1 INFO-VAX	Mon, 04 Sep 2000	Volume 2000 : Issue 495       Contents:- Re: Alpha 200 4/233 firmware upgrade question  Re: analysing object libraries Re: dcps9 Re: DEC 3000 model 300, help with firmware/serial console 9 Re: DEC 3000 model 300, help with firmware/serial console  Re: IMAP-Server for VMS?# Re: Is there any new Alpha CPU out? # Re: Is there any new Alpha CPU out? # RE: Is there any new Alpha CPU out? # RE: Is there any new Alpha CPU out? ) RE: Oracle/VMS (in the US) teleconference ) Re: Oracle/VMS (in the US) teleconference  Re: RTR and DECdtm Re: RTR and DECdtm Re: RTR and DECdtm srm pc164 oddities! Re: Sun Hardware problems persist ! Re: Sun Hardware problems persist ! Re: Sun Hardware problems persist ! RE: Sun Hardware problems persist ( Re: TCP/IP 5.0A PatchList or ECO Summary( Re: TCP/IP 5.0A PatchList or ECO Summary Re: TCPIP$FTP Bugchecking!& Urgent help with backup restore needed- Urgent, Help to remove a SOH or ETX hex code. 1 Re: Urgent, Help to remove a SOH or ETX hex code.  Re: VMS and 100 meg Ethernet Re: WORD viewer for VMS   F ----------------------------------------------------------------------  " Date: Mon, 4 Sep 2000 08:01:57 GMT( From: Terry Kennedy <terry@gate.tmk.com>6 Subject: Re: Alpha 200 4/233 firmware upgrade question' Message-ID: <G0CrnA.IL3@spcuna.spc.edu>   + James L. Wiley <Wiley@tarleton.edu> writes: M > Thanks for all the help.  The problem did turn out to be that I didn't have H > enough memory to run the newer firmware.  I put another 16Mb in and itM > works.  Now I think I have a video card problem.  VMS comes up to the point K > of saying "OpenVMS (TM) Alpha Operating System, Version 7.2" and seems to N > stop there.  I have tried a Diamond Multimedia VGA and a generic Trident VGA? > cards with the same result.  I assume I need a specific card?   K   In my experience, random VGA-compatible cards will work in text mode, but L you won't be able to configure DECwindows. Early in the startup, if you haveK a VGA card VMS recognizes, the screen font will change from the regular one M to a larger one (done in graphics mode), with an inverse-video "OpenVMS Oper- 2 ator Console" message at the bottom of the screen.  M   It sounds as if you have a card that's close enough to a supported one that 2 VMS tries to put it into graphics mode but wedges.  J   One "generic" card that VMS does support is the #9 Motion 330 - in fact,8 I had some AS200's that shipped from DEC with that card.  L > I will connect a terminal to the com1 port and see if I can get it to work > that way.   K   Disconnect the keyboard from the AS200 as well. You might need to also do ' "set console serial" at the >>> prompt.   - 	Terry Kennedy             http://www.tmk.com 5         terry@tmk.com             Jersey City, NJ USA    ------------------------------  # Date: Mon, 04 Sep 2000 12:21:01 GMT  From: waarom_nok@my-deja.com' Subject: Re: analysing object libraries ) Message-ID: <8p0434$mib$1@nnrp1.deja.com>   G > Well, you didn't say which version of VMS your running, but on my 7.2 " > system that is not legal syntax. > $ We're running VMS version  7.1 here.  C > Try extracting the module from the library and then analyze it; I  believe < > your are confusing object modules with library modules.... >   7 Yes it's confusing indeed, I just don't get grip on it. D Indeed I solved the problem by extracting de object from library and deleting afterwards.G But it should've been nice when I could've accessed the lib without the  use of temporary files. F I also downloaded the sourses of gnm and sdl but can't get those tools  to work, maybe some day later...   Thanks everybody.    Norbert     & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  $ Date: Mon, 4 Sep 2000 15:19:37 +0200- From: "Sandrine Treheux" <satr@memo.ikea.com>  Subject: Re: dcps , Message-ID: <8p07o9$cup$1@mailgate.ikea.com>   Hej,  H     I had the same problem, and after checking in DCPS docs, I found outJ that this can be due to a the firmware of the printer. After upgrading it, no problem anymore.    Br/Sandrine       ! Bob Ricci a crit dans le message 3 <00d601c0141c$38fdf0a0$585b5cc0@socrates.Subway>... G >i have a hp laserjet 40tn and am trying to print landscape- set up the : >ip_rawtcp/192.92.91.68:9100 port and on the generic queueG >page_orientation=landscape....the queus set up properly, but i have to  reset L >the queues every time i print as they remain in a starting state....until iJ >reset then all prints normally- i have tried manually setting the printer toI >autosense and ps - but neither helps...aanyone else having this problem?  >Robert V. Ricci >Systems Manager >Drs. Associates (SUBWAY)  >325 Bic Dr. >Milford, Ct 06460 > tel  203 877 4281 ext 1144 >fax to pc 203 783 7144  > fax 203 876 6682 >email ricci_r@subway.com   > or     maxx0623@concentric.net >http://www.subway.com >  >    ------------------------------  $ Date: Mon, 4 Sep 2000 10:49:31 +02005 From: "philippe bocher" <philippe.bocher@euriware.fr> B Subject: Re: DEC 3000 model 300, help with firmware/serial console$ Message-ID: <39b352b4@news.euriware>  K "Tobin Fricke" <tobin@bombay.CS.Berkeley.EDU> a crit dans le message news: C Pine.GSO.4.10.10009032135240.14448-100000@bombay.CS.Berkeley.EDU...  > J > My roommate and I have come across a DECstation 3000 model 300, which weL > am trying to network boot.  We do not have a keyboard or display, so we'reJ > using the serial console only, and we have no removable media (CD-ROM orJ > floppy) so we need to boot over the ethernet. We've updated the firmwareK > successfully using the MOP protocol (something like bootp but proprietary F > to DEC) and mopd for linux. Now, however, with the new firmware, theI > serial console has stopped working fully; although we receive data from K > the machine, we can't seem to send anything, and thus we can't accomplish K > anything. However, if we boot into the existing operating system (OSF/1), K > the serial connection works in both directions.  Here is what we get over  > the serial console:  >  > DEC 3000 - M300  > Digital Equipment Corporation K >      VPP PAL V5.56-80800101/OSF PAL V1.45-80800201 - Built on 28-JAN-1997 
 > 10:54:25.34 # > TCINFO      DEVNAM        DEVSTAT $ > ------      --------      --------J >                  CPU      OK KN16-AA -V7.0-S889-I21F-sV2.0-DECchip 21064 P3.0# >                  OSC      150 MHz  >                 ASIC      OK >                  MEM      OK >                FEROM      OK > 6  >                  CXT      OK > 5  >                  NVR      OK  >                  SCC      ? 60 >                   NI      OK >                 ISDN      OK > 4  >                 SCSI      OKB > 0) Dansk                             8) Franais (Suisse Romande)2 > 1) Deutsch                           9) Italiano5 > 2) Deutsch (Schweiz)                 10) Nederlands 0 > 3) English                           11) Norsk3 > 4) English (British/Irish)           12) Portugus / > 5) Espaol                           13) Suomi 1 > 6) Franais                          14) Svenska 0 > 7) Franais (Canadien)               15) Vlaams >  > 3 >>>  >  > L > If anyone could help us out, perhaps by pointing us to where we may obtain@ > manuals or a description of the SCC error, it would be greatly > appreciated. > 	 > Thanks,  > Tobin <tobin@sji.org>  >   A     You can try http://www6.compaq.com/alphaserver/index.html ...    Philippe Bocher  philippe.bocher@euriware.fr    ------------------------------  $ Date: Mon, 4 Sep 2000 02:39:37 -0700& From: Ryan Moore <rmoore@qualcomm.com>B Subject: Re: DEC 3000 model 300, help with firmware/serial consoleL Message-ID: <Pine.GSO.4.10.10009040222300.21654-100000@illyana.qualcomm.com>  ' On Sun, 3 Sep 2000, Tobin Fricke wrote: 0 > We do not have a keyboard or display, so we'reJ > using the serial console only, and we have no removable media (CD-ROM orJ > floppy) so we need to boot over the ethernet. We've updated the firmwareK > successfully using the MOP protocol (something like bootp but proprietary F > to DEC) and mopd for linux. Now, however, with the new firmware, theI > serial console has stopped working fully; although we receive data from K > the machine, we can't seem to send anything, and thus we can't accomplish K > anything. However, if we boot into the existing operating system (OSF/1), K > the serial connection works in both directions.  Here is what we get over  > the serial console:  > 
 > [delete] >   >                  SCC      ? 60
 > [delete] > L > If anyone could help us out, perhaps by pointing us to where we may obtain@ > manuals or a description of the SCC error, it would be greatly > appreciated.  I I have a manual titled "DEC 3000 Model 300 Hardware Reference Guide".  It J doesn't say specifically what "? 60" means, but it does say that the "SCC"I test has to do with the keyboard/mouse connector.  So I suspect it's just J telling you it can't talk to the keyboard and mouse.  But you already know that.   C There's a jumper on the motherboard that will change the machine to E "alternate console mode" (ie. use the serial port as a console).  The > jumper is labelled "W2" and is located toward the front of the< motherboard.  It has 4 pins labelled 1-4 from left to right.  A If the jumper is on pins 1 and 2, it means use the graphics port. @ If the jumper is on pins 2 and 3, it means use the serial port.   F Maybe you should try changing the position of this jumper to see if itI helps.  If you don't have a keyboard plugged in, it's supposed to use the   serial console (as you've seen).  ' The pin-out of the 25-pin connector is:   
 1 = ground 2 = transmit 3 = receive  4 = RTS  5 = CTS  6 = DSR 
