1 INFO-VAX	Mon, 30 Jul 2001	Volume 2001 : Issue 419       Contents:  Re: (auto|sys)gen and params.dat% Re: Advanced Server and virtual disks ' Re: Alpha: an invitation to communicate ! Re: benchmarking disk performance & Re: Compaq FUD and lack of information Re: CSA  [was LPs on the Web] $ Re: Few People in DEC Understood....% Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance % Re: IA64 volume and low-end dominance : Re: IPF Console Bootstrap (was: Re: No chance for OpenVMS)  login-f-clisymtbl error on Alpha Re: No chance for OpenVMS  Re: Ont "The Inquirer" today. ' Re: Passing a socket to another process P Re: problem deallocating a CD rom device when reflection used to allocate it is P Re: problem deallocating a CD rom device when reflection used to allocate it is $ Re: Selling VMS to another company ?3 RE: TCPIP v5.1 startup on dial-up service provider.  Re: V7.3 XFCACHE-W-DATALOSS ?   F ----------------------------------------------------------------------    Date: 29 Jul 2001 11:32:46 -0700- From: afeldman@gfigroup.com (Alan E. Feldman) ) Subject: Re: (auto|sys)gen and params.dat = Message-ID: <af1e4ce6.0107291032.7f748d7a@posting.google.com>   ^ "Brian M. Kelley" <bkelley@vizdom.com> wrote in message news:<3B63854D.2B543075@vizdom.com>...7 > Hey all. VMS-admin newbie here. We're running OpenVMS  >   7.2-1 on an Alpha 433a.  > : >   I did a bad (?) thing: I deleted sys$system:params.dat; >   (don't ask; little sleep the previos night). Here's the = >   catch: I don't have a backup. The question is, of course, = >   how can I recover/regenerate this file? Is this possible? ? >   I can see the system param values running 'paramaters/show' = >   in sys$system:sysman, so I know regenerating this file is : >   possible in some manner (that, and booting the machine0 >   doesn't seem to depend on this file either).  	 Just run          @SYS$UPDATE:AUTOGEN GETDATA   ' to recover PARAMS.DAT. But I'd just do    +     @SYS$UPDATE:AUTOGEN GETDATA <end-phase>   A where <end-phase> is the last phase I want done. PARAMS.DAT is an C intermediate file. It is the output of GETDATA and is used as input B for GENFILES, GENPARAMS, and TESTFILES. So I don't see it as a big deal that you deleted it.    / >   I need the file to be in place so I can run : >   sys$system:autogen to reconfig some system parameters.9 >   Apologies if this question has an obvious answer that  >   I'm not seeing. Thanks!    C You don't need it to do that. Please read the chapter on AUTOGEN in < the System Utilities manual and the chapter "Managing System+ Parameters" in the system manager's manual.    Disclaimer: JMHO   &-) Alan E. Feldman  afeldman@gfigroup.com   	 >   Brian    ------------------------------  % Date: Mon, 30 Jul 2001 14:14:36 +0930 % From: Jeremy Begg <jeremy@vsm.com.au> . Subject: Re: Advanced Server and virtual disks* Message-ID: <3B64E634.CD30D0B1@vsm.com.au>  
 Hello all,  H My virtual disk came up after I followed Brad's suggestion of restarting2 Advanced Server after the LDA1 device was mounted.  < Thanks for the AUTOSHARE tip -- I'll try that one next time.   Regards,   	Jeremy Begg   David Jones wrote: > , > In message <3B626BB6.B7F84CBF@vsm.com.au>,+ >   Jeremy Begg <jeremy@vsm.com.au> writes: L > >We want to use this machine as the primary file server for our PCs, so itP > >seemed like a sensible thing to set up an ODS-5 device for this purpose.  NotQ > >having any spare physical drives, I decided the way to go would be to set up a C > >virtual disk on one of the shadow sets and format that as ODS-5.  > M > I agree, Pathworks serving an ODS-5 volume causes fewer headaches than when 9 > serving an ODS-2 volume.  Virtual disks work just fine.  > M > >But when I try to create the corresponding file service I get a rather odd 	 > >error:  > >  > >    SLICKA> admin5 > >    SIL\\SLICKA> add share/dir j_test d2:[s_drive] 7 > >    %PWRK-E-ERRADDSHARE, error adding share "J_TEST" D > >    -LM-E-NERR_UNKNOWNDEV, the device or directory does not exist > >  > >    SIL\\SLICKA>  > G > Did you synchronize the autoshares after you created the virtual disk M > (admin set computer/autoshare_synchronize)?  Pathworks has this infuriating J > restriction that it will only create a share on disks that have a hidden8 > administrative share on the underlying volume as well. > > > David L. Jones               |      Phone:    (614) 292-6929/ > Ohio State University        |      Internet: N > 140 W. 19th St. Rm. 231a     |               jonesd@er6s1.eng.ohio-state.edu< > Columbus, OH 43210           |               vman+@osu.edu > 3 > Disclaimer: I'm looking for marbles all day long.    ------------------------------  + Date: Sun, 29 Jul 2001 19:42:41 +0000 (UTC) / From: Sander Vesik <sander@haldjas.folklore.ee> 0 Subject: Re: Alpha: an invitation to communicate2 Message-ID: <996435766.338869@haldjas.folklore.ee>  8 In comp.arch David B. Chorlian <davidc@panix.com> wrote:Y > In <9jslh5$cq6$1@mail.pl.unisys.com> "Brig Campbell" <brig.campbell@unisys.com> writes:     B >>"Fred Kleinsorge" <kleinsorge@star.zko.dec.com> wrote in message- >>news:oMj87.274$Yx2.3846@news.cpqcorp.net... J >>> The stock market is pure black magic, legalized gambling.  It may once >>have? >>> had some rational basis, but that has long since been lost.   N >>It is not black magic, it's the foundation for capitalism.  It the mechinismK >>for which capital flows from old to new.  At times markets are irrational 1 >>but that never lasts, they are self correcting.   J > Not so; the (U.S)  government (in one form or another) steps in when theC > market cannot correct itself: The Fed's actions in October 1987,  L > and later in bailing out a failing hedge firm; other governmental actions I > with regard to foreign markets (bailout of Mexico) when this is thought @ > to be disruptive to the U.S. economy.  The last time the U.S. A > government let the market correct itself was 1929.  Capitalism  ? > requires support from non-capitalist institutions to survive.   