1 INFO-VAX	Mon, 07 Jun 2004	Volume 2004 : Issue 315       Contents: Re: AST level I/O  Re: AST level I/O  Re: Cluster disk mount problem Re: Cluster disk mount problem. Re: Computer foul-up has bank customers fuming DECnet-Plus DECdts Management  Re: few questions ! Re: Free surplus SW Prod Lib kits 	 Re: Merci 5 Re: More on Processors from May issues of "PROCESSOR" 1 Re: New IA64 binaries run on Old IA64 computers ?  Re: Not adding up? Re: Not adding up?& Re: OpenVMS Primer for System Managers& Re: OpenVMS Primer for System Managers& Re: OpenVMS Primer for System Managers9 Re: pdp11/05 complete system available; John  Wisniewski. 9 Re: pdp11/05 complete system available; John  Wisniewski. 9 Re: pdp11/05 complete system available; John  Wisniewski.  Re: Storing system files in CMS 3 Re: Sun To Open Source Solaris ... sort of ...maybe  Re: T4 and friends Re: Welcome FujitSUN  F ----------------------------------------------------------------------   Date: 7 Jun 2004 07:16:54 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: AST level I/O3 Message-ID: <IPfa1RE3+YBO@eisner.encompasserve.org>   b In article <d56d1c2d.0406061125.13954a15@posting.google.com>, cstranslations@msn.com (Joe) writes:H > I am wondering about I/O to SYS$OUTPUT from nonAST code which might beG > interrupted by AST code (which might generate an I/O to SYS$OUTPUT of H > its own). I know that if the synchronous thread is in the middle of anE > I/O, is interrupted, then the AST thread goes and issues an I/O the ' > program is going to fall down an die.  > E > I can block ASTs in the synchronous thread or is I need to write to H > SYS$OUTPUT in the AST open my own channel but what's bugging me is isF > there some non-zero probability that in the middle of the I/O in theD > synchronous thread is RMS going to go call $FLUSH so that I end up > with something like: > G > this I/O wasTHIS I/O WAS ISSUED FROM AN AST issued from the main loop   A   I don't know exactly what you're doing, but unless you're doing C   byte-by-byte output and forcing it to flush any buffers, you have B   two I/O requests, and the I/O subsystem will complete one before2   starting the other, so you'll always get one of:  & this I/O was issued from the main loop THIS I/O WAS ISSUED FROM AN AST    or   THIS I/O WAS ISSUED FROM AN AST & this I/O was issued from the main loop   ------------------------------  # Date: Mon, 07 Jun 2004 16:27:02 GMT # From: hoff@hp.nospam (Hoff Hoffman)  Subject: Re: AST level I/O2 Message-ID: <qH0xc.3076$r%6.1697@news.cpqcorp.net>  b In article <d56d1c2d.0406061125.13954a15@posting.google.com>, cstranslations@msn.com (Joe) writes:G :I am wondering about I/O to SYS$OUTPUT from nonAST code which might be F :interrupted by AST code (which might generate an I/O to SYS$OUTPUT ofG :its own). I know that if the synchronous thread is in the middle of an D :I/O, is interrupted, then the AST thread goes and issues an I/O the& :program is going to fall down an die.  C   That looks more like a programming bug, as I'd not expect that to    trigger the stated failure.   D :I can block ASTs in the synchronous thread or is I need to write toF :SYS$OUTPUT in the AST open my own channel but what's bugging me is isE :there some non-zero probability that in the middle of the I/O in the C :synchronous thread is RMS going to go call $FLUSH so that I end up  :with something like:  : F :this I/O wasTHIS I/O WAS ISSUED FROM AN AST issued from the main loop  F   Certainly possible given some forms of I/O in various languages, butE   the usual approach of dedicating a thread or the mainline code to a G   particular task such as terminal I/O does apply.  Treat your terminal J   I/O like any other task that you don't want interrupted and/or otherwise9   garbled -- this is standard parallel-processing coding.   C   I usually use a hardware queue or a mailbox or such to pass along G   such tasks to a server thread or to a loop in the mainline, as having C   an AST doing comparatively slow tasks such as terminal I/O is not ,   something I prefer to implement in an AST.  D   The AST drops work packets onto a queue processed by the mainline,A   and the mainline chugs aware until there is no more work in the B   queue.  The mainline then hibernates.  (I tend to also implementA   a scheduled wakeup call, in addition to the $wake issued by the A   AST.  This means the code can recover should a $wake get lost.) A   When the code awakens, it starts pulling entries off the queue, D   of course, and (usually) releasing packets back onto a free queue.  G   I also prefer to avoid blocking the AST activities via the big-hammer H   $setast operation, of course.  (There are a few cases where I will useJ   this as an expedience, but I tend to prefer to use queues or safe-update   sequences where I can.)     F   Ask The Wizard (1661) has pointers to the usual sorts of programmingI   problems, and pointers from there off to parallel-processing and shared     memory management discussions.  