 7 = ground 15 = transmit timing 17 = receive timing  20 = DTR    J I think you really only need to worry about 2, 3, and ground.  But I'm not sure.   # Console should be 9600 baud, 8-N-1.   ? When you flashed the ROM, how was the write-protect jumper set? C Jumper W-1.  Jumper on pins 1-2 means write-protect.  Pins 2-3 mean  write-enabled.   -Ryan    ------------------------------  % Date: Mon, 04 Sep 2000 15:45:21 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk> ! Subject: Re: IMAP-Server for VMS? ) Message-ID: <39B3B581.A5C9A582@bbc.co.uk>    John Santos wrote:  * > On Fri, 1 Sep 2000, Tim Llewellyn wrote: >  > >  > > ) > > Steve.Spires@yellowpages.co.uk wrote:  > > C > > > Contact:   Tel: 3063  -  VSSG, 1st Floor, Bridge Street Plaza  > > > 9 > > > I would like to be the first, but I'd be fibbing...  > > > U > > >  Just as a sanity check, is there anybody out there that can report a situation Q > > > where the network people aren't running the whole show, and are expected to L > > > provide required services, not dictate what protocols can or cannot be. > > > transmitted over the network?  Even one? > > >  > > M > >  Oh, bow down to the great God Cisco for the skills required to configure K > > it are a highly guarded secret, especially from those who could run the O > > network in the free time they have watching those oh so reliable VMS boxes, P > > and configure it to run as many other protocols as well as IP as are needed. > > * > > Call it job security? Call me a cynic? > I > Actually, I thought the online help on the Cisco boxes I've played with C > was pretty good.  Reminded me a lot of VMS HELP.  A lot easier to ) > understand and navigate than man pages.  > B > Or are you saying that managing Ciscos isn't as difficult as theE > people who do it for a living pretend?  They just claim it's arcane  > so they can keep their jobs?  I No, I was just  being slightly sarcy on a friday PM and taking a slightly F tangential shot at agency-driven IT job market where current skill set+ rather an aptitude and talent is important.   G Oh, and all those "Get-your-non-IP-traffic-off-my-network" type network   admins do wind me up, sometimes.   > A > Oh, and DECNET on Ciscos is easy enough, just expensive.  (They 6 > charge a lot extra for protocols other than TCP/IP.) > B > Also, BTW, I've never dealt with anything but static routing.  IG > think this is where all the complexity lies.  But that has nothing to & > do with which protocols are allowed. >   9  Aha, I did setup RIP on a DECNis 600 once, that was fun.    --6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk   A I speak for myself only and my views in no way represent those of  MedAS or the BBC.    ------------------------------  % Date: Mon, 04 Sep 2000 12:24:21 +0100 B From: Andrew Harrison SUNUK Consultancy <andrew.nospam@uk.sun.com>, Subject: Re: Is there any new Alpha CPU out?* Message-ID: <39B38664.506281E2@uk.sun.com>  B Sorry Kerry I don't want to rain on your parade (well to be honest? I do) but the SPECint/fp advocacy you are engaged in is costing  your customers money.   9 Do you know why. Well rather absurdly SPECint/SPECfp tend 7 to show that the Alpha 21264 running at say 731+ Mhz is ? as quick as its clock speed and that fact that it is a RISC CPU  would lead you to expect.   < Now why you are going to say is this costing customers money= and the answer is simple, because SPECint/fp are not at least ? for the Alpha good predictors of commercial applications server 7 performance but people are being charged on that basis.   8 Many of the major software vendors now charge by CPU and: they charge more for faster CPU's. Some measure this using< SPECint which grossly over estimates the capabilities of the> 21264 if the current crop of benchmark results are anything to> go on. Some and Oracle is one say real applications throughput> is higher for a given clock speed for RISC machines than Intel< boxes (a rule that generally works for HP's, Sun's and IBM's6 but does not for Alphas) so they charge on this basis.   RISC system + 100 dollars * (Mhz * 1.5 * Number of CPU's)   	 Intel Box % 100 dollars * (Mhz * Number of CPU's)   " http://www.oraclestore.oracle.com/  9 Now if I was you I would be arguing that people should be ; charged for Alpha using either the Intel model or a special 
 Alpha one.  8 You see if you use Oracle applications benchmark results6 what you will see from the following table that Compaq5 WildFire customers are paying way, way more for their 9 licenses than Sun or even Intel  customers to ge the same  throughput.   D These are based on the published Oracle apps benchmark results from.  A http://www.oracle.com/apps_benchmark/html/index.html?results.html   < And remember the benchmark has been developed to help people
 size systems.   G                         Oracle users/CPU      CPU Mhz      Oracle costs  dollars 8 Sun                 488                              400 245,900 I CPQ/Alpha    350                               731                626,570 K CPQ/Intel      448                               500                223,210   < The costs based on the cost for 2000 users using a perpetual@ power unit cost of 100 dollars. Remember the cost per processing8 unit is the same for each platform. You then get charged; maintenance as a % of license cost so this is an additional  wammy.  ; Ironically if as you are advocating you subsituted relative 6 SPECint performance for Mhz the story would get worse.; If Oracle had used relative SPECint then the Alpha licences 7 would have cost 3x the Sun's rather than 2.5 using Mhz.   > So I guess you are between a rock and a hard place, you either; want to be in marketing and make marketing points that cost 9 you customers real money or you want to help them out and A save them very substantial cash but you will  ruin your marketing  campaign in the mean time.  ? What you should be doing is lobbying Oracle to create a special 9 category for Alpha servers which would have the following ? formula. Even better for you would be to lobby Oracle to charge ; on the basis of their apps_server benchmarks but this would : mean everyone doing the benchmark and Compaq eating a huge: amount of humble pie neither of which I can see happening.   Alpha Server cost 7 Fixed processing unit cost * Mhz * .5 * Number of CPU's   9 You should also be lobbying all the other vendors who use < this method or relative SPECint to create a special case for4 Alpha boxes where you are allowed to discount either value.  5 As a half way house you could be lobbying vendors whoE; charge on a relative SPECint basis to change their charging : to relative Mhz, it sounds dreadfull but it will save your customers money.  8 Most of all you should not be trying to defeat Intel Mhz7 arguments with SPECint results because its costing your # customers loads and loads of money.    Regards( Andrew Harrisonn Enterprise IT Architect(     "Main, Kerry" wrote:  	 > Rudolf,  >dH > You are correct in that marketing is a huge component of "the fastest" > cpu's. >mL > Here is the actual latest industry standard benchmarks for CPU, memory and > compiler optimizations:lG > <http://ideasinternational.com/benchmark/spec/specfp2000.html> SpecFPaK > <http://ideasinternational.com/benchmark/spec/specint2000.html> Spec Int.e >rK > Also, keep in mind that the 1Ghz x86 designs are WS / dept server designs " > only ie. single / dual cpu only. > K > The greater than 2 cpu x86 Server designs (Xeon etc) are still at 550 and-	 > 700Mhz.-N > <http://commerce.www.ibm.com/cgi-bin/ncommerce/CategoryDisplay?cntrfnbr=1&cg4 > menbr=1&cgnbr=Netfinity+7600&cntry=840&lang=en_US> >25 > <http://www.compaq.com/products/servers/platforms/>p >a
 > Regards, >a > Kerry Main > Senior Consultant, > Compaq Canadaf > Professional Servicesb > Voice : 613-592-4660 > FAX   : 819-772-7036 > Email : kerry.main@compaq.comw >. > -----Original Message-----/ > From: Rudolf Wingert [mailto:win@FOM.FGAN.DE]e) > Sent: Thursday, August 31, 2000 2:45 AM  > To: Info-VAX@Mvb.Saic.Com . > Subject: Re: Is there any new Alpha CPU out? >I > Hello, > F > you are all right: MHz is not all. But it is marketing. Why does AMDB > and Intel fight the MHz war? AMD does not say we have the betterE > systembus, the better architecture implementation, the smaller chiplC > design. The did say: "We do have the fastestes CPU (MHz)." Now doaD > Intel say: "We do have the fastestes CPU (MHz)." The OpenVMSler do@ > know the truth, but do the normal PC user also know the truth.E > My question did have an other direction. I did see, that Compaq didtD > sold 1.25GHz Alphas to the NNSA, and so I would like to know, when > do they relise them. >n > Regards Rudolf Wingert   --   ------------------------------  % Date: Mon, 04 Sep 2000 15:01:44 +0200 = From: Arne =?iso-8859-1?Q?Vajh=F8j?= <arne.vajhoej@gtech.com> , Subject: Re: Is there any new Alpha CPU out?) Message-ID: <39B39D38.CD768CDD@gtech.com>   ( Andrew Harrison SUNUK Consultancy wrote:; > Do you know why. Well rather absurdly SPECint/SPECfp tend 9 > to show that the Alpha 21264 running at say 731+ Mhz isoA > as quick as its clock speed and that fact that it is a RISC CPU- > would lead you to expect.a > > > Now why you are going to say is this costing customers money? > and the answer is simple, because SPECint/fp are not at least@A > for the Alpha good predictors of commercial applications serverr9 > performance but people are being charged on that basis.. > : > Many of the major software vendors now charge by CPU and< > they charge more for faster CPU's. Some measure this using> > SPECint which grossly over estimates the capabilities of the@ > 21264 if the current crop of benchmark results are anything to@ > go on. Some and Oracle is one say real applications throughput@ > is higher for a given clock speed for RISC machines than Intel> > boxes (a rule that generally works for HP's, Sun's and IBM's8 > but does not for Alphas) so they charge on this basis.  < Everyone knows that commercial apps are usually more IO than CPU intensive.  B But what is your point ? That Compaq should let Alpha CPU's run at( half speed to be more like IBM and SUN ?  < (It is not Compaqs fault that their systems are so fast that' ISV's think they can charge too much !)$   Arne   ------------------------------  % Date: Mon, 04 Sep 2000 11:53:47 -0400i+ From: "Main, Kerry" <Kerry.Main@compaq.com>e, Subject: RE: Is there any new Alpha CPU out?J Message-ID: <910612C07BCAD1119AF40000F86AF0D80528475B@kaoexc4.kao.dec.com>   Andrew, Andrew ...  I As Arne pointed out, the Specint shows the native speed of the Alpha chipaJ sets and architecture. Software that is optimized for the Alpha can reallyH take advantage of this high speed. It is likely the reason why Alpha has4 been winning some of the recent supercomputer wins.   
