1 INFO-VAX	Thu, 05 May 2005	Volume 2005 : Issue 250       Contents: A C programming question Re: A C programming question Re: Appletalk on Alphas  Re: Appletalk on Alphas  Re: Appletalk on Alphas % Cluster IP, LAN failover and Failsafe ) Re: Cluster IP, LAN failover and Failsafe = Re: Eliminating files with extraneous characters in filenames - Re: Finally, some good news (at least for me) - Re: Finally, some good news (at least for me) % Re: finding the (biggest) bottle neck % Re: finding the (biggest) bottle neck & Re: how, exactly, is SYS$NODE defined?& Re: how, exactly, is SYS$NODE defined?/ RE: HP hostility to alpha holding back itanium? / Re: HP hostility to alpha holding back itanium? / Re: HP hostility to alpha holding back itanium? / Re: HP hostility to alpha holding back itanium? 8 Re: Maybe HP should get out of the hardware business....8 Re: Maybe HP should get out of the hardware business.... Message definition files Re: Message definition files Re: Message definition files Re: Message definition files Re: More VAX help needed Re: More VAX help needed Re: OpenVMS using RAID drives  Re: OpenVMS using RAID drives  Re: OpenVMS using RAID drives  Re: OpenVMS using RAID drives  ordering OpenVMS 8.2 Re: ordering OpenVMS 8.2 Re: ordering OpenVMS 8.2K Re: Time to vote on your favorite slogan for the next OpenVMS License plate   F ----------------------------------------------------------------------  % Date: Thu, 05 May 2005 10:13:41 -0600 0 From: Mark Berryman <mark.berryman@mvb.saic.com>! Subject: A C programming question % Message-ID: <4279e3cb@cpns1.saic.com>   ? A program written for the GNU C environment does the following:    typedef short DCTELEM;   and then, inside a structure,   2 DCTELEM block[6][64] __attribute__((aligned (8)));  = This causes the short array to aligned on an 8-byte boundary.   I DECC does appear to have any mechanism for specifying alignment within a  I structure.  For the structure itself, and standalone variables, yes, but  1 not within a structure (that I can find, anyway).   A The original authors of the package will not accept anythink like   
 #ifdef __DECC  long long dummy; #endif  I being placed in front of such declarations everytime they are used (they  C are used extensively) although it is certainly a way to force such  
 alignment.  B The alignments are necessary.  Does anyone know of a way to align I variables within a structure, at an alignment greater than their natural  C alignment, using the VMS C compiler?  If it helps, this particular  I package usually places the alignment in a macro similar to the following:   - #define __align8 __attribute__((aligned (8)))   
 Mark Berryman    ------------------------------  # Date: Thu, 05 May 2005 17:26:45 GMT 5 From: "Ed Vogel" <edward.vogel_stop_the_spam.@hp.com> % Subject: Re: A C programming question 0 Message-ID: <pHsee.4953$%I.278@news.cpqcorp.net>  > "Mark Berryman" <mark.berryman@mvb.saic.com> wrote in message  news:4279e3cb@cpns1.saic.com... M >     Does anyone know of a way to align variables within a structure, at an  K > alignment greater than their natural alignment, using the VMS C compiler?   	     Mark,   %     (I think the following will work)   M     As you pointed out, there is a way to specify the alignment of a struct,   soL     the solution would be to wrap the field in a variant_struct (so the nameJ     does not get propagated)  SO...your field declaration would look like:  %         #pragma member_alignment save +         #pragma nomember_alignment quadword          variant_struct {!             DCTELEM block[6][64];          } dummy_name; (         #pragma member_alignment restore       See if this works....        Ed Vogel#     HP/Compaq/DEC C/C++ Engineering    ------------------------------   Date: 5 May 2005 08:05:16 -0700 $ From: "Ed Wilts" <ewilts@ewilts.org>  Subject: Re: Appletalk on AlphasB Message-ID: <1115305516.704342.83710@f14g2000cwb.googlegroups.com>  B Is it just me or does the Google web interface to comp.os.vms just suck?   E >Get an old VAX, put V7.3 on it and put it in your cluster.  Then you C >can upgrade the Alphas to your heart's content because you'll have F >AppleTalk running on the VAX.  That's a lot cheaper than $337,000 for >new printers.  E Theory is great, but in practise this isn't a good alternative.  I've F got a disaster tolerant cluster located in 2 data centers.  That meansD that I have to get at least 2 systems just for Appletalk.  Each willE need a bunch of network and SAN connectivity.  The price adds up fast G even if I start with low-end Alpha like a DS10 (probably cheaper than a  Vax anyway).  A When I'm doing putting an old system into my cluster, I get an OS G environment that likely isn't supported for full production.  I haven't = checked the SPD, but I'd guess that a mixed-version 7.3-2/8.2 G environment is migration-only support.  When 8.3 or 8.4 comes out, will C this be supported?  I doubt it.  How long before DCPS says that the  minimum OS release is 8.2?  C Again, their are a bunch of kludgy alternatives, but going forward, D none of the options are great.  I was disappointed when DEC sold theB DECps product set to CA, and now I'm disappointed to see that theyE killed off another good product by selling it to somebody who doesn't  know what to do with it.  E >> I've been a VMS admin since the blue wall (V2.4 over 20 years ago) > >> but unfortunately this is the end of the line for VMS here.  A >Just because there's no more AppleTalk on the current version of 	 >OpenVMS?   F VMS was declining here as it was and this will accelerate it.  At bestD for HP, I'll likely cancel all my software contracts when they're upG for renewal at year-end since I'll never be upgrading and trying to get E various vendors to fix bugs in older releases doesn't work very well.   