must be some UK agencies which areB > > superior when dealing with OVMS contracts, any suggestions? As. > > Jobserve is a interesting form of fiction. > >4 > C > I did a year or two back read about an agency that handled (then)n	 Digital'srG > own requirements, but I didn;t keep the article. I'd be interested in* knowing*: > if such an agency does exist. Jobserve is fun, isn;t it. >* >n  F I just did a quick look on http://www.geekfinder.com for InternationalG jobs for "OVMS VMS OpenVMS" and found three jobs - one in Saudi Arabia, D two in the Netherlands.  When I selected "All US States", I got 1006D hits (naturally I didn't look at them all.)  Checking in New Jersey,> area code 201, I found 17 hits.  It might be worth looking at," particularly if you're in the USA.   --H Henry T.   "What shall it profit a man if he shall gain the whole world,(  and lose his own soul?" - Jesus Christ./ http://home.att.net/~henryandmartha/  And check5< http://www.computronsoftware.com  and http://home.arachne.cz    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Fri, 07 Apr 2000 10:45:00 -07000! From: Shane.F.Smith@healthnet.com0" Subject: Re: Re submit a batch job8 Message-ID: <882568BA.006185A1.00@WHDOM99.HEALTHNET.COM>  M Also if logicals are being used to indicate where the logfile should go, make P sure those logicals are visible to the batch job as it's starting. I once had toM work out why only half the jobs being submitted to a queue vanished without a P trace: It was a generic queue on a 2-node cluster, and the logfile logicals were only set up on one node.   Shane                 8 "Dale A. Marcy" <dqm@ornl.gov> on 04/07/2000 10:14:57 AMP                                                                                 P                                                                                 P                                                                                     >                                                               >                                                               >                                                               >  To:      Info-VAX@Mvb.Saic.Com                               >                                                               >  cc:      (bcc: Shane F Smith/WH/HN/FHS)                      >                                                               >                                                               >                                                               >  Subject: Re: Re submit a batch job                           >                                                                              Stephen,  M      Generally when I see a batch job just "disappear", it is because the log L file cannot be created or it cannot access the command procedure that drivesK the batch job.  For the log file, check access to the directory where it is L being created and diskquotas for the username on that disk.  For the commandM procedure, check the access.  I think I have witnessed this same problem once L when the account could not be logged into, but I cannot recall the specifics on that one.   Hope this helps,  
 Dale A. MarcyN. Science Applications International Corporation   "Ste C." wrote:*   > Thanks Bernard,  >RH > That seems to work ok except the job I am submitting goes on the batchD > queue but when it is scheduled to run it dissapears and doesn't doC > anything. If I execute the job on its own (ie @aqa010bgd) then itP# > executes as it should. Any ideas?  >  > Thanks again >2	 > StephenT > , > In article <38D63922.69755902@azisoft.ch>,5 > Bernard Schluep <Bernard.Schluep@azisoft.ch> wrote: ( > > I hope it is enough for your request > > 9 > > $ submit/notif/noprint/log login.com/after="tom+7:00" J > > Job LOGIN (queue BATCH$XXX, entry 896) holding until 21-MAR-2000 07:00 > >* > > Bernard* > >E > > "Ste C." wrote:  > > G > > > Ive got a batch job that processes batches of data files one at a0G > > > time, if there is more than two files in a certain directory thenP > the J > > > job processes the file then re submits its self to process the otherI > > > file(s) straight away, if there is only one file there then the jobVI > > > will resubmit itself and run in an hours time (just incase the fileUF > > > that is there is still being written to), that is if the time is > after B > > > 7.00am but before midnight. If the time plus 1 hour is after