 Reference:L <http://www.compaq.com/newsroom/pr/2000/0,1494,wp~14583_2!ob~33266_1_1,00.htJ ml> (U.S. Department of Energy Selects Compaq to Build World's Fastest andF Most Powerful Supercomputer - 375 GS320 systems with 600Tb of storage)L <http://www.compaq.com/newsroom/pr/2000/0,1494,wp~14583_2!ob~33046_1_1,00.htI ml> (National Science Foundation Selects Pittsburgh Supercomputing CenterlE and Compaq to Offer Scientists Access to World's Largest Non-Militaryw supercomputer)  L Note - you seem really high on using Oracle benchmarks. Thats great, becauseE based on some recent tidbits I heard, there is some really good stuff J happening with Oracle being much more optimized for the Alpha (OpenVMS and Tru64) architecture..   L Perhaps this is because of all the attention Alpha has been getting recentlyJ ? Perhaps its because the load on many Customers is going through the roof3 and they absolutely need very high end performance?    :-)m  L re: Oracle pricing .. Andrew, you stated that Oracle charges less than AlphaK systems for the lower speed Sun and HP boxes. Thats very true. However, anyyE comments on what Oracle is planning to charge for new SPARC III basedu	 systems? u  F I am of course assuming they will be considered RISC systems ...please correct me if they are not.r  @ re: your comments about Wildfire systems being slow and late ..   L Late, yes. Slow, nope. Nice try. The TPC benchmark you keep referring to wasJ an initial one that was put up for the announcement. As I am sure you knowK Andrew, it takes time to optimize the compilers to the fullest. Stay tuned.i :-)o  J It is also why I do not expect SPARC III TPC numbers for a good time after it is introduced either.  J I think the first url above for a DOE order for 375 GS320 Wildfire systemsL (approximately 12,000 Alpha cpus's) and 600Tb of storage speaks for itself.    Slow and late? -  J Hey - any update on the Oct '97 announcement for SPARC III (those in glass	 houses..) L http://www.sun.com/smi/Press/sunflash/9710/sunflash.971006.1.html (hopefullyH the specs for SPARC III have been updated since this press announcement)   :-)    Regards,  
 Kerry Main Senior Consultant,
 Compaq Canada  Professional Services  Voice : 613-592-4660 FAX   : 819-772-7036 Email : kerry.main@compaq.com        -----Original Message-----' From: Andrew Harrison SUNUK Consultancys! [mailto:andrew.nospam@uk.sun.com] ( Sent: Monday, September 04, 2000 7:24 AM To: Info-VAX@Mvb.Saic.Com), Subject: Re: Is there any new Alpha CPU out?    B Sorry Kerry I don't want to rain on your parade (well to be honest? I do) but the SPECint/fp advocacy you are engaged in is costing  your customers money.4  9 Do you know why. Well rather absurdly SPECint/SPECfp tend 7 to show that the Alpha 21264 running at say 731+ Mhz isa? as quick as its clock speed and that fact that it is a RISC CPUi would lead you to expect.l  < Now why you are going to say is this costing customers money= and the answer is simple, because SPECint/fp are not at leastr? for the Alpha good predictors of commercial applications server 7 performance but people are being charged on that basis.0  8 Many of the major software vendors now charge by CPU and: they charge more for faster CPU's. Some measure this using< SPECint which grossly over estimates the capabilities of the> 21264 if the current crop of benchmark results are anything to> go on. Some and Oracle is one say real applications throughput> is higher for a given clock speed for RISC machines than Intel< boxes (a rule that generally works for HP's, Sun's and IBM's6 but does not for Alphas) so they charge on this basis.   RISC systemt+ 100 dollars * (Mhz * 1.5 * Number of CPU's)n  	 Intel Boxv% 100 dollars * (Mhz * Number of CPU's)l  " http://www.oraclestore.oracle.com/  9 Now if I was you I would be arguing that people should be ; charged for Alpha using either the Intel model or a speciale
 Alpha one.  8 You see if you use Oracle applications benchmark results6 what you will see from the following table that Compaq5 WildFire customers are paying way, way more for their 9 licenses than Sun or even Intel  customers to ge the samen throughput.u  D These are based on the published Oracle apps benchmark results from.  A http://www.oracle.com/apps_benchmark/html/index.html?results.htmla  < And remember the benchmark has been developed to help people
 size systems.   G                         Oracle users/CPU      CPU Mhz      Oracle costss dollarst8 Sun                 488                              400 245,900oI CPQ/Alpha    350                               731                626,570 K CPQ/Intel      448                               500                223,210a  < The costs based on the cost for 2000 users using a perpetual@ power unit cost of 100 dollars. Remember the cost per processing8 unit is the same for each platform. You then get charged; maintenance as a % of license cost so this is an additionalc wammy.  ; Ironically if as you are advocating you subsituted relativet6 SPECint performance for Mhz the story would get worse.; If Oracle had used relative SPECint then the Alpha licencesu7 would have cost 3x the Sun's rather than 2.5 using Mhz.u  > So I guess you are between a rock and a hard place, you either; want to be in marketing and make marketing points that cost 9 you customers real money or you want to help them out and=A save them very substantial cash but you will  ruin your marketingu campaign in the mean time.  ? What you should be doing is lobbying Oracle to create a specialo9 category for Alpha servers which would have the followingn? formula. Even better for you would be to lobby Oracle to charge ; on the basis of their apps_server benchmarks but this would : mean everyone doing the benchmark and Compaq eating a huge: amount of humble pie neither of which I can see happening.   Alpha Server costc7 Fixed processing unit cost * Mhz * .5 * Number of CPU's   9 You should also be lobbying all the other vendors who uses< this method or relative SPECint to create a special case for4 Alpha boxes where you are allowed to discount either value.  5 As a half way house you could be lobbying vendors whos; charge on a relative SPECint basis to change their charging : to relative Mhz, it sounds dreadfull but it will save your customers money.  8 Most of all you should not be trying to defeat Intel Mhz7 arguments with SPECint results because its costing yourr# customers loads and loads of money.    Regardsb Andrew Harrisone Enterprise IT Architectl     "Main, Kerry" wrote:  	 > Rudolf,h >lH > You are correct in that marketing is a huge component of "the fastest" > cpu's. >eL > Here is the actual latest industry standard benchmarks for CPU, memory and > compiler optimizations: G > <http://ideasinternational.com/benchmark/spec/specfp2000.html> SpecFPyK > <http://ideasinternational.com/benchmark/spec/specint2000.html> Spec Int.d >tK > Also, keep in mind that the 1Ghz x86 designs are WS / dept server designsn" > only ie. single / dual cpu only. > K > The greater than 2 cpu x86 Server designs (Xeon etc) are still at 550 andp	 > 700Mhz.t >rL <http://commerce.www.ibm.com/cgi-bin/ncommerce/CategoryDisplay?cntrfnbr=1&cg4 > menbr=1&cgnbr=Netfinity+7600&cntry=840&lang=en_US> > 5 > <http://www.compaq.com/products/servers/platforms/>n > 
 > Regards, >e > Kerry Main > Senior Consultant, > Compaq Canadai > Professional Services  > Voice : 613-592-4660 > FAX   : 819-772-7036 > Email : kerry.main@compaq.comp >o > -----Original Message-----/ > From: Rudolf Wingert [mailto:win@FOM.FGAN.DE]e) > Sent: Thursday, August 31, 2000 2:45 AMr > To: Info-VAX@Mvb.Saic.ComT. > Subject: Re: Is there any new Alpha CPU out? >g > Hello, > F > you are all right: MHz is not all. But it is marketing. Why does AMDB > and Intel fight the MHz war? AMD does not say we have the betterE > systembus, the better architecture implementation, the smaller chiptC > design. The did say: "We do have the fastestes CPU (MHz)." Now do.D > Intel say: "We do have the fastestes CPU (MHz)." The OpenVMSler do@ > know the truth, but do the normal PC user also know the truth.E > My question did have an other direction. I did see, that Compaq did0D > sold 1.25GHz Alphas to the NNSA, and so I would like to know, when > do they relise them. >u > Regards Rudolf Wingert   --   ------------------------------  % Date: Mon, 04 Sep 2000 12:37:56 -0400'+ From: "Main, Kerry" <Kerry.Main@compaq.com>e, Subject: RE: Is there any new Alpha CPU out?J Message-ID: <910612C07BCAD1119AF40000F86AF0D80528475D@kaoexc4.kao.dec.com>   Andrew,   L For a company that likes to promte itself as "open", it would appear you areF running out of benchmarks to measure your systems with repect to other systems.  K First you criticize Spec numbers (which Sun is not even on top 20) and then I the whole issue of TPC benchmarks is criticized in the 156K TPC benchmarko& press release that Sun just released.   
 Reference:= http://ideasinternational.com/benchmark/spec/specint2000.html   
 Reference:F http://www.sun.com/smi/Press/sunflash/2000-08/sunflash.20000831.1.htmlK "Recently, the TPC-C benchmark has fallen under sharp criticism for failing I to keep pace with the times. Two separate cluster results reported by IBMoL and Compaq confirm that it can no longer be trusted as an accurate measuring for ad-hoc workloads.C  F It's well-understood in the technical communities that TPC-C no longerI represents current customer workloads since the transaction load that itssL models are made of are small, primitive and disconnected transactions. WhileL this model was acceptable for the workloads of the late 1980s, it misses theI mark for the object-based, integrated application environments of today'siF dot-com world that are part of the system design criteria at Sun." end quote.  L [Note - this 156Kb benchmark was also done with Sybase 12 (w Solaris7) - not Oracle..very interesting]p   :-)a   Regards,  
 Kerry Main Senior Consultant,
 Compaq Canadao Professional Servicesh Voice : 613-592-4660 FAX   : 819-772-7036 Email : kerry.main@compaq.coms       -----Original Message-----' From: Andrew Harrison SUNUK Consultancy ! [mailto:andrew.nospam@uk.sun.com]b( Sent: Monday, September 04, 2000 7:24 AM To: Info-VAX@Mvb.Saic.Com , Subject: Re: Is there any new Alpha CPU out?    B Sorry Kerry I don't want to rain on your parade (well to be honest? I do) but the SPECint/fp advocacy you are engaged in is costing  your customers money.P  9 Do you know why. Well rather absurdly SPECint/SPECfp tend 7 to show that the Alpha 21264 running at say 731+ Mhz iss? as quick as its clock speed and that fact that it is a RISC CPUc would lead you to expect.   < Now why you are going to say is this costing customers money= and the answer is simple, because SPECint/fp are not at leastc? for the Alpha good predictors of commercial applications servere7 performance but people are being charged on that basis.a  8 Many of the major software vendors now charge by CPU and: they charge more for faster CPU's. Some measure this using< SPECint which grossly over estimates the capabilities of the> 21264 if the current crop of benchmark results are anything to> go on. Some and Oracle is one say real applications throughput> is higher for a given clock speed for RISC machines than Intel< boxes (a rule that generally works for HP's, Sun's and IBM's6 but does not for Alphas) so they charge on this basis.   RISC system + 100 dollars * (Mhz * 1.5 * Number of CPU's)d  	 Intel BoxO% 100 dollars * (Mhz * Number of CPU's)   " http://www.oraclestore.oracle.com/  9 Now if I was you I would be arguing that people should be ; charged for Alpha using either the Intel model or a special 