J Strange, I thought 1929 was teh year the US govermnent discovered the hardF way that having some regulations that are enforced on what gets listedE and just what can the company founders do with the money they receive  from the shareholders...     > [remainder snipped]      > --   > David B. Chorlian  > Neurodynamics Lab  SUNY/HSCB  > chorlian@spot.cns.hscbklyn.edu > davidc@panix.com   --   	Sander    +++ Out of cheese error +++    ------------------------------    Date: 29 Jul 2001 23:01:59 -0500+ From: young_r@encompasserve.org (Rob Young) * Subject: Re: benchmarking disk performance3 Message-ID: <9HbrV7Zxfl6m@eisner.encompasserve.org>   U In article <9jtci1$abo8@imsp212.netvigator.com>, "Kenneth" <best@hotmail.com> writes: H > Is there any tools for benchmarking the disk performance in VMS 7.2-1? >   < 	Yes.  Check the freeware... it is called TESTDEV.  I dumped/ 	a small .com file out here to act as a driver.   > http://www.openvms.compaq.com/freeware/FREEWARE40/TESTDEV/SRC/   				Rob    ------------------------------    Date: 29 Jul 2001 22:59:58 -0500+ From: young_r@encompasserve.org (Rob Young) / Subject: Re: Compaq FUD and lack of information 3 Message-ID: <gn8FS7EUqCBb@eisner.encompasserve.org>   S In article <3B61FFCA.56B6C96C@dplanet.ch>, John McLean <mcleanj@dplanet.ch> writes:  > 9 > Why is Compaq incapable of making public comments about  > this ? > % > What is so difficult about it all ?  >    	Legal.     G > Why should *we* be begging the supplier for information ?  Any of the J > competition would be bending over backwards to make public reassurances.  = 	This presumes that the agreement is a simple one.  From what < 	I am guessing (no secret knowledge here) is that it isn't a 	simple one.  ? > Compaq are really showing their true colors in respect to the  > Alpha-to-Intel transition.  ? 	Hmmm...  maybe after we know more by mid-August as Fred states 7 	we revisit this and applaud you for your insight John.    				Rob    ------------------------------  % Date: Mon, 30 Jul 2001 14:21:52 +0930 % From: Jeremy Begg <jeremy@vsm.com.au> & Subject: Re: CSA  [was LPs on the Web]* Message-ID: <3B64E7E8.95EA61CA@vsm.com.au>  	 Hi David,    David B Sneddon wrote: > B > At 10:50 AM 27/07/01 +0000, Brian Schenkenberger, VAXman- wrote:F > >In article <3B60A9A0.28541.10D0D76A@localhost>, "Stanley F. Quayle" > ><stan@stanq.com> writes: 0 > > >On 26 Jul 2001, at 3:01, John Santos wrote:N > > >> Hmmm.  If you are eligible and can join CSA (~$100-$150/year), then youL > > >> can get the VMS and/or Tru64 CD distributions (LP+DOCs 4 times a yearM > > >> plus current O/S kit + any new O/S releases) for about $500/year.  Say 3 > > >> one VMS kit + 4 LP kits for about $100 each.  > > > ? > > >Actually, joining CSA is $600 per year.  Assuming that the J > > >distribution subscriptions are $500 per year for CSA members, joining > > >would make sense. > > > I > > >Unfortunately, I've received no response to my email to Compaq about E > > >the cost of the distribution subscriptions for CSA members, so I ! > > >don't know if I should join.  > > G > >John may have been wrong about the yearly fee for CSA membership but D > >the fee for the SDK is about $500/year ($495/year to be precise). > B > I am in the process of registering our company for the local CSAH > program (Australia) and the registration is $0 and the SDK cost is $0.  M Are you sure about that?  The web site and the local CSA manager (Ian Bowley) M both advise that the SDK is approx $800 per annum (we've been it for a couple  of years now).   Who told you it was free?            Jeremy Begg   =   +---------------------------------------------------------+ =   |            VSM Software Services Pty. Ltd.              | =   |                 http://www.vsm.com.au/                  | =   |       "OpenVMS Systems Management & Programming"        | =   |---------------------------------------------------------| =   | P.O.Box 402, Walkerville, |  E-Mail:  jeremy@vsm.com.au | =   | South Australia 5081      |   Phone:  +61 8 83592155    | =   |---------------------------|  Mobile:  0414 422 947      | =   |  A.C.N. 068 409 156       |     FAX:  +61 8 82231777    | =   +---------------------------------------------------------+    ------------------------------  % Date: Sun, 29 Jul 2001 23:11:07 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> - Subject: Re: Few People in DEC Understood.... , Message-ID: <3B64D04A.A9D3E082@videotron.ca>   Bill Todd wrote:L > All-in-one was, as best I can remember, regarded as an abomination when itM > first appeared:  large, kludgey, and slow.  In retrospect, it seems to have 4 > been the logical precursor to Microsoft bloatware:  N Consider that I was able to support 12 office workers on ALL-IN-1 running on aN Microvax II with 16 meg of memory back in 1987. Consider that image activation1 (essentially once per day) took about 15 seconds.   N Compare this to MS office and you wouldn't apply the term "bloat" to ALL-IN-1.N  And unfortunatly few people at Digital understood that ALL-IN-1 was an officeM application framework and not just the menus that come with the installation.   M Consider that the A1 engineers were able to take the product from an industry M leading character cell stuff in the early 1980s, to a client server with both M Teamlinks and MAPI (for MS clients) support, and later on added web , POP etc K support without having to change the core of the product, and all the while / still supporting the character cell interfaces.   L Meanwhile, the PC industry flipped flopped between a handful of trendy emailK packages, each requiring a totally new installation, training and usually a J sudden switchover from one to the other. The most interesting is Microsoft9 ending MS mail and forcng folks over to Outlook/Exchange.   K Having designed a product back in the 1980s whose architecture allowed very K large scaling, and the evolution of the product through all the trends over E the years is a testament to Digital. It is very unfortunate that even I internally Digital never understood ALL-IN-1 and itself never really took ) advantage of its tailoring capabilities.    I Had Digital put in as much effort into training its own supporrt staff on L ALL-IN-1 as it did on Exchange, perhaps Digital's use of ALL_IN-1 would haveE been far more extensive with lots of integrated applications etc etc.    ------------------------------  + Date: Sun, 29 Jul 2001 20:25:50 +0000 (UTC) / From: Sander Vesik <sander@haldjas.folklore.ee> . Subject: Re: IA64 volume and low-end dominance2 Message-ID: <996438355.577188@haldjas.folklore.ee>  9 In comp.arch Ben Franchuk <bfranchuk@jetnet.ab.ca> wrote: 
