1 INFO-VAX	Sat, 21 Dec 2002	Volume 2002 : Issue 703       Contents: Re: backup question  RE: backup question  Re: backup question  Re: backup question 2 Re: Compaq's phenomonal stupidity in killing Alpha4 Re: Demise of VMS (and most other operating systems) Digital server 5000  RE: Digital server 5000  Re: Digital server 5000  Re: Digital server 5000  Re: Digital server 5000  Re: Digital server 5000  RE: Digital server 5000  Re: Digital server 5000  Re: Digital server 5000  Re: Digital server 5000 . Re: How many messages/second can a CI rail do?3 RE: HP Advocacy "me too"'s - Hammer port, marketing ( Re: initialize/erase does no useful work( Re: initialize/erase does no useful work( Re: initialize/erase does no useful work6 Re: Is it possible to add a cluster node w/o decnet???( Re: Issue with Working Set in SYS$CREPRC( Re: Issue with Working Set in SYS$CREPRC Re: just a couple of things  Re: just a couple of things  LCD monitor for VAX?1 Re: looking for CD-RW or DVD-RW capability on VMS 1 Re: looking for CD-RW or DVD-RW capability on VMS 4 Re: Loss of Enet Performance on DS10 -> DS25 UpgradeG Re: Machine checks not always hardware failures, was: Re: Machine Check G Re: Machine checks not always hardware failures, was: Re: Machine Check ) Re: MIME extension recognition under CSWS - Re: OT: Whoa! Is Sun aiming at VMS's jugular? ' Partial memory option needed - PB7MA-AE + Re: Partial memory option needed - PB7MA-AE  Slightly OT- Need VT-525 info 2 Re: TECSys Consoleworks on VMS - first impressions# Re: Tracking down a TOP CPU Problem  Re: VAX 7810's.  VAXstation 4000 cd-rom bezel  Re: VAXstation 4000 cd-rom bezel  RE: VAXstation 4000 cd-rom bezel Vol Shadowing QuestionF Re: Why does BACKUP/IMAGE need write access to INDEXF.SYS, BITMAP.SYS?  F ----------------------------------------------------------------------  % Date: Fri, 20 Dec 2002 20:01:57 +0100 " From: "Hans Vlems" <hvlems@iae.nl> Subject: Re: backup question5 Message-ID: <atvpfg$36p9u$1@ID-143435.news.dfncis.de>   0 "Tom Linden" <tom@kednos.com> schreef in bericht3 news:CIEJLCMNHNNDLLOOGNJIIEJAGEAA.tom@kednos.com... 7 > I like to keep a current backup copy online of all my > > disks, since I have a large spare drive using the following: >  > $ type copyback.com  > $ set verifyH > $ define/user back_disk node$dka300:['P1'.'P2'.]/translation=concealed* > $ backup/verify/record  disk$'P2':[*...]" back_disk:[*...]/by_owner=original > D > Now this doesn't copy the files at the [000000] level so I suppose > the command should be  > 1 > $ backup/verify/record  disk$'P2':[000000...] - ) > back_disk:[000000...]/by_owner=original  > 7 > where P1 and P2 are node and disk name, respectively.  >  > Is this correct? > --- I AFAIK only an image backup will copy those files as well (you do mean the # various [00000]*.sys files, right?)    Hans   ------------------------------  % Date: Fri, 20 Dec 2002 11:31:04 -0800 # From: "Tom Linden" <tom@kednos.com>  Subject: RE: backup question9 Message-ID: <CIEJLCMNHNNDLLOOGNJICEJKGEAA.tom@kednos.com>   = but the problem with image is that you have to mount foreign. 9 All I want to do is to copy an entire disk to some rooted C directory on a spare drive, and to copy only those files that don't 5 exist on the target, or have a higher version number.    >-----Original Message----- ( >From: Hans Vlems [mailto:hvlems@iae.nl]) >Sent: Friday, December 20, 2002 11:02 AM  >To: Info-VAX@Mvb.Saic.Com >Subject: Re: backup question  >  >  > 1 >"Tom Linden" <tom@kednos.com> schreef in bericht 4 >news:CIEJLCMNHNNDLLOOGNJIIEJAGEAA.tom@kednos.com...8 >> I like to keep a current backup copy online of all my? >> disks, since I have a large spare drive using the following:  >> >> $ type copyback.com >> $ set verify I >> $ define/user back_disk node$dka300:['P1'.'P2'.]/translation=concealed + >> $ backup/verify/record  disk$'P2':[*...] # >back_disk:[*...]/by_owner=original  >>E >> Now this doesn't copy the files at the [000000] level so I suppose  >> the command should be >>2 >> $ backup/verify/record  disk$'P2':[000000...] -* >> back_disk:[000000...]/by_owner=original >>8 >> where P1 and P2 are node and disk name, respectively. >> >> Is this correct?  >> ---J >AFAIK only an image backup will copy those files as well (you do mean the$ >various [00000]*.sys files, right?) >  >Hans  >  >---' >Incoming mail is certified Virus Free. ; >Checked by AVG anti-virus system (http://www.grisoft.com). B >Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002 >  --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002    ------------------------------  # Date: Sat, 21 Dec 2002 01:08:19 GMT + From: Jeff Cameron <JCam90502@jcameron.com>  Subject: Re: backup question2 Message-ID: <BA28FEFD.2919%JCam90502@jcameron.com>  I On 12/20/02 11:55 AM, in article atvsjq$37mhk$1@ID-143435.news.dfncis.de, # "Hans Vlems" <hvlems@iae.nl> wrote:    > 2 > "Tom Linden" <tom@kednos.com> schreef in bericht5 > news:CIEJLCMNHNNDLLOOGNJICEJKGEAA.tom@kednos.com... @ >> but the problem with image is that you have to mount foreign.< >> All I want to do is to copy an entire disk to some rootedF >> directory on a spare drive, and to copy only those files that don't8 >> exist on the target, or have a higher version number. >>  < > This will work with disks that are both mounted correctly: > , > $ backup/image/ignore=interlock $3$dka200:' > $2$dkc100:[backup]disk_label.bck/save  > > > Which is what you wanted, if I understand your initial post. >  > Hans > K Hans is correct. You should backup using /IMAGE to a saveset on your target J disk. You can still do selective restores from an image saveset if needed.L Plus there is no real need to backup the .sys files of a volume unless it isK an image backup, because these files are specific to the file system of the L volume they reside on. Restoring a .sys file from one volume to another will% most likely corrupt your file system.    Try : 0 $ backup/image/verify/record  disk$'P2':[*...] -!     back_disk:['p1']'p2'.bck/save B Where P1 is your node and p2 is your disk as you specified before.   Jeff Cameron   ------------------------------  % Date: Fri, 20 Dec 2002 20:43:02 -0500 ( From: David Froble <davef@tsoft-inc.com> Subject: Re: backup question, Message-ID: <3E03C726.5050406@tsoft-inc.com>  L What's the problem with using /IMAGE.  Eyesight is fading, but I didn't see , anything about INCREMENTAL in your commands.   Dave     Tom Linden wrote:   7 > I like to keep a current backup copy online of all my > > disks, since I have a large spare drive using the following: >  > $ type copyback.com  > $ set verifyH > $ define/user back_disk node$dka300:['P1'.'P2'.]/translation=concealedM > $ backup/verify/record  disk$'P2':[*...] back_disk:[*...]/by_owner=original  > D > Now this doesn't copy the files at the [000000] level so I suppose > the command should be  > 1 > $ backup/verify/record  disk$'P2':[000000...] - * > 	back_disk:[000000...]/by_owner=original > 7 > where P1 and P2 are node and disk name, respectively.  >  > Is this correct? > --- ( > Outgoing mail is certified Virus Free.< > Checked by AVG anti-virus system (http://www.grisoft.com).C > Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002  >  >    ------------------------------  % Date: Fri, 20 Dec 2002 22:55:21 -0500   From: John Santos <JOHN@egh.com>; Subject: Re: Compaq's phenomonal stupidity in killing Alpha 4 Message-ID: <1021220223404.400D-100000@Ives.egh.com>  ) On Fri, 20 Dec 2002, Peter Flunger wrote:    > , > "Bob Ceculski" <bob@instantwhip.com> wrote > > < > > I wish you had a sun sparky box to run also so you could) > > prove Andrew the idiot that he is ...  > 6 > Maybe there is somebody out there with a SUN running > SETI ?= > We do not have any, and as long as it is my saying, we will 