 Alpha one.  8 You see if you use Oracle applications benchmark results6 what you will see from the following table that Compaq5 WildFire customers are paying way, way more for their 9 licenses than Sun or even Intel  customers to ge the sameo throughput.a  D These are based on the published Oracle apps benchmark results from.  A http://www.oracle.com/apps_benchmark/html/index.html?results.html,  < And remember the benchmark has been developed to help people
 size systems.   G                         Oracle users/CPU      CPU Mhz      Oracle costsc dollarsc8 Sun                 488                              400 245,900rI CPQ/Alpha    350                               731                626,570fK CPQ/Intel      448                               500                223,210s  < The costs based on the cost for 2000 users using a perpetual@ power unit cost of 100 dollars. Remember the cost per processing8 unit is the same for each platform. You then get charged; maintenance as a % of license cost so this is an additionalc wammy.  ; Ironically if as you are advocating you subsituted relatives6 SPECint performance for Mhz the story would get worse.; If Oracle had used relative SPECint then the Alpha licencesn7 would have cost 3x the Sun's rather than 2.5 using Mhz./  > So I guess you are between a rock and a hard place, you either; want to be in marketing and make marketing points that cost 9 you customers real money or you want to help them out and A save them very substantial cash but you will  ruin your marketingf campaign in the mean time.  ? What you should be doing is lobbying Oracle to create a specialn9 category for Alpha servers which would have the followingo? formula. Even better for you would be to lobby Oracle to charge,; on the basis of their apps_server benchmarks but this wouldI: mean everyone doing the benchmark and Compaq eating a huge: amount of humble pie neither of which I can see happening.   Alpha Server costh7 Fixed processing unit cost * Mhz * .5 * Number of CPU'su  9 You should also be lobbying all the other vendors who use < this method or relative SPECint to create a special case for4 Alpha boxes where you are allowed to discount either value.  5 As a half way house you could be lobbying vendors whoe; charge on a relative SPECint basis to change their chargingt: to relative Mhz, it sounds dreadfull but it will save your customers money.  8 Most of all you should not be trying to defeat Intel Mhz7 arguments with SPECint results because its costing your-# customers loads and loads of money.0   Regards  Andrew Harrison? Enterprise IT Architectj     "Main, Kerry" wrote:  	 > Rudolf,  > H > You are correct in that marketing is a huge component of "the fastest" > cpu's. >sL > Here is the actual latest industry standard benchmarks for CPU, memory and > compiler optimizations:2G > <http://ideasinternational.com/benchmark/spec/specfp2000.html> SpecFPcK > <http://ideasinternational.com/benchmark/spec/specint2000.html> Spec Int.  >nK > Also, keep in mind that the 1Ghz x86 designs are WS / dept server designs " > only ie. single / dual cpu only. > K > The greater than 2 cpu x86 Server designs (Xeon etc) are still at 550 ande	 > 700Mhz.i >cL <http://commerce.www.ibm.com/cgi-bin/ncommerce/CategoryDisplay?cntrfnbr=1&cg4 > menbr=1&cgnbr=Netfinity+7600&cntry=840&lang=en_US> >r5 > <http://www.compaq.com/products/servers/platforms/>e >p
 > Regards, >e > Kerry Main > Senior Consultant, > Compaq Canadae > Professional Servicesm > Voice : 613-592-4660 > FAX   : 819-772-7036 > Email : kerry.main@compaq.comc >p > -----Original Message-----/ > From: Rudolf Wingert [mailto:win@FOM.FGAN.DE],) > Sent: Thursday, August 31, 2000 2:45 AM) > To: Info-VAX@Mvb.Saic.Com.. > Subject: Re: Is there any new Alpha CPU out? >  > Hello, > F > you are all right: MHz is not all. But it is marketing. Why does AMDB > and Intel fight the MHz war? AMD does not say we have the betterE > systembus, the better architecture implementation, the smaller chipnC > design. The did say: "We do have the fastestes CPU (MHz)." Now doeD > Intel say: "We do have the fastestes CPU (MHz)." The OpenVMSler do@ > know the truth, but do the normal PC user also know the truth.E > My question did have an other direction. I did see, that Compaq didnD > sold 1.25GHz Alphas to the NNSA, and so I would like to know, when > do they relise them. >e > Regards Rudolf Wingert   --   ------------------------------  % Date: Mon, 04 Sep 2000 13:21:26 -0400.+ From: "Main, Kerry" <Kerry.Main@compaq.com>w2 Subject: RE: Oracle/VMS (in the US) teleconferenceJ Message-ID: <910612C07BCAD1119AF40000F86AF0D80528475E@kaoexc4.kao.dec.com>  
 G'day Ken,   Small world eh ?   :-)y  K The Oracle 11.5 App Server will not be on OpenVMS, but Oracles new strategy $ for APP Servers called IAS will be.   J The Oracle folks will need to confirm what the future strategy for 11.5 is% with respect to the new IAS strategy.s  
 Reference:B http://www.zdnet.com/eweek/stories/general/0,11011,2596660,00.htmlK http://platforms.oracle.com/com/index_com.htm (IAS on all Compaq platforms)i   Regards,  
 Kerry Main Senior Consultant,
 Compaq Canadae Professional Servicesr Voice : 613-592-4660 FAX   : 819-772-7036 Email : kerry.main@compaq.come       -----Original Message-----8 From: forank@DFO-MPO.GC.CA [mailto:forank@DFO-MPO.GC.CA]( Sent: Friday, September 01, 2000 3:32 PM To: Info-VAX@Mvb.Saic.Como2 Subject: RE: Oracle/VMS (in the US) teleconference    L I tuned in to the teleconference and it was quite interesting.  It concernedL the Oracle e-business suite iAS on VMS.  Unfortunately from another source IJ found that while Oracle has V11.3 on VMS they do not plan to port V11.5 toH VMS.  I have not been able to confirm this one way or another.  If V11.5D will not be ported to VMS then in the long term the viability of theE e-business of Oracle on VMS will be dead.  Does anyone have any otherg information?   Ken    	-----Original Message-----.5 	From:	Dave Gudewicz [SMTP:david.gudewicz@abbott.com]l) 	Sent:	Thursday, August 31, 2000 11:24 AMA 	To:	Info-VAX@Mvb.Saic.Com/ 	Subject:	Oracle/VMS (in the US) teleconferencef  9 	Anyone tune in to this week's Oracle/VMS teleconference?S  	 	Commentsn   	Dave...   	    ------------------------------  $ Date: Mon, 4 Sep 2000 18:51:01 +0100- From: "Rob van Lopik" <lopik@mail.telepac.pt>o2 Subject: Re: Oracle/VMS (in the US) teleconference+ Message-ID: <8p0ncd$ake$1@venus.telepac.pt>   6 "Main, Kerry" <Kerry.Main@compaq.com> wrote in messageD news:910612C07BCAD1119AF40000F86AF0D80528475E@kaoexc4.kao.dec.com... > G'day Ken, >h > Small world eh ? >c > :-)P > D > The Oracle 11.5 App Server will not be on OpenVMS, but Oracles new strategy% > for APP Servers called IAS will be.d  L These are two different kind of things. The first one is Oracle ApplicationsH (Oracle's ERP package) that will not be continued on OpenVMS. The secondJ used te be called Oracle Application Server and is now renamed to InternetI Application Server. This some middleware server, not really a transactionrI server. It will be ported to OopenVMS (actually Oracle application Serverd% version 3.x is available for OpenVMS.C  
 rob van lopiko   ------------------------------  % Date: Mon, 04 Sep 2000 07:15:16 -0700l5 From: Richard  <maher_rjNOmaSPAM@hotmail.com.invalid>e Subject: Re: RTR and DECdtm39 Message-ID: <0f7e580b.619cf19e@usw-ex0108-062.remarq.com>i   Hi Jim,   A Please get in touch with John Apps, if you havenâ€™t done soh7 already. I know that he would be very interested in thea5 information that you have provided here regarding thee5 effort involved in upgrading DECdtm. Especially as it 1 appears that he has been previously supplied withy: questionable reports, of varying veracity, on this subject: from other quarters.  I hope that he is also interested in5 your availability. (Whatever happened to Alan Potter? 6 Porter?) John is a good bloke to talk to and I hope he= wonâ€™t be too annoyed with me dropping his name here oroE identifying him as being on the â€œIf it ainâ€™t RTR it just K ainâ€™t!â€ side of the fence. (For example, â€œJohn, you needrK TIP if youâ€™re ever going to have COM+/DCOM on VMS!â€ â€œNo,1* Iâ€™m sure RTR could do that!â€)  6 I hope you/someone can find a different angle into VMS; management so that DECdtm can receive the upgrade budget itl: so desperately needs and deserves!  I canâ€™t imagine; watching the RTR tail wagging a VMS dog in a world that hasc8 gone TIP and XA. Unfortunately, Rich Roscoe â€“ whoA undoubtedly knows what heâ€™s talking about when it comes to > implementing a TIP compatible product on VMS â€“ was last< seen banging his head against a brick wall saying  â€œOfH course TIP belongs in DECdtm! Why canâ€™t you people see it?â€; Did he end up leaving? This is purely my version of his ownb9 opinion, but Rich did say that he thought the engineering 6 effort to make DECdtm TIP compliant would be 3 months!  ? And donâ€™t even bother trying to tell Rich Marcello about ; what Microsoft, let alone Tandem, is doing with MTS and TIP ? because it appears that once again â€œDigital is a software < company!â€ (I think *not*) Windows 2000 teller machinesA backed up by fault tolerant Tandem TIP compliant widgets â€“d7 sounds like a pretty good combination. Why was it againe that Compaq needed VMS?h  6 I really thought that Oracle Rdb would save the day by: rolling up to VMS engineering with a fat cheque to get the9 ball rolling!  After squandering tens, if not hundred(s), ; of millions (How much was it Norm?) of Rdb/*V.M.Sâ€™s*nA license fees on Rdb/NT and Rdb/DUNIX youâ€™d think that selfn7 preservation, let alone embarrassment/shame, would have : sent them racing to build bridges from the island that hasE become Rdb/VMS.  Unfortunately â€œSplendid Isolationâ€ seemsp5 to be de rigueur over at OracleRdb and if you look at = attachment (1) youâ€™ll see that at least one senior Rdby7 engineer believes that Rdb engineering has already doneo: everything humanly possible to further the cause of DECdtm$ and is happy to rest on its laurels.  @ Regarding JF Mezeiâ€™s reply, I have to say that it saddens: me to see that the Menendez spirit is still alive and well: in the VMS layered product crÃ¨che. Let me make two very quick points: -n  I 1)	RTR is a â€œproprietaryâ€ solution! (Unless youâ€™ve got 9 a different dictionary or overnight Microsoft, Sun or IBM : started selling their version of RTR!) If you want RTR you7 must buy it from Compaq and by all the definitions that B Iâ€™ve seen, that makes it proprietary! If you want to use an8 RTR solution then RTR must exist (be purchased) on every machine/os involved the txn.6 2)	It is ludicrous to suggest that Compaq says this or; Compaq thinks that, let alone that Compaq is suffering from 9 conviction on the issue of txn management. Please confineu; such claims to VMS engineering/management or Rich Marcello.o9 What Compaq does do is make a whole lot of ProLiants that 9 run Windows2000 which contains MTS/DTC which, you guessed 9 it, is TIP compatible.  Compaq also makes Tandem machinesu9 that also talk TIP and curiously enough Keith B Evans who 6 was one of the authors of this open standard works for; Compaq. Can you see a pattern developing here? What we, thei3 loyal VMS users, are being left with is a backwaterc5 database and a backwater operating system governed byg7 backwater management. This may explain how a tiny self-l9 serving myopic engineering group could presume to be in a : position to control the destiny of VMS. How can we get the; message across to those poor misguided fools out there that=6 are using VisualBasic and MTS on Windows2000 that Rich. Marcello expects them to use DECforms and RTR?  5 But I think JF is right about one thing. RTR would berB better off being spun off into itâ€™s own company or sold off1 to someone else! I can see its true potential andr: development being stifled by the bureaucracy at Compaq and? the legacy of VMS. Look at DMQ and BEAâ€™s share price. OMu9 Gruppen appears to be flushed with cash at the moment andc: uses a lot of RTR. The time may be right! The revenue from/ the RTR sale could then go straight into DECdtme3 development. Sounds like a win-win situation to me!D  G What I donâ€™t want to see is VMS usersâ€™ license fees beinge0 diverted to the RTR (NT, MVS, Solaris, IMac . .)7 development/slush fund. And when nobody will buy RTR on > these platforms then obviously itâ€™ll be given away as aN loss leader. Hey, itâ€™s only VMS usersâ€™ money. Theyâ€™ll payH for anything. Donâ€™t worry about it.  Weâ€™ve been doing this