N  ---------------------------- #include <rtfaq.h> -----------------------------K     For additional, please see the OpenVMS FAQ -- www.hp.com/go/openvms/faq N  --------------------------- pure personal opinion ---------------------------E         Hoff (Stephen) Hoffman   OpenVMS Engineering   hoff[at]hp.com    ------------------------------  % Date: Mon, 07 Jun 2004 11:25:40 +0100 * From: Nic Clews <sendspamhere@[127.0.0.1]>' Subject: Re: Cluster disk mount problem ' Message-ID: <ca1fv1$o8u$1@lore.csc.com>    Alfred Falk wrote: > / > Nic Clews <sendspamhere@[127.0.0.1]> wrote in   > news:c9p9ld$el$1@lore.csc.com: >   & > > Go into SDA on the running system: > >  > > $ ANAL/SYS > >  > > SDA> SH LAN/COU  > >    > I > Current symptom:  erratically bad performance at times.  At bad times I J > have seen processes show up either RWSCS or PFW.  (Latter is puzzling toD > me as physical memory is more than sufficient and working sets are > adequate.)  D RWSCS is *largely* caused by cluster communication delays - ResourceE Wait (on) System Communication Services. Typically heavy locking, and G particularly if your system is doing lock directory lookups to a remote G node, this is either due to luck of the draw or inappropriate parameter G settings, but I'd rather try to steer clear of fiddling with parameters 	 just yet.   B $ MONITOR DLOCK and look for the OUTGOING DIRECTORY function, (DIRH FUNCTN RATE) and also the figures for the outgoing locking rates. If youC are seeing comparative to the local locking, more than 20% remote / F versus 80% (or less) local, then this is a probable cause of slowdown.  G You say PROCESSES, so we're outside the disk rebuild area I'm assuming.   H To further muddy the issue, I understand that some CISCO routers fail toE properly prioritize traffic. On your SHOW LAN/COU display, you'd have A seen the SCA traffic, there is also a packet type (that a network E manager can relate to) which is 60-07 (IP is 08-00) so make sure that E the "packet priority" is correct on the router. e.g. you could have a A fancy graphic display running over IP, but it has dependencies on D locking completing in a timely fashion, but if the two protocols areC fighting over the same bit of cable, yet one is given priority... I  think you know what I'm saying.   H So that's two things to check, the distributed locking behaviour, and if$ the router is introducing a problem.  G If we are seeing unwanted locking behaviour, we can probably advise how H to address it, but there's not usually a guarantee, and it may also meanE that the application needs rethinking how it uses the locking system.   H To the PFW. Page faults are caused when a process wants to access memoryB it does not have. Included in this are image files on disk, globalF shared memory and sections, and similar for global buffers for indexedF files with buffers (and also interchanging locking information!). It'sB quite normal to have a working set of a process up to the ideal orF maximum size, but if it maps to global pages, then it will page fault.H Image run up and run down is quite complex, and also often is the sourceH of high page faults. A wait can happen particularly for shared memory asG the locking co-ordinates access. De-access is also a component of this.   8 I'd hesitate a guess that the two effects are related...    ! > The only odd item is I see from  >   SDA> SHOW LAN COUNTER  > is >   CRC errors           10642( > (All other VMS nodes show 0 for this.) > F > Network manager has noticed "Late collision errors" for this node onG > CISCO switch.  Description suggests hardware problem, possibly cables E > and connectors.  Latter is unlikely.  I suppose I can try replacing  > DE600 with old DE500.   B This is also suggesting a reasonably busy network. If you can swap2 cables, then you can see if the behaviour moves...  ? If you like, post a MONITOR DLOCK after, oh, at least an hour's E monitoring. What are the respective LOCKDIRWT values for the nodes in H the cluster? How much are you using installed images, indexed files with global buffers?    --  ? Regards, Nic Clews a.k.a. Mr. CP Charges, CSC Computer Sciences  nclews at csc dot com    ------------------------------  $ Date: Mon, 7 Jun 2004 11:39:07 +0100* From: "Richard Brodie" <R.Brodie@rl.ac.uk>' Subject: Re: Cluster disk mount problem , Message-ID: <ca1gkc$1as2@newton.cc.rl.ac.uk>  6 "Alfred Falk" <falk@arc.REMOVE.ab.ca> wrote in message3 news:Xns94FE7FC5F2EC8falkarcabca@198.161.157.145...   F > Network manager has noticed "Late collision errors" for this node on > CISCO switch.   M A possible cause for this is having the node set to full duplex operation and - the switch to autonegotiate (or half duplex).    MC LANCP SHOW DEVICE/PARAM   / will show you what the duplex settings are now.    ------------------------------  $ Date: Mon, 7 Jun 2004 12:23:22 -0400< From: "Peter Weaver" <WeaverConsultingServices@sympatico.ca>7 Subject: Re: Computer foul-up has bank customers fuming * Message-ID: <2ijj3rFo06o7U1@uni-berlin.de>   JF Mezei wrote:  >...D > For them to have a foul up lasting a whole week, preventing direct@ > deposits (depriving customers of their money), blocking  debit@ > transactions because pension/paychecks were not deposited intoF > customer's accounts etc etc is a HUGE problem for a bank. This isn'tH > an outage of a few hours, it is something lasting almost a whole week. >...  