B What sucks the most is that the applications are being migrated toG Solaris, Windows, and MacOS.  One year of my software support contracts > will exceed the profit you'll make on all the Proliant servers	 combined.   G I appreciate all the comments and suggested workarounds from everybody.   	    .../Ed    ------------------------------  # Date: Thu, 05 May 2005 16:22:21 GMT # From: hoff@hp.nospam (Hoff Hoffman)   Subject: Re: Appletalk on Alphas1 Message-ID: <1Lree.4946$gG.3224@news.cpqcorp.net>   U In article <d585ao$ome$1@naig.caltech.edu>, David Mathog <mathog@caltech.edu> writes:  :Hoff Hoffman wrote:f :> In article <jUZyyreGECjE@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: : I :>   To recap something I researched for one of the previous discussions  L :>   of this particular topic and this particular product, the code involvedK :>   is encumbered by at least one third-party vendor, and I am accordingly G :>   unable to add the package onto the OpenVMS Freeware.  I don't have > :>   clearance to provide it via the Freeware, in other words. : I :Would they have to shoot you if you divulged which part of the code and   :which vendor?    L   I'd have to dig back into this to determine which group(s) had contractualM   and/or royalty and/or redistribution agreements for the PATHWORKS Macintosh K   package.  For obvious reasons, I'm not particularly included to dig into  K   it (a half-decade on) as this was for release of something that I recall  H   receiving a clear response of "no way" from the then-current legal andK   contract folks; this from when I last asked about its potential inclusion    on the Freeware.    3   You might check with the folks at SRI, of course.   I   Try a Google for "pathworks macintosh sri" or such, and you should turn -   up a few references to David Cotton at SRI.   2   I do not know the status of this product at SRI.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------   Date: 5 May 2005 12:15:12 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)   Subject: Re: Appletalk on Alphas3 Message-ID: <INf27Aqx8tQB@eisner.encompasserve.org>   d In article <4278EDE4.20533.2ACBD64D@localhost>, "Stanley F. Quayle" <squayle@insight.rr.com> writes: > C > I think a PC running CHARON-VAX would make a great print server.  2 > [Shameless Plug (tm) from a CHARON-VAX reseller]  -    How does that fix the OP's AppleTalk need?    ------------------------------  $ Date: Thu, 5 May 2005 11:54:41 -0400) From: "Scott Greig" <jsgreig@geminaq.com> . Subject: Cluster IP, LAN failover and Failsafe9 Message-ID: <9iree.6217$VL3.622880@news20.bellglobal.com>   
 Hello all:  - I have an ES47 model 4, dual instance galaxy. $ Each instance has a pair of DE600's.. LAN failover on both instances, so that device LLA0 (IP LE0) exists.  Node A is 10.20.200.4  Node B is 10.20.200.5 3 Because this is a cluster, my (external) DNS server ' has a dual-homed cluster alias name as: " node C is 10.20.200.4,10.20.200.5.  ) Using ifconfig fo provide failsafe IP as:   $ on A: ifconfig le0 alias 10.20.200.5$ on B: ifconfig le0 alias 10.20.200.4  2 Works great, but the ifconfig stuff is lost when a2 node is rebooted.  How do I get the ifconfig stuff to survive across boots?   Scott    ------------------------------  % Date: Thu, 05 May 2005 13:02:23 -0400 6 From: Brad Hamilton <braMdhamAiltPonS@comMcasAt.nPeSt>2 Subject: Re: Cluster IP, LAN failover and Failsafe0 Message-ID: <xMGdnS5Y-ZkCzOffRVn-iA@comcast.com>   Scott Greig wrote: > Hello all: <snip>+ > Using ifconfig fo provide failsafe IP as:  > & > on A: ifconfig le0 alias 10.20.200.5& > on B: ifconfig le0 alias 10.20.200.4 > 4 > Works great, but the ifconfig stuff is lost when a4 > node is rebooted.  How do I get the ifconfig stuff > to survive across boots? >   I It's been awhile since I used TCP/IP Services, (aka UCX) but I think you  F need to put the ifconfig "stuff" into SYS$STARTUP:TCPIP$SYSTARTUP.COM 3 (which is called by SYS$STARTUP:TCPIP$STARTUP.COM).    ------------------------------   Date: 5 May 2005 12:18:33 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) F Subject: Re: Eliminating files with extraneous characters in filenames3 Message-ID: <qAixwaAmjBYB@eisner.encompasserve.org>   c In article <1115243226.298899.136140@g14g2000cwa.googlegroups.com>, JimStrehlow@data911.com writes: @ > We get "strange file names" on NFS (network file system) filesH > sometimes when you use mIxEd CaSe in the filename. I notice additional. > characters every time the case is different.H > Building the file using all lowercase or all UPPERCASE has reduced our > particular application woes. > G    If you mean $, your using Multinet's support for ODS-2, or something     like it.    ------------------------------  $ Date: Thu, 5 May 2005 10:37:36 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> 6 Subject: Re: Finally, some good news (at least for me)9 Message-ID: <Wcqee.14300$3U.948511@news20.bellglobal.com>   5 "Dave Froble" <davef@tsoft-inc.com> wrote in message  * news:117fm4nap41m702@corp.supernews.com... [...snip...] > M > Just a guess.  Right to use new versions, and telephone support, VMS only.  @ > Maybe only the first.  Usually when you get media updates and I > documentation the price is very hefty.  Paper documentation may be the  % > biggest part of the price.  Ask HP.  >  > --  6 > David Froble                       Tel: 724-529-04506 > Dave Froble Enterprises, Inc.      Fax: 724-529-0596@ > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com > 170 Grimplin Road  > Vanderbilt, PA  15486  > 0 Question? Can you still buy paper documentation?K Our OpenVMS docs come on 6 CD-ROMs. Three for use with Windows systems and   three for use with OpenVMS.   