 for years!  L Whatâ€™s that other â€œstrategicâ€ product DCE RPC doing aboutA transactional RPC? Let me guess: â€œRTR could do that!â€0  3 Here are a couple of pages that may be of interest:M  < http://www.openvms.compaq.com/commercial/acms/htm_docs/tip.h tm  9 http://www.openvms.compaq.com/commercial/acms/1000new.htm.  5 I like the following because it contains the heading:e  / â€œCompaqâ€™s leadership in TIPâ€   6 http://www.tandem.com/info/inform/issues/issue27/ha02- integrate_nt-ns.htms  ) If the one above doesnâ€™t work try:t  , http://himalaya.compaq.com/view.asp?ioid=549  ; In fact, go anywhere on the Tandem site and do a search fore; Transaction Internet Protocol. Thereâ€™s heaps of infoe
 available.  < http://msdn.microsoft.com/library/psdk/cossdk/pgfuncref_8ylq htms   Regards Richard Maher.  : (Apologies to John Apps and Rich Roscoe for dropping their< names here without checking first, but this is dragging on.)   ATTACHMENT (1)  9 Unofficial Rdb position: (The first bit is part of what I - placed in the Rdb List server on August 17th)u   > If Rdb lifted a finger: > to help with a *decent* XA version of DECdtm then we may be getting somewhere	 > on VMS.t  ; By "lift a finger" you don't mean the 10 years of effort wes have put into DECdtm2 validation and testing or our many discussions and suggestions for XA on OpenVMS  through DECdtm.l  ; Please talk from fact and not imagination.  It neither does  you nor us any goodn* to make such statements on the listserver.  8 The fact that we can't convince COMPAQ to invest in that interface is probablyw9 because there are no customers (who pay money) asking fory such functionality9 from COMPAQ.  We would be only too willing to help COMPAQn if they asked.   --: Ian Smith                               Read the Technical
 Corner columna6 Oracle Rdb Engineering Group            in the Rdb Web JournalU email: Ian.E.Smith@oracle.comw http://www.oracle.com/rdb 2 (Standard disclaimer:  The statements and opinions expressed here are my own andr9 do not necessarily represent those of Oracle Corporation)a       * Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautifuls   ------------------------------  % Date: Mon, 04 Sep 2000 07:41:44 -0700+5 From: Richard  <maher_rjNOmaSPAM@hotmail.com.invalid>t Subject: Re: RTR and DECdtmd9 Message-ID: <0ca0b8b7.6884698e@usw-ex0108-062.remarq.com>s   Hi Jim,e  A Please get in touch with John Apps, if you havenâ€™t done sos7 already. I know that he would be very interested in theo5 information that you have provided here regarding the 5 effort involved in upgrading DECdtm. Especially as itc1 appears that he has been previously supplied with.: questionable reports, of varying veracity, on this subject: from other quarters.  I hope that he is also interested in5 your availability. (Whatever happened to Alan Potter?i6 Porter?) John is a good bloke to talk to and I hope he= wonâ€™t be too annoyed with me dropping his name here orh? identifying him as being on the "If it ainâ€™t RTR it justp? ainâ€™t!" side of the fence. (For example, "John, you need ? TIP if youâ€™re ever going to have COM+/DCOM on VMS!" "No,*$ Iâ€™m sure RTR could do that!")  6 I hope you/someone can find a different angle into VMS; management so that DECdtm can receive the upgrade budget itt: so desperately needs and deserves!  I canâ€™t imagine; watching the RTR tail wagging a VMS dog in a world that hase1 gone TIP and XA. Unfortunately, Rich Roscoe - whoeA undoubtedly knows what heâ€™s talking about when it comes toc7 implementing a TIP compatible product on VMS - was last 6 seen banging his head against a brick wall saying  "OfB course TIP belongs in DECdtm! Why canâ€™t you people see it?"; Did he end up leaving? This is purely my version of his owne9 opinion, but Rich did say that he thought the engineeringU6 effort to make DECdtm TIP compliant would be 3 months!  ? And donâ€™t even bother trying to tell Rich Marcello about ; what Microsoft, let alone Tandem, is doing with MTS and TIPf9 because it appears that once again "Digital is a software ; company!" (I think not) Windows 2000 teller machines backede: up by fault tolerant Tandem TIP compliant widgets - sounds5 like a pretty good combination. Why was it again thatl Compaq needed VMS?  6 I really thought that Oracle Rdb would save the day by: rolling up to VMS engineering with a fat cheque to get the9 ball rolling!  After squandering tens, if not hundred(s),t; of millions (How much was it Norm?) of Rdb/*V.M.Sâ€™s*nA license fees on Rdb/NT and Rdb/DUNIX youâ€™d think that selfu7 preservation, let alone embarrassment/shame, would haved: sent them racing to build bridges from the island that has9 become Rdb/VMS.  Unfortunately "Splendid Isolation" seemst5 to be de rigueur over at OracleRdb and if you look atu= attachment (1) youâ€™ll see that at least one senior Rdb : engineer believes that Rdb engineering has done everything6 humanly possible to further the cause of DECdtm and is happy to rest on its laurels.f  @ Regarding JF Mezeiâ€™s reply, I have to say that it saddens: me to see that the Menendez spirit is still alive and well: in the VMS layered product crÃ¨che. Let me make two very quick points: -   = 1)	RTR is a "proprietary" solution! (Unless youâ€™ve goto9 a different dictionary or overnight Microsoft, Sun or IBMy: started selling their version of RTR!) If you want RTR you7 must buy it from Compaq and by all the definitions thatoB Iâ€™ve seen, that makes it proprietary! If you want to use an8 RTR solution then RTR must exist (be purchased) on every machine/os involved the txn.  6 2)	It is ludicrous to suggest that Compaq says this or; Compaq thinks that, let alone that Compaq is suffering fromi9 conviction on the issue of txn management. Please confineu; such claims to VMS engineering/management or Rich Marcello. 9 What Compaq does do is make a whole lot of ProLiants thatM9 run Windows2000 which contains MTS/DTC which, you guessedt9 it, is TIP compatible.  Compaq also makes Tandem machines.9 that also talk TIP and curiously enough Keith B Evans whor6 was one of the authors of this open standard works for; Compaq. Can you see a pattern developing here? What we, theS3 loyal VMS users, are being left with is a backwatere5 database and a backwater operating system governed byd7 backwater management. This may explain how a tiny self- 9 serving myopic engineering group could presume to be in a2: position to control the destiny of VMS. How can we get the; message across to those poor misguided fools out there thatC6 are using VisualBasic and MTS on Windows2000 that Rich. Marcello expects them to use DECforms and RTR?  5 But I think JF is right about one thing. RTR would beeB better off being spun off into itâ€™s own company or sold off1 to someone else! I can see its true potential and-: development being stifled by the bureaucracy at Compaq and? the legacy of VMS. Look at DMQ and BEAâ€™s share price. OM,9 Gruppen appears to be flushed with cash at the moment andX: uses a lot of RTR. The time may be right! The revenue from/ the RTR sale could then go straight into DECdtm 3 development. Sounds like a win-win situation to me!e  G What I donâ€™t want to see is VMS usersâ€™ license fees beingb0 diverted to the RTR (NT, MVS, Solaris, IMac . .)7 development/slush fund. And when nobody will buy RTR on(> these platforms then obviously itâ€™ll be given away as aN loss leader. Hey, itâ€™s only VMS usersâ€™ money. Theyâ€™ll payH for anything. Donâ€™t worry about it.  Weâ€™ve been doing this
 for years!  @ Whatâ€™s that other "strategic" product DCE RPC doing about5 transactional RPC? Let me guess: "RTR could do that!"   3 Here are a couple of pages that may be of interest:0  < http://www.openvms.compaq.com/commercial/acms/htm_docs/tip.h tm  9 http://www.openvms.compaq.com/commercial/acms/1000new.htm   5 I like the following because it contains the heading:.  # "Compaqâ€™s leadership in TIP"   6 http://www.tandem.com/info/inform/issues/issue27/ha02- integrate_nt-ns.htmr  ) If the one above doesnâ€™t work try:   , http://himalaya.compaq.com/view.asp?ioid=549  ; In fact, go anywhere on the Tandem site and do a search fore; Transaction Internet Protocol. Thereâ€™s heaps of infol
 available.  < http://msdn.microsoft.com/library/psdk/cossdk/pgfuncref_8ylq html   Regards Richard Maher.  : (Apologies to John Apps and Rich Roscoe for dropping their< names here without checking first, but this is dragging on.)   ATTACHMENT (1)  9 Unofficial Rdb position: (The first bit is part of what Ik- placed in the Rdb List server on August 17th)e   > If Rdb lifted a finger: > to help with a *decent* XA version of DECdtm then we may be getting somewhere	 > on VMS.   ; By "lift a finger" you don't mean the 10 years of effort we. have put into DECdtm2 validation and testing or our many discussions and suggestions for XA on OpenVMSw through DECdtm.1  ; Please talk from fact and not imagination.  It neither doesb you nor us any goodh* to make such statements on the listserver.  8 The fact that we can't convince COMPAQ to invest in that interface is probably 9 because there are no customers (who pay money) asking for? such functionality9 from COMPAQ.  We would be only too willing to help COMPAQw if they asked.   --: Ian Smith                               Read the Technical