F Those are just the points they are admitting to publicly. Not only wasD my mortgage payment processed three days late, a credit card paymentE showed correctly on Tuesday and Wednesday, then it appeared twice for H one day before it was corrected again on Friday. Two days last week theyE did not publish their Mutual Funds prices by their usual time and one F day the Mutual Funds section was not available at all. Tuesday, June 1F they sent out a message through their Internet Banking system but thatH message had a date of July 3, 2008. The next day the message was deleted from my read box.    --   Peter Weaver Weaver Consulting Services Inc.  Canadian VAR for CHARON-VAX  www.weaverconsulting.ca    ------------------------------  $ Date: Mon, 7 Jun 2004 13:07:39 -0400< From: "Peter Weaver" <WeaverConsultingServices@sympatico.ca>& Subject: DECnet-Plus DECdts Management* Message-ID: <2ijlmuFnou92U1@uni-berlin.de>  E Does anyone know where I can find the "DECnet-Plus DECdts Management" C manual in PDF (or HTML if PDF is not available) format? The various F DECNet-Plus manuals at http://h71000.www7.hp.com/doc/decnetplus73.htmlA mention this manual but the only on-line version I can find is at J http://www.sysworks.com.au/disk$vaxdocdec022/network/dcntp73/dn_dtsm1.bkb.  E I am trying to sync the clocks at a site that has a two node cluster, F and two other standalone machines. NTP is not an option for this site.   --   Peter Weaver Weaver Consulting Services Inc.  Canadian VAR for CHARON-VAX  www.weaverconsulting.ca    ------------------------------   Date: 7 Jun 2004 07:08:11 -0600 ; From: koehler@eisner.nospam.encompasserve.org (Bob Koehler)  Subject: Re: few questions3 Message-ID: <s5PtTEj6Nddp@eisner.encompasserve.org>   m In article <f401eb7f.0406041616.255bd228@posting.google.com>, soccer13player@yahoo.com (Nom de Plume) writes:  > D > Barry, yes and no.  I have to respectfully disagree.  UNIX and VMSD > perform mostly the same functionality but do it in a different wayF > (e.g.: filesystem, user interface, etcetera).  Take the command line > interface for instance:  > C    Get real.  UNIX was designed from the bottom up as a timesharing E    system.  VMS was designed as a general pupose system:  it has both C    timesharing and realtime capabilities designed in.  Anyone who's ?    done "realtime" on UNIX knows what functionality is lacking.    ------------------------------   Date: 7 Jun 2004 10:04:44 -0700 & From: jordan@ccs4vms.com (Rich Jordan)* Subject: Re: Free surplus SW Prod Lib kits< Message-ID: <cc5619f2.0406070904.7c5529c@posting.google.com>  h "Stanley F. Quayle" <squayle@insight.rr.com> wrote in message news:<40C3A0B5.5498.14756D1A@localhost>..., > On 4 Jun 2004 at 15:26, Rich Jordan wrote:C > > Heh!  I'm looking at my vari-colored wall of (almost) every VAX J > > CONDIST from late 1990 to September 2001, and Alpha CONDIST from MarchH > > 1997 to present, plus most of the VAX and Alpha VMS binaries kits onI > > TK50 and CDs... we've moved three times, each time the powers that be H > > tell me to trim them down, and each time I just can't make myself do> > > it, move them myself, and sneak them back onto my shelves. > H > Actually, Tom's sending the lot to me.  I have a 137 GB disk with all C > the CONDIST's I could find from V5.5 and on.  Unfortunately, the  F > original source of much of this dropped VAX support in the mid 90's A > ("dropped" isn't the right word, more like "mostly went out of  E > business").  The Alpha stuff came from another source, and is only   > the late 90's. >  .....  > --Stan Quayle  > Quayle Consulting Inc. >  > ----------/ > Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363 5 > 8572 North Spring Ct., Pickerington, OH  43147  USA 2 > stan-at-stanq-dot-com       http://www.stanq.com  E Since I got a 160GB peecee disk (with an Acard adapter) working in my E PWS600au I've been archiving the freeware disks plus all my own CD-Rs C and 'old work'.  Backing up the CONDISTS as well seems a fine idea, B though to be honest the media they are on now is probably far moreC durable than a consumer IDE drive, the DDS-2 tapes I back up to, or D any CD-R I might generate from them.  I'm considering finding a usedC media rated firebox to store them in.  There is no way I'd have the C bandwidth to serve anything to the net, so its personal use only on  this stuff.   E The only third party stuff I have is older version of WordPerfect/VAX 1 and Lotus 123, which require license PAKs to use.   D Maybe its time to look at the old TK50 stack; I know I have many preD V5.5 releases, though I doubt I've got anything from pre 5.0.  Those3 would be worth archiving to disk and to good CD-Rs.    