 > never have. : > Some of our customers are running SUNs as well, and they5 > are not impressed by what they get in comparison to  > an Alpha.  > Peter   ; This may be an apples-to-oranges situation.  ISTR that some 9 time ago, someone drastically improved the performance of = seti@home on Alphas by either rewriting the Fourier transform 7 routine or possibly by using a stock FFT routine from a ? system math library.  (Don't know the details; it's been almost , 30 years since I did this kind of stuff...)   < The SETI people weren't interested.  Apparently they thought6 it was too much trouble to validate the output vs. the8 limited number of systems involved.  So it just uses the: FFT built into the seti@home program, which is designed to> work anywhere.  However, they may very well be using optimized9 FFT's on other platforms which they perceive to have more < users (i.e. x86 Windows, x86 Linux, Solaris.)  No one should< jump to any conclusions without knowing for sure about this.   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Sat, 21 Dec 2002 00:56:05 +0000 + From: John Laird <john@laird-towers.org.uk> = Subject: Re: Demise of VMS (and most other operating systems) 8 Message-ID: <oqd70v8op60mo274kpis74fqh1ffc3fgei@4ax.com>  K On 20 Dec 2002 11:15:18 -0600, koehler@eisner.nospam.encompasserve.org (Bob  Koehler) wrote:   g >In article <rkk60volem5bgmqmnsrqsd0nq8u20db920@4ax.com>, John Laird <john@laird-towers.org.uk> writes: N >> On 20 Dec 2002 07:22:30 -0600, koehler@eisner.nospam.encompasserve.org (Bob >> Koehler) wrote: >>  f >>>In article <HiWVK9XMOhRq@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes: >>>>  9 >>>> RMS merely does what it is told by your FTP program.  >>> D >>>   Ture, but they're all consistent.  I'll have to go dig throughH >>>   my RMS manual, unless someone knows off the top of their head:  isD >>>   setting the last byte used supported for fixed length 512 byteH >>>   records?  (I'd assume not, or the records wouldn't truely be fixedG >>>   length.)  Is there a better record format for a true byte stream  3 >>>   that does support setting the last byte used?  >>  M >> Your assumption is wrong.  If you FTP a random-length file in binary mode, O >> then fixed-length 512-byte records is the structure you get and the EOF byte M >> is correctly set.  A small Fortran test program confirms - a partial write D >> to the last block results in a partial read when it is read back. > F >   I am making no assumption.  This is based on experience.  We FTP aF >   random length file in binary.  We get a collection of fixed lengthH >   512 byte records.  The last record is 0 filled.  A C program readingI >   the file as a stream file gets the whole last record.  There's no way E >   to tell the correct length unless you know that the original file  >   did not end in bytes of 0. > D >   If you'ld like to post your example, I'm willing to work out why >   you think it's working.    Because I screwed up :-(  L Having tried again, I confirm you are quite correct.  The easiest example isJ any old .EXE file - despite having fixed-length records, the last block isJ generally not complete, and the file header reflects this (FFB is not zeroK in the last block +1).  However, $GET will return 512 bytes read.   Fortran  and C reads do the same.  G I know I have been around this one in the past, way back when I used to H transfer seismic data about from one architecture to another.  I guess IG must have peeked at the file header and remembered to correct the final  block size.   I UCX FTP transfers appear to correctly set the EOF position when receiving J and then honour it when transmitting.  Looks like you have to be as smart.  L One final possibility is block i/o.  I think it *may* return a partial block on the last read.    Sorry for any confusion.   	John    ------------------------------  % Date: Fri, 20 Dec 2002 20:09:05 +0100 " From: "Hans Vlems" <hvlems@iae.nl> Subject: Digital server 50005 Message-ID: <atvpsl$38onb$1@ID-143435.news.dfncis.de>   I I have a temporary guest: a Digital Server 5000 system. It might become a H permanent resident provided it will behave and run VMS. The system bootsI from a VMS 7.3 cd, so that's encouraging. The problem is that I'm used to G VAXes, VAXstations actually so this is all rather new, please have some & patience with the following questions.  H Unfortunately the SCSI controller is missing, which leads to question 1.* What kind of SCSI controller is supported?G I have two white flatcables with black connectors on them; 1"x1/8", are  these SCSI cables?  H The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware available on-line?; If so how do I install the firmware? (I have no books...:-(   
 Hans Vlems   ------------------------------  % Date: Fri, 20 Dec 2002 11:28:43 -0800 # From: "Tom Linden" <tom@kednos.com>   Subject: RE: Digital server 50009 Message-ID: <CIEJLCMNHNNDLLOOGNJIOEJJGEAA.tom@kednos.com>   - I thought the 5000 were mips/ultrix machines?    you might find something at 7 ftp://ftp.digital.com/pub/digital/DECinfo/semiconductor   9 If the cables have 50 conductors, they are probably scsi.      >-----Original Message----- ( >From: Hans Vlems [mailto:hvlems@iae.nl]) >Sent: Friday, December 20, 2002 11:09 AM  >To: Info-VAX@Mvb.Saic.Com >Subject: Digital server 5000  >  > J >I have a temporary guest: a Digital Server 5000 system. It might become aI >permanent resident provided it will behave and run VMS. The system boots J >from a VMS 7.3 cd, so that's encouraging. The problem is that I'm used toH >VAXes, VAXstations actually so this is all rather new, please have some' >patience with the following questions.  > I >Unfortunately the SCSI controller is missing, which leads to question 1. + >What kind of SCSI controller is supported? H >I have two white flatcables with black connectors on them; 1"x1/8", are >these SCSI cables?  > I >The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware  >available on-line? < >If so how do I install the firmware? (I have no books...:-( >  >Hans Vlems  >  >---' >Incoming mail is certified Virus Free. ; >Checked by AVG anti-virus system (http://www.grisoft.com). B >Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002 >  --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002    ------------------------------  % Date: Fri, 20 Dec 2002 20:37:27 +0100 6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>  Subject: Re: Digital server 5000) Message-ID: <3E037177.5080800@vajhoej.dk>    Hans Vlems wrote:   K > I have a temporary guest: a Digital Server 5000 system. It might become a J > permanent resident provided it will behave and run VMS. The system bootsK > from a VMS 7.3 cd, so that's encouraging. The problem is that I'm used to I > VAXes, VAXstations actually so this is all rather new, please have some ( > patience with the following questions. > J > Unfortunately the SCSI controller is missing, which leads to question 1., > What kind of SCSI controller is supported?I > I have two white flatcables with black connectors on them; 1"x1/8", are  > these SCSI cables? > J > The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware > available on-line?= > If so how do I install the firmware? (I have no books...:-(     Firmware can be downloaded from:5    http://ftp.digital.com/pub/Digital/Alpha/firmware/ 6 (I belive installation instructiosn come with the kit)  8 VMS is not supported on Digital Server 5000 systems. But it may work anyway.    Arne   ------------------------------  % Date: Fri, 20 Dec 2002 20:57:10 +0100 " From: "Hans Vlems" <hvlems@iae.nl>  Subject: Re: Digital server 50005 Message-ID: <atvsn1$3a8a3$1@ID-143435.news.dfncis.de>   0 "Tom Linden" <tom@kednos.com> schreef in bericht3 news:CIEJLCMNHNNDLLOOGNJIOEJJGEAA.tom@kednos.com... / > I thought the 5000 were mips/ultrix machines?  >  > you might find something at 9 > ftp://ftp.digital.com/pub/digital/DECinfo/semiconductor  > ; > If the cables have 50 conductors, they are probably scsi.  >  > I The DEC 5000 series were MIPS machines all right. But this is a white box J Alpha system, labeled Digital Server 5000 model 5305. It has a 5/533 alpha. processor in it. Definitely not a MIPS box ...   ------------------------------  % Date: Fri, 20 Dec 2002 21:01:21 +0100 " From: "Hans Vlems" <hvlems@iae.nl>  Subject: Re: Digital server 50005 Message-ID: <atvsun$3a78e$1@ID-143435.news.dfncis.de>   2 "Arne Vajhj" <arne@vajhoej.dk> schreef in bericht# news:3E037177.5080800@vajhoej.dk...  > Hans Vlems wrote:  > K > > I have a temporary guest: a Digital Server 5000 system. It might become  a L > > permanent resident provided it will behave and run VMS. The system bootsJ > > from a VMS 7.3 cd, so that's encouraging. The problem is that I'm used toK > > VAXes, VAXstations actually so this is all rather new, please have some * > > patience with the following questions. > > L > > Unfortunately the SCSI controller is missing, which leads to question 1.. > > What kind of SCSI controller is supported?K > > I have two white flatcables with black connectors on them; 1"x1/8", are  > > these SCSI cables? > > L > > The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware > > available on-line?? > > If so how do I install the firmware? (I have no books...:-(  > " > Firmware can be downloaded from:7 >    http://ftp.digital.com/pub/Digital/Alpha/firmware/ 8 > (I belive installation instructiosn come with the kit) > : > VMS is not supported on Digital Server 5000 systems. But > it may work anyway.  >  > Arne >  Arne,   H thanks for the URL, I had a look at it, but couldn't figure out how that site works, not yet at least. I You are right: VMS is not supported on a Dig. Srvr 5000. But it boots VMS K from the installation CD. The problem is that there's no SCSI adapter, so I I cannot start the installation. I tried this once succesfully on a Digital E Server 3000, a smaller (EV4) nephew of the 5000. Ans that works well, 0 provided you invoke the correct incantations :-)G The question remains: what SCSI controlers does VMS support, say, on  a  1000(A) or 1200?   Hans   ------------------------------  % Date: Fri, 20 Dec 2002 21:15:46 +0100 6 From: Arne =?ISO-8859-1?Q?Vajh=F8j?= <arne@vajhoej.dk>  Subject: Re: Digital server 5000) Message-ID: <3E037A72.9040100@vajhoej.dk>    Hans Vlems wrote:   4 > "Arne Vajhj" <arne@vajhoej.dk> schreef in bericht% > news:3E037177.5080800@vajhoej.dk...  >>Hans Vlems wrote: K >>>The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware  >>>available on-line? > >>>If so how do I install the firmware? (I have no books...:-(    " >>Firmware can be downloaded from:7 >>   http://ftp.digital.com/pub/Digital/Alpha/firmware/ 8 >>(I belive installation instructiosn come with the kit)    J > thanks for the URL, I had a look at it, but couldn't figure out how that > site works, not yet at least.     9 If you pick Digital Server 5000 in the combo box with the = label Digital Server, then you is immediatetly redirected to: H    http://ftp.digital.com/pub/Digital/Alpha/firmware/readmes/as1200.html  ; (at least if you use a recent MS Windows based web-browser)    Arne   ------------------------------  % Date: Fri, 20 Dec 2002 12:14:48 -0800 # From: "Tom Linden" <tom@kednos.com>   Subject: RE: Digital server 50009 Message-ID: <CIEJLCMNHNNDLLOOGNJIMEJLGEAA.tom@kednos.com>    >-----Original Message----- + >From: Arne Vajhj [mailto:arne@vajhoej.dk] ) >Sent: Friday, December 20, 2002 12:16 PM  >To: Info-VAX@Mvb.Saic.Com! >Subject: Re: Digital server 5000  >  >  >Hans Vlems wrote: > 5 >> "Arne Vajhj" <arne@vajhoej.dk> schreef in bericht & >> news:3E037177.5080800@vajhoej.dk... >>>Hans Vlems wrote:L >>>>The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware >>>>available on-line?? >>>>If so how do I install the firmware? (I have no books...:-(  >  > # >>>Firmware can be downloaded from: 8 >>>   http://ftp.digital.com/pub/Digital/Alpha/firmware/9 >>>(I belive installation instructiosn come with the kit)  >  > K >> thanks for the URL, I had a look at it, but couldn't figure out how that   >> site works, not yet at least. >  > : >If you pick Digital Server 5000 in the combo box with the> >label Digital Server, then you is immediatetly redirected to:I >   http://ftp.digital.com/pub/Digital/Alpha/firmware/readmes/as1200.html  > < >(at least if you use a recent MS Windows based web-browser)  	 or Opera.    >  >Arne  >  >---' >Incoming mail is certified Virus Free. ; >Checked by AVG anti-virus system (http://www.grisoft.com). B >Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002 >  --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.431 / Virus Database: 242 - Release Date: 12/17/2002    ------------------------------  % Date: Fri, 20 Dec 2002 20:54:42 -0500 ( From: David Froble <davef@tsoft-inc.com>  Subject: Re: Digital server 5000, Message-ID: <3E03C9E2.6080909@tsoft-inc.com>   Hans Vlems wrote:   K > I have a temporary guest: a Digital Server 5000 system. It might become a J > permanent resident provided it will behave and run VMS. The system bootsK > from a VMS 7.3 cd, so that's encouraging. The problem is that I'm used to I > VAXes, VAXstations actually so this is all rather new, please have some ( > patience with the following questions.    N Alphas, while a bit different, are actually easier to work with.  You can run Q VMS from the CD.  Installs and upgrades are run from the CD, and are easier than  	 on a VAX.     J > Unfortunately the SCSI controller is missing, which leads to question 1.    	 Not good.   , > What kind of SCSI controller is supported?    M Probably quite a few.  When I say this, I'm refering to the SCSI controllers  M supported by the system and VMS, not just any old ADAPTEC board you may find.   < Go to the SOC archives, and get the doc on SCSI controllers.  I > I have two white flatcables with black connectors on them; 1"x1/8", are  > these SCSI cables?    ) Don't know what these are.  Wide SCSI????     J > The firmware is at level 5.4 and VMS rather prefers 5.9. Is the firmware > available on-line?    * Yes, and usually comes with any VMS media.  = > If so how do I install the firmware? (I have no books...:-(     M You boot the firmware disk.  It's pretty self explanatary from that point on.    Dave   ------------------------------    Date: 20 Dec 2002 19:55:01 -0600- From: Kilgallen@SpamCop.net (Larry Kilgallen)   Subject: Re: Digital server 50003 Message-ID: <TZgJsf7pyo3W@eisner.encompasserve.org>   W In article <3E03C9E2.6080909@tsoft-inc.com>, David Froble <davef@tsoft-inc.com> writes:   P > Alphas, while a bit different, are actually easier to work with.  You can run S > VMS from the CD.  Installs and upgrades are run from the CD, and are easier than   > on a VAX.   K While one does not do VAX installs and upgrades when booted from the CDROM, C one can boot VAX/VMS from the distribution CD for utility purposes.    ------------------------------  # Date: Sat, 21 Dec 2002 06:03:55 GMT + From: Jeff Campbell <jcampbell@ins-msi.com>   Subject: Re: Digital server 5000+ Message-ID: <3E03F7D4.54A51BD1@ins-msi.com>    > Hans Vlems wrote: Y > >"Arne Vajhj" <arne@vajhoej.dk> schreef in bericht news:3E037177.5080800@vajhoej.dk...    [snip]  " > Firmware can be downloaded from:7 >    http://ftp.digital.com/pub/Digital/Alpha/firmware/ 8 > (I belive installation instructiosn come with the kit) > : > VMS is not supported on Digital Server 5000 systems. But > it may work anyway.  >  > Arne >  >Arne, > I >thanks for the URL, I had a look at it, but couldn't figure out how that  >site works, not yet at least.J >You are right: VMS is not supported on a Dig. Srvr 5000. But it boots VMSL >from the installation CD. The problem is that there's no SCSI adapter, so IJ >cannot start the installation. I tried this once succesfully on a DigitalF >Server 3000, a smaller (EV4) nephew of the 5000. Ans that works well,1 >provided you invoke the correct incantations :-) H >The question remains: what SCSI controlers does VMS support, say, on  a >1000(A) or 1200?  >  >Hans   B The SCSI controllers you want are the KZPBA or  KZPSC / KZPAC. The? later two are DEC compatible Mylex DAC960 RAID controllers. All > three of these SCSI interfaces will work with both the SRM and Alphabios consoles.   = The Adaptec AHA-[23]940 will work with the Alphabios console.   = The above controllers are single-ended. The KZPSA, which is a @ differential-ended interface, will also work with both consoles.  B A caution about the KZPBA, which is Qlogic 1040B based. Use eitherB the internal HD68 connector or the external HD68 connector but not@ both. So if you want devices in both the 5000's StorageWorks bay@ and an external BA356 (say), use a second KZPBA for the external StorageWorks box.   