 > mulp wrote:  >>  ? >> "Jordan Henderson" <jordan@lisa.gemair.com> wrote in message ' >> news:9ibj2h$9nt$1@lisa.gemair.com... D >> >   What really drove the move from IA16 to IA32 was not customer@ >> >   demand for faster systems, but rather customer demand forE >> >   the new system software (Windows386, OS/2, Unix) that required A >> >   IA32.  This is what will drive the move from IA32 to IA64.  >>  > >> What software exactly was it that drove the demand for 386? >>  N >> Windows 3.0 only used the memory management of the 386 to provide access toM >> more memory to run 16 bit code; this was just a cheaper solution than some N >> of the earlier memory expansion kludges that had been available previously.  N > The fact that DOS was never modified to run 32 bit code and there was was noQ > FREE 32 bit assembers/compilers forced windows on everbody.The other factor was 	 > the 386 R > had 1) larger memory than 16MEG, 2) Built in floating point 3) Could run old dos  J 2) is wrong - 386 didn't have built in floating point, or why do you thinkG there was 387? 486DX was AFAIK the first intel-made x86 with an on-chip  FPU.  N PLain vanill a386 didn't even have on-chip cache, the AMD 386-40 changed that.  Q > programs unchanged. I think they could of  made the 286 to have real new 32 bit E > instruction set as well as toggle to the 8086 instruction set. Ben.  > --  @ > Standard Disclaimer : 97% speculation 2% bad grammar 1% facts.> > "Pre-historic Cpu's" http://www.jetnet.ab.ca/users/bfranchuk > Now with schematics.   --   	Sander    +++ Out of cheese error +++    ------------------------------  + Date: Sun, 29 Jul 2001 20:32:58 +0000 (UTC) / From: Sander Vesik <sander@haldjas.folklore.ee> . Subject: Re: IA64 volume and low-end dominance2 Message-ID: <996438783.418841@haldjas.folklore.ee>  : In comp.arch mulp <michaelpettengill@earthlink.net> wrote:  > > "Jordan Henderson" <jordan@lisa.gemair.com> wrote in message& > news:9ibj2h$9nt$1@lisa.gemair.com...B >>   What really drove the move from IA16 to IA32 was not customer> >>   demand for faster systems, but rather customer demand forC >>   the new system software (Windows386, OS/2, Unix) that required ? >>   IA32.  This is what will drive the move from IA32 to IA64.   = > What software exactly was it that drove the demand for 386?   D Most software that did drive it was 16 bit. It was well into the 486E times that 32 bit software became semi-important. Of course, ports of @ various OS-s became a lot easier to x86 machines by having a new; (and one might say more sensible) VM usage model available.   M > Windows 3.0 only used the memory management of the 386 to provide access to L > more memory to run 16 bit code; this was just a cheaper solution than someM > of the earlier memory expansion kludges that had been available previously.   L Right. Also, unix ports became way easier. But what really mattered was thatK esp. on the machines where there was on-motherboard cache, old software ran ) a lot (to hughesly) faster than on a 286.    --   	Sander    +++ Out of cheese error +++    ------------------------------  % Date: Sun, 29 Jul 2001 17:15:12 -0400 ' From: "Bill Todd" <billtodd@foo.mv.com> . Subject: Re: IA64 volume and low-end dominance( Message-ID: <9k1ue5$fb1$1@pyrite.mv.net>  ; "Kurt Shoens" <shoens@lenny.sfrn.dnai.com> wrote in message % news:9k1h91$8lc$1@bob.news.rcn.net... H > In article <9k1e5g$lic@web.nmti.com>, Peter da Silva <peter@abbnm.com> wrote:E > >In the 64-bit arena, it's large sparse address spaces for high-endl
 databases. >iE > Please explain why this helps high-end databases.  The direct value.I > I can think of is the ability to address a large buffer pool that stillaG > fits in physical memory.  Sparse addressing isn't necessary for that.   K And for that matter, Xeon-like 'physical address extension' mechanisms go a K long way toward handling large memory-resident buffer pools without needingnK to increase the virtual address space (i.e., it becomes largely a matter of-2 minor coding convenience rather than performance).   - bill   ------------------------------   Date: 29 Jul 2001 21:25:17 GMT& From: peter@abbnm.com (Peter da Silva). Subject: Re: IA64 volume and low-end dominance% Message-ID: <9k1uvt$sfd@web.nmti.com>   + In article <9k1h91$8lc$1@bob.news.rcn.net>, / Kurt Shoens <shoens@lenny.sfrn.dnai.com> wrote: O > In article <9k1e5g$lic@web.nmti.com>, Peter da Silva <peter@abbnm.com> wrote:oP > >In the 64-bit arena, it's large sparse address spaces for high-end databases.  E > Please explain why this helps high-end databases.  The direct value I > I can think of is the ability to address a large buffer pool that still:G > fits in physical memory.  Sparse addressing isn't necessary for that.@  J I don't know any of the details... I'm just quoting notes I took at UsenixG or LISA a year or so ago. It could be that there's an "or" relationshipeE and not an "and" between the two points, and I don't recall the talk.   I The point is that applications that really use 64 bits are high end. ThattK wasn't true for 32-bit applications... there were already people using MacsnF and Amigas and the like to run applications that took advantage of the? larger address space on the 68000: Photoshop, Lightwave, etc...r   --  +  `-_-'   In hoc signo hack, Peter da Silva.aE   'U`    "A well-rounded geek should be able to geek about anything."mL                                                        -- nicolai@esperi.org          Disclaimer: WWFD?   ------------------------------  % Date: Sun, 29 Jul 2001 17:54:18 -0400D' From: "Bill Todd" <billtodd@foo.mv.com>t. Subject: Re: IA64 volume and low-end dominance( Message-ID: <9k20nd$h97$1@pyrite.mv.net>  3 "Peter da Silva" <peter@abbnm.com> wrote in messageT news:9k1uvt$sfd@web.nmti.com...b- > In article <9k1h91$8lc$1@bob.news.rcn.net>,t1 > Kurt Shoens <shoens@lenny.sfrn.dnai.com> wrote:mJ > > In article <9k1e5g$lic@web.nmti.com>, Peter da Silva <peter@abbnm.com> wrote:G > > >In the 64-bit arena, it's large sparse address spaces for high-end 