 Neil Rieck Kitchener/Waterloo/Cambridge,  Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  $ Date: Thu, 5 May 2005 11:22:12 -0400 From: norm.raphael@metso.com6 Subject: Re: Finally, some good news (at least for me)Q Message-ID: <OF60EF4F23.7CB4544F-ON85256FF8.00546402-85256FF8.005445C0@metso.com>   D "Neil Rieck" <n.rieck@sympatico.ca> wrote on 05/05/2005 10:37:36 AM:   > 6 > "Dave Froble" <davef@tsoft-inc.com> wrote in message, > news:117fm4nap41m702@corp.supernews.com... > [...snip...] > > H > > Just a guess.  Right to use new versions, and telephone support, VMS only. A > > Maybe only the first.  Usually when you get media updates and J > > documentation the price is very hefty.  Paper documentation may be the' > > biggest part of the price.  Ask HP.  > >  > > --8 > > David Froble                       Tel: 724-529-04508 > > Dave Froble Enterprises, Inc.      Fax: 724-529-0596B > > DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com > > 170 Grimplin Road  > > Vanderbilt, PA  15486  > > 2 > Question? Can you still buy paper documentation?  , Of course, and as an update service as well./ Check the website for the ordering number  ;) .   H > Our OpenVMS docs come on 6 CD-ROMs. Three for use with Windows systems and  > three for use with OpenVMS.   ) Do you print it out duplexed and bind it?    >  > Neil Rieck > Kitchener/Waterloo/Cambridge,  > Ontario, Canada.: > http://www3.sympatico.ca/n.rieck/links/cool_openvms.html >  >    ------------------------------  # Date: Thu, 05 May 2005 09:09:04 GMT + From: Jack Patteeuw <jjpatteeuw@nospam.net> . Subject: Re: finding the (biggest) bottle neckA Message-ID: <Qolee.4025$pe3.568@newsread3.news.atl.earthlink.net>    Duncan Macdonald wrote: J > If you have a spare SCSI interface (or better still 2) available on the M > Alpha then you might get a significent speed up by moving some of the JBOD  K > disks to the SCSI interface (and if there is a second SCSI interface use  ' > this for some of the TZ tape drives).  > N > The CI is (by modern standards) a not very fast interface and all the disks O > share its bandwidth (if my memory is correct it is about 70MB/sec). Also the  @ > HSJ40 is again by modern standards a not very fast controller. > M > Using a SCSI interface on the Alpha should increase the data transfer rate  M > from the disks, (it may well increase its CPU loading on the Alpha but you  ) > said that the system is not CPU bound).   F The HSJ40s are heading out the door as fast as I can get rid of them. G What we really need is 2 pair of HSJ80s but I don't know how to afford   to them.  I At the moment we have about 100 disk drives on line so going direct SCSI  G is no realistic.  Many are 2, 4 or 9 GB.  We hope to get everything up  E to 18 GB so that  we would haves much fewer drives to deal with.  We  D also want to RAID 5 most of the data and HSJ50 and HSJ80 require no   additional licensing to do RAID.  H We have benchmarked TZ89 against TZ88 and they are about 50% faster (no E surprise).  Interesting, the TZ89 hanging off of a HSJ50 is a simdge  H faster that a TZ89 hanging off of the KZPSA controller that is directly  connected to the Alpha GS140 !  9 Does any one know if a KZPCA is any faster than a KZPSA ?    ------------------------------  % Date: Thu, 05 May 2005 07:05:39 -0700 # From: "Tom Linden" <tom@kednos.com> . Subject: Re: finding the (biggest) bottle neck( Message-ID: <opsqa2rpjezgicya@hyrrokkin>  I On Thu, 05 May 2005 09:09:04 GMT, Jack Patteeuw <jjpatteeuw@nospam.net>    wrote:  L > At the moment we have about 100 disk drives on line so going direct SCSI  * > is no realistic.  Many are 2, 4 or 9 GB.H >   We hope to get everything up to 18 GB so that  we would haves much  K > fewer drives to deal with.  We also want to RAID 5 most of the data and   = > HSJ50 and HSJ80 require no additional licensing to do RAID.  >   L How much total storage?  You might be better off going with, say, 74 Gbyte   drivesK which you can get with either 160 or 320 MBs transfer rates for about $200@   K I am sure you can replace the drives in the cannisters, at least I did on    a BA356.   ------------------------------   Date: 5 May 2005 09:57:46 -0700 $ From: "AEF" <spamsink2001@yahoo.com>/ Subject: Re: how, exactly, is SYS$NODE defined? C Message-ID: <1115312266.196521.281320@f14g2000cwb.googlegroups.com>   / Phillip Helbig---remove CLOTHES to reply wrote: E > SYS$NODE is normally defined by DECnet.  I'm not running any DECnet  atF > the moment, but even in the future I probably won't run it on all myF > systems.  However, SYS$NODE is needed by some other applications, inF > particular DECwindows (perhaps only to say "Welcome to SYS$NODE" andC > "Session Manager on SYS$NODE") and (I'm not sure about this) LAT.  Thus, F > I have been defining it /SYSTEM/EXEC relatively early in the startupB > sequence.  Is this the proper way to define it?  I'm pretty sure1 > /SYSTEM/EXEC is correct, but what about /TRANS?      $ SHOW LOG/FUL SYS$NODE E    "SYS$NODE" [exec,crelog] = "IDS15::" [terminal] (LNM$SYSTEM_TABLE)   G Yes, /SYS/EXEC is correct, but to match the real SYS$NODE you also need G /TRAN=TERMINAL. But you'll still be missing the crelog part and I don't G think you can do that part with DCL! But if your app is only asking for " the node name it shouldn't matter.  F Whether it is a good idea or not to define SYS$NODE outside of running DECnet I can't say.    ------------------------------  # Date: Thu, 05 May 2005 17:23:01 GMT # From: hoff@hp.nospam (Hoff Hoffman) / Subject: Re: how, exactly, is SYS$NODE defined? 1 Message-ID: <VDsee.4952$gG.3806@news.cpqcorp.net>   w In article <d5df13$22t$2@online.de>, helbig@astro.multiCLOTHESvax.de (Phillip Helbig---remove CLOTHES to reply) writes: H :SYS$NODE is normally defined by DECnet.  I'm not running any DECnet at F :the moment, but even in the future I probably won't run it on all my F :systems.  However, SYS$NODE is needed by some other applications, in F :particular DECwindows (perhaps only to say "Welcome to SYS$NODE" and J :"Session Manager on SYS$NODE") and (I'm not sure about this) LAT.  Thus, F :I have been defining it /SYSTEM/EXEC relatively early in the startup A :sequence.  Is this the proper way to define it?  I'm pretty sure 0 :/SYSTEM/EXEC is correct, but what about /TRANS?  F   You should not need to define it.  If DECwindows thinks it needs it,E   then that would arguably be a bug in either DECwindows (for needing B   it) or in OpenVMS (for not defining it if DECnet is not around).  *   What versions of OpenVMS and DECwindows?  