Rich CCS    ------------------------------  $ Date: Mon, 7 Jun 2004 10:22:42 +03007 From: "Oleksii Krykun" <krikun@academy.kiev.ua.no.spam>  Subject: Re: MerciS Message-ID: <1037270357C4D411A1C900A0C9D4BFCB010E190A@hqnts40div01.academy.kiev.ua>   : "JF Mezei" <jfmezei.spamnot@teksavvy.com> wrote in message: news:850774cc9c4abc3564e5e57d8f55f42d@news.teranews.com... > Rich Alderson wrote: > > Il n'y a point des mots... > > + > > Vive la France!  Vivent les Etats-Unis!  > I > Vive le Canada, Vive l'Angleterre. Vive tout autre pays qui a particip  mais > qui n'est pas mention.    ??????? ?? ???!!!    ------------------------------  % Date: Mon, 07 Jun 2004 10:28:36 +0100 9 From: Andrew Harrison <andrew_._remove_harrison@su_n.com> > Subject: Re: More on Processors from May issues of "PROCESSOR"0 Message-ID: <ca1cg5$fv6$1@new-usenet.uk.sun.com>   David Svensson wrote: c > Paul Repacholi <prep@prep.synonet.com> wrote in message news:<87hdtuz23z.fsf@prep.synonet.com>...  > = >>Andrew Harrison <andrew_._remove_harrison@su_n.com> writes:  >> >>F >>>Tejas and Jayhawk are dead, Prescott looks very much like a dud andD >>>all Intel can show for their new very very expensive 90 nanometerE >>>process is a hotter, more complex and slower set of units. Imagine E >>>that Intel has just written of a 1.5 billion dollar investment for F >>>the 3 CPU's and their 2.5 billion and upwards investment in the FAB7 >>>isn't delivering a return as early as they expected.  >> >>   >>C >>>And then they have to fixup a new strategy, getting some designs F >>>done, goto their ISV's and saying start doing threading because theC >>>cores are just not going to go on getting faster at the rate you A >>>expected and fix the marketing none of which are cheap options 
 >>>either. >>G >>Intel went to IBM to liscence a pile of IBM FAB proscess tricknology. G >>The balked at IBMs price of cross licences to itanic etc, and are now E >>having to spend many years getting the heat and leakage out without  >>tripping over IBMs patents.  >  > F > Looks like Intel is not alone on 90nm problems.  IBM is having largeH > yield problems on 90nm and AMDs 90nm CPUs will draw 104W. Intel on theF > other hand has a killer in Pentium M, no one is near the performance > of Pentium M at 21W.   Well except for Opteron EE.    Regards  Andrew Harrison    ------------------------------  * Date: Mon, 7 Jun 2004 10:23:04 +0000 (UTC) From: david20@alpha2.mdx.ac.uk: Subject: Re: New IA64 binaries run on Old IA64 computers ?) Message-ID: <ca1fm8$sft$1@news.mdx.ac.uk>   h In article <c9vlu3$1f6$1@hercules.btinternet.com>, "Richard Maher" <maher_rj@hotspamnotmail.com> writes:
 >Hi David, > D >I live in Middlesex (East Twickers). Where is Middlesex University? >    Spread across North London.     A Main campuses in Hendon, Trent park, Cat Hill, Enfield, Tottenham F + a number of smaller campuses/halls of residence and also associated 
 hospitals.   See   % http://www.mdx.ac.uk/campus/index.htm         
 David Webb VMS and Unix team leader CCSS Middlesex University             >Regards Richard Maher.  > , ><david20@alpha2.mdx.ac.uk> wrote in message$ >news:c9ve5e$8bg$1@news.mdx.ac.uk...> >> In article <40C0C524.4070805@MMaz.com>, "Barry Treahy, Jr." ><Treahy@MMaz.com> writes: >> >JF Mezei wrote:  >> >> >> >>(Lets assume for the moment that IA64 survives 13 years). >> >>  >> >> M >> >>I can run the latest version of VAX-VMS on my teenage allmighty Microvax  >II. IL >> >>think that it is simply because only a couple of support files specific >to the @ >> >>MVII are needed early in the boot, and after that, VMS runs >transparently.  >> >>(Correct ?) >> >> L >> >>However, with compilers so tighly tied to the IA64 CPU to take advatage >of CPU H >> >>changes/imrprovemenst for performance, will a version of VMS OS andM >> >>applications compiled for Itanium 4  with Itanium 4 compilers be able to  >runI >> >>on Merced boxes ? (eg: in the future, will binaries for recent IA64s  >still  >> >>work on much older IA64s ?) >> >> G >> >>Won't IA64 compilers in the future generate code that assumes that  >certainL >> >>tricks are possible on IA64 ? What happens when that code executes on aG >> >>generation of IA64 that precedes the addition of those performance 	 >tricks ?  >> >> J >> >>Or is IA64 now a "stable" architectire with no more changes coming in >the wayL >> >>compilers generate code to provide about half the performance of IA64 ? >> >>  >> >> E >> >I'm not a Itanic fan, in fact I think it is the 21st centuries IT F >> >disaster just wanting to happen.  None the less, the same problemsE >> >impacted VAX & Alpha too, but this was dealt with at the hardware J >> >abstraction layer, so if Intel/HP use a HAL for IA64 to protect the OSD >> >from odd differences in chipsets that basically perform the sameC >> >function, it shouldn't be an issue...  If they don't then prior G >> >discussions of binary portability may come into play, but the other G >> >factor is that over 13 years, when considering how much VMS changes J >> >during the past decades, we've 'lost' binary portability several times? >> >with new hardware but also at major dot releases when major / >> >functionality was integrated into the OS...  >> >J >> Apart from the VAX -> Alpha transition I'm not aware of any lost binary2 >> portability of VMS due to changes in the chips.L >> There has of course always been the potential for lost binary portability >for@ >> user programs utilising kernel structures when upgrading VMS. >>I >> The possibility of tying VMS to the chip version has always been there  >sinceL >> the compilers such as C include /archicture qualifiers. However as far as >I am A >> aware VMS has never been complied with anything other than the  >/ARCH=GENERIC