 > midnightH > > > then I want to submit the job to run at 7:00am the next morning. I > can G > > > add an hour to the time easily enough but is there an easy way to  > add 1 H > > > day ? is there a way to check if its the last day of the month and' > > > therefore add another month etc ?  > > > ) > > > Hope this makes sense to someone :)B > > >  > > > Thanks > > > 
 > > > Stephen  > > > -- > > > ArrghO > > >g, > > > Sent via Deja.com http://www.deja.com/ > > > Before you buy.  > >  > >  > -- > Arrgh  > ( > Sent via Deja.com http://www.deja.com/ > Before you buy.a   ------------------------------  % Date: Fri, 07 Apr 2000 18:30:58 -0500 ( From: yyyc186.illegaltospam@flashcom.net" Subject: Re: So who will buy VMS ?9 Message-ID: <38ee6fd5$2$lllp186$mr2ice@news.flashcom.com>   4 In <senu2950rfg174@corp.supernews.com>, on 04/07/00 ;    at 06:30 PM, "David Turner" <d_b_turner@yahoo.com> said:=  J >> >* a large group of 30-40 year olds which learned VMS is the golden VAX >> >VMS days in 6 >> >  the 80's - this is undoubtetly the largest groupI >> >* a smaller group of 40-60 year olds which started with PDP-somethingO >> >and  >> >  migrated to VAX VMSRI >> >* a very small group of 20-30 year olds which learned VMS in the 90's   F How about us 30 somethings that started with PDP-11 and RSTS/E because: that's what school had, then went on to learn VMS?  <Grin>   Roland   --  ; -----------------------------------------------------------0D yyyc186@flashcom.net              To Respond delete ".illegaltospam"6                             MR/2 Internet Cruiser 1.528                             For a Microsoft free univers; -----------------------------------------------------------    ------------------------------  " Date: Fri, 7 Apr 2000 21:09:59 GMT0 From: "Terry C. Shannon" <shannon@world.std.com>/ Subject: Re: Sun considers VMS a "mainframe" OSH& Message-ID: <Fsnzpp.AJo@world.std.com>  2 "Bill Todd" <billtodd@foo.mv.com> wrote in message! news:8cifa0$rk$1@pyrite.mv.net...EL > Considering VMS a 'mainframe' system is not limited to Sun.  In fact, backG > when the most powerful VAXen were in development, there was sentimentP within- > DEC that VMS was headed for that territory.  > I > VMS certainly has far closer to mainframe-level support for things like*B > batch processing than Unices do, and in general may look like an inscrutable K > mainframe-style OS to people raised on Unix.  Which is not to say that itAG > competes directly with S/390 system variants across their full range._  E Remember the VAX 9000 (Aquarius/Aridus)? That box was positioned as a  mainframe...   ------------------------------  " Date: Fri, 7 Apr 2000 21:08:57 GMT0 From: "Terry C. Shannon" <shannon@world.std.com>/ Subject: Re: Sun considers VMS a "mainframe" OSC& Message-ID: <Fsnzo2.A6o@world.std.com>  3 Well, they have to consider it to be something, no?D  7 "Craig Paul" <paul@kuhub.cc.ukans.edu> wrote in message ' news:JD+aBN9gf+uG@kuhub.cc.ukans.edu... ? > Employees of Sun consider VMS a "mainframe" operating system.  > 0 > That says alot to me about what Solaris isn't.   ------------------------------  " Date: Fri, 7 Apr 2000 21:12:35 GMT0 From: "Terry C. Shannon" <shannon@world.std.com>/ Subject: Re: Sun considers VMS a "mainframe" OSi& Message-ID: <Fsnzu1.BrC@world.std.com>  = "JF Mezei" <jfmezei.spamnot@vl.videotron.ca> wrote in message ) news:38ECD343.19D395A3@vl.videotron.ca...EI > Ok, someone please explain to me why VMS/ALPHA is not considered in the  same! > class as IBM/390/MVS machines ?C >R" > Isn't Alpha a faster processor ? In most cases, yes.   4 > Do IBM disk drives have far superior performance ?  0 Perhaps... which is why CPQ buys disks from IBM.  J > Is is a question of the bus which is much faster on IBM than on Alphas ?  / Have never really figured out parallel sysplex.*    D MS is at a performance disadvantage compared to the 1960s's IBM 360, > 1- how come ?E  I Ask His Gateness. Advertising might have a lot to do with it. Advertising/- foisted Bill Clinton on us, so why not an OS?    > 2- will wildfire fix that ?S   I fervently hope so!   ------------------------------  % Date: Fri, 31 Mar 2000 17:45:40 -0500S* From: David A Froble <davef@tsoft-inc.com>9 Subject: Re: VMS today: % users x-widnows vs VT terminalse- Message-ID: <38E52A94.D04A0EFB@tsoft-inc.com>L  I All VT-500 series terminals, the 510, 520, and 525 use PS-2 type keyboard N connections.  I will not claim that the 'feel' of the later keyboards is quiteP as good as a VT-420, which was/is in my opinion the best 'feeling' keyboard I'veN ever user, and the way things are going, probably the best the world will everF have had.  Still, the LK411-AA I am typing on right now isn't too bad.   Dave     Gord Coulman wrote:  > a > I don't understand.  You kept the VT-500 series keyboard?  It plugs into your PC?  I have neverFW > seen a VT-500, but the 420's we have around here don't have anything like an AT-style0, > keyboard.  I wonder where I could get one? >  > Gord.e >  > David A Froble wrote:  > I > > I can scrool back a thousand lines or more, I can cut/copy and paste,1O > > and I have retained the VT-500 series keyboard.  A few function and numeric S > > keypad keys are not mapped correctly, but I believe they could be, I'm just too T > > lazy.  For this purpose it's a better VT terminal.  Of course, if I had to use aS > > peecee keyboard you'd hear a different story.  (You got to wonder, watching allEK > > the keyboard design advances through the 80s, just to revert to junk on ) > > peecees!  One of my biggest bitches!)    --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596; 170 Grimplin Road               E-Mail: davef@tsoft-inc.comL Vanderbilt, PA  15486M   ------------------------------  # Date: Sat, 01 Apr 2000 00:08:30 GMT  From: jbecker@syssol.com9 Subject: Re: VMS today: % users x-widnows vs VT terminals ) Message-ID: <8c3elm$v51$1@nnrp1.deja.com>T  . In article <he+hXvGeMCnQ@tachxxsoftxxconsult>,< wayne@tachyon.xxx.424646.killspam.064d (Wayne Sewell) wrote:  F > I find the ability to roll back decterms to see lines that have been scrolledC > off the screen one of the most important features. Can any of theP straightC > terminal emulators do this? The real terminals they are emulating 	 certainly  > can't.  ? You bet -- Reflection, SmarTerm, PowerTerm, and probably others @ have scroll bars, and you don't need their X emulators to do it.A Their just-plain VT emulators can do it. I depend on this featureL@ routinely. I wish Kermit95 supported scroll bars too. Using Page; Up/Page Down to step through screens gets in the way of theE" host-specific uses for those keys.  B > Also, the ability to cut and paste between decterms is extremely	 useful as8G > well. Since the decterms in question could be on different computers,  thisG > could be considered a poor man's sysman "do" command. In other words,  you canoF > paste the command into windows on different nonclustered systems and	 issue the  > command on each.  A I also depend routinely on cut-and-paste between terminal windows ) using the same emulators mentioned above.   E > True, I am talking about decterms directly on a vms system, but you 	 should beE+ > able to do the same stuff with x on a pc.   ? There are times when I wish I had an X emulator instead of a VT ? emulator, but X emulators are sometimes more trouble to set up, @ especially if you switch back and forth between separate systems a lot.   --
 Jim Becker6 System Solutions Incorporated (http://www.syssol.com/) Rockville MD & Herndon VAu3 DECUS ESILUG (http://eisner.decus.org/lugs/esilug/)=    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  $ Date: Fri, 7 Apr 2000 17:40:34 -0400= From: "Mike Foley" <mike.foley@technologist-dot-com-spamspam> " Subject: Re: VRC16 Monitor Problem. Message-ID: <sesleksicqv41@corp.supernews.com>  ?     I think that's the VRT16. The VRC16 is a multisync monitor. ?     When you start to hear pops or cracks, it means it's on its      way out.     mike  : "David B Sneddon" <dbsneddon@bigpond.com> wrote in message% news:38ED5C84.37C32772@bigpond.com... E > The VRC16 uses three BNC connectors (R, G and B) at the monitor endDJ > and a rather non-standard connector at the video card end... not readily! > connectable to your average PC.C > --