4   The usual definition is /SYS/EXEC/TRAN=TERM, FWIW.  N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------  $ Date: Thu, 5 May 2005 08:20:22 -0400' From: "Main, Kerry" <kerry.main@hp.com> 8 Subject: RE: HP hostility to alpha holding back itanium?R Message-ID: <FD827B33AB0D9C4E92EACEEFEE2BA2FB5ED072@tayexc19.americas.cpqcorp.net>   > -----Original Message-----3 > From: Dave Froble [mailto:davef@tsoft-inc.com]=20  > Sent: May 4, 2005 9:38 PM  > To: Info-VAX@Mvb.Saic.Com : > Subject: Re: HP hostility to alpha holding back itanium? >=20     [snip..]  A > > What CEO's want today is not "rip and replace", but rather=20  > "integrate= > > new technologies and make better use of what we have".=20  >=20$ > Did I remember to ask about Alpha? >=20H > > From an OpenVMS perspective, that is why the Itanium/Alpha/VAX mixed< > > clusters support is a very good investment protection=20 > strategy. And fromD > > an obviously biased vendor viewpoint, this mixed cluster supportF > > represents a brilliant OS architecture i.e. what other OS platformE > > supports 3 different architectures (CISC, RISC, EPIC) in the same B > > cluster running different OS versions and all accessing the=20 > same data? >=20J > Uh, I'm sure some people are going to ask why they can't just stay on=20J > Alpha, with the architecture continuing to advance, rather than doing=20? > all the things you mentioned above.  Without all the costs=20  > forced upon them.  >=205 > You should never have responded, bad, bad judgment!  >=20  F I am sure you are right. Some people will always be wondering what theB "real" story is behind all of the Alpha decisions [insert whatever conspiracy theory you like].=20   A Some people will also still say that Alpha was a bad decision and 3 Digital should have continued enhancing the VAX.=20   D Some people will also state IBM should have continued to enhance and
 support OS/2.   D You can speculate all you want, but the decisions have been made and time marches on.=20   D Other Customers will look at the cheaper Intel based OpenVMS systemsF which can also run Linux, HP-UX and Windows, realize that with a small= amount of work to port/test, they can keep their existing OS, G applications, run the new servers in the same cluster as their existing G systems, protect their investments in business logic, keep their staff, G custom code, and have the same level of support as they are used to and 3 think "well, perhaps this is not such a bad thing".   H As I stated earlier in the thread, the Server HW is important, but it isH a small piece of the pie when looking at overall IT costs. And that is aB fact that is common to every Server HW platform from every vendor.   Regards   
 Kerry Main Senior Consultant  HP Services Canada Voice: 613-592-4660  Fax: 613-591-4477  kerryDOTmainAThpDOTcom (remove the DOT's and AT)=20  $ "OpenVMS has always had integrity .. Now, Integrity has OpenVMS .."   ------------------------------  % Date: Thu, 05 May 2005 11:32:52 -0400 ' From: Dave Froble <davef@tsoft-inc.com> 8 Subject: Re: HP hostility to alpha holding back itanium?0 Message-ID: <117kf551cla4g61@corp.supernews.com>   Main, Kerry wrote: >>-----Original Message-----1 >>From: Dave Froble [mailto:davef@tsoft-inc.com]   >>Sent: May 4, 2005 9:38 PM  >>To: Info-VAX@Mvb.Saic.Com : >>Subject: Re: HP hostility to alpha holding back itanium? >> >  >  > 
 > [snip..] >  > > >>>What CEO's want today is not "rip and replace", but rather  >> >>"integrate >>: >>>new technologies and make better use of what we have".  >>$ >>Did I remember to ask about Alpha? >> >>G >>>From an OpenVMS perspective, that is why the Itanium/Alpha/VAX mixed 9 >>>clusters support is a very good investment protection   >> >>strategy. And from >>C >>>an obviously biased vendor viewpoint, this mixed cluster support E >>>represents a brilliant OS architecture i.e. what other OS platform D >>>supports 3 different architectures (CISC, RISC, EPIC) in the same? >>>cluster running different OS versions and all accessing the   >> >>same data? >>H >>Uh, I'm sure some people are going to ask why they can't just stay on H >>Alpha, with the architecture continuing to advance, rather than doing = >>all the things you mentioned above.  Without all the costs   >>forced upon them.  >>5 >>You should never have responded, bad, bad judgment!  >> >  > H > I am sure you are right. Some people will always be wondering what theD > "real" story is behind all of the Alpha decisions [insert whatever > conspiracy theory you like].    I No conspiracy on Compaq's part.  They just didn't want to be in the chip   business, period.   G Now, if Carly and Curly were talking earlier, it's possible that Carly  > indicated that she didn't want the Alpha, no way, no how, etc.  C > Some people will also still say that Alpha was a bad decision and 3 > Digital should have continued enhancing the VAX.    G There was, possibly still would be, a market.  Not everybody needs top  % end stuff.  Price would be the issue.   F > Some people will also state IBM should have continued to enhance and > support OS/2.   G But, was the OS2 base a large and established group such as VMS users?  C Trying something and giving up is one thing.  Having something for  F years, where people and companies have based their future, is another  matter.   F > You can speculate all you want, but the decisions have been made and > time marches on.    G Just don't try to gloss over the mistakes.  That's the only issue left.   F > Other Customers will look at the cheaper Intel based OpenVMS systemsH > which can also run Linux, HP-UX and Windows, realize that with a small? > amount of work to port/test, they can keep their existing OS, I > applications, run the new servers in the same cluster as their existing I > systems, protect their investments in business logic, keep their staff, I > custom code, and have the same level of support as they are used to and 5 > think "well, perhaps this is not such a bad thing".    We can hope.   --  4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road  Vanderbilt, PA  15486    ------------------------------  % Date: Thu, 05 May 2005 11:44:56 -0400 ( From: Bill Todd <billtodd@metrocast.net>8 Subject: Re: HP hostility to alpha holding back itanium?= Message-ID: <VtSdnfsQPcHkouffRVn-tw@metrocastcablevision.com>    Dave Froble wrote: > Main, Kerry wrote:   ...   G >> You can speculate all you want, but the decisions have been made and  >> time marches on.  >  > I > Just don't try to gloss over the mistakes.  That's the only issue left.   