 Corner columnt6 Oracle Rdb Engineering Group            in the Rdb Web Journali email: Ian.E.Smith@oracle.comn http://www.oracle.com/rdb.2 (Standard disclaimer:  The statements and opinions expressed here are my own and 9 do not necessarily represent those of Oracle Corporation):       * Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is BeautifulD   ------------------------------  % Date: Mon, 04 Sep 2000 08:19:08 -0700l5 From: Richard  <maher_rjNOmaSPAM@hotmail.com.invalid>v Subject: Re: RTR and DECdtmt9 Message-ID: <2d56739d.7246240a@usw-ex0108-062.remarq.com>e  6 It's all Greek to me :-( If anyone knows why my single4 quotes are coming out like that then please tell me.     Hi Jim,o  : Please get in touch with John Apps, if you haven't done so7 already. I know that he would be very interested in the 5 information that you have provided here regarding thea5 effort involved in upgrading DECdtm. Especially as it 1 appears that he has been previously supplied withe: questionable reports, of varying veracity, on this subject: from other quarters.  I hope that he is also interested in5 your availability. (Whatever happened to Alan Potter?R6 Porter?) John is a good bloke to talk to and I hope he6 won't be too annoyed with me dropping his name here or8 identifying him as being on the "If it ain't RTR it just8 ain't!" side of the fence. (For example, "John, you need8 TIP if you're ever going to have COM+/DCOM on VMS!" "No, I'm sure RTR could do that!")c  6 I hope you/someone can find a different angle into VMS; management so that DECdtm can receive the upgrade budget itb3 so desperately needs and deserves!  I can't imaginet; watching the RTR tail wagging a VMS dog in a world that hast1 gone TIP and XA. Unfortunately, Rich Roscoe - whoI: undoubtedly knows what he's talking about when it comes to7 implementing a TIP compatible product on VMS - was lastt6 seen banging his head against a brick wall saying  "Of; course TIP belongs in DECdtm! Why can't you people see it?"; Did he end up leaving? This is purely my version of his owno9 opinion, but Rich did say that he thought the engineeringu6 effort to make DECdtm TIP compliant would be 3 months!  8 And don't even bother trying to tell Rich Marcello about; what Microsoft, let alone Tandem, is doing with MTS and TIPt9 because it appears that once again "Digital is a softwareP; company!" (I think not) Windows 2000 teller machines backed : up by fault tolerant Tandem TIP compliant widgets - sounds5 like a pretty good combination. Why was it again thata Compaq needed VMS?  6 I really thought that Oracle Rdb would save the day by: rolling up to VMS engineering with a fat cheque to get the9 ball rolling!  After squandering tens, if not hundred(s),w4 of millions (How much was it Norm?) of Rdb/*V.M.S's*: license fees on Rdb/NT and Rdb/DUNIX you'd think that self7 preservation, let alone embarrassment/shame, would haveM: sent them racing to build bridges from the island that has9 become Rdb/VMS.  Unfortunately "Splendid Isolation" seems5 to be de rigueur over at OracleRdb and if you look at 6 attachment (1) you'll see that at least one senior Rdb: engineer believes that Rdb engineering has done everything6 humanly possible to further the cause of DECdtm and is happy to rest on its laurels.a  9 Regarding JF Mezei's reply, I have to say that it saddensr: me to see that the Menendez spirit is still alive and well7 in the VMS layered product crche. Let me make two verya quick points: -b  8 1) RTR is a "proprietary" solution! (Unless you've got a7 different dictionary or overnight Microsoft, Sun or IBMr: started selling their version of RTR!) If you want RTR you7 must buy it from Compaq and by all the definitions thaty; I've seen, that makes it proprietary! If you want to use ane8 RTR solution then RTR must exist (be purchased) on every machine/os involved the txn.  6 2) It is ludicrous to suggest that Compaq says this or; Compaq thinks that, let alone that Compaq is suffering fromo9 conviction on the issue of txn management. Please confine ; such claims to VMS engineering/management or Rich Marcello. 9 What Compaq does do is make a whole lot of ProLiants thate9 run Windows2000 which contains MTS/DTC which, you guessed 9 it, is TIP compatible.  Compaq also makes Tandem machinesn9 that also talk TIP and curiously enough Keith B Evans whor6 was one of the authors of this open standard works for; Compaq. Can you see a pattern developing here? What we, the 3 loyal VMS users, are being left with is a backwaterm5 database and a backwater operating system governed byn7 backwater management. This may explain how a tiny self-i9 serving myopic engineering group could presume to be in a : position to control the destiny of VMS. How can we get the; message across to those poor misguided fools out there thate6 are using VisualBasic and MTS on Windows2000 that Rich. Marcello expects them to use DECforms and RTR?  5 But I think JF is right about one thing. RTR would bet; better off being spun off into it's own company or sold offp1 to someone else! I can see its true potential anda: development being stifled by the bureaucracy at Compaq and8 the legacy of VMS. Look at DMQ and BEA's share price. OM9 Gruppen appears to be flushed with cash at the moment ande: uses a lot of RTR. The time may be right! The revenue from/ the RTR sale could then go straight into DECdtmt3 development. Sounds like a win-win situation to me!e  9 What I don't want to see is VMS users' license fees beingo0 diverted to the RTR (NT, MVS, Solaris, IMac . .)7 development/slush fund. And when nobody will buy RTR on 7 these platforms then obviously it'll be given away as al9 loss leader. Hey, it's only VMS users' money. They'll payi: for anything. Don't worry about it.  We've been doing this
 for years!  9 What's that other "strategic" product DCE RPC doing aboute5 transactional RPC? Let me guess: "RTR could do that!"t  3 Here are a couple of pages that may be of interest:t  < http://www.openvms.compaq.com/commercial/acms/htm_docs/tip.h tm  9 http://www.openvms.compaq.com/commercial/acms/1000new.htma  5 I like the following because it contains the heading:s   "Compaq's leadership in TIP"  6 http://www.tandem.com/info/inform/issues/issue27/ha02- integrate_nt-ns.htms  " If the one above doesn't work try:  , http://himalaya.compaq.com/view.asp?ioid=549  ; In fact, go anywhere on the Tandem site and do a search forh4 Transaction Internet Protocol. There's heaps of info
 available.  < http://msdn.microsoft.com/library/psdk/cossdk/pgfuncref_8ylq html   Regards Richard Maher.  : (Apologies to John Apps and Rich Roscoe for dropping their< names here without checking first, but this is dragging on.)   ATTACHMENT (1)  9 Unofficial Rdb position: (The first bit is part of what I!- placed in the Rdb List server on August 17th)    > If Rdb lifted a finger: > to help with a *decent* XA version of DECdtm then we may be getting somewhere	 > on VMS.m  ; By "lift a finger" you don't mean the 10 years of effort wet have put into DECdtm2 validation and testing or our many discussions and suggestions for XA on OpenVMSr through DECdtm.h  ; Please talk from fact and not imagination.  It neither doeso you nor us any goodm* to make such statements on the listserver.  8 The fact that we can't convince COMPAQ to invest in that interface is probablyo9 because there are no customers (who pay money) asking for. such functionality9 from COMPAQ.  We would be only too willing to help COMPAQ. if they asked.   --: Ian Smith                               Read the Technical
 Corner column,6 Oracle Rdb Engineering Group            in the Rdb Web Journals email: Ian.E.Smith@oracle.comf http://www.oracle.com/rdb 2 (Standard disclaimer:  The statements and opinions expressed here are my own and 9 do not necessarily represent those of Oracle Corporation)n       * Sent from AltaVista http://www.altavista.com Where you can also find related Web Pages, Images, Audios, Videos, News, and Shopping.  Smart is Beautiful    ------------------------------  % Date: Mon, 04 Sep 2000 10:01:09 -0400o) From: yyyc186.illegaltospam_@flashcom.net  Subject: srm pc164 odditiese9 Message-ID: <39b3abef$1$lllp186$mr2ice@news.flashcom.com>n   All,  I Have some oddities with the SRM console and my Alpha pc164.  When the ARCaE console is installed I can access both my SCSI CD rom and SCSI drive,iJ partition the drive, etc.  When I install the SRM console so I can installJ a USABLE OS it cannot find the Can't-Adapt-Crap 2940UW controller.  I haveE also noticed that it doesn't swat the Diamond-Stealth video BIOS intow displaying either.  J Is there a console environment variable I must set to tell it Plug&Pray isE on this system?  Not much doc about this in the download from Compaq.   G Would really hate to have to install IDE components just to install VMSt and boot it.   Roland   -- i; ------------------------------------------------------------D yyyc186@flashcom.net              To Respond delete ".illegaltospam"6                             MR/2 Internet Cruiser 1.528                             For a Microsoft free univers; -----------------------------------------------------------h   ------------------------------  % Date: Mon, 04 Sep 2000 10:22:07 +0100e/ From: Nigel Arnot <sysmgr@maxwell.ph.kcl.ac.uk>v* Subject: Re: Sun Hardware problems persist6 Message-ID: <009EF9E2.70F8C2AD.4@maxwell.ph.kcl.ac.uk>   > Dirk Munk wrote:E > > I once had a similar problem. No airco and the temperature in theb > > computerroom was > 40C !E > > RJ45 plugs were soft etc, but the Alphas kept running, with theire7 > > temperature alarm beepers making a lot of noice ...  > O > Tandem had/has machines called "Cyclones".  They were noisy beasts because ofnM > all the fans they had. But the tech had told me that because of those fans,dI > the machines resist temperature "extremes" much better since the higheroD > ventilation volume allows more than adequate cooling even when air > conditioning fails.o > N > Remember that heat buildup is what is dangerous. As long as the chips aren'tP > allowed to get too hot, you're OK. And I suspect that chips internally sustainF > temperatures that are significantly higher than the recommended roomD > temperature for computers (unless the later is poorly ventilated).  F Yes, but that's irrelevant. To a first approximation, for a particularJ chip and heatsink in operation, the heatsink will be X degrees hotter thanJ the air passing through it, and the chip will be Y degrees hotter than theJ heatsink. If the air gets Z degrees hotter, so does the chip. Pumping moreJ air reduces X, and gains some extra safety margin at the expense of noise.L I imagine that Tandem owners care less about machine-room noise than abilityF to ride-out a partial or temporary air-con failure without a shutdown.  J As many folks have observed, most manufacturers' boxen will work fine withL air temperature pushing 30C and many have been observed to continue doing soK at 40C or even 50C. Incidentally CMOS does not suffer thermal run-away likelK TTL used to; a CMOS chip that crashes because it gets too hot will normallyyK work fine once it cools down again. (Of course thermal abuse sometimes willAL cause permanent damage, it will shorten component life expectancy, and is in general a BAD THING. s > N > I beleive disk drives are also in the same case. You need proper ventilation( > or they will fail due to heat buildup.  K Similar, but there's another factor. Above a certain temperature the disk's M seek algorithm may start malfunctioning. The disk seeks, finds by reading theaN servo track that it's not quite in the right place, and has to do a correctionL re-seek. The result is that your disk drives appear to have suddenly startedF operating at half-performance. I'm told that this happens at about theC same temperature at which the staff also reach half-performance :-)   G And another, a disk has moving parts on bearings. Extended operation atrK over-temperature will shorten those bearings' life and reduce MTBF to quitep a noticeable extent.   	Yours,