 Jeff Campbell  n8wxs@arrl.net   ------------------------------    Date: 20 Dec 2002 11:37:45 -08001 From: keithparris_NOSPAM@yahoo.com (Keith Parris) 7 Subject: Re: How many messages/second can a CI rail do? < Message-ID: <cf15391e.0212201137.17afd31@posting.google.com>  i "James Gessling" <jgessling@yahoo.com> wrote in message news:<at4tqt$10m4hi$1@ID-46415.news.dfncis.de>... I > I'm investigating a performance issue that occured in our cluster.  One N > morning, access to a certain file was really slow.  So I looked in the decpsK > data in the performance evaluation display/CI statistics and see that one M > rail (of six) was doing 15,000+ messages per second.  This was (I suppose)  @ > due to the lock requests between other nodes reading the file.  @ One can use the DECamds SCS display (drill deep) to see, for theA problem node(s), which SYSAP the majority of the SCS messages are ? going through (i.e. MSCP$DISK for MSCP serving or (most likely) @ VMS$VAXcluster for connection manager, lock manager, OPCOM, etc.	 traffic).   D If the majority of the SCS messages are on the VMS$VAXcluster SYSAP,C one could then use the SDA extension CNX to take a trace of all the E SCS sequenced messages going through the VMS$VAXcluster SYSAP and see D what type of traffic they represent.  A type of "LCK" indicates lockE request stuff.  (I came across one cluster recently with huge amounts F of "LKI" type traffic indicated here -- which turned out to be from an" application's overuse of $GETLKI.)   > There are no credit waits.  D But then I'm guessing you still run CLUSTER_CREDITS at a high value,& like 128, not the default value of 10.  N > So management asked the obvious question, is that too high a rate?  How many# > messages/second can a CI rail do?   F I'm surprised you got 15K/second.  I would have considered 7K or 8K to
 be saturated.   B In Verell Boaen's latest slides on Cluster Interconect PerformanceF from the OpenVMS Technical Symposium, he got 6,725 lock-sized messagesC per second through a CI adapter in a single-stream test, and 18,134 C per second with 32 streams (and I doubt you have 32 remote nodes in D the cluster).  (There's an older version of this presentation postedE at http://vmsone.com/dfwdays/Presentations/CI_MEAS_V73_2001_03.PPT ).   | jones.computer.srv@worldnet.att.net (Daryl Jones) wrote in message news:<8a646952.0212101305.31c5418e@posting.google.com>...F > Please go back to the DECps analyze data and look at the MB/sec rateH > for the CI channel. It might not have anything to do with the messagesF > per sec more with MB/sec. Once upon a time a single init of a Oracle. > AIJ file suck up the entire CI channels (2).  E My experience has been that with small messages like lock requests, a @ CI adapter will saturate long before it reaches its rated MB/sec
 bandwidth.  E Before we had Verell's data on how many lock-request-sized messages a C CI adapter could really do, the best we could do was something like D multiply the number of messages per second by 3 (approximating theirD size as 1/3 KByte) and add that to the block data transfer KB/sec toE get a total (and then perhaps double that figure to account for peaks F that get averaged away by DECps' 2-minute default sampling interval).  This was pretty crude.  D Now that we have actual numbers for small-packet performance, we can# look at the message rates directly.   i "James Gessling" <jgessling@yahoo.com> wrote in message news:<at7g94$111cui$1@ID-46415.news.dfncis.de>... M > I'll answer my own question.  I set up a test case where I had 5 gs140's in  > a ci cluster.  ... E > So I did have a ci saturation problem and we are considering how to 
 > address it.   ' I'd attack this on more than one front.   A 1) First, one can spread lock mastership for the most-active lock D trees across multiple nodes (and thus across multiple CI adapters atE the same time).  By default, sometimes OpenVMS allows more lock trees F to be mastered by a given node than the node eventually has CPU [oftenB the Primary] and/or interconnect resources to handle when the load hits its peak.  E You would measure lock tree request rates using the SDA extension LCK B SHOW ACTIVE or my freeware LOCK_ACTV tool, and find out which lockB trees are getting hit the hardest on the problem node(s), and thenA decide how one might spread mastership of those lock trees across C [more?] nodes to more-evenly balance the workload and avoid any one E node (or CI adapter) from getting overloaded.  This involves inducing < OpenVMS to move lock trees where you want them (typically byA artificially generating high request rates temporarily so OpenVMS D wants to move the tree) while using a non-zero value of PE1 on otherD nodes to prevent OpenVMS from moving them back, and at least do this( for times where performance is critical.  @ Taking control away from OpenVMS in this manner does mean you in; effect take responsibility yourself for load balancing lock > mastership, instead of relying on the default lock remasteringA facilities.  This has to be tracked, monitored, and adjusted on a E periodic basis to adapt to shifts and increases in workload.  And you < have to be prepared to handle the case where one of your keyC lock-master nodes fails, and OpenVMS moves lock mastership to other E nodes in an arbitrary manner that may be non-optimal, and you'll have E to fix that up, at the same time you're dealing with the node failure  itself.   C But by spreading the lock mastership workload across more nodes, or A more evenly across the set of nodes it's already on, one can also E minimize the risk of saturation of the Primary CPU in interrupt state  on any one node.  C 2) CI port load-sharing code was introduced in OpenVMS VAX 6.0, but @ never got ported to Alpha.  So you will have to do the balancingF yourself.  You mentioned the MOVE_REMOTENODE_CONNECTIONS tool from the; CSC -- you can  balance the load across CI adapters (and by C implication, CI rails) using that tool.  This program allows you to E specify which CI rail will be used for lock requests (as it moves the F VMS$VAXcluster SYSAP connection over which lock requests flow) to/fromF each and every remote node.  Knowing the request rate from each remoteB node for the hottest lock trees, you should be able to balance theD lock message workload fairly evenly across CI adapters/rails.  DECpsD data could also help you balance the message and MB/sec rates acrossE all the available CI rails.  (Static load balancing -- basically just F spreading the VMS$VAXcluster connections evenly across all CI adaptersD and rails within the cluster -- is relatively easy to do, and I haveF DCL to do that which I'd be willing to share with sites which need it;A dynamically adjusting the paths would even be possible, but would  require more work).   E Because by default OpenVMS tends to put all connections onto a single E rail at boot time (typically to the one rail with the disk controller @ it booted from), you'll need to re-adjust this every time a node reboots.   ------------------------------  % Date: Fri, 20 Dec 2002 12:46:02 -0800 $ From: Shane Smith <ssmith@icius.com>< Subject: RE: HP Advocacy "me too"'s - Hammer port, marketing0 Message-ID: <01C2A825.C1634540@sulfer.icius.com>  . You have to sign up, and drill down from here:   http://www.hpuseradvocacy.org/  C Go to "submit issues", then there's a "popular issues" link on that  page.    Shane    -----Original Message-----* From: Mariuz [mailto:mariuz@stop.spam.org]' Sent: Friday, December 20, 2002 3:06 AM2 To: Info-VAX@Mvb.Saic.Come< Subject: Re: HP Advocacy "me too"'s - Hammer port, marketing     Where is the link ?  pls    ------------------------------    Date: 20 Dec 2002 12:10:39 -0600 From: briggs@encompasserve.org1 Subject: Re: initialize/erase does no useful work 3 Message-ID: <cRzIxarSpfUg@eisner.encompasserve.org>e  c In article <PZTsx20QzMdV@eisner.encompasserve.org>, Kilgallen@SpamCop.net (Larry Kilgallen) writes:e] > In article <xWwM9.234141$pN3.19022@sccrnsc03>, "Mark E. Levy" <levy@sysman-inc.com> writes:?? >> "David J. Dachtera" <djesys.nospam@fsi.net> wrote in messagea$ >> news:3E028FF4.2F161443@fsi.net... > L >>> Hhmmm... Funny. When I did something similar on an ES40 with twin HSZ80sG >>> and 18GB disks, my process did no I/O, there was no I/O activity ingG >>> MONITOR SYSTEM and the lights on the SBBs stayed on solid the wholec	 >>> time.i >>  G >> You won't see any I/O on your process because the work is being done M >> elsewhere. Your process merely requested an ACP operation and it (the ACP)  >> took off and did the work.o > G > Is this an ODS-1 disk (for which the ACP is relevant) ?  For ODS-2/5, H > the ACP was long ago (as in, for any version that can run on an Alpha,G > much less an ES40) replaced with the XQP which issues IO from processh