 Not quite.  D Mistakes can be made by anyone.  Shattered solemn 'commitments' and A overt lies about the reasons why fall into an entirely different  $ category of 'business relationship'.   - bill   ------------------------------   Date: 5 May 2005 11:51:20 -0500 4 From: kaplow_r@encompasserve.org.TRABoD (Bob Kaplow)8 Subject: Re: HP hostility to alpha holding back itanium?3 Message-ID: <wkmcoDHygAaK@eisner.encompasserve.org>   [ In article <slrnd7ivfe.4tp.rivie@Stench.no.domain>, Roger Ivie <rivie@ridgenet.net> writes: ? > On 2005-05-04, JF Mezei <jfmezei.spamnot@teksavvy.com> wrote: I >> Sorry. The original 8086 was nothing more than a fancy toy controller,  >> 16 bit CPU with 8 bit bus,  > G > The original 8086 had a 16-bit bus. It was the cut down 8088 that had  > the 8-bit bus. >  >> And not long after the 1 >> 8086 came the vax with full 32 bit addressing    K When was the 8086 first released? I don't recall seeing 8086 based stuff on ' the market until AFTER the VAX and VMS.   1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD" & 		>>> To reply, remove the TRABoD! <<<K Kaplow Klips & Baffle:	http://nira-rocketry.org/LeadingEdge/Phantom4000.pdf L     www.encompasserve.org/~kaplow_r/    www.nira-rocketry.org    www.nar.org  @         You [should] not examine legislation in the light of theD         benefits it will convey if properly administered, but in theF         light of the wrongs it would do and the harm it would cause ifF         improperly administered -- Lyndon Johnson, former President of         the U.S.   ------------------------------   Date: 5 May 2005 00:47:25 -0700  From: icerq4a@spray.seA Subject: Re: Maybe HP should get out of the hardware business.... C Message-ID: <1115279245.415669.114430@g14g2000cwa.googlegroups.com>    Bill Gunshannon wrote:- > In article <4278316D.9F9AAC8F@comcast.net>, 7 > 	David J Dachtera <djesys.nospam@comcast.net> writes:  > > Dave Froble wrote: > >> [snip] G > >> Me, I think that Intel doesn't like to lose face.  That may be the  > >> savior of VMS.  > > A > > Dunno, Dave... A "savior" that can't humble itself to admit a F > > misjudgement might very well lead itself - and by extension, VMS - toF > > ruin. History is rife with the exploits of those who tried to save face, E > > and ended up destroying themselves and/or their allys in the end.  > F > How true.  VMS is already slower at some tasks (Note, I said slower,C > not poorer) because of things like guaranteed IO (which I am sure  veryG > few users understand).  Now put it on an inferior processor resulting E > in it being even slower than it's peers.  How long would you expect  itE > to survive?  How many people will even understand enough to know it B > isn't the OSes fault at all?  All they will see is a really slow system.   . The performance is not a problem with Itanium,9 there are lots of other things I would worry about first.    ------------------------------   Date: 5 May 2005 12:08:19 -0500 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler) A Subject: Re: Maybe HP should get out of the hardware business.... 3 Message-ID: <3V+XLbaAUZcw@eisner.encompasserve.org>   U In article <3dstdpF20ouU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:  >   B > I was just responding to the expressed idea that somehow ItaniumB > was going to save VMS.  If VMS on Itanium is seen as inferior toC > other OSes (and remember, we have already been shown that reality A > is unimportant and perception is everything.) then it will die. @ > VMS needs to run on the best platform available or at the very3 > least on the same one as all of the popular OSes.   ;    Was VMS on VAX considered inferior to Ultirx, or VAXeln? 9    Was VMS on Alpha considered inferior to Tru64, or WNT?   =    I/O perfomance is not everyone's first consideration.  And ?    on VMS you do always have the choice of turning on buffering     at a variety of levels.   ------------------------------   Date: 5 May 2005 08:21:00 -0700 ( From: "Lawrie" <stroker_ace@hotmail.com>! Subject: Message definition files C Message-ID: <1115306460.188186.120170@o13g2000cwo.googlegroups.com>    Hi,   C I am working on an application that uses user defined VMS messages.    The build procedure uses   $message filename    to generate a file filename.obj   C This .obj is then linked when building the C++ application. Message 0 types are declared as external in a header file.  E I have added a few new message types but when I come to use them they , appear as the wrong message type or unknown.  E Apart from building and linking the message definitions .obj file and G declaring the messages as external is there anything else I have to do?    Cheers   Lawrie   ------------------------------  # Date: Thu, 05 May 2005 15:56:18 GMT + From: Jeff Chimene <jchimene@earthlink.net> % Subject: Re: Message definition files B Message-ID: <Cmree.6010$GQ5.5732@newsread1.news.pas.earthlink.net>  