 >> option. >>F >> Some thirdparty applications do utilise these architecture compiler >options for9 >> instance Oracle recently stopped supporting EV5 chips.  >>C >> Given that Itanium is much more dependent for performance on the  >optimisationsM >> performed by its compilers than Alpha I would suspect that both thirdparty L >> applications and the base VMS operating system will probably need to make >useB >> of such architecture qualifiers in the future in order to boost
 >performance. D >> Since as Oracle has demonstrated thirdparty suppliers do not like >supporting L >> multiple versions compiled with different compiler options it is probable >that M >> the lifcycle of a VMS itanium system will be much shorter than that for an I >> Alpha system if the users wish to use the latest versions of software.  >>
 >> David Webb  >> VMS and Unix team leader  >> CCSS  >> Middlesex University  >> >> >> >> >	 >> >Barry  >> > >> >-- >> >B >> >Barry Treahy, Jr                       E-mail: Treahy@MMaz.comB >> >Midwest Microwave                          Phone: 480/314-1320B >> >Vice President & CIO                         FAX: 480/661-7028 >> > >> > >> > >  >    ------------------------------  % Date: Mon, 07 Jun 2004 11:07:04 +0100 9 From: Andrew Harrison <andrew_._remove_harrison@su_n.com>  Subject: Re: Not adding up? 0 Message-ID: <ca1eo9$gpa$1@new-usenet.uk.sun.com>   perfnerd wrote:   f > "Jack Peacock" <peacock@simconv.com> wrote in message news:<8ImdncfsSKbepiLdRVn-uw@mpowercom.net>... > 1 >>>"John Smith" <a@nonymous.com> wrote in message  >>+ >> news:<U92dnaFADoTYrSDdRVn-jg@igs.net>...  >>A >>>>http://www.theregister.co.uk/2004/05/28/itanium_opteron_q104/  >>>>I >>>>....The latest server data from Gartner shows that only 6,281 Itanium  >> >> boxes >>L >>>>and 31,184 Opteron boxes were shipped in the first quarter of this year. >>> J >>The numbers for Itanium don't seem to be adding up.  I thought Intel hadN >>shipped 100,000+ Itanium CPUs last year?  What are the year on year shippingN >>numbers?  Are all those 6281 boxes last quarter fitted with at least 4 CPUs? >>  Jack Peacock >  > < > Since the average price for the Itanium box is $45,000 andE > 100K/6281=~16, so I would think that puts the average box in the 16 C > CPU range.  While the Opteron sales appear to be in the 1 - 2 CPU  > range.    ? The cheapest way of buying a 4 way HP Itanium box is to buy the A rx4640 and a 4 way system with 8 GB of RAM lists for 60K dollars.   = The rx5670 which was until recently the only way of getting a ? 4 way unit from HP starts is about 12K more than the rx4640 for  the same config.  > Even the rx4640 is priced at over 2x the price of a equivalent; config but faster Opteron based system even the HP DL585 no = price beater lists at ~27K under half the price of the rx4640 > and its quicker for most of the things that people care about.  > The last numbers I saw said that in fact the average number of. CPU's in Itanium systems was just less than 3.   Regards  Andrew Harrison    ------------------------------   Date: 7 Jun 2004 07:38:05 -0700 ' From: icerq4a@spray.se (David Svensson)  Subject: Re: Not adding up? = Message-ID: <734da31c.0406070638.4b2a7ce1@posting.google.com>   q Andrew Harrison <andrew_._remove_harrison@su_n.com> wrote in message news:<ca1eo9$gpa$1@new-usenet.uk.sun.com>...  > perfnerd wrote:  > h > > "Jack Peacock" <peacock@simconv.com> wrote in message news:<8ImdncfsSKbepiLdRVn-uw@mpowercom.net>... > > 3 > >>>"John Smith" <a@nonymous.com> wrote in message  > >>- > >> news:<U92dnaFADoTYrSDdRVn-jg@igs.net>...  > >>C > >>>>http://www.theregister.co.uk/2004/05/28/itanium_opteron_q104/  > >>>>K > >>>>....The latest server data from Gartner shows that only 6,281 Itanium  > >>
 > >> boxes > >>N > >>>>and 31,184 Opteron boxes were shipped in the first quarter of this year. > >>> L > >>The numbers for Itanium don't seem to be adding up.  I thought Intel hadP > >>shipped 100,000+ Itanium CPUs last year?  What are the year on year shippingP > >>numbers?  Are all those 6281 boxes last quarter fitted with at least 4 CPUs? > >>  Jack Peacock > >  > > > > > Since the average price for the Itanium box is $45,000 andG > > 100K/6281=~16, so I would think that puts the average box in the 16 E > > CPU range.  While the Opteron sales appear to be in the 1 - 2 CPU 