 databases. >aG > > Please explain why this helps high-end databases.  The direct valuehK > > I can think of is the ability to address a large buffer pool that stillaI > > fits in physical memory.  Sparse addressing isn't necessary for that.  >-$ > I don't know any of the details...  H One application is to in-memory meta-data structures that may frequentlyJ need to be touched (hence could noticeably impede performance if placed inL 'extended physical memory' that needed to be explicitly mapped before use) -J e.g., the volume of lock-related meta-data can threaten to exceed a 32-bitD address space in large installations.  There are, as always, ways toH partition the problem among multiple 32-bit address spaces to avoid suchH issues, but since they add complexity to already-complex code the larger address space has some value.r   - bill   ------------------------------  % Date: Sun, 29 Jul 2001 22:28:31 -0000S- From: glc+spam@patriot.net (George Coulouris) . Subject: Re: IA64 volume and low-end dominance7 Message-ID: <slrn9m93gf.64f.glc+spam@adams.patriot.net>w  D On 29 Jul 2001 16:38:08 GMT, Peter da Silva <peter@abbnm.com> wrote: [snip] >cN >In the 64-bit arena, it's large sparse address spaces for high-end databases. >s  J I'd go so far as to generalize this to dense address spaces as well. It's J handy to use mmap() to pull in data. When said data exceeds 4G, PAE won't @ help you, and one must resort to sliding a window over the data.   ------------------------------  % Date: Sun, 29 Jul 2001 18:22:02 -0600e From: yyyc186@mindspring.com. Subject: Re: IA64 volume and low-end dominance; Message-ID: <3b64a8f6$4$lllp186$mr2ice@nntp.mindspring.com>t  C In <fdecfe0a.0107171156.1eef297@posting.google.com>, on 07/17/2001 e6    at 12:56 PM, rudy@edpstaff.com (Rudy de Haas) said:  I >Will the Itantic's lack of backwards compatibility have the same effect? J >Personally I think it will, but it is clear that Intel and Microsoft haveJ >something different in mind. It's no coincidence - at least I don't thinkE >it is- that the alphacide announcement came on the same day that the  >appeals courtG >set aside the penalities Judge Jackson had assessed against Microsoft.     I If it won't run _any_ 16 bit instructions, then Windows 2000 won't run on G it since its kernel (as all revs of windows) still relies on 16 bit DOSn and BIOS instructions.   Roland --  ; -----------------------------------------------------------B yyyc186@mindspring.com; -----------------------------------------------------------m   ------------------------------  % Date: Sun, 29 Jul 2001 19:17:18 -0400o' From: "Bill Todd" <billtodd@foo.mv.com>>. Subject: Re: IA64 volume and low-end dominance( Message-ID: <9k25j2$l65$1@pyrite.mv.net>  : "George Coulouris" <glc+spam@patriot.net> wrote in message1 news:slrn9m93gf.64f.glc+spam@adams.patriot.net...iF > On 29 Jul 2001 16:38:08 GMT, Peter da Silva <peter@abbnm.com> wrote: > [snip] > >aE > >In the 64-bit arena, it's large sparse address spaces for high-endh
 databases. > >h >aK > I'd go so far as to generalize this to dense address spaces as well. It'se& > handy to use mmap() to pull in data.  K Handy, possibly.  Inefficient, usually, since mmap doesn't know the size ofaL the chunk of data you need and unless the system's page-fetch algorithm justK happens to correspond with that size you'll either perform multiple fetches.L from disk for a single chunk or fetch more than you need.  And when it comesJ time to write the data, you almost always need fine control over when, and how much, hits the disk.  H Such use of mmap is a Unixism, and Unix has typically lagged a decade orH three behind when it comes to high-performance disk access (extent-basedH file systems, which date at least back to the early 1970s and likely farE earlier, are still the exception rather than the rule).  I doubt thataG high-end databases can throw away performance to that degree and remainaJ competitive.  Low-end databases may be able to, but then low-end databases2 far less frequently need the larger address space.   - bill   ------------------------------  % Date: Mon, 30 Jul 2001 02:49:58 -0000p- From: glc+spam@patriot.net (George Coulouris)c. Subject: Re: IA64 volume and low-end dominance7 Message-ID: <slrn9m9iqm.fe1.glc+spam@adams.patriot.net>   J On Sun, 29 Jul 2001 19:17:18 -0400, Bill Todd <billtodd@foo.mv.com> wrote: > ; >"George Coulouris" <glc+spam@patriot.net> wrote in message 2 >news:slrn9m93gf.64f.glc+spam@adams.patriot.net...G >> On 29 Jul 2001 16:38:08 GMT, Peter da Silva <peter@abbnm.com> wrote: 	 >> [snip]- >> >F >> >In the 64-bit arena, it's large sparse address spaces for high-end >databases.  >> > >>L >> I'd go so far as to generalize this to dense address spaces as well. It's' >> handy to use mmap() to pull in data.e >BL >Handy, possibly.  Inefficient, usually, since mmap doesn't know the size ofM >the chunk of data you need and unless the system's page-fetch algorithm justiL >happens to correspond with that size you'll either perform multiple fetchesM >from disk for a single chunk or fetch more than you need.  And when it comesoK >time to write the data, you almost always need fine control over when, andn >how much, hits the disk.a  J I agree that it's probably inefficient in the general case. The particularJ app I'm working on, however, has the nice property that pages are accessedG sequentially, are touched only once, and are read-only, so it works outt okay.d  I >Such use of mmap is a Unixism, and Unix has typically lagged a decade or I >three behind when it comes to high-performance disk access (extent-based I >file systems, which date at least back to the early 1970s and likely far F >earlier, are still the exception rather than the rule).  