 > context.  B But the relevant I/O is likely to be a single $QIO with an IO$_DSE@ function code.  (Or maybe more than one -- I'm not sure how high0 the block count on an IO$_DSE operation can go).  A In any case, you're not going to see the process direct I/O countA@ incrementing wildly as 127 block chunks of zeroes are repeatedly< written to the disk.  The disk class driver is going to take@ responsibility for the whole erase effort and parcel out work to the port driver as it sees fit.-  < Regardless of whether the erase was requested by an XQP, ACPB or directly as a user mode QIO, you won't get a massive I/O count.   	John Briggs   ------------------------------  # Date: Fri, 20 Dec 2002 23:40:34 GMT@, From: Frank Troth <Frank.E.Troth@boeing.com>1 Subject: Re: initialize/erase does no useful work * Message-ID: <3E03AA72.EABAE2AC@boeing.com>  @ I assume it's ods-2/5, we're running VMS 7.2-1 on an Alpha 8400.   Larry Kilgallen wrote: > ] > In article <xWwM9.234141$pN3.19022@sccrnsc03>, "Mark E. Levy" <levy@sysman-inc.com> writes:p@ > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message% > > news:3E028FF4.2F161443@fsi.net...e > M > >> Hhmmm... Funny. When I did something similar on an ES40 with twin HSZ80sfH > >> and 18GB disks, my process did no I/O, there was no I/O activity inH > >> MONITOR SYSTEM and the lights on the SBBs stayed on solid the whole
 > >> time. > > H > > You won't see any I/O on your process because the work is being doneN > > elsewhere. Your process merely requested an ACP operation and it (the ACP) > > took off and did the work. > G > Is this an ODS-1 disk (for which the ACP is relevant) ?  For ODS-2/5, H > the ACP was long ago (as in, for any version that can run on an Alpha,G > much less an ES40) replaced with the XQP which issues IO from processt
 > context.   ------------------------------  # Date: Fri, 20 Dec 2002 23:42:47 GMTp, From: Frank Troth <Frank.E.Troth@boeing.com>1 Subject: Re: initialize/erase does no useful workr* Message-ID: <3E03AAF7.BDF13E99@boeing.com>  @ I assume it's ods-2/5, we're running VMS 7.2-1 on an Alpha 8400.   Larry Kilgallen wrote: > ] > In article <xWwM9.234141$pN3.19022@sccrnsc03>, "Mark E. Levy" <levy@sysman-inc.com> writes::@ > > "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message% > > news:3E028FF4.2F161443@fsi.net...  > M > >> Hhmmm... Funny. When I did something similar on an ES40 with twin HSZ80syH > >> and 18GB disks, my process did no I/O, there was no I/O activity inH > >> MONITOR SYSTEM and the lights on the SBBs stayed on solid the whole
 > >> time. > >.H > > You won't see any I/O on your process because the work is being doneN > > elsewhere. Your process merely requested an ACP operation and it (the ACP) > > took off and did the work. > G > Is this an ODS-1 disk (for which the ACP is relevant) ?  For ODS-2/5,kH > the ACP was long ago (as in, for any version that can run on an Alpha,G > much less an ES40) replaced with the XQP which issues IO from process]
 > context.   ------------------------------  % Date: Fri, 20 Dec 2002 23:39:45 -0600C! From: Dan Moore <dmoore@sosu.edu>t? Subject: Re: Is it possible to add a cluster node w/o decnet??? ( Message-ID: <3E03FEA1.C07F0675@sosu.edu>   Jim,  J   Our decnet seems to be working fine. We also use CSLG. I'm not sure whatH the product description is in the show lic listing. We are using OpenVMS 7.1-1H2  and  OpenVMS 7.3-1.   Dan    Jim Agnew wrote:  A > subject says it all... the vms 5.5-2 clusterconfig.com seems tocI > want/need decnet running, but with the advent of the new cslg, it looks < > like decnet is gone (or i had too much Christmas cheer...) >dB > So before twiddling with this, has anyone been there, done that? >< > jimn      K -----------== Posted via Newsfeed.Com - Uncensored Usenet News ==---------- G    http://www.newsfeed.com       The #1 Newsgroup Service in the World!IM -----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----r   ------------------------------    Date: 20 Dec 2002 12:14:06 -08007 From: patricia.beck@verizonwireless.com (Patricia Beck)w1 Subject: Re: Issue with Working Set in SYS$CREPRC < Message-ID: <da59dcbc.0212201214.d6113d1@posting.google.com>  @ Thanks everybody for your help - I finally got the relationshipsD straight between WSMAX, WSExtent, and PQL_MWSEXTENT. Now I can start> the detached process using $CREPRC, with NOUAF on, and get the6 override value of WSExtent Size just for that process.   ------------------------------  % Date: Fri, 20 Dec 2002 20:31:30 -0500 ( From: David Froble <davef@tsoft-inc.com>1 Subject: Re: Issue with Working Set in SYS$CREPRC , Message-ID: <3E03C472.7000304@tsoft-inc.com>   Patricia Beck wrote:  B > Thanks everybody for your help - I finally got the relationshipsF > straight between WSMAX, WSExtent, and PQL_MWSEXTENT. Now I can start@ > the detached process using $CREPRC, with NOUAF on, and get the8 > override value of WSExtent Size just for that process.    
 Hmmmm........0  M When you change a sysgen parameter, such as PQL_MWSEXTENT, it applies to the sM entire system.  Personally, I don't like to see the *MINIMUM* PQL parameters EP played with.  I once had a customer that thought the 'M' stood for MAXIMUM, and M made the number very high.  Turned out that every process was eating up hugh oQ amounts of memory, meeting the MINIMUM parameter.  Don't remember the parameter,   but it wasn't WSEXTENT.       PQL_MWSEXTENT  G         PQL_MWSEXTENT sets the minimum working set extent for a processaI         created by the Create Process ($CREPRC) system service or the DCL          command RUN (Process).  A         This value overrides a smaller quantity set for a user in          AUTHORIZE.  C         PQL_MWSEXTENT has the AUTOGEN, DYNAMIC, and GEN attributes.n      PQL_DWSEXTENT  G         PQL_DWSEXTENT sets the default working set extent for a process I         created by the Create Process ($CREPRC) system service or the DCLt         command RUN (Process).  C         PQL_DWSEXTENT has the AUTOGEN, DYNAMIC, and GEN attributes.     Q Normally, if I wanted the default (startup) value for a PQL parameter to be some oO minimum amount, then that's the parameter I'd use.  Note that the system could lO then decrease the usage, when appropriate.  When setting a MINIMUM, the system ,. cannot decrease the value, regardless of need.  I Not sure what you are doing, but, you might have chosen the wrong system aQ parameter.  Using the DEFAULT for initial process values may be more appropriate.s  N In the case of the WSEXTENT, it is normally set at process creation time, and ; not changed again.  Definitely a DEFAULT and not a MINIMUM.u   Dave   ------------------------------  % Date: Fri, 20 Dec 2002 22:53:15 -0500 2 From: rdeininger@mindspring.com (Robert Deininger)$ Subject: Re: just a couple of thingsL Message-ID: <rdeininger-2012022253150001@user-2ive15h.dialup.mindspring.com>  O In article <IMEM9.22240$Sy3.202091@news.chello.at>, peter@langstoeger.at wrote:u  I >The new version of the docset does still no bring TCPIP V5.3 docs in alllH >cases. Most of them are still 5.1 though listed under the 5.3 column...  E I expect those are manuals that were not updated for the 5.3 release.    ------------------------------  % Date: Fri, 20 Dec 2002 23:17:06 -0400 0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>$ Subject: Re: just a couple of things/ Message-ID: <3E03DD0D.52EBDA18@vl.videotron.ca>t  K > >The new version of the docset does still no bring TCPIP V5.3 docs in all J > >cases. Most of them are still 5.1 though listed under the 5.3 column...  N The 5.3 release notes provide some of the additional features of 5.3 that were5 not documened in those manuals that are still at 5.1.-   ------------------------------  % Date: Fri, 20 Dec 2002 22:38:29 +0100d5 From: "GWDVMS::MOELLER" <moeller@gwdvms.dnet.gwdg.de>o Subject: LCD monitor for VAX?g. Message-ID: <E18PUqj-0005JT-00@mailer.gwdg.de>   Merry Xmas everybody,5  : it's about time for my (office) VAX to find a LCD monitor,3 as it's become fashionable all around the place ...q6 (traditional connection is via 3 (RGB) BNC connectors, with "sync on green").  ? Did anyone test LCD monitors for "sync on green" compatibility?    Grateful for any leads,>M Wolfgang J. Moeller, Tel. +49 551 201-1516/-1510, moeller@gwdvms.dnet.gwdg.derM GWDG, D-37077 Goettingen, F.R.Germany     |    Disclaimer: No claim intended!pM http://www.gwdg.de/~moeller/ ---- <moeller@gwdg.de> ---- <w.moeller@ieee.org>r   ------------------------------  % Date: Fri, 20 Dec 2002 23:30:40 +0100 I From: "Eberhard Heuser-Hofmann" <Eberhard.Heuser-Hofmann@uni-konstanz.de>>: Subject: Re: looking for CD-RW or DVD-RW capability on VMS0 Message-ID: <006c01c2a877$6e6e6b60$0201a8c0@WS1>  7 For CD-RW use cdrecord (Version 1.11 coming very soon). / This program is free, but you cannot burn DVDs.p  I I'm offering for DVD-+R(W) recording my DVDwrite program (The price is 50e Euro):   http://home.tiscali.de/dvdwritey   regardst eberhard   ----- Original Message -----+ From: "Irving F. Snurd" <dacadc0@lycos.com>r To: <Info-VAX@Mvb.Saic.Com>s' Sent: Friday, December 20, 2002 4:50 PMc6 Subject: looking for CD-RW or DVD-RW capability on VMS    J > We're running several DS10s (Alpha, OpenVMS 7.3, PCI bus, IDE) and wouldH > like to be able to write files (or do backup) to CD or DVD.  Are thereI > any driver/hardware combinations out there yet?  We could go SCSI if weT
 > need to. >         Thanks for any help." >             Bill Manwaring, IUCF >l   ------------------------------  # Date: Sat, 21 Dec 2002 00:56:39 GMTx+ From: Jeff Cameron <JCam90502@jcameron.com>h: Subject: Re: looking for CD-RW or DVD-RW capability on VMS2 Message-ID: <BA28FC41.2917%JCam90502@jcameron.com>  G On 12/20/02 7:50 AM, in article 3E033C42.40740992@lycos.com, "Irving F.e! Snurd" <dacadc0@lycos.com> wrote:e  J > We're running several DS10s (Alpha, OpenVMS 7.3, PCI bus, IDE) and wouldH > like to be able to write files (or do backup) to CD or DVD.  Are thereI > any driver/hardware combinations out there yet?  We could go SCSI if wes