 > > range. >  > A > The cheapest way of buying a 4 way HP Itanium box is to buy the C > rx4640 and a 4 way system with 8 GB of RAM lists for 60K dollars.  > ? > The rx5670 which was until recently the only way of getting a A > 4 way unit from HP starts is about 12K more than the rx4640 for  > the same config. > @ > Even the rx4640 is priced at over 2x the price of a equivalent= > config but faster Opteron based system even the HP DL585 no ? > price beater lists at ~27K under half the price of the rx4640 @ > and its quicker for most of the things that people care about. > @ > The last numbers I saw said that in fact the average number of0 > CPU's in Itanium systems was just less than 3.  E I can buy a 4-way rx4640 with 8GB for 30K on the online store and get ) even better deals through other channels.    ------------------------------  # Date: Mon, 07 Jun 2004 15:26:54 GMT , From: "warren sander" <warren.sander@hp.com>/ Subject: Re: OpenVMS Primer for System Managers 1 Message-ID: <2P%wc.3060$817.144@news.cpqcorp.net>   D in answer the to question.. Yes I know what's wrong.. I'm an idiot..  H well not that much but I put our metrics javascript on every page and it broke something.L It's my fault for only checking the netscape version and not the IE version.  I I've backed the code out (actually restored a copy of it from my archives  and it seems to  be working again).  I I don't know what the status of the primer will be in the future as it is  very 'compaq' and weK are now 'HP'. I've requested to have it re-written, updated and branded but  that is not happening.  0 If you do like the primer please respond back...   -warren     B "Bradford J. Hamilton" <brad@rabbit.dnsalias.org> wrote in message& news:O09wc.52226$Ly.11802@attbi_s01...? > In article <eb55ac2a.0406040402.5e2e08e3@posting.google.com>, 5 pa@it.singer-friedlander.com (Piyush Avichal) writes: I > !Anyone know whats wrong with the OpenVMS Primer for System Managers at 1 > !http://h71000.www7.hp.com/wbt/pc/welcome.htm ?  > !Most of the pages are blank.  > ! ! > !Has anyone used this recently?  > L > Despite Hoff's reply, I just visted this site yesterday, and had no issuesE > navigating it using Mozilla Firefox on a W2K PC (/ducks for cover).  > L > I just re-visited the site now, using the same browser/HW combo, and still had I > no issues navigating through the site (Java and Javascript are enabled.  /ducks > again for more cover...) :-) >  > ! 
 > !Thanks, > ! 	 > !Piyush  > L > __________________________________________________________________________C > Bradford J. Hamilton                    "All opinions are my own" L > bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-'2 >                                          with @"   ------------------------------  # Date: Mon, 07 Jun 2004 15:09:44 GMT 5 From: brad@rabbit.dnsalias.org (Bradford J. Hamilton) / Subject: Re: OpenVMS Primer for System Managers - Message-ID: <Yy%wc.22204$%F2.20292@attbi_s04>   ` In article <2P%wc.3060$817.144@news.cpqcorp.net>, "warren sander" <warren.sander@hp.com> writes:E !in answer the to question.. Yes I know what's wrong.. I'm an idiot..  ! I !well not that much but I put our metrics javascript on every page and it  !broke something. M !It's my fault for only checking the netscape version and not the IE version.  ! J !I've backed the code out (actually restored a copy of it from my archives !and it seems to !be working again).  ! J !I don't know what the status of the primer will be in the future as it is !very 'compaq' and we L !are now 'HP'. I've requested to have it re-written, updated and branded but !that is not happening.  ! 1 !If you do like the primer please respond back...   
 Hi Warren,  O I recommend the primer to folks who are "new" to VMS, or returning after a long  hiatus.	:-)   M I found it useful when first starting out as a System Manager, among the many C other "tools" available to use.  I think it's worth saving in 'HP'.    !  !-warren !snip!  J __________________________________________________________________________A Bradford J. Hamilton                    "All opinions are my own" K bMradAhamiPltSon-at-coMmcAast.nPeSt     "Lose the MAPS, and replace '-at-'  0                                          with @"   ------------------------------  % Date: Mon, 07 Jun 2004 12:16:32 -0400 2 From: "Stanley F. Quayle" <squayle@insight.rr.com>/ Subject: Re: OpenVMS Primer for System Managers / Message-ID: <40C45CA0.18648.17537B17@localhost>   , On 7 Jun 2004 at 15:26, warren sander wrote:H > I don't know what the status of the primer will be in the future as itA > is very 'compaq' and we are now 'HP'. I've requested to have it < > re-written, updated and branded but that is not happening. > 2 > If you do like the primer please respond back...  @ The primer is incredibly useful.  I point prospective customers  towards it.   D Perhaps you could get the training people to invest in the rewrite, F since the last section steers them into the Compaq training program.  D And since HP has recently begun to offer training on OpenVMS again,  this would be a great tie-in.   B Might I suggest some third-party links for reference material and  also some success stories?  