I doubt thatH >high-end databases can throw away performance to that degree and remainK >competitive.  Low-end databases may be able to, but then low-end databases.3 >far less frequently need the larger address space.l  E I use MapViewOfFile() as a drop-in replacement for mmap() under w2k.   Does pretty well.-   ------------------------------  % Date: Sun, 29 Jul 2001 23:33:46 -0400(- From: JF Mezei <jfmezei.spamnot@videotron.ca> . Subject: Re: IA64 volume and low-end dominance, Message-ID: <3B64D597.A6FA3182@videotron.ca>   Peter da Silva wrote: K > The point is that applications that really use 64 bits are high end. ThattM > wasn't true for 32-bit applications... there were already people using MacseH > and Amigas and the like to run applications that took advantage of theA > larger address space on the 68000: Photoshop, Lightwave, etc...   G What is considered high end today becomes common in a few years. FlightaL simulators used to be considered high end, require very fancy VAXes and lotsN of anciliary processors (graphics etc), but then came MS Flight simulator thatG ran on a wintel cardboard box. (granted nowhere near as sophisticated).   H Consider Nintendo games, and the X-BOX. It is my understanding that theyM render "live" the scenes as you move around with an ever increasing number ofpK polygons.  This stuff used to be relegated to the very high end, and now ityN sits under a TV set with the kids/teenagers spending all day in front of this.J (Some of these game consoles do claim to be 64 bits, but I am not sure hwo that claim applies).  N The 32 bit limit may not have YET been reached for MOST PC applications, but II think that for some, it has been reached or is very near, and ironically,-J games will be the first ones to need the extra performance. Digital shouldM have pushed Alpha down the throats of Nintendo and Sega because those are the / ones who would have made use of its power ASAP.2   ------------------------------  % Date: Sun, 29 Jul 2001 23:17:50 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>-C Subject: Re: IPF Console Bootstrap (was: Re: No chance for OpenVMS)u, Message-ID: <3B64D1DC.96F6875F@videotron.ca>   Rich Jordan wrote:N > thing on the bus barfs in a system-stopping fashion.  A plain serial consoleM > is that wonderful lowest common denominator, reliable, well known, comfort- I > inducing usable as a last ditch fallback to re-establishing control ands > operation of the system.  J Compaq considers serial ports "legacy". And their use of the word "legacy"K indicates to me that they really think that serial ports are very very veryx> bad. (legacy has such a negative connotation in the industry).  L The whole console thing brings up interesting culture clash at Compaq. It isK the PC folks who will now be called to build "true" enterprise systems, notlJ the glorified game controllers with Proliant label on it. So their bias isG towards what they see as industry standard, whereas the true entreprises! requirements are quite different.e  M It will be most interesting to see how the Tandem folks manage to get control  over their hardware.  I And it will also be interesting to see if , once the PC guys learn how toeK build true enterprise boxes, if they go back to Microsoft and pass on theirnI experience so that MS can tailor NT to work as an enterprise machine (eg:h serial console etc etc).   ------------------------------  # Date: Mon, 30 Jul 2001 04:20:22 GMTt) From: rob.buxton@wcc.govt.nz (Rob Buxton)l) Subject: login-f-clisymtbl error on Alphae1 Message-ID: <3b64df41.284326289@news.wcc.govt.nz>l   Hi All,   7 I vaguely seem to recall a thread on this a while back.c  ; Saw the error LOGIN-F-CLISYMTBL on one of our Alpha Users. l  D Running VMS 7.2-1 fairly up to date on Patches. (more scheduled this	 weekend!)t  F Help / Message (from an Alpha) indicates on VAX that you need to check9 VIRTUALPAGCNT which is all very well except I'm on Alpha!d  2  What are the areas that might need to be checked?  ! As usual, many thanks in advance.    Rob.   ------------------------------  % Date: Sun, 29 Jul 2001 18:06:57 -0600f From: yyyc186@mindspring.com" Subject: Re: No chance for OpenVMS; Message-ID: <3b64a52f$1$lllp186$mr2ice@nntp.mindspring.com>   9 In <29Y47.833$rc5.60629@news.cpqcorp.net>, on 07/17/2001 VE    at 10:42 AM, "Fred Kleinsorge" <kleinsorge@star.zko.dec.com> said:d    > >JF Mezei wrote in message <3B53E787.52AD90A0@videotron.ca>...  L >>Actions speak louder than words. Compaq may have bribed its key customers, >butD >>the effect will be temporary and I think that those customers will >eventuallysF >>come to the same complusion as Gartner unless Compaq takes some real >actions >>to push VMS.    * >Uh, I don't think anyone has been bribed.  ) Snowed with fraud would be more accurate.w       --  ; -----------------------------------------------------------h yyyc186@mindspring.com; -----------------------------------------------------------n   ------------------------------  % Date: Sun, 29 Jul 2001 22:30:49 +0100 + From: "antonio.carlini" <arcarlini@iee.org>6& Subject: Re: Ont "The Inquirer" today.' Message-ID: <3B648089.43EC6229@iee.org>e  & paddy.o'brien@zzz.tg.nsw.gov.au wrote:O > Is this really how this works?  An ex-colleague of mine was in correspondence N > with Barbara Burri at EDS regarding enhancements (not just maintenance) someM > 2/3 years ago.  I missed out on a beta version because I have been away fortK > about 10 months over the last 18.  I would have guessed that the beta wastP > going to include new features, and I would have guessed that their money comes4 > from software support contracts and new purchases. > L > I would put this in the same league as when Hoff has asked for suggestionsM > here for customers' wishes in the next version of (Open)VMS.  