 		Nigel Arnot - 		NRA@MAXWELL.PH.KCL.AC.UK                   S  7 		"In the beginning there was nothing, which exploded."a   ------------------------------  % Date: Mon, 04 Sep 2000 12:12:59 +0100B From: Andrew Harrison SUNUK Consultancy <andrew.nospam@uk.sun.com>* Subject: Re: Sun Hardware problems persist* Message-ID: <39B383BB.801BF18D@uk.sun.com>   Rob Young wrote:     >hD >         For those of us that may be getting a bit turned around byB >         your constant twisting and turning in the wind I need to >         recap: >oH >                 1)  In December 1999 you claimed the CPU problems thatM >                     Gartner wrote an advisory about were "old news" and Sun M >                     actually told them about them.  They were fixed.  Yeah!a > H >                 2)  You now claim the [majority] of these problems areF >                     environmental in nature.  This is absolutely notJ >                     the case as the Gartner folks are quick to point outD >                     and which you consistently seem to overlook in1 >                     your very LAME refutations.  >p" >         To recap for the masses: >eL >         1) and 2) above are clearly incorrect views.  They are clearly not >         the case.s >dC >         And yes we can read and yes we have re-read the articles.0 >a. >         One data point does not a case make! >TB >         Little to do with environmentals.  I hang on the Gartner? >         statement and I also trust these folks know about thes >         environmental issues:o >tO > My group runs a mid-size to large server farm at a (very) major ISP and we'reoQ > constantly replacing processors that crash due to "Ecache Writeback Data Parity P > Errors." We were told that even replacing the processor isn't a remedy becauseN > the new one is just as likely to eat itself as the old one if you left it in > place (As far as they know.) >hQ > In addition, the latest fix is a software patch that is supposed to massage the P > Ecache so that it never finds itself in the condition that they believe causesQ > the error. Remember, they're still guessing at this point. 18 months later. HowlM > many of those 400Mhz are now used up with self-checks and Ecache scrubbing?n >i  L Blimey Rob do you ever learn. The software patch impliments memory scrubbingS for the e-Cache, this is hardly revolutionary or strange. Most large systems that ItN know of use this for main memory and have done for some time, S390's, HP's andN Sun's all scrub main memory. The Sun patch impliments this for e-Cache and forI all I know Alpha's running OpenVMS/Tru64 may have this facility allready.l  I As a matter of interest though Rob don't you get pretty bored of this andPG dispirited. Lets just examine your track record on OpenVMS boosting andtI Sun FUD which seem sadly for you to go hand in glove. I say sadly becauseoE you accuse me of missrepresentation on spin when your track record isn! truly dreadfull in both respects.e  
 First the FUDe  F eBay well do I have to say more, well OK I will, you jumped on the FUDE Sun over eBays outages bandwagon. Sadly all your speculation over theiF eBay 22 hour outage and attempts to suggest that it was a E10K/SolarisM was totally wrong. The final analysis on this one is long gone and the outages wasn't caused by either.  D You suspected a conspiracy because Sun wouldn't  say what caused theD outage when in fact from a technology standpoint (but not a customerI relations standpoint) Sun had the most to gain from the causes being madem public.0  L You then found some new outages to pin you FUD Sun flag to only to find that. they were caused by NT WEB servers going down.  ) Second the absurd OpenVMS/Alpha boosting.   E Spiralog, well before its announcment you were predicting that it anda@ Galaxies (covered next) would deliver massive performance boostsB for DBMS applications. How embarrasing it didn't did it in fact it? missed its design goals so spectacularly that Compaq dumped it.MF But this does not stop you suggesting that my postings are inaccurate.  B Galaxies, again long before they were available you announced that> they would revolutionise DBMS app performance and that because; of this Sybase and Oracle were real interested in them. Howo9 embarassing, Sybase responded by dropping OpenVMS support D and Oracle responded by dropping many of their apps from OpenVMS andC releasing the rest a year after they become available on any almostaE all the other Oracle ports. To date there is no benchmark information ; or examples from Compaq of people making the kind of use of-? Galaxies that you claimed they would. How very embarassing. Buth@ you still persist in suggesting that my postings are inaccurate.  A And now we get to the really good one WildFire. You pre announced K it 18 months before it actually shipped. You claimed it was a revolutionaryu= technology that would blow the socks of Sun, HP etc. So whats B happened, WildFire isn't revolutionary its just a NUMA system withB similar memory latency to the old Origin 2000. It hasn't blown theE socks off anything in fact its performance is unless you only want to ? run SPECint on it is, lets be kind underwhelming. In otherwordsiF its late, its slow, its expensive (go to the withdrawn results sectionF of TPC-C if you want to know just how slow and costly). Finally people@ have not bought them in droves (another prediction) according to@ Compaq sales have been slow. How very very very very embarassingG but again this does not seem to stop you popping up and suggesting thati my postings are inaccurate.,  A So we get neatly back to eBay, eBay as you may know are what they A auction the more they auction the more their revenues are so your @ suggestion that eBay would be interested in WildFire's even fromC a performance reason was again absurdly over optimistic. Given thatmA WildFires actual performance is way lower than expected you wouldt@ be asking eBay to take a drop in revenue simply for the "joy" of> owning a shiny new WildFire, I mean look at its performance on5 Oracle apps/TPC or any other standard OLTP benchmark.h  > So come on Rob you have been wildly wrong about almost all theA major technology that Compaq has introduced in the last year nonesE of it has remotely delivered on the independent marketing  you pushedhM out prior to its introduction and yet you still persist with the line that myn postings are inaccurate.  E Incedentally I dissagreed you each of your posts publically (actuallywC I could not be bothered with Spiralog, I get SPAM like that all thefA time) and got well flamed for each posting by inhabitants of this 
 newsgroup.   RegardsT Andrew Harrisonn Enterprise IT Architecta   ------------------------------  # Date: Mon, 04 Sep 2000 15:06:27 GMT 4 From: "Terry C. Shannon" <terryshannon@mediaone.net>* Subject: Re: Sun Hardware problems persist< Message-ID: <TTOs5.68493$_s1.798757@typhoon.ne.mediaone.net>  G "Andrew Harrison SUNUK Consultancy" <andrew.nospam@uk.sun.com> wrote in , message news:39B383BB.801BF18D@uk.sun.com... > Rob Young wrote: So whatsD > happened, WildFire isn't revolutionary its just a NUMA system with0 > similar memory latency to the old Origin 2000.  J Ah. The ancient Origin 2000 had a maximum local to remote latency ratio of 1:3?   > It hasn't blown thelG > socks off anything in fact its performance is unless you only want togA > run SPECint on it is, lets be kind underwhelming. In otherwordstH > its late, its slow, its expensive (go to the withdrawn results section9 > of TPC-C if you want to know just how slow and costly).p  L Ah, you must be referrring to the ~120K tpmC number that briefly appeared onJ www.tpc.org about a month and a half ago. That number was posted by CompaqI in order to win a bid (the purchasing manager at the client site required J posted TPC numbers) and once the bid was won, the numbers were taken down.   >Finally peoplenB > have not bought them in droves (another prediction) according to > Compaq sales have been slow.  G Hmmm... I would love to see the CPQ quote on slow GS-Series sales. LastlF numbers I saw were 237 (plus 100 Marvel systems) on Day One, and about% shipments in the last month of 2FQ00.v  G But you may be right... sales may be slow since CPQ is concentrating oneF piddling, irrelevant, rinky-dink little Mom and Pop customers like theE French CEA (Marvel complex) and ASCI (the 30TFLOPS bid that a certaini: Mountain View vendor tried like hell to wrest from the Q).   ------------------------------  % Date: Mon, 04 Sep 2000 12:18:24 -0400r+ From: "Main, Kerry" <Kerry.Main@compaq.com>d* Subject: RE: Sun Hardware problems persistJ Message-ID: <910612C07BCAD1119AF40000F86AF0D80528475C@kaoexc4.kao.dec.com>  	 Andrew ..a  & Just catching up on various threads ..  F >>> However a lot of customers have turned around and said after theirJ datacenter assessment that Sun was the first vendor who has ever explainedH to them what the issues are and why they should take the actions we have recommended. <<<<a  G ROTFL ... are you seriously trying to state (with a straight face) thatsF those in charge of the datacenters with Sun systems did not know about( static or proper environmental controls?  I Are you saying that Sun systems are not properly grounded so that personseD walking by might accidentally touch a system and thus bring it down?  D Based on all of the notes you have presented here, are Sun technicalD documents being updated to reflect Suns requrements for much tighterL environmental controls than other vendors? IBM, HP, Compaq, DG do not appearK to be having these issues, so it appears their systems can operate reliablytD at higher (what some would consider normal) datacenter temperatures.   Wow ..   :-)I   Regards,  
 Kerry Main Senior Consultant,
 Compaq CanadaR Professional Servicesh Voice : 613-592-4660 FAX   : 819-772-7036 Email : kerry.main@compaq.comr       -----Original Message-----' From: Andrew Harrison SUNUK Consultancyn! [mailto:andrew.nospam@uk.sun.com]t( Sent: Friday, September 01, 2000 6:02 AM To: Info-VAX@Mvb.Saic.Comm* Subject: Re: Sun Hardware problems persist     Rob Young wrote:  B > In article <39AE62FF.5D19D867@uk.sun.com>, Andrew Harrison SUNUK. Consultancy <andrew.nospam@uk.sun.com> writes: >/ > >rK > > As I said re-read the last paragraph and note the action the datacenterg% > > manager took and note the result.t > >wJ > > What did he say, we were having problems we reduced the temperature in oureB > > datacenter and the problem went away, what does that tell you. > >  >nK > > So here is a question, do you know what causes the majority of hardwaren; > > failures on all systems not just Sun's but all systems.. >rJ >         Ummmmm... using the same crappy design as a Sun high-end server?J >         Just a guess.  Actually, you're probably after "heat."  I'll buyK >         that.  Funny how much of this Sun hardware is surrounded by otheriF >         hardware that works fine.  Must be a tough sell to push thatI >         "heat" angle in a 75 degree computer room where everything elsee! >         is just fine.  Spin on!m >   @ Sorry wrong the biggest cause of hardware failure on all systemsD not just Sun's is static discharge. Some studies have put it as high as 70% of all failures.o  < Now people who know what they are talking about will now say? well Rob was sort of right though he didn't realise why because F heat is a contributing factor to making static discharges more likely.  J But it isn't the only cause, the kind of floor you have in your datacenterG is a contributory factor and a whole host of other variables. Where ther machinesI are in a datacenter is also an issue, being on the row next to any of them