 Lawrie wrote:  > Hi,  > E > I am working on an application that uses user defined VMS messages.  >  > The build procedure uses >  > $message filename  > ! > to generate a file filename.obj  > E > This .obj is then linked when building the C++ application. Message 2 > types are declared as external in a header file. > G > I have added a few new message types but when I come to use them they . > appear as the wrong message type or unknown. > G > Apart from building and linking the message definitions .obj file and I > declaring the messages as external is there anything else I have to do?  >  > Cheers >  > Lawrie > F Wrong message type is a little vague. Are you defining the facility? A; snippet from the .msg file might help diagnose the problem.    Cheers,  jec    ------------------------------   Date: 5 May 2005 12:14:29 -0500  From: briggs@encompasserve.org% Subject: Re: Message definition files 3 Message-ID: <cYSZgNZSOiwe@eisner.encompasserve.org>   n In article <1115306460.188186.120170@o13g2000cwo.googlegroups.com>, "Lawrie" <stroker_ace@hotmail.com> writes: > Hi,  > E > I am working on an application that uses user defined VMS messages.  >  > The build procedure uses >  > $message filename  > ! > to generate a file filename.obj  > E > This .obj is then linked when building the C++ application. Message 2 > types are declared as external in a header file. > G > I have added a few new message types but when I come to use them they . > appear as the wrong message type or unknown. > G > Apart from building and linking the message definitions .obj file and I > declaring the messages as external is there anything else I have to do?   ? You aren't expecting sys$exit(message-code), exit(message-code) F or return(message-code) to display the message text on standard output are you?  > They won't.  Standard image rundown leaves it up to the CLI toH produce the text output corresponding to the image termination conditionH code.  The CLI doesn't have access to image-private message definitions.  @ On the other hand, if you blow your image up inadvertently (thusC triggering the traceback handler), kill your stack catastrophicallyu= (thus triggering the last chance handler), blow your image up > intentionally with LIB$STOP or LIB$SIGNAL (thus triggering the> traceback handler) or explicitly call SYS$GETMSG or SYS$PUTMSGA then you can expect that your private message definitions will bee used.    	John Briggs   ------------------------------  # Date: Thu, 05 May 2005 17:20:25 GMT # From: hoff@hp.nospam (Hoff Hoffman)c% Subject: Re: Message definition filesc1 Message-ID: <tBsee.4951$gG.3959@news.cpqcorp.net>,  n In article <1115306460.188186.120170@o13g2000cwo.googlegroups.com>, "Lawrie" <stroker_ace@hotmail.com> writes:D :I am working on an application that uses user defined VMS messages. :The build procedure usesd :$message filename  :to generate a file filename.obj  C   If that is the exact command and not an abbreviated command, theniE   it's generating a message object, and not a message pointer object.e  D   There is a subtly different mechanism, the message pointer object.E   That approach involves the message compilation and a second messagepC   compilation with the /FILENAME qualifier on the MESSAGE compiler. E   That is how message object files are built and (eventually) linked.   D :This .obj is then linked when building the C++ application. Message1 :types are declared as external in a header file.-  B   See the MESSAGE/SDL stuff, too.  That's easier than what you are@   doing here, and is how OpenVMS does this sort of thing -- withC   the resulting SDL file and the Freeware SDL compiler, you can get.D   the include files similar to those seen throughout OpenVMS itself.  F :I have added a few new message types but when I come to use them they- :appear as the wrong message type or unknown.;  E   That would imply you're not linking against what you think you are.CF   (It's also possible there is an installed image around, too, as I've,   seen that affect the message output, too.)  B   If you can post a short reproducer, that would be quite helpful.  F :Apart from building and linking the message definitions .obj file andH :declaring the messages as external is there anything else I have to do?  A   See the build procedure for GNM on Freeware V7.0, both for whatnB   the tool does and for how the tool is built.  That's how I buildA   several components of OpenVMS, including the SDML documentation    and the help files.r    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faqiN  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com,   ------------------------------  # Date: Thu, 05 May 2005 16:08:29 GMT@# From: hoff@hp.nospam (Hoff Hoffman)g! Subject: Re: More VAX help needed$0 Message-ID: <1yree.4943$gG.941@news.cpqcorp.net>  W In article <3dpbekF6rifjsU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes:n@ :...When I asked here a few days ago I got the idea that I couldA :connect my storageworks cabinet to a sikngle 6640 without a starpA :coupler.  I have had no luck doing it with any combination of CI D :cable connections.  I asked the local VMS guru and he said it won'tE :work without a star coupler.  Sooooo......  Can anyone confirm this?r  %   Your local OpenVMS guru is correct.a  I   In addition to the attenuation provided by the SC008 Star Coupler -- asaI   was mentioned by other responders -- and assuming you do not simply frynH   something as the signal at the CI receivers will be rather hotter thanE   expected; the signal definitely won't match expectations -- each CI H   controller expects to hear itself -- its echo -- as part of its normal   CI cluster operations.    