 --Stan Quayle  Quayle Consulting Inc.  
 ----------- Stanley F. Quayle, P.E. N8SQ  +1 614-868-1363 3 8572 North Spring Ct., Pickerington, OH  43147  USA 0 stan-at-stanq-dot-com       http://www.stanq.com   ------------------------------  % Date: Mon, 07 Jun 2004 18:22:33 +0800 , From: Paul Repacholi <prep@prep.synonet.com>B Subject: Re: pdp11/05 complete system available; John  Wisniewski.- Message-ID: <87aczffz12.fsf@prep.synonet.com>   6 patrick jankowiak <nounwantedemail@swbell.net> writes:  E > I did refer to that page, but found the page ambiguous because note = > #3 on the same page seems to indicate this unless read very  > carefully:  D > The 11/20 docs referred to a stripped down model, the 11/10, whichE > became the OEM version of the 11/05.  So therefore, the 11/05 begat C > the 11/10(oem), which was mentioned later in the 11/20 documents.   C That is a KA-11, NOT the KD-11 based second 11/10! That is the only ; 11 that as far as I know, used a re-tread model designater.    ? > Or it can be taken that way. The written mention in the 11/20 ? > document, and then the actual numbering of, 'pdp11/10', being : > applied to two different machines, has caused confusion.  F Yep. The original 11/10 used a KA-11 in a half length BA-11. The otherF members of thr originals was the 11/15 with a KC-11 CPU, and the 11R20+ that was a rugidized version of the 11/20.     --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076* comp.os.vms,- The Older, Grumpier Slashdot. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Mon, 07 Jun 2004 10:21:41 -0600 4 From: Norman Lastovica <norman.lastovica@oracle.com>B Subject: Re: pdp11/05 complete system available; John  Wisniewski.* Message-ID: <40C49615.54DDF9D7@oracle.com>  2 The 11/20 was the first PDP 11 model produced.  In5 school, I used serial number 166 (ie, the 66th PDP11   off the line).   Renaissance wrote: >  > patrick jankowiak wrote: > : > > If you ever wanted to own a pdp11/05, the first pdp11, >  > uhm... the first?b0 > http://www.psych.usyd.edu.au/pdp-11/table.html > 
 > bye G.L. >  > --3 >  E' assolutamente evidente che l'arte del cinemae; >   si ispira alla vita, mentre la vita si ispira alla TV b > (Woody Allen)d   -- o	 - - - - - 0  opinions expressed here are mine and mine alone.  and certainly are not intended in any way to 0  express or represent any opinions or commitment  of oracle corporation.c  *  norman lastovica / oracle rdb engineering   ------------------------------  % Date: Mon, 07 Jun 2004 19:03:38 +0200m( From: Renaissance <glucasole@tiscali.it>B Subject: Re: pdp11/05 complete system available; John  Wisniewski.* Message-ID: <2ijlfbFmu9m9U1@uni-berlin.de>   patrick jankowiak wrote:  * > Since I am not a computer history expert  A Oh, it's only a small note.  :-) I'm not a retrocomputing expert,@: too. However i have a little passion for computer history.   Bye, G.L.   -- g1  E' assolutamente evidente che l'arte del cineman9   si ispira alla vita, mentre la vita si ispira alla TV  
 (Woody Allen)m   ------------------------------   Date: 7 Jun 2004 00:59:41 -0700o( From: pulley_p@hotmail.com (Phil Pulley)( Subject: Re: Storing system files in CMS= Message-ID: <7041fc41.0406062359.54f662a8@posting.google.com>a  e David J Dachtera <djesys.nospam@comcast.net> wrote in message news:<40C38785.E15FC0D3@comcast.net>...a > David J Dachtera wrote:d > >  > > Phil Pulley wrote: > > >y? > > > I am trying to store the following system files in CMS...S > > >m > > > Sysuaf.dat > > > Vms$audit_server.dat > > > vue$profile.vue$datl > > >t% > > > ...but it doesn't seem to work.c > > > J > > > I can (apparently)create the element OK but, when I try and fetch orK > > > reserve it CMS terminates & the file is not retrieved (I am using theoJ > > > Decwindows front end to access CMS) however - I can store & retrieve > > > text files OK.J > > > I suspect that CMS does not handle this type of file but the manuals* > > > do not confirm that this is the case > > ; > > I'd be curious as to what the value of this would be...  > A > Am I to conclude by the lack of response that this is merely anW) > exercise, and holds no practical value?  >  > D.J.D.   Apologies for late response-  E My reason for attempting to store these files in CMS is as follows...f  E We have several stand-alone Alpha workstations at different locationsn' running a Satcomms related application.0C These all need to be configured the same way in terms of user id's,wE audit setup and applications available during Decwin sessions - thus,m? to achieve this,  the files listed need to be installed on eacho workstation.  B As all the application source code etc. comprising the baseline isD stored in a set of CMS libraries I thought it would be neat to store# these system files in the same way.a  C Obviously this is not essential, they could be stored outside CMS -h? however I could not see any reason why I should NOT be able to.e   Hope this clarifies things.d  A I shall now try some of the suggestions in earlier posts to try &u( resolve the problems & post the results.   