1 INFO-VAX	Sun, 11 May 2003	Volume 2003 : Issue 259       Contents:) Re: AlphaStation 255/500 power up problem ) Re: AlphaStation 255/500 power up problem 9 as2000 power connectors (was: Re: as2100 inside pictures)  Re: Backup question  Re: Backup question  Re: Backup question . Re: Computerworld: HP continues to support VMS. Re: Computerworld: HP continues to support VMS. Re: Computerworld: HP continues to support VMS) Re: DEC TCP/IP Services 5.3 for Hobbyist? A Re: DECnet-Plus (DECnet PhaseV) (DECnet/OSI) migration necessary? ! Re: Go to an AS/400 from VAX/VMS? < Re: How Alpha will save Itanium - must reading for Bill Todd< Re: How Alpha will save Itanium - must reading for Bill Todd< Re: How Alpha will save Itanium - must reading for Bill Todd* Re: How to determine boot device from DCL?* Re: How to determine boot device from DCL?* Re: How to determine boot device from DCL?" Re: INDEX.SYS size and performance" Re: INDEX.SYS size and performance" Re: INDEX.SYS size and performance" Re: INDEX.SYS size and performance" Re: INDEX.SYS size and performanceP Re: Not fixed yet but still trying: Problems changing from serial port to DecSerP Re: Not fixed yet but still trying: Problems changing from serial port to DecSerP Re: OT - PVS (was Re: DECnet-Plus (DECnet PhaseV) (DECnet/OSI) migration necessa Re: Splitting cluster   F ----------------------------------------------------------------------  # Date: Sat, 10 May 2003 19:23:14 GMT ' From: "Mark E. Levy" <melevy@attbi.com> 2 Subject: Re: AlphaStation 255/500 power up problem? Message-ID: <Ckcva.281675$Si4.225644@rwcrnsc51.ops.asp.att.net>   5 "Bob Ceculski" <bob@instantwhip.com> wrote in message 7 news:d7791aa1.0305100425.61cda074@posting.google.com... ; > you are thinking right ... and you are batting 2 for 2 on  > power supplies ...  G Bob, just how do you do it? You are the king of helpful answers. With a K response like that, he should be able to get immediately to the root of the  problem and solve it.   H There are two things that can cause the power supplies to shut down likeJ that. Either they're overloaded (too many disk drives, for example) or theH fans are bad. It's also quite possible that the sensing circuitry in theE supplies is bad, making them "think" that there is an overload or fan  problem.  K Try disconnecting everything but the motherboard. If the power stays on and J the fans appear to spin at the right speed, try reconnecting other devices: one at a time. If nothing works, you might have a bad p/s.   --   Mark E. Levy" System Management Associates, Inc. www.sysman-inc.com www.openvms.com    ------------------------------    Date: 10 May 2003 18:25:51 -0700( From: bob@instantwhip.com (Bob Ceculski)2 Subject: Re: AlphaStation 255/500 power up problem< Message-ID: <d7791aa1.0305101725.8ff6376@posting.google.com>  n "Mark E. Levy" <melevy@attbi.com> wrote in message news:<Ckcva.281675$Si4.225644@rwcrnsc51.ops.asp.att.net>...7 > "Bob Ceculski" <bob@instantwhip.com> wrote in message 9 > news:d7791aa1.0305100425.61cda074@posting.google.com... = > > you are thinking right ... and you are batting 2 for 2 on  > > power supplies ... > I > Bob, just how do you do it? You are the king of helpful answers. With a M > response like that, he should be able to get immediately to the root of the  > problem and solve it.   = I deal with these boxes all the time and 99% of the time it's 9 the power supply!  Just don't go giving the guy any legal " advice or he'll end up in jail ...   ------------------------------    Date: 10 May 2003 23:23:42 +03009 From: costello@iki.fi (Antti =?iso-8859-1?q?J=E4rvinen?=) B Subject: as2000 power connectors (was: Re: as2100 inside pictures)2 Message-ID: <m37k8ytusx.fsf_-_@muikku.katiska.org>  > "Peter Weaver" <WeaverConsultingServices@sympatico.ca> writes: > You might want to check out @ > http://h18002.www1.hp.com/alphaserver/archive/2100/2100_tech.h= > tml, that page should tell you everything you need to know.   E There is some troubleshooting documentation package, do I understand  , right or does it require ms-windows to run?   I I was under impression that the machine I'd be getting would be as2100 ;  D it turned out to be as2000 and after having somewhat assembled that A thing it after reboot lights up OPC, says "test cpu_00 xyz", then G "test mem_00 xyz" where xyz are changing numbers and then powers itself - off. Anybody has clue what might be going on?   G One thing that comes to my mind is fan. The thing came without original G cooling propeller(s?) and would it be that this machine has some clever F safety mechanism that if fan is missing or out of order it won't run? G Where does original cooling fan get its power from? In system backplane ! there is one empty connector, see ? http://muikku.katiska.org/~costello/