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------   Date: 5 May 2005 16:51:06 GMT ( From: bill@cs.uofs.edu (Bill Gunshannon)! Subject: Re: More VAX help neededw* Message-ID: <3dv17qFb9fbU1@individual.net>  0 In article <1yree.4943$gG.941@news.cpqcorp.net>,& 	hoff@hp.nospam (Hoff Hoffman) writes:Y > In article <3dpbekF6rifjsU1@individual.net>, bill@cs.uofs.edu (Bill Gunshannon) writes: A >:...When I asked here a few days ago I got the idea that I couldsB >:connect my storageworks cabinet to a sikngle 6640 without a starB >:coupler.  I have had no luck doing it with any combination of CIE >:cable connections.  I asked the local VMS guru and he said it won'tuF >:work without a star coupler.  Sooooo......  Can anyone confirm this? > ' >   Your local OpenVMS guru is correct.o > K >   In addition to the attenuation provided by the SC008 Star Coupler -- asHK >   was mentioned by other responders -- and assuming you do not simply fryrJ >   something as the signal at the CI receivers will be rather hotter thanG >   expected; the signal definitely won't match expectations -- each CICJ >   controller expects to hear itself -- its echo -- as part of its normal >   CI cluster operations. >   A I should have known better than to doubt him.  I have yet to havet" him be wrong when it comes to VMS.  B The good news is that Island Computers has come thru with an SC008C for the cost of shipping and without the rack (I have got plenty ofOD them) it's going to be cheap enough that I got my boss to spring forB it.  Many thanks to David Turner.  With luck I will be loading VMSB next week and may be building my first cluster shortly after that.  > Thanks to everyone for your help and be prepared for even more- questions when I get to the cluster part. :-)s   bill   -- pJ Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolvesD bill@cs.scranton.edu     |  and a sheep voting on what's for dinner. University of Scranton   |A Scranton, Pennsylvania   |         #include <std.disclaimer.h>       ------------------------------  % Date: Thu, 05 May 2005 02:52:03 -0400h' From: Dave Froble <davef@tsoft-inc.com>h& Subject: Re: OpenVMS using RAID drives0 Message-ID: <117jgkk84dssmcd@corp.supernews.com>   JF Mezei wrote:  > "Barry Treahy, Jr." wrote: > E >>more.  An RMS file with the easy potential of more than 100 million I >>records is nuts, and there is even debate about needing to double that?aE >>This is certainly an example of using the wrong tool for the job...r  D Well, just for kicks, lets say that we're keeping a record for each  person in the US.a  J > Pardon my ignorance, But in what way would RMS be so much less efficient? > than some Oracle or RDB database (assuming a mostly read only  > application) ? e  ! Why do you assume less efficient?r  D > Is there some document somewhere that describres the philosophicalG > diofferences between RDB and RMS in terms of building and maintaining  > indices ?i > , > I can understand RDB having better locking   Where do you get such an idea?   > and transaction rollback for> > multiple record update transactions, as well as field formatF > definitions, but in terms of the actualy storage and indices, is RMS# > really that much less efficient ?a  E Actually, RMS, doing less work, would be more efficient in some ways.l  A But really, who wants to run CONVERT on that file to clean it up?a  I I don't know, but am guessing that RDB and such expect to use such large ,A amounts of storage, and have utilities for managing such storage.h   -- o4 David Froble                       Tel: 724-529-04504 Dave Froble Enterprises, Inc.      Fax: 724-529-0596> DFE Ultralights, Inc.              E-Mail: davef@tsoft-inc.com 170 Grimplin Road- Vanderbilt, PA  15486i   ------------------------------  % Date: Thu, 05 May 2005 09:20:24 +0200-+ From: Karsten Nyblad <nospam@nospam.nospam>-& Subject: Re: OpenVMS using RAID drives= Message-ID: <4279c973$0$67257$157c6196@dreader2.cybercity.dk>-   Barry Treahy, Jr. wrote: > David J Dachtera wrote:01 >> Imagine managing an RMS indexed file of 1TB...B >> >> *SHUDDER* >>   >>J > If the application is so dumb to create a single file of that size, one D > should fix the application...  Presuming that it had record sizes F > pressing the max limit of 32k, that would represent over 33 million G > records at 1TB.  If the record size was more reasonable in size, the XK > number of records could easily double, triple, or even quadruple, if not -F > more.  An RMS file with the easy potential of more than 100 million K > records is nuts, and there is even debate about needing to double that?   E > This is certainly an example of using the wrong tool for the job...:  G First, such a file would most likely be indexed.  Then you have should rH be glad if you can use more than 1/3 of the file for storing data.  You K should be glad to have more than 300GB data in an indexed file of size 1TB.   D Secondly, you can get disk storage at less than $1000 per TB if you F chose to use ATA or SATA disks.  E.g. put 16 400GB SATA disks on on anE Adaptec controller and you should have something like 5TB for files. lH Google think that they can store E-mail at a price of a couple of bucks B per years per GB.  Many organization will find it easier to throw 3 storage at a problem than to redesign the programs.-  I Thirdly, the company may not have the source for the application, either PI because it is lost or because the program is closed source provided from e> a software vendor, that does not want to change their program.   ------------------------------  $ Date: Thu, 5 May 2005 12:11:42 -0400) From: "Neil Rieck" <n.rieck@sympatico.ca> & Subject: Re: OpenVMS using RAID drives9 Message-ID: <9Bree.14676$3U.958746@news20.bellglobal.com>   @ "David J Dachtera" <djesys.nospam@comcast.net> wrote in message % news:4279810E.DF1F6BF2@comcast.net...e > Bob Koehler wrote: >>? >> In article <42790964.7000503@MMaz.com>, "Barry Treahy, Jr."   >> <Treahy@MMaz.com> writes: >> >J >> > Perhaps this is a dumb question, but does any one really have any oneJ >> > application that uses a single file of that size?  Presuming there isG >> > any type of underlying databases, most databases will allow you tonK >> > spread the areas across multiple spindles, or RAID partitions, therebyoH >> > making the individual file size, or volume size limit irrelevant... >> > >>F >>    There are lots of applications that are too dumb to spread thier >>    files around.  >e0 > Imagine managing an RMS indexed file of 1TB... >c > *SHUDDER*  >hM Since the RFA is a 24 bit structure, wouldn't there be an upper record limit cJ of ~ 16 million? Depending on the record size you might not ever get to 1  TB.o  
 Neil Rieck Kitchener/Waterloo/Cambridge,D Ontario, Canada.8 http://www3.sympatico.ca/n.rieck/links/cool_openvms.html   ------------------------------  # Date: Thu, 05 May 2005 17:01:01 GMT # From: hoff@hp.nospam (Hoff Hoffman)s& Subject: Re: OpenVMS using RAID drives0 Message-ID: <hjsee.4948$gG.510@news.cpqcorp.net>  V In article <JhVde.16$It1.10@lakeread02>, "stinehelferw" <stinehelferw@cox.net> writes: :We have an OpenVMS network.I :We are purchasing an archival system (re: several RAID drives with smartbJ :controller).  The RAID drives would appear as a single 2 Terra Byte disk.L :Can OpenVMS recognize a "disk" of that size or would we just have to settle% :for a SCSI bay full of 400GB drives?i  G   One terabyte (1 TB) is the volume capacity limit for a single volume ,F   (whether an actual spindle, or synthetic) on OpenVMS; please see theF   OpenVMS FAQ for more on this and on other related dscussions, and on9   system-specific limits and ECOs that may be applicable.p    N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faqdN  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.coma   ------------------------------  $ Date: Thu, 5 May 2005 12:14:01 -0400? From: "David Turner, Island Computers US Corp" <david@hpaq.net>0 Subject: ordering OpenVMS 8.2.4 Message-ID: <Gyree.28$0i3.14@bignews3.bellsouth.net>  E  have called HP, Two authorized resellers and HP financial services.mF None of them can find out the ordering part number for the OpenVMS 8.2 distribution media.e  & Can anyone advise of the part number ?   Thanks   -- t   David B Turner Island Computers US Corp 2700 Gregory St, Suite 180 Savannah GA 31404t Tel: 912 447 6622 X201 Skype ID: islandco Fax: 912 201 0402c Email: dbturner@icusc.com  Web: http://www.islandco.com% =====================================o< All orders are subject to the following terms and conditions. of sale. These should be read before ordering.% http://www.islandco.com/warranty.htmli   ------------------------------   Date: 5 May 2005 10:09:39 -0700o! From: kenneth.randell@verizon.nete! Subject: Re: ordering OpenVMS 8.2oC Message-ID: <1115312979.545047.129070@f14g2000cwb.googlegroups.com>   ? > =CC have called HP, Two authorized resellers and HP financialV	 services.yD > None of them can find out the ordering part number for the OpenVMS 8=2E2l > distribution media.i >C( > Can anyone advise of the part number ? >n   Which platform are you after?-  3 Itanium looks like (deciphering an old quote here):0  * BA322AA listed as HP OpenVMS I64 FOE media   option AJR says DVD media.   ------------------------------   Date: 5 May 2005 10:19:14 -0700e! From: kenneth.randell@verizon.netc! Subject: Re: ordering OpenVMS 8.2 C Message-ID: <1115313554.183204.252870@f14g2000cwb.googlegroups.com>   C While are at this, this information is in table 2-2 of the recently  posted VMS FAQs.   ------------------------------   Date: 5 May 2005 06:26:41 -0700r From: dooleys@snowy.net.auT Subject: Re: Time to vote on your favorite slogan for the next OpenVMS License plateC Message-ID: <1115299601.660434.283980@o13g2000cwo.googlegroups.com>)  " susan_skonet...@hotmail.com wrote: > Dear Newsgroup >rC > Time to vote on your favorite slogan for the next OpenVMS Licensei > plate. >aG > Please visit:  www.openvms.org on the left nav bar the poll is there.n >nD > The license plate will have a large OpenVMS in the middle with the, > slogan underneath.  Please vote only once. >nE > This poll will only be up on the web site until next Friday May 13.  >w > Warm Regards,  > Sue., Will "Goodbye and thanks for all the $" fit? Phil   ------------------------------   End of INFO-VAX 2005.250 ************************thing up  E to 18 GB so that  we would haves much fewe?_磦يazLE
$Ah@:a|̪<C84`8Z.T䂰qW[ljM͵v?k@yG\`qJ+H,'-c2No9qƷǵP//-^?<t卋o֛o믿k/ΎeYKKKsssN:y0iW'+Gmyq~cyxc5<ÛWy`kGGPO>vώw懾9rTo}뵗G?V~+s׆ŗ=8ʻl?7>14{яb?/.R9A*:b$߿>HKu{WիЫ36?/~O>T~lCo;~k/~̵#؃|?<5ο|q_S7ssC_>
{>~Ea_{`8'^}Z[?ʿzow㍻??p_7k/o];sNO\;C:]zICsVvW?F(7ؑj{
Ҿs?arC?iN{tLULHtx~yk
ݛSFZ,.G\y|Cr}DƹbaEP|Ac-,{*,,[w6B`Wf2i'_dHxx"E/
)3]Lǎ`*PIYa6S42L((42L(AeZD4-"zK<1I,FpDW2OO.up]\^h	'=A~
hw뱱"dI*~%rq+Ȼ
r	Ѷν@pw1A:r=[YT򭞭q)3"ߍL4f.,y#
F3C^/g[wXt\`-R\'lRFc*FxT'OQ0n<aI
zH"mXI9"M>UMQEry[*TJJxE>Q#R1v,[0aΣk	fh0{g,l8/@H(7Qrԣ4`+[!)倪湶n/kO2O9tbX;&R\@mi%
ĨƌWMJsƹ(ƸBLLLLY-YZght7BǎԤQBX$T I	qB7,فڶaaûl2((}h2rJ͑F]^2*ht[KlO[S
+	y ˖r=8S12^щl$~Es