 > need to. >       Thanks for any help.  >           Bill Manwaring, IUCF > D The following page gives a full description including how to make an ISO9660/Files-11 Hybrid CD.   ? http://www.geocities.com/SiliconValley/Lakes/9999/vmscdwri.html    Jeff Cameron   ------------------------------  % Date: Fri, 20 Dec 2002 18:47:37 -0800s0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>= Subject: Re: Loss of Enet Performance on DS10 -> DS25 Upgrade , Message-ID: <3E0365C9.1F821B18@Mvb.Saic.Com>   Robbie Benton wrote: > D > I have two DS10 computers in the following configuration: (this isF > part of a larger configuration - but I think this is enough for this
 > discussion)t > * >                      +-----------------+* >                      |  Enet Switch A  |* >                      +-----------------+( >                        |             |( >                        |             |0 >                +----------+       +----------+0 >                |  CPU #1  |       |  CPU #2  |0 >                +----------+       +----------+( >                        |             |( >                        |             |* >                      +-----------------+* >                      |  Enet Switch B  |* >                      +-----------------+ > E > Both Ethernet Switches are identical and are 3Com SuperStacker 3300?E > 10/100 models.  The DS10s (OpenVMS V7.3-1, TCP/IP Services 5.3 with-E > the MUP, DECnet IV) are configured to use TCP on both networks, andn > DECnet only on Network B.l > E > A simple copy of a large file on CPU #1 to the NL: Device on CPU #2 F > provides an indication that the two networks are fairly equal:  WhenG > copying the file via FTP, both Network A and Network B provide a foureC > second copy time, while the DECnet copy on Network B completes inn > about 1.5 seconds. > E > I then upgraded CPU #1 to a DS25.  The DS25, like the DS10, has two9C > ethernet ports on the motherboard.  Since the DS10 was already at H > 7.3-1, I simply loaded an image copy from the DS10 to the DS25 for theD > system disk, and made the required changes for device names in theE > TCP/IP and DECnet configurations.  I connected the 10/100 Port A toh? > Network A, and connected the 10/100/1000 Port B to Network B.r > D > The DS10s are set up at the console mode to auto-negotiate on bothH > ports, so I configured the DS25 to do the same.  Both ports, accordingB > to the console messages, negotiated to 100 Full Duplex. This was > expected.c  G Here is your first issue.  You must check the devices at both ends of a E connection to verify that they both autoconfigure to the same thing. eE The fact that the Alpha reported that it negotiated 100/Full does not(= mean that the switch did as well.  You must explicitly check.   F For example, only the latest DS10 firmware (the version that came withF V7.3-1, i.e., V6.2-1) is any good at auto-negotiation with any varietyF of vendors.  All previous versions I tested failed with every switch IF have available (Cisco, Cabletron, Foundry, Extreme, Milan, etc.).  TheF most common symptom being that the Alpha would claim it had negotiatedC 100/Full and the switch would claim 100/Half.  The current version,rD however, works correctly with all the above vendor's switches I have tested.@  G In other words, it is quite likely that your DS10s are auto-negotiating 
 correctly.  H I have not had a chance to perform similar tests with any of our DS25s. G Again, the fact that the DS25 said that it negotiated 100/Full, or evenhH the fact that you hard set it to 100/Full, does not mean that the switchF is running 100/Full.  If you are not running the latest DS25 firmware, it probably isn't.  G > After the first few minutes of uptime, we noticed the Ethernet Switch C > Port Link Light for Network B (DS25) would occasionally drop out.bA > Expecting a problem with negotiation, we shut down the DS25 and E > configured Port B for 100 Full Duplex and rebooted.  We are in thate > state now.  C Auto-negotiation only happens at port initialization, i.e., once at C system startup, once again at VMS startup.  It does not continue toi@ negotiate during operations.  Some platforms will auto-negotiateH everytime you plug the ethernet cable in but I do not believe VMS is one of them.  A The symptoms you describe can indicate a bad cable, bad NIC, or a $ mismatch between the two end-points.   > F > Performing the same test, the file transfers on Network A via FTP inG > about the same four seconds.  However, the same file takes 34 secondseB > to transfer via FTP on Network B, and a whopping 1 minute and 50! > seconds to transfer via DECnet.i  F You've either got a bad NIC in the DS25 or the DS25 and the switch are* not in agreement on the ethernet settings.   > ( > So my questions to the group would be: > D > 1.  Is there additional configurations that I should have done for? > Port B, even though I am not trying to obtain gigabit speeds?e  B Yes, but on the switch not the DS25.  You can set the DS25 and theH switch to auto-negotiate and then compare both devices to make sure theyF came to the same settings.  If they do, you're done.  If not, then youG need to explicitly set the speed and duplex settings and you must do ita# on both devices, not just the DS25.r   > D > 2.  Even though I am not trying to obtain Gigabit speeds, must the > Ethernet Switch be upgraded?  ? The 3Com Superstack 3300 is supposed to be manageable via a webpH interface.  If yours is, you shouldn't need to upgrade.  Just get in andD check the switch settings.  If it is not manageable, at least to theE point of being able to manually define the speed and duplex settings,  then, yes, you need to upgrade.i  
 Mark Berrymann   ------------------------------  % Date: Fri, 20 Dec 2002 22:42:57 -0000 * From: "John Travell" <john@travell.uk.net>P Subject: Re: Machine checks not always hardware failures, was: Re: Machine Check5 Message-ID: <au06dp$3bfbt$1@ID-120847.news.dfncis.de>5  G "Simon Clubley" <clubley@remove_me.eisner.decus.org-Earth.UFP> wrote in 5 message news:OXUtF2p6V58I@eisner.encompasserve.org....7 > In article <3E00E7FB.8060100@Free.fr>, Didier Morandi/' <Didier.Morandi.nospam@Free.fr> writes:e > > JF,b > > * > > A machine check IS a hardware failure. > L > I would like to address the myth that a machine check is always a hardware0 > failure. This is usually true, but not always. >   - But it is *ALWAYS* a hardware DETECTED error!h  E There was once a problem with one of the early FDDI drivers. An errornH handler routine tried to access a register that was present on the earlyJ development boards, but which had been deleted from the production boards.K The resulting NXM invariably took the system out, a fatal machine check due J to software mapping, and trying to use, a register that no-longer existed.  J I also remember a situation from my first internals course, the instructorJ wrote a program, compiled it, ran it, deleted the evidence, logged out andE went for a coffee break. Soon after the Vax11780 crashed with a fatalr) machine check... Hardware ? Absolutely... K All the program did was to increment the base address of the second massbusAJ adapter. As soon as someone tried to access any device on that massbus theL hardware reported a fatal unaligned read to a register that must be accessed& by a longword aligned read or write... (This was 1985...!)>       ---d& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.426 / Virus Database: 239 - Release Date: 02/12/2002a   ------------------------------  % Date: Fri, 20 Dec 2002 20:26:08 -0400k0 From: JF Mezei <jfmezei.spamnot@vl.videotron.ca>P Subject: Re: Machine checks not always hardware failures, was: Re: Machine Check/ Message-ID: <3E03B509.9CA1F267@vl.videotron.ca>t   John Travell wrote: G > There was once a problem with one of the early FDDI drivers. An error J > handler routine tried to access a register that was present on the earlyL > development boards, but which had been deleted from the production boards.M > The resulting NXM invariably took the system out, a fatal machine check dueEL > to software mapping, and trying to use, a register that no-longer existed.  M ALL-IN-1 version 3.0 would always generated 12 errors on the disk drive where6I it resided during startup. (Something about odd byte count). These drives L appeared as RA82s to VMS. This stopped at 3.1 when they fixed whatever funky; stuff they were doing during the initialisation of allin-1.t   ------------------------------  # Date: Sat, 21 Dec 2002 06:08:12 GMTi1 From: Michael Austin <maustin@firstdbasource.com>i2 Subject: Re: MIME extension recognition under CSWS2 Message-ID: <3E040524.9BE03071@firstdbasource.com>   Martin Vorlaender wrote: >  > Hi,w > E > I just stumbled over an annoyance with CSWS 1.2 and didn't find any>4 > notice in the 1.3 release notes, so I'll ask here: > I > I have a few .PCSI-ZIP_AXPEXE files on my site, and would like to servegE > them as application/zip (or /octet-stream). So I started out adding  > * > AddType application/zip .pcsi-zip_axpexe > H > to conf/httpd.conf, and restarted. Didn't work - file gets served with > the DefaultType. > D > I entered the extension in conf/mime.types, and restarted. No luck	 > either.i > # > Am I just thick or is this a bug?  >  > Environment:$ > OpenVMS 7.2-1H1 with loads of ECOs$ > System disk ODS-2, data disk ODS-5 > CSWS 1.2 + UPDATE-4.0  >  > cu,m