 main walkways H for example particularly if you have the wrong flooring and other issues like low relative J humidity will expose those machines to a higher chance of failure than theL machines in the same datacenter that are not next door to the main walkways.  L Paradoxically Rob you may think that the publicity about Sun's Ecache issuesD has caused a negative impact on Sun's image with its major customers and for some this is true.  C However a lot of customers have turned arround and said after their-
 datacenterL assessment that Sun was the first vendor who has ever explained to them whatH the issues are and why they should take the actions we have recommended. FromK you posting it would appear that although you are unlikely to thank Sun for  puttingpH you in the picture you would at leat have learnt something that would be usefull L to help you keep all the systems in your datacenter running without hardware	 failures.   D Most customers seem to assume that if they keep systems within their	 operatingtJ temperature ranges and humidity ranges they will be reliable, when in fact the.J systems will be more reliable (from all vendors) if they are operated in a much- tighter range of temperatures and humidities.o  I Your post reveals that you like the majority of customers were completelyuL unaware of the fact that the majority of the hardware failures you have seen onI all the computer systems you have used have been caused by static, Sun's,d Compaqs,J Digital boxes etc. Why didn't you know, if you had known earlier would youI (even with Compaqs etc) have taken more care of the environment that your- systems-L run in ? would you for example have carpet tiles in you datacenter ? you may not L but people do ? would you monitor the relative humidity of the envrionment ? you:K may but many people don't ? would you make sure that the sub-floor pressurec is: is correct ? that the sub-floor is clean ? etc etc etc etc  K Your datacenter practice may well be exemplary in which case you don't needpL to worry about these issues but the fact that you didn't appear to know whatK the single most important cause of hardware failure is implies that if yourdL datacenter is well managed it is not because people are aware of the issues,( or they havn't communicated this to you.   Regards  Andrew Harrisons Enterprise IT Architectm   ------------------------------  % Date: Mon, 04 Sep 2000 09:29:18 +0100e2 From: Chris Sharman <Chris.Sharman@CCAgroup.co.uk>1 Subject: Re: TCP/IP 5.0A PatchList or ECO SummaryA. Message-ID: <39B36B6D.54FCFE69@CCAgroup.co.uk>   Hoff Hoffman wrote:sC >If folks are having specific problems with TCP/IP Services, pleasec5 >contact the Compaq Customer Support Center directly.o > 7 >"Ruslan R. Laishev" <Laishev@SMTP.DeltaTel.RU> writes:bN > :       Sorry, Hoff, but CCS can't offer anything, I have contacted with UCXA > :engineering a couple month ago and they was going to silence !i > @ >To the point: Details of the specific problem(s) or request(s)?  G I called CSC (UK) a week or two back, preparatory to installing 5.0A in @ the hope of fixing lpd problems we're having (can't see anythingE relevant in the FAQ). I was told no patch yet, & after discussion wass7 recommended to upgrade 4.2 from eco 2 to eco 4 instead.n   Regards, Chriss   ------------------------------  % Date: Mon, 04 Sep 2000 19:19:52 +0400k4 From: "Ruslan R. Laishev" <Laishev@SMTP.DeltaTel.RU>1 Subject: Re: TCP/IP 5.0A PatchList or ECO Summary50 Message-ID: <39B3BD98.FF285479@SMTP.DeltaTel.RU>   Hoff Hoffman wrote:; > i > In article <39B2CD20.189992C1@SMTP.DeltaTel.RU>, "Ruslan R. Laishev" <Laishev@SMTP.DeltaTel.RU> writes:eI > :>   If folks are having specific problems with TCP/IP Services, pleasep; > :>   contact the Compaq Customer Support Center directly. N > :       Sorry, Hoff, but CCS can't offer anything, I have contacted with UCXA > :engineering a couple month ago and they was going to silence !o > C >   To the point: Details of the specific problem(s) or request(s)?sT 	The problem is in different behaviour of binding a UDP socket (in particulary) withV INADDR_ANY under UCX 4.x, PSC TCPWare/Multinet and TCPIP 5.0A (with "PTR" ECO, and w/oR it). Under TCPIP 5.0a calling of sys$qio(IO$_SETMODE) cause to %x0134. I asked UCXI guys - "what behaviour is correct? Under UCX 4.x or under TCPIP 5.0A?"...a  P 	I found workaround, but ... no answer from UCX guys for the past 2-3 month. :-) 	Why?a  : 	In the same time a PSC staff are opening for any contact. > P >  --------------------------- pure personal opinion ---------------------------N >    Hoff (Stephen) Hoffman   OpenVMS Engineering   hoffman#xdelta.zko.dec.com   -- p Cheers, F +OpenVMS [Sys|Net] HardWorker........................................+E  Russia,Delta Telecom Inc,                    Cel:  +7 (901) 971-3222 E  191119,St.Petersburg,Transportny per. 3                     116-32221E                                               Fax:  +7 (812) 115-1099hG +http://www.levitte.org/~rlaishev/ .......... SysMan rides HailStorm +
g   ------------------------------  % Date: Mon, 04 Sep 2000 15:48:20 +0100 - From: Tim Llewellyn <tim.llewellyn@bbc.co.uk>m# Subject: Re: TCPIP$FTP Bugchecking! ) Message-ID: <39B3B634.CC565089@bbc.co.uk>    Cthulhu wrote:  2 > Mark D. Jilson <jilly@clarityconnect.com> wrote: >.& > > $ ANA/CRASH SYS$SYSTEM:SYSDUMP.DMP
 > > READ/EXECm > > READ SYS$SYSTEM:SYSDEF > > SET LOG {file.ext}3 > > SHOW CRASH ! Make sure to press <CR> thru every . > > SHOW STACK ! screen that will be presented >r, > Ok, thanks, it said a lot of funny things.@ > Now, I suppose I should refer to "OpenVMS System Dump Analyze" > manual...A  F Nah, post the output an Mark or someone else will try to diagnose your problem.    -- 6 Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project0 MedAS at the BBC, Whiteladies Road, Bristol, UK.A Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.ukd  A I speak for myself only and my views in no way represent those of  MedAS or the BBC.n   ------------------------------  % Date: Mon, 04 Sep 2000 12:19:31 +0400 $ From: Mikhail Nosov <mike@marsat.ru>/ Subject: Urgent help with backup restore needede9 Message-ID: <NDBBLKMFABODFJIBDBAMMEPIDEAA.mike@marsat.ru>   	 Dear all,e, I've got trouble restoring my recent backup. Error message is the following: ? %BACKUP-F-POSITERR, error positioning MKC500:[000000]<filename>k -SYSTEM-F-PARITY, parity error.i  J I've got two other backups done the same day, to the same tape, everything is OKs5 with them. Commands to create file was the following:c= $mount /for MKC500: /ov=id /media=compaction /cache=tape_dataeK $backup DRA3:[000000...] mkc500:<filename> /media=compact /ignore=interlock  /label=<label>  0 OS version - OVMS 7.1-1H1, tape drive TZ-89N-TA.  0 Is there any way to get this data back to life ? Or at least part of them ?A Is this tape usable for future backups or I'd better put it off ?    I need any suggestions badly !   With best regards,   Mikhail Nosov, Morsviazsputnik IT Manager.r Phone +7 095 967-1878s ---------------------------m http://www.marsat.ru E-Mail: mike@marsat.ru   ------------------------------  $ Date: Mon, 4 Sep 2000 17:05:41 +0800, From: "Little White" <Trishgoff@hotmail.com>6 Subject: Urgent, Help to remove a SOH or ETX hex code.( Message-ID: <8ovom0$9gf$1@m5.att.net.hk>  	 Dear All,   F Please give a helping hand to remove a file that generated by run time application that consists ofH start of header (Hex: 01") in the beginning and End of Text (Hex: 03). I don't have the source code of I the run time application and I have try to convert, copy, rename the file  that also failed to remove the 01 and 03 hex code.i   Many Thanks.         -- Where were you in '82 ?g! I was in France, where were you ?   I was in school, where was Lou ? Lou wasn't born in '82 :Pd   ------------------------------  # Date: Mon, 04 Sep 2000 12:57:40 GMTu= From: system@SendSpamHere.ORG (Brian Schenkenberger, VAXman-)a: Subject: Re: Urgent, Help to remove a SOH or ETX hex code.0 Message-ID: <009EF9D6.A4B5FB7E@SendSpamHere.ORG>  W In article <8ovom0$9gf$1@m5.att.net.hk>, "Little White" <Trishgoff@hotmail.com> writes:n
 >Dear All, >rG >Please give a helping hand to remove a file that generated by run time  >application that consists offI >start of header (Hex: 01") in the beginning and End of Text (Hex: 03). It >don't have the source code ofJ >the run time application and I have try to convert, copy, rename the file >that also failed to removeo >the 01 and 03 hex code. >m
 >Many Thanks.h  + Use one of the editors to remote the codes.s   Using EDT, edit the file.   - Then <GOLD><KP7> and at the 'Command:' prompte' enter DEFINE MACRO NOSOHETX and <ENTER>   - Then <GOLD><KP7> and at the 'Command:' promptO! enter FIND = NOSOHETX and <ENTER>    In this buffer enter:    SUB/^A//WHOLE/NOTYPE SUB/^C//WHOLE/NOTYPE  * You enter the ^A and ^C with the sequence:) <GOLD>1<GOLD><KP3> and <GOLD>1<GOLD><KP3>  repectively.  ( When you've entered the two SUBs, enter ( <GOLD><KP7> and at the 'Command:' prompt enter FIND = MAIN and <ENTER>m  # You can now activate the macro withe( <GOLD><KP7> and at the 'Command:' prompt enter: NOSOHETX <ENTER>u     --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM             oO city, n., 1. a place where trees are cut down and streets are named after them.c   ------------------------------  / Date: Mon, 04 Sep 2000 08:28:34 +0200 (MET DST)S& From: Rudolf Wingert <win@fom.fgan.de>% Subject: Re: VMS and 100 meg Etherneti6 Message-ID: <200009040625.IAA10020@sinet1.fom.fgan.de>   Hello,  C did you check the console value EWA0_MODE? You have to set this one C to "AUTO". After channging this value you must "init"Alpha or cucle 
 the power.   Regards Rudolf Wingert   ------------------------------  % Date: Mon, 04 Sep 2000 12:48:25 +0100dB From: Andrew Harrison SUNUK Consultancy <andrew.nospam@uk.sun.com>  Subject: Re: WORD viewer for VMS* Message-ID: <39B38C08.70846910@uk.sun.com>   "David J. Dachtera" wrote:  * > Andrew Harrison SUNUK Consultancy wrote: > [snip]G > > 7.    Doing the port might help someone else apart from the generaluD > >         OpenVMS community so hey we don't want to do that do we. >v
 > Um, Andrew,u > J > YOU seem to have copious free time, given how much we see of you here in > cov...  > Sorry I don't have an Alpha box running OpenVMS lurking in the lab :)   Regards  Andrew Harrisone Enterprise IT ArchitectB   ------------------------------   End of INFO-VAX 2000.495 ************************