The engineers P > would consider all requests but surely no individual or their company was ever > directly billed.  ' As far as COMPAQ and EDS goes - I neverp% saw any paperwork. But maintenance isn' easily measured and metrics can be set.n& New development is typically something% that has to be handled via a proposal  to do the work for a set price.U  + Remember that EDS do not own these productsr. and (AFAIK) cannot do anything with them other% than maintenance and development thata( COMPAQ pay for. EDS (again AFAIK) do not& generally have support contracts with ( customers and do not receive any service revenue that way.a  ' Things might be different if you wanted + (for example) VAX DOCUMENT to acquire some I* specific feature: that product was (AFAIK)' sold, so the new owner can do what they * want with it and can choose to  charge you( for the work or spread the cost over the whole customer base.  J > As a reader of the plethora of threads on this porting subject, I remainK > skeptical that the port will ever happen, or if it does it will be a bareo3 > bones operating system with limited applications.r >   . I'd be surprised if the port is not attempted,$ otherwise there are a lot of people # feeding false information! The portM( could always be pulled later on I guess,% but that would almost certainly cause3' an immediate tremendous loss of currento service revenue.  s Antonio    -- e   ---------------e- Antonio Carlini             arcarlini@iee.orgt   ------------------------------  # Date: Sun, 29 Jul 2001 21:30:01 GMTs= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)e0 Subject: Re: Passing a socket to another process0 Message-ID: <009FFBDC.8FB8E9C2@SendSpamHere.ORG>  f In article <3B644DE0.F68824B1@vista-control.com>, Eugene A Zharkov <zharkov@vista-control.com> writes: >Larry Kilgallen wrote:B >> oi >> In article <3B632497.EACDA07B@vista-control.com>, Eugene A Zharkov <zharkov@vista-control.com> writes:e? >> > I have found in comp.os.vms an example of passing a sockete9 >> > to another process that is based on the sys$creprc'se? >> > prc$m_netwrk option. But there is still a couple of thingst) >> > that I am not sure how to deal with.s >> >A >> > 1. The prc$m_netwrk is not documented very well. Maybe there C >> > is a better (documented) way of doing this (on VMS 7.1/UCX 5.0  >> > and higher).a >> eA >> Generally (and I believe specifically in this case) techniquessB >> that are not documented are not supported.  They may break withA >> VMS 7.next.  Posts in comp.os.vms are often made by people whogA >> are not particularly concerned with something being supported.h> >> But certainly VMS Development does not document unsupported >> methods.c >eA >Too bad. (Not that VMS Development does not document unsupported-? >methods, but that VMS Development does not provide a supported ? >method. Even Winsocks seem to have a supported method of doing  >this!).  > Why do you want to create another process and pass the socket?  J Many things you might do in one OS are not always the best ways to do themL in VMS.  When Weendoze can do ASTs, come back here and we will discuss this 
 at length.   --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMe            dJ   "And of course, I'm a genius, so people are naturally drawn to my fiery I   intellect.  Their admiration overwhelms their envy!" -- Calvin & Hobbes    ------------------------------    Date: 29 Jul 2001 13:27:49 -0700 From: stelth@iname.com (sunil)Y Subject: Re: problem deallocating a CD rom device when reflection used to allocate it is n= Message-ID: <11331f03.0107291227.613bfa10@posting.google.com>     Thanks a lot for your responses.  7 I tried the STOP /id=(id reported by show device /full)a but the response i got was e   $ stop /id=2D40166A & %SYSTEM-W-NONEXPR, nonexistent process  @ I tried the approach suggested by andrew, but from a non priv'ed account     $ dismount/nounload/abort dka500< %SYSTEM-W-DEVALLOC, device already allocated to another user  > I tried this from the same account with which i had originallyF allocated the drive. But looks like VMS allocates devices to processes and not users.  E i guess this might work from a priv'ed account, but is there a way toD@ retrieve the drive as the user who originally allocated it ( non	 priv'ed).e  A The shell in which the allocation was done has terminated, so thei: process to which the allocation was done no longer exists.     Thanks.b    [ Andrew Robert <arobert@townisp.com> wrote in message news:<3B641385.7000409@townisp.com>...e3 > From a priv'ed account, do the following command:a > & > $ dismount/nounload/abort   {device} >  >  > Jerry Leslie wrote:s > # > > sunil (stelth@iname.com) wrote:.3 > > : I use reflections to connect to a VMS system.b > > :TQ > > : when i accidentally closed a reflection from which i allocated and mounted tL > > : a CD rom drive with out deallocating it. I could not allocate it from  > > : a different reflection.yN > > : I got a message to the effect of device already allocated to a different > > : process. > > : K > > : Is there a way for me ( with out admin priviledges) to get access to  G > > : the CD drive. (other than requesting a reboot, which is the only  & > > : suggestion i could get till now) > > :e) > > : any help will be  much appreciated.e > > M > > Issue the " SHOW DEVICE/FULL DKxxx" command, where "DKxxx" is the device 3# > > name of the CD-ROM drive; e.g.:s > >   > >    $ show device/full dka400 > > R > >    Disk SCCVXG$DKA400:, device type RRD42, is online, allocated, file-oriented7 > >        device, shareable, error logging is enabled.S > > Q > >        Error count               0    Operations completed                  0mQ > >        Owner process      "LESLIE"    Owner UIC                      [SYSTEM]1Q > >    >>  Owner process ID   0000DA7D    Dev Prot            S:RWPL,O:RWPL,G:R,WeQ > >        Reference count           1    Default buffer size                 512s > >  > > N > > Then issue the "STOP/ID=xxxxxxx" command to stop the process whose "owner N > > process ID" appears in the output of the "SHOW DEVICE/FULL DKxxx" command; > > e.g. > >  > >    $ STOP/ID=0000DA7D- > >  > > --Jerry Leslie   ------------------------------  % Date: Sun, 29 Jul 2001 20:00:48 -0400R  From: John Santos <JOHN@egh.com>Y Subject: Re: problem deallocating a CD rom device when reflection used to allocate it is .6 Message-ID: <1010729194250.36974A-100000@Ives.egh.com>   On 29 Jul 2001, sunil wrote:  " > Thanks a lot for your responses. > 9 > I tried the STOP /id=(id reported by show device /full)l > but the response i got was e >  > $ stop /id=2D40166A ( > %SYSTEM-W-NONEXPR, nonexistent process > B > I tried the approach suggested by andrew, but from a non priv'ed	 > accounte > " > $ dismount/nounload/abort dka500> > %SYSTEM-W-DEVALLOC, device already allocated to another user > @ > I tried this from the same account with which i had originallyH > allocated the drive. But looks like VMS allocates devices to processes > and not users. > G > i guess this might work from a priv'ed account, but is there a way to-B > retrieve the drive as the user who originally allocated it ( non > priv'ed).-  B You could argue that since you are logged into the same account asE the process that originally allocated the CD, you should have control.C access to the drive, but this might have security implications that B prevent this change from being made.  It wouldn't hurt to ask yourB support contact (assuming you have a support contract) for this as a security enhancement.1  C > The shell in which the allocation was done has terminated, so the < > process to which the allocation was done no longer exists.  E That is not actually true, or the CD would be available.  The processnG is actually in some kind of process run-down state, where it is waiting E for something (possibly I/O completion of some sort, maybe to the CD,iE to complete.)  If you have physical access to the CD drive, you couldlB try removing the CD from the drive.  This might put the drive intoG mount-verify state, which will eventually timeout (default is one hour,oE determined by SYSGEN parameter MVTIMEOUT), at which time you might bee able to access the drive.a  E A better solution would be to have your system mangler enable virtualaB terminals and configure whatever transport method you are using toC connect to the system (probably TCP/IP's telnet, since you said yous= are using Reflections) to use virtual terminals.  Then if youe@ close the window without logging out first, you can reconnect to? the session and not lose anything.  You have several minutes tot@ reconnect before the process is deleted.  (Time is determined by@ TTY_TIMEOUT, default value 900 seconds.)  You can reconnect fromE the same PC or from anywhere else that has enabled virtual terminals,pF including a directly-connected terminal, LAT terminal, or telnet from  another PC.d  A You can even reconnect from a non-virtual terminal by using a VMS ? telnet client to "telnet 0" to connect to the same host.  AftercA logging in, it will let you reconnect to any disconnected virtual * terminals logged into the current account.  > The exact commands to enable virtual terminals depends on your@ TCP/IP stack, and maybe the version.  (It changed between UCX V4< and TCP/IP V5.)  It won't help you in your current situationA because the process that owns the CD already exists, but a prived < user can load the virtual terminal driver and enable virtualC terminals for telnet without rebooting.  Depending on the IP stack,wC it might require stopping and restarting the telnet server, though.e   >  > 	 > Thanks.i >  > ] > Andrew Robert <arobert@townisp.com> wrote in message news:<3B641385.7000409@townisp.com>... 5 > > From a priv'ed account, do the following command:  > > ( > > $ dismount/nounload/abort   {device} > >  > >  > > Jerry Leslie wrote:c > > % > > > sunil (stelth@iname.com) wrote:b5 > > > : I use reflections to connect to a VMS system.  > > > :nS > > > : when i accidentally closed a reflection from which i allocated and mounted eN > > > : a CD rom drive with out deallocating it. I could not allocate it from  > > > : a different reflection.yP > > > : I got a message to the effect of device already allocated to a different > > > : process. > > > :mM > > > : Is there a way for me ( with out admin priviledges) to get access to tI > > > : the CD drive. (other than requesting a reboot, which is the only  ( > > > : suggestion i could get till now) > > > :n+ > > > : any help will be  much appreciated.e > > > O > > > Issue the " SHOW DEVICE/FULL DKxxx" command, where "DKxxx" is the device  % > > > name of the CD-ROM drive; e.g.:a > > > " > > >    $ show device/full dka400 > > > T > > >    Disk SCCVXG$DKA400:, device type RRD42, is online, allocated, file-oriented9 > > >        device, shareable, error logging is enabled.u > > > S > > >        Error count               0    Operations completed                  0eS > > >        Owner process      "LESLIE"    Owner UIC                      [SYSTEM]aS > > >    >>  Owner process ID   0000DA7D    Dev Prot            S:RWPL,O:RWPL,G:R,WgS > > >        Reference count           1    Default buffer size                 512e > > >  > > > P > > > Then issue the "STOP/ID=xxxxxxx" command to stop the process whose "owner P > > > process ID" appears in the output of the "SHOW DEVICE/FULL DKxxx" command;