Phil P   ------------------------------  % Date: Mon, 07 Jun 2004 11:12:30 +0100c9 From: Andrew Harrison <andrew_._remove_harrison@su_n.com>t< Subject: Re: Sun To Open Source Solaris ... sort of ...maybe0 Message-ID: <ca1f2f$gth$1@new-usenet.uk.sun.com>   Bill Todd wrote:  0 > "John Smith" <a@nonymous.com> wrote in message( > news:pemdnQneyfJ_bV3dRVn-jg@igs.net... > N > http://story.news.yahoo.com/news?tmpl=story&ncid=1209&e=5&u=/nf/24366&sid=95 >  >>573734 >  > I > That's nice, I guess.  But I'm more interested in the new 'dynamic fileoN > system' that they just announced, which at least claims to support the kindsJ > of features I've been suggesting for years:  anyone seen any substantive8 > information on it yet (I didn't find any more than the> > press-release/high-level-manager comments stuff at sun.com)? >   ? Access to Solaris Dynamic Filesystem AKA ZFS can be had through,= the Software Express Program which gives people with existing ; Solaris licenses access to the latest builds of Solaris 10.o  = The Sofware Express distributions are for SPARC and x86 whichs also means Opteron.   > Also in Solaris 10 is dtrace, which is a very very interesting new tracability capability.p   Regardsi Andrew Harrison    ------------------------------  % Date: Mon, 07 Jun 2004 08:26:41 +0200 * From: Paul Sture <nospam@sture.homeip.net> Subject: Re: T4 and friends * Message-ID: <2iig53Fn4gtoU1@uni-berlin.de>   JF Mezei wrote:i > David J Dachtera wrote:e > F >>One key item to remember: in order to be truly valuable, performanceJ >>reporting solutions *MUST* be 100% automated. They *MUST* be able to run4 >>in batch, unattended, scheduled by whatever means. >  > > > The collection and perhaps some processing of data should beG > "background/automated. But it should be viewable on-line very easily.  > N > But accessing graphs and displays should be available interactively at will.M > It is rather pointless to automatically print reports that you may not use. P > Much better to have a list of available reports on-line, and view the ones youC > wish, and possibly print the one or two you really need on paper.   I Moving the emphasis away from performance reporting to reporting for the e whole business here...  I I came across the following product nearly 10 years ago. The idea behind 1I it then was that you archive your ASCII spoolfiles to optical disk (COLD eF - Computer Outper to Laser Disk), and then you can not only call them G back for auditing purposes, but can reconstruct data long since purged  I from databases, or simply analyze it. Today's large disk capacities mean zI you don't need the often expensive hardware/software solutions that were a on offer back then of course.a  I http://www.datawatch.com/dataconversionsoftware/monarch-report-mining.aspm  G "Monarch software, the world's #1 Report Mining software solution, may sF be the only data access and analysis tool you ever need. With Monarch G software, there's no need to flip through thick printed reports, rekey hC data into spreadsheets, or struggle with complex SQL data analysis OH tools. Instead, Monarch transforms your report files into data, without ? programming! Work smarter, with Monarch report mining software!r  E A powerful data analysis tool, Monarch lets you easily sort, filter, s? summarize and graph report data, export to Excel, Access, etc.  E Additionally, Monarch Pro can access additional data sources: Excel, g Access, ODBC, HTML, etc.  C Keep your data, lose the paper, with Monarch software: the premier z& report mining and data analysis tool!"   ------------------------------  % Date: Mon, 07 Jun 2004 09:00:19 +0100w9 From: Andrew Harrison <andrew_._remove_harrison@su_n.com>o Subject: Re: Welcome FujitSUNr0 Message-ID: <ca17b0$edu$1@new-usenet.uk.sun.com>   Karsten Nyblad wrote:p= > "Fabio Cardoso" <fabiopenvms@yahoo.com.br> wrote in messagee9 > news:f30679fb.0406011721.315076f3@posting.google.com...h > I >>http://www.sun.com/smi/Press/sunflash/2004-06/sunflash.20040601.14.htmlw >  > N > In the old days there was a UK computer company called ICL.  It did not haveM > the money to develop its own CPUs, so it started buying Fujitsi.  Later, itnI > got in even more trouble, and because of ICL buying Fujitsi, ICL had nouM > choice but to let Fujitsi buy ICL.  It seems like the links www.icl.com andoK > www.icl.co.uk are not working, at least when I tried this morning.  Try ar7 > search on Google for Fujitsi and ICL.  Very few hits.i > I > I hope for SUN that McNeally has been reading every single comma in the K > contract, or SUN will be the next company eaten by Fujitsi.  I wonder howc > Andrew will spin this. >  > Karsten Nyblad! > ibpit1202 at sneakemail dot com  >  > D Sun has had contracted deals with Fujitsu for years. Fujitsu resellsF Sun systems, OEMs Solaris and is a SPARC licensee. If you can find anyH parallels at all in the Sun Fujitsu relationship and the ICL Fujitsu one" then please feal free to air them.   The spin seems all yours.e   Regards  Andrew Harrisona   ------------------------------   End of INFO-VAX 2004.315 ************************