 >   Martin > I > P.S.: For now, I've changed my DefaultType to application/octet-stream.e > --K > One OS to rule them all       | Martin Vorlaender  | VMS & WNT programmer-9 > One OS to find them           | work: mv@pdv-systeme.derK > One OS to bring them all      |  http://www.pdv-systeme.de/users/martinv/s@ > And in the Darkness bind them.| home: martin@radiogaga.harz.de      5 try changing the type to application/x-zip-compressedm7 try adding this filetype to the browser not the server.a  E I have attempted to use  a cgi script to accomplish this, but takes aiE few gyrations to get the file to look like it is supposed to... It istH late and I need some sleep, so, here is the start. see if you can get it to work:  C $ write sys$output f$fao("!AS!/!/", "Content-type: application/uu")o $ uuencode :== $TCPIP$UUENCODE) $ uuencode   <filename>.ZIP <filename>.uud $ exit  G I tried using copy and type to "download" the file, but it gets munged.      -- d Regards,  6 Michael Austin            OpenVMS User since June 19847 First DBA Source, Inc.    Registered Linux User #261163    ------------------------------    Date: 20 Dec 2002 11:20:38 -0800( From: bob@instantwhip.com (Bob Ceculski)6 Subject: Re: OT: Whoa! Is Sun aiming at VMS's jugular?= Message-ID: <d7791aa1.0212201120.568eeeba@posting.google.com>e   Andrew Harrison SUNUK Consultancy <Andrew_No.Harrison_No@nospamn.sun.com> wrote in message news:<3E033DB4.9020908@nospamn.sun.com>...d >> > tA > > well, if sparky and all these other junk cpus are better thand@ > > alpha, why did intel buy the alpha design and the alpha team? > > instead of sun or ibm?  because even intel knows a superiork@ > > chip when they see it, and they use it everyday on their fab< > > line to make their chips ... intelligent answers please! > 3 > The more I re-read your post the funnier it gets.c > 6 > Question for you, if Intel knows a superior CPU when# > they see one, then why have they:E >  > 1.	Continued with IA-64D0 > 2.	Bought according to you Alpha Team and then1 > 	not used them to do any more Alpha processors.  > 	 > regards  > Andrew Harrisonm   1. answer to number 2 is why  > 2. because they are using the alpha team to improve on itanium<    and implement alpha features over time to make it the cpu5    leader ... first alpha influences due out 2005 ... =    I talked to one of the hp engineers at a vms seminar, theye=    believe epic will be more efficient then an EV8 design ...l=    in other words, they are using the alpha team to implement3=    alpha goodies and create an EV8 like processor using epic.e8    Of course, we are all waiting to see the outcome, but:    itanium 2 looked rather decent, and with the alpha team?    contributions, I believe they will succeed ... but you stilla:    evaded my question, why not the sparky team and chip or?    ibm power?  They choose alpha ... almost every chip guru outt<    there states the same, alpha was a superior chip with bad>    marketing, where sparky and power are bad chips with glitzy<    marketing ... this isn't the first time inferior products9    outsold a superior one due to marketing strategies and .    lies and stupidity on the customer side ...   ------------------------------    Date: 20 Dec 2002 17:17:25 -0800" From: seadraem@hotmail.com (Susan)0 Subject: Partial memory option needed - PB7MA-AE= Message-ID: <dc6db063.0212201717.5a13914b@posting.google.com>   F I have an Alpha 1000 4/200 at home. I was trying to upgrade the memoryD but read too late an SOC and thought I could use the PB7MA-CE. Now IB have four sticks of memory that I can't use because I need five. IE have not had much luck in finding a single stick to upgrade with, thewA original seller isn't willing to take it back, and I don't have a48 spare $500-700 to throw away on the right set of memory.  D Does anyone know where I could find a single stick to round out thisD set? It is the Kingston version of the PB7MA-CE. As far as I've beenA able to tell, all I need is another stick that is the same memoryc< speed and contact type (gold contacts) and this should work.  
 Any ideas?   ------------------------------  % Date: Fri, 20 Dec 2002 22:49:53 -0500 2 From: rdeininger@mindspring.com (Robert Deininger)4 Subject: Re: Partial memory option needed - PB7MA-AEL Message-ID: <rdeininger-2012022249530001@user-2ive15h.dialup.mindspring.com>  = In article <dc6db063.0212201717.5a13914b@posting.google.com>,i# seadraem@hotmail.com (Susan) wrote:1  G >I have an Alpha 1000 4/200 at home. I was trying to upgrade the memoryoE >but read too late an SOC and thought I could use the PB7MA-CE. Now I C >have four sticks of memory that I can't use because I need five. I F >have not had much luck in finding a single stick to upgrade with, theB >original seller isn't willing to take it back, and I don't have a9 >spare $500-700 to throw away on the right set of memory.s >rE >Does anyone know where I could find a single stick to round out thishE >set? It is the Kingston version of the PB7MA-CE. As far as I've beenlB >able to tell, all I need is another stick that is the same memory= >speed and contact type (gold contacts) and this should work.s   I would try clearpoint.M   www.clearpoint.com  G Don't expect to find info for old or oddball memory on their web site. tI Give them a call.  If they don't have it, they'll make it.  If they can'tAD get the parts to make it, they'll try to find it on the used market.   ------------------------------  % Date: Fri, 20 Dec 2002 01:14:15 -0600r2 From: "Stuart Johnson" <ssj152 AT charter DOT net>& Subject: Slightly OT- Need VT-525 info/ Message-ID: <v05gq9a1ukh9f1@corp.supernews.com>e  I I recently got a used VT-525 terminal, provided with only the power cord. K Can anyone point me to a user's manual online, or is anyone willing to loan_L me such a manual for a short time? I haven't a clue how to setup and operateK this unit so as to take advantage of its advanced features. I wouldn't mindgF having access to a programmers manual for the VT-525 either. I will beB willing to pay postage both ways if someone has manual(s) to loan.  K I played with the thing and managed to get it into setup via the F3 key (bymK pressing the key shortly after power-up), and reset it to factory defaults.aJ That is as far as I have gotten. It would be VERY nice if I knew how to at least switch sessions...  K I am aware than Boundless currently makes these terminals, but haven't seen:E a place on their site to download or order a manual for the terminal..  F I am hoping to connect the VT-525 to a KVM switch so I can eliminate aJ VT-420 terminal and a computer monitor and keyboard. I have a Deserver 300L that I will connect the terminal serial I/O to that will allow me to control* the consoles of several VAXen and Alpha's.  J In case anyone wonders, yes, I run OpenVMS on all of the VAXen and all butL one of the Alpha's. I eagerly await Montgar's forthcoming 7.3 release of the
 Hobbyist kit.c   Thanks,  Stuart Johnson ssj152 AT charter DOT net    ------------------------------  # Date: Sat, 21 Dec 2002 01:21:34 GMTw+ From: "Bill Johnson" <res0xcil@verizon.net>t; Subject: Re: TECSys Consoleworks on VMS - first impressions 8 Message-ID: <ymPM9.68994$4W1.13346@nwrddc02.gnilink.net>   Guys:rL     Who Do I pay? Wow, from a vendors point of view, I really appreciate it!H We strive VERY VERY hard, everyone at TDi to get a potential customer or customers to say those things!  G     I will answer Robert in one word for ALL his request - YES, it doesaF that! Robert, please call support and speak with them or email them atG support@tditx.com, they will be glad to give you the CLI interfaces andgE commands for the things you want. Regarding console definition on theoJ command line use the old PCM console definitions format and import it fromJ the ADMIN section as a quick import of many definitions - maybe that works as well for you.  K     Really, check the support team out - I think you will find them to very  helpful!   Best Regards to you all! --Bill Johnson President - TECSys Development b_johnson at tditx dot com 972-881-1553  & <dittman@dittman.net> wrote in message2 news:uerM9.25738$3t6.25161@nwrddc03.gnilink.net...4 > Michael Austin <maustin@firstdbasource.com> wrote:G > > I currently am using ConsoleWorks by TECSys (an obvious play on thee> > > pronunciation as they are located in Austin, TEXAS :) )... >T- > Actually, they are located in Plano, Texas.  > -- o > Eric Dittman > dittman@dittman.netn? > Check out the DEC Enthusiasts Club at http://www.dittman.net/r& <dittman@dittman.net> wrote in message2 news:uerM9.25738$3t6.25161@nwrddc03.gnilink.net...4 > Michael Austin <maustin@firstdbasource.com> wrote:G > > I currently am using ConsoleWorks by TECSys (an obvious play on theh> > > pronunciation as they are located in Austin, TEXAS :) )... >a- > Actually, they are located in Plano, Texas.  > --   > Eric Dittman > dittman@dittman.neto? > Check out the DEC Enthusiasts Club at http://www.dittman.net/    ------------------------------  % Date: Fri, 20 Dec 2002 23:01:22 -0500   From: John Santos <JOHN@egh.com>, Subject: Re: Tracking down a TOP CPU Problem4 Message-ID: <1021220225604.400E-100000@Ives.egh.com>  % On Wed, 18 Dec 2002, Nic Clews wrote:a   > Jeff Zide wrote:> > > I have several users that at random times consume 90% plus< > > of the CPU on an Alpha 1200 running V7.1-1H1 of OpenVMS.8 > > The users only have one file open when this happens,# > > [VMS$COMMON.SYSLIB]FDLSHR.EXE;1 U > > I run a show dev/file on all of my disks and this is the only file open/attached.t6 > > I stop their PID and the system returns to normal.J > > All the users that this happens to are at a remote location connecting > > over a WAN.e >  > Let me guess, over TCPIP ? > I > I saw this issue some while ago, and because a user chose to switch offUG > their PC rather than end sessions correctly, it managed to force somesH > image into a loop on the VMS box. It also happened when disconnects or' > some other networking event happened.Y > G > And what about the LAT and DECnet users? No problems. Still, you've aaE > WAN to cover, so its just a hazard of the protocol you've selected.gH > TCPIP has totally different design principles to native DEC protocols. > G > Anyway, a suggestion to you is to use virtual terminals for these WANw > sessions, it should help.r  B V7.1-1H1???  Are you also running TCP/IP V5.1?  There was a bug in? the telnet server doing options negotiation that could cause ann; infinite loop (while allowing the telnet session to proceeds@ "normally", so the user had no idea he was causing the problem.)A Only some clients triggered this; I know that some Linux versionscA could cause it. I think this was fixed in the 1st ECO kit.  It ise+ definitely fixed in V5.3 (current version).t   -- t John Santost Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Fri, 20 Dec 2002 19:25:56 -0000 8 From: "Leigh Bowden" <LGBowden@bowdenfamily.fsnet.co.uk> Subject: Re: VAX 7810's./ Message-ID: <atvqqd$eps$1@newsg1.svr.pol.co.uk>l  ! Does that apply with FDDI though?h   -- Leigh G. Bowden.# http://www.bowdenfamily.fsnet.co.uk ! LGBowden#bowdenfamily.fsnet.co.uk  +44 161 477 2526   Opinions are mine alone.   ------------------------------  # Date: Fri, 20 Dec 2002 17:46:05 GMTc3 From: Vance Haemmerle <vance@toyvax.Glendale.CA.US>r% Subject: VAXstation 4000 cd-rom bezel.5 Message-ID: <3E02E6E4.3902AB73@toyvax.Glendale.CA.US>d  < I recently purchased a field installable 32x SCSI CD-ROM for; a VAXstation 4000 (RRD47-EN) from Compurex.  I was suprized ; that it didn't come with a front panel bezel to replace thez> blank one with the system.  MPRS says that spares for VAX 4000> are no longer for sale.  Does anyone have a spare one of these, I can have or know of a place I can buy one?   Thanks.    -- Vance Haemmerle-   ------------------------------  % Date: Sat, 21 Dec 2002 01:02:34 +0000a+ From: John Laird <john@laird-towers.org.uk>s) Subject: Re: VAXstation 4000 cd-rom bezel@8 Message-ID: <q4f70vk04kl7d6hq7p3hbqtp4vmne53olb@4ax.com>  C On Fri, 20 Dec 2002 21:57:57 GMT, VAXman-  @SendSpamHere.ORG wrote:e  l >In article <3E02E6E4.3902AB73@toyvax.Glendale.CA.US>, Vance Haemmerle <vance@toyvax.Glendale.CA.US> writes:> >>I recently purchased a field installable 32x SCSI CD-ROM for= >>a VAXstation 4000 (RRD47-EN) from Compurex.  I was suprizedm= >>that it didn't come with a front panel bezel to replace the @ >>blank one with the system.  MPRS says that spares for VAX 4000@ >>are no longer for sale.  Does anyone have a spare one of these. >>I can have or know of a place I can buy one? >e* >Time to break out the dremel tool, Vance.  K S'what I had to do on mine.  I removed the CDROM front panel, slid the bare K chassis up to the VS bezel from inside and marked the corners.  Drilled andeH cut between the holes.  That way, the drive slides in from the front andJ stops up against the panel. Otherwise you end up cutting a hole big enoughF for the whole drive to slide through, which looks even uglier.  I alsoH vaguely recall having to drill a locating hole in the plastic tray for a side screw.      	Johnh   ------------------------------  % Date: Fri, 20 Dec 2002 17:09:25 -0800o$ From: Shane Smith <ssmith@icius.com>) Subject: RE: VAXstation 4000 cd-rom bezel 0 Message-ID: <01C2A84A.93953270@sulfer.icius.com>  H Being something of a case modder I've been up against this kind of thingF before. I don't know how your bare CD rom looks, but you might want toF consider cutting the mounts off the bezel and mounting the whole thingF to the front of the CD tray. You'll need a hole over the eject button,G and possibly have to extend the button itself. When you hit the button, D the whole bezel comes forward. It's a relatively simple mod, and youG can't easily bork it as long as you're careful. People use it when theyuH buy fancy brushed aluminium cases and don't want ugly beige CDs sticking
 out of it.  9 I'm only happy when I'm programming, or dremeling.... :-)    Shanev   -----Original Message-----2 From: John Laird [mailto:john@laird-towers.org.uk]' Sent: Friday, December 20, 2002 5:03 PMi To: Info-VAX@Mvb.Saic.Com ) Subject: Re: VAXstation 4000 cd-rom bezelS    C On Fri, 20 Dec 2002 21:57:57 GMT, VAXman-  @SendSpamHere.ORG wrote:a  F >In article <3E02E6E4.3902AB73@toyvax.Glendale.CA.US>, Vance Haemmerle% <vance@toyvax.Glendale.CA.US> writes: > >>I recently purchased a field installable 32x SCSI CD-ROM for= >>a VAXstation 4000 (RRD47-EN) from Compurex.  I was suprizedp= >>that it didn't come with a front panel bezel to replace the @ >>blank one with the system.  MPRS says that spares for VAX 4000@ >>are no longer for sale.  Does anyone have a spare one of these. >>I can have or know of a place I can buy one? >e* >Time to break out the dremel tool, Vance.  F S'what I had to do on mine.  I removed the CDROM front panel, slid the bareG chassis up to the VS bezel from inside and marked the corners.  Drilled  andnH cut between the holes.  That way, the drive slides in from the front andC stops up against the panel. Otherwise you end up cutting a hole bige enoughF for the whole drive to slide through, which looks even uglier.  I alsoH vaguely recall having to drill a locating hole in the plastic tray for a side screw.m     	John    ------------------------------  % Date: Fri, 20 Dec 2002 23:13:42 -0600e! From: Dan Moore <dmoore@sosu.edu>H Subject: Vol Shadowing Questionr( Message-ID: <3E03F886.C92DB021@sosu.edu>  
 Greetings!  1     I would like a sanity check for our shadowingdH strategy/configuration. We have been using a starnd alone system for theE last few years and now we are moving to a 2 node cluster connected bytF gig Ethernet Fiber. Both are Alpha Systems running OpenVMS 7.3-1. Each8 Alpha contains one shadow set member for each data disk.  F    We are forced to use  a sperate system disk on each node because ofD the fiber connections. I would like any user to be able to log on toC either system and have the exact same access to the same data. This G requires that the system common files be available to each system, event% if the other system is not available.c  H   One thought is to mount  the system disk of our primary system on eachH node as a shadowset member. Then point the other  Alpha at the shawdowedH system disk of the primary system. This way, each Alpha will have a copyG of the system files (And the primary system will have a shadowed systemcF disk). We tested this and it seems to work well; however, I know thereF could be some down sides on this strategy depending upon what fails in what sequence.  A   Another thougth is to move these file to one of the data disks,tG therefore avoiding the issues that can occor with shadowed system diskso when things go wrong.   #   Any advice or insight is welcome.    Thanks,.   Dand        K -----------== Posted via Newsfeed.Com - Uncensored Usenet News ==----------UG    http://www.newsfeed.com       The #1 Newsgroup Service in the World!-M -----= Over 100,000 Newsgroups - Unlimited Fast Downloads - 19 Servers =-----o   ------------------------------    Date: 20 Dec 2002 14:22:50 -0800. From: spamsink2001@yahoo.com (Alan E. Feldman)O Subject: Re: Why does BACKUP/IMAGE need write access to INDEXF.SYS, BITMAP.SYS?s= Message-ID: <b096a4ee.0212201422.504ac8c7@posting.google.com>n  e Jeff Cameron <JCam90502@jcameron.com> wrote in message news:<BA21D1B4.246E%JCam90502@jcameron.com>...u [...]tJ > I believe what John has said is correct. The INDEXF.SYS file is open forL > write access as to prevent any changes to the file system while the backupL > image makes a complete mapping of the file system in it's internal tables. [...]a  E So what happens if some files are deleted or added to the disk duringtE the writing-to-the-save-set phase of the save operation? What happenscE to those files as regards the backup save set and what error messagest would one get, if any?  : And why does BACKUP/IMAGE need write access to BITMAP.SYS?  A And why does BACKUP/FAST need write access to INDEXF.SYS, but noto BITMAP.SYS?e  
 Thanks again.l   Alan E. Feldman  spamsink2001 yahoo com  = [read the following quote slowly in a moderately deep voice:]'  C Benjamin Franklin: <pause> The only president of the United States, A <pause> who never was president of the United States.  --FiresignM Theater (IIRC)   ------------------------------   End of INFO-VAX 2002.703 ************************