 > Regards, > Dave. K > ------------------------------------------------------------------------- K > David B Sneddon (dbs)  OpenVMS Systems Programmer   dbsneddon@bigpond.comOK > DBS software at ...   http://www.users.bigpond.com/dbsneddon/software.htmEK > "Life is what happens to you while you're busy making other plans" Lennon-   ------------------------------  $ Date: Sat, 8 Apr 2000 15:06:04 +1000- From: David B Sneddon <dbsneddon@bigpond.com>U" Subject: Re: VRC16 Monitor Problem0 Message-ID: <05060411176431@domain6.bigpond.com>  / At 05:40 PM 07-04-2000 -0400, Mike Foley wrote:- >-@ >    I think that's the VRT16. The VRC16 is a multisync monitor.@ >    When you start to hear pops or cracks, it means it's on its
 >    way out.4 >0 >3 >mike7 >z; >"David B Sneddon" <dbsneddon@bigpond.com> wrote in messagei& >news:38ED5C84.37C32772@bigpond.com...F >> The VRC16 uses three BNC connectors (R, G and B) at the monitor endK >> and a rather non-standard connector at the video card end... not readily " >> connectable to your average PC. >> --t  5 The monitor I have (I'm not the one with the problem)t is a VRC16-C4.4 There are actually 5 BNC connectors R,G,B,H/C and V.6 The cable I have on mine only uses the RGB connectors.   Regards, Dave.oI -------------------------------------------------------------------------eI David B Sneddon (dbs)  OpenVMS Systems Programmer   dbsneddon@bigpond.com F DBS software is at http://www.users.bigpond.com/dbsneddon/software.htmI "Life is what happens to you while you're busy making other plans" Lennon.   ------------------------------  % Date: Fri, 07 Apr 2000 20:56:40 -0400c From: mhr <mhr@sunline.net> 1 Subject: Re: Warning: circuitsurgeon is a ripoff!X+ Message-ID: <38EE83C7.EFC6774C@sunline.net>   F Tom, you should indicate we had to (entirely) replace one system after less then 90 days.(                                     Marc   ------------------------------  # Date: Sat, 08 Apr 2000 03:41:54 GMTi! From: Beyonder <beyonder@vrx.net>71 Subject: Re: Warning: circuitsurgeon is a ripoff!n' Message-ID: <38EE9CE7.819D5BA7@vrx.net>-   coy@eisner.decus.org wrote:   F > I think I've lost track of some things here.  Could you answer a few > questions, please? >I8 > 1.  When did you buy this equipment, and for how much?  F I don't honestly remember, it's been several months. maybe 6 at least.$ I think I paid about $600 US for it.  0 > 2.  When did you find out that it didn't work?   the day I got it!.  G > 3.  What parts have been replaced so far, and what did they cost you?n  F circuitsurgeon replaced the hard drive, but I had to pay shipping both ways. D he said only fedex was acceptable. so there is another $70. lets not forget customs which cost me about $200-$300.  E I later had problems with the kb/mouse about two days after I got theeG replacement hard drive. Since that time I have emailed him, and called.,A But as I said, I was hung up on, and my emails have been ignored.t  @ Also more recently the "alternate console" connection has ceased functioning.  I Today I took the unit into DEC (Compaq) service center and found out thatgD BOTH the kb/mouse adapter cable *AND* the motherboard are defective.  I Compaqs cost for a new cable $22 US (which I purchased -warranty and all)0  H Compaqs cost for a new motherboard $3,400.00 US (no, that is NOT a typo)  H I have enroute a new motherboard and cpu board with cpu, total cost willH be about $100 US (including shipping but not customs). Gil was trying to' charge me $200 US just for an IO board.   I The person I am dealing with in california really knows his stuff, unlikedC Gil, and is also pleasant to talk to, and a really great guy, again  unlike Gil-who is just a jerk.  H So total cost to me for all this (let's include Compaqs bench test costs7 which were about $100) and we come to a grand total of:l System - $600 US! Shipping on replacement HD - $100  Customs & duties - $250 US  Replacement board, etc - $100 US Compaq testing - $100: Replacement cable - $22 US< Gas costs driving all over the damn place - 200miles of fuel! Phone costs (long distance) - $50n& My time and trouble - priceless (grin)  I so let's see that comes to about $1,300.00 for a system that is worth, asa$ it turns out, about $200.00 ah well.  G seeing as I just picked up two vaxstations (3100 m98) for $70 for both!eE I also got a micro-vax ii for $100, and two dozen vt320 terminals for  $100G (that's total, not each - roughly $5 a piece!) although most of this isC