 > > > e.g. > > >  > > >    $ STOP/ID=0000DA7Dh > > >  > > > --Jerry Leslie >  >    -- h John Santosw Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------    Date: 29 Jul 2001 18:08:06 -00004 From: Doc.Cypher <Use-Author-Address-Header@[127.1]>- Subject: Re: Selling VMS to another company ?y5 Message-ID: <20010729180806.8857.qmail@nym.alias.net>   " -----BEGIN PGP SIGNED MESSAGE-----  6 On Sat, 28 Jul 2001, jlsue <jlsuexxxz@home.com> wrote:  2 >And being successful financially is what we need. >tE >We'll all just have to wait and see if this is the path to financialv	 >success.r  J A path to financial success is - to many in c.o.v. - a simple prerequisiteJ for the continued existence of the VMS operating system. [With Compaq that is].  K Since the deed has been done and the Alpha platform must now serve its timeaB on death row, perhaps we should return to berating the Q for theirD marketing incompetence. The handling of the Alpha decision is simply9 another example of their ability to fire off footbullets.o     Doc. - -- R6 The bigger the humbug, the better people will like it.K ~ Phineas Taylor Barnum.                              http://vmsbox.cjb.netn   -----BEGIN PGP SIGNATURE-----o Version: 2.6.2  @ iQEVAwUBO2ND8sriC3SGiziTAQGLbAf5AXhllmJWCcIyMoLSItuWEI6iMBkD1WZ5@ TNqPDQKcJZQoG5EGfcpyl4kfVW9XFlQ2PNwhGBLN9G5q33L2BgJWaxmR8fsNxL7c@ NeBADHPTaULX4o6lFypv8TRWzgZdTUrRfWvvwaPZx2InU99evSpUg+rS5SPwuoY2@ ugMH/JOOVEv6om4FlEmqwTlvoeuICoz7F/Td6ZNNs53jnIyDXfcxC8MQEV7Ytz/I@ UPNpn9hk/zKcpmnrmgMeZPu89HU7op+/aa4bnTMEBiJiBwphe2tyAo0Vjtpu0KwI8 Zwo6uW1tI1p8+hLobRcv/FDJKkwSyTPkFRF4MQE5+s73lSX85LxvrQ== =Mykzc -----END PGP SIGNATURE-----    ------------------------------  % Date: Mon, 30 Jul 2001 12:52:56 +1000 6 From: "Arena, Steve" <Steve.Arena@pacificpower.com.au>< Subject: RE: TCPIP v5.1 startup on dial-up service provider.Q Message-ID: <D750FFBD4936D111842000805F15EFA4043A879E@meppb1.pacificpower.com.au>   K I'm wanting to use an Alpha 255 VMS V7.1 ,UCX V4 or V5, as a firewall for amG dial-up connection to an ISP. Any suggestions as to where to start? Fort< example, how do you configure the modem and the serial port?   Thanks Stevel   > -----Original Message-----I > From:	Simon Clubley [SMTP:simon_clubley@remove_me.excite.com-Earth.UFP]h" > Sent:	Friday, 27 July 2001 22:25 > To:	Info-VAX@Mvb.Saic.Coml> > Subject:	Re: TCPIP v5.1 startup on dial-up service provider. > 0 > On Fri, 27 Jul 2001 07:44:51 +0100, in article@ > <3B611BF3.37F654C3@ipl.demon.co.nospam.uk>, Steve Reece wrote: > >-D > >I'm now a subscriber to a dial-up ISP but there is (at least) one > >problem : > >aE > >As part of my configuration of TCPIP Services for OpenVMS v5.1 (onuC > >Alpha) I have smtp configured to use my ISP's smtp server as itsmJ > >alternate gateway.  When TCP/IP starts it expects to be able to see the: > >server (as it did with v4.2 as well) so I have to do anI > >@SYS$STARUTP:TCPIP$STARTUP, wait for it to pause and then do my dialuph+ > >in order for TCP/IP Services to startup. K > >Any ideas what I might do to ease the situation and have TCP/IP Services0& > >start as part of my system startup? > >RG > >Environment is a DEC 3000 model 600 workstation running OpenVMS v7.3dH > >with TCP/IP Services for OpenVMS Version 5.1.  Dialup connection usesB > >PPP and the number is stored in the configuration on the modem. > >  > >Thanks in advance.m	 > >Steve.3 > J > [Based on your address, I'm assuming that the ISP in question is Demon.] > G > My production machine is VMS Alpha 7.1, UCX 4.2, ECO 4. I used to usei
 > PPP/SLIPG > dialup, but now use a ISDN LAN based router. I also have my alternateu	 > gatewayqK > set to post.demon.co.uk, but I have never experienced the problem you are 1 > describing, either now or when I used SLIP/PPP.u > L > I suspect that it is because I have a different routing setup. How are youI > defining your route to Demon ? Are you using a default gateway, or just-I > enabling specific routes ? Are your routes defined all the time or just4 > when > you need to dialup ? > J > In my setup, routes are defined at the time of dialup and removed by the > batchCK > job that initiated the dialup when the job has finished collecting E-Mailw > orJ > FTPing files. I do not use a default gateway, but only define routing toI > the required subnets. If I want to web browse from this machine, I have. > LynxJ > setup to use Demon's proxy server, so I only need to set a subnet to the > proxy server.I > H > I think your problem may be that you have a route to Demon permanently > setup,J > so startup hangs. If you can define the route only when you need to, the > problem may go away. >  > Simon. > J > PS: Be aware that the only time I have ever crashed a VMS box was when IJ > tried to get PPP working. I think that there are patches for the various > problems that I uncovered. >  > -- i= > Simon Clubley, simon_clubley@remove_me.excite.com-Earth.UFPEK > In the task of removing Microsoft from the marketplace, I have discoveredn > arG > truly remarkable plan, but this signature is too small to contain it.h   ------------------------------  % Date: Mon, 30 Jul 2001 13:40:35 +0930D/ From: Mark Daniel <Mark.Daniel@wasd.vsm.com.au>d& Subject: Re: V7.3 XFCACHE-W-DATALOSS ?/ Message-ID: <3B64DE3B.E62691B2@wasd.vsm.com.au>a  ( Thanks Ian, only just noticed your post.  E When there was little c.o.v response to the original query I logged ay@ call with CSC.  "The" TIMA has been applied with no effect on myH particular problem.  Over this last weekend I found out how to reproduce? it at will and sent (via my CSC contact) a snippet of code thatsB demonstrates this.  I will post some sort of summary when it's all	 resolved..   IanPercival wrote: > G > You need to install the latest XFC TIMA kit which should address thiseM > problem.  Please contact your support organization - or obtain the kit fromtN > the web.   There are/will be two XFC TIMA kits,  the first is for standaloneM > machines,  the second is a few weeks away from release for clustered nodes.M > 	 > Regardsb >  > Ian Percival+ > (OpenVMS Engineering,  writing from home)3 > K > "> > Recently (just) installed V7.3 on DS20 (1024MB) has started spittinge > > > OPA0 messages reportinga > > > L > > >   %XFCCACHE-W-DATALOSS, read I/O completetion with <less bytes that it > > > thinks it should>  > > >yN > > > We have been COPYing a *lot* of small files around (documentation CDs to > > > hard-disk).o > > >eM > > > This is not in the HELP message database (yet).  Anyone else seen this?@ > > >o > > > As always, TIA.a   -- h Illegitimis nil carborundum.   ------------------------------   End of INFO-VAX 2001.419 ************************ugene A Zharkov <zharkov@vista-control.com> writes:e? >> > I have found in c