 already sold.   H I have another source of a dozen vt320s but the cost will be about $150.   Dan.   ------------------------------  $ Date: Fri, 7 Apr 2000 15:09:49 -0400) From: "Ebinger . Eric" <EEbinger@drc.com>a? Subject: RE: Why are there no French people in this newsgroup ? B Message-ID: <7162F87E9EF4D311BA9900805FC1D3AE7A6116@and02.drc.com>  G Something to remember is that in France there is a public system calledaE Minitel that predates the development of HTTP/HTML and is a "free"=20tC service of the telephone monopoly.  To put it in perspective, 1.8 =u millionsG of France's 26 million homes are hooked up to the internet; 8.6 million F are hooked up to the Minitel network.  The Minitel network has been=20 in existence for 17 years.     Eric Ebinger       > -----Original Message-----: > From: morrisj@epsilon3.com [mailto:morrisj@epsilon3.com]& > Sent: Friday, April 07, 2000 1:31 PM > To: Info-VAX@Mvb.Saic.Com A > Subject: Re: Why are there no French people in this newsgroup ?s >=20 >=20> > On Fri, 7 Apr 2000 14:03:48 +0200, "Jean-Fran=E7ois Marchal") > <jean-francois.marchal@x9000.fr> wrote:b >=20 > >l> > >"Allan Hawdon" <udaa460@axolotl.kcl.ac.uk> wrote in message+ > >news:8ckdv0$ndb$1@willow.cc.kcl.ac.uk...  > >aF > >> Take a look on vmsnet.sysmgt - there seems to be a French landing > >> party there :-) > >r$ > >Yes , and completely off topic... > >Jean-Fran=E7ois > >  > >h? > I'd been wondering about that.  I don't speak French so I wasa6 > wondering if vmsnet.sysmgt meant "The party's here". >=20 > Ahh, what is their topic?  >=20 > -- > Jay E. Morris' > System Software Specialist=20T2 > (confuses the PHBs to call us managers or admin)( > General Dynamics Communication Systems) > for/Epidemilogical Surveilence Division  > Brooks AFB, TX >=20   ------------------------------  # Date: Fri, 07 Apr 2000 20:57:05 GMTeM From: =?iso-8859-1?Q?Andr=E9?= Bastien <bastien.andre.p@hydro-no-spam-.qc.ca>s? Subject: Re: Why are there no French people in this newsgroup ?n4 Message-ID: <38EE4BBD.D5E45CFF@hydro-no-spam-.qc.ca>  ' Merci pour la rfrence  DECUS France.c   Un cousin d'outremer.   ! "Jean-Franois Marchal" a crit :"   > Bonjour  tous ... >  > L > A WEBnotes variant has been implemented too (info at http://www.decus.fr ) >u > Cordialement > Jean-Franois Marchaln > X9000 - LYON (FR)d >sB > "Larry D Bohan, Jr" <LBohan@dbc.spam_less..com> wrote in message. > news:AzztODjHT2DShecd+hu9KKrIiYjg@4ax.com...5 > > On Thu, 6 Apr 2000 20:57:30 -0400, "David Turner"o! > > <d_b_turner@yahoo.com> wrote:  > >  > > >This is a weird deal....XN > > >Netherlands, Germany, US UK, Ozzy's;  pretty much everyone, but no French > >e" > > there's always  fr.comp.os.vms > > 5 > > It doesn't look too busy/active, but I recognizedr) > > one ot two names from C.O.V, even so..   ------------------------------   Date: 7 Apr 2000 16:33:09 CSTe; From: wayne@tachyon.xxx.264252.killspam.03a8 (Wayne Sewell)d? Subject: Re: Why are there no French people in this newsgroup ?a. Message-ID: <O9S8$bKpw8Mj@tachxxsoftxxconsult>  l In article <8ckipl$eo8$1@minus.oleane.net>, "Jean-Franois Marchal" <jean-francois.marchal@x9000.fr> writes:= > "Allan Hawdon" <udaa460@axolotl.kcl.ac.uk> wrote in message1* > news:8ckdv0$ndb$1@willow.cc.kcl.ac.uk... > D >> Take a look on vmsnet.sysmgt - there seems to be a French landing >> party there :-) > # > Yes , and completely off topic...  > Jean-Franois3    N I'll take your word for it, since I don't speak French and have no idea *what* they are talking about.  :-)  O This appears to be the same bunch that has been hanging out in vmsnet.groups ingO past weeks.  Since that has to do with forming/disbanding newsgroups in vmsnet,-K I would be surprised if their discussion, whatever it is, is relevant therer either.  :-)     -- nO ===============================================================================tK Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachyon.xxxo8 http://www.tachyon.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-)RO ===============================================================================0N Butler:"Gentlemen!"  Curly(as he and other Stooges look around):"Who came in?"   ------------------------------   Date: 7 Apr 2000 19:37:45 CSTs; From: wayne@tachyon.xxx.254471.killspam.03b4 (Wayne Sewell)k? Subject: Re: Why are there no French people in this newsgroup ?g. Message-ID: <oPJZu2C$9rMh@tachxxsoftxxconsult>  b In article <38EDFE47.70B866FC@vl.videotron.ca>, JF Mezei <jfmezei.spamnot@vl.videotron.ca> writes:L > Si les francophones commen=E7aient =E0 envoyer des messages en fran=E7ais= > , le reste< > des usagers sur comp.os.vms se plaindraient tr=E8s fort... > L > (If french speaking users started to post in french here, the other users= >  > would complain loudly).r > L > When when a french speaking poster posts in english, he becomes "invisibl= > e") > and you don't really notice him/her/it.S >   M Well, not *quite* invisible.  Some of us do understand what an address endingnM in ".fr" means.  I think the original poster meant that you don't see that ask" often as ".uk", ".au", ".dk", etc.      L > Also, I think that the number of users of comp.os.vms has gone down quite= >  aL > bit, whcih really reflects the VMS user base in the world (more so that t= > heL > sales that Compaq brags about since those sales don't mention the number = > ofL > existing users who shut down their last VMS machine or have remaining VMS= > D > machines simply running with minimal maintanence and no upgrades). -- dO ===============================================================================gK Wayne Sewell, Tachyon Software Consulting  (281)812-0738  wayne@tachyon.xxxn8 http://www.tachyon.xxx/www/tachyon.html and wayne.html  K change .xxx to .com in addresses above, assuming you are not a spambot  :-)rO ===============================================================================sN Butler:"Gentlemen!"  Curly(as he and other Stooges look around):"Who came in?"   ------------------------------  # Date: Sat, 08 Apr 2000 03:34:25 GMT.2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>. Subject: Re: Wierd Elsa Gloria Synergy Problem6 Message-ID: <5NxH4.1733$h81.30718@typhoon.aracnet.com>  4 Fred Kleinsorge <kleinsorge@star.zko.dec.com> wrote:L > If you are not doing heavy graphics output, and have a god quality monitorL > (I have a nice DEC 21" monitor) I haven't had many problems -- at least at > 70Hz.7  K > If you can live with the beat frequency flicker at 60Hz, it "should" workpJ > better... but this also tends to exacerbate any power supply glitches as > well.-  K Unfortunatly the results at 60Hz and 70Hz, while not as bad, were the same  I as at say 75Hz.  At 1280x1024 24-bit I get 'wobbles' on the left side of cM the screen.  I've tried this on both my DEC 17" and my VeiwSonic 21".  Hmmm, c9 I take that back, I've not tried 60Hz on the ViewSonic.  u  J On the DEC at 60Hz the wobble was *almost* gone.  In fact the only thing IG was able to do at 60Hz to bother it was to page through a web page with K Mozilla.  At higher frequencies I was able to effect it by paging through a.2 directory listing with the DECwindows filemanager.  L It's really wierd since at 1024x768 70Hz 8-bit or 24-bit it's rock solid on  both monitors.  I Actually I think the best solution for me is to leave it at 1024x768 70Hz,J 24-bit and simply start up another session at 1280x1024 24-bit on my Mac. L That way I've got everything I want to run on the same desktop.  I don't seeH me needing to do anything with really heavy graphics on VMS, and the twoB machines are on a 100Mbit switch, so it seems to work pretty good.   				Zane   ------------------------------  % Date: Fri, 07 Apr 2000 22:59:11 -0400y* From: David A Froble <davef@tsoft-inc.com>. Subject: Re: Wierd Elsa Gloria Synergy Problem- Message-ID: <38EEA07F.7D7F5206@tsoft-inc.com>r   Fred Kleinsorge wrote: > L > If you are not doing heavy graphics output, and have a god quality monitor<                                                          ^^^O I really hope you don't consider a 21" monitor, even a DEC 21" monitor, in this0K class.  If so, then I don't have much to look forward to. :-(  Now 3 walls,eM ceiling and floor of a generous sized room,  That would be a serious monitor!bP :-)  (I'd ask for the back wall also, but as I get older, I find turning my head that far is becoming harder.)s  L > (I have a nice DEC 21" monitor) I haven't had many problems -- at least at > 70Hz.s   Dave   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596; 170 Grimplin Road               E-Mail: davef@tsoft-inc.com4 Vanderbilt, PA  15486    ------------------------------   End of INFO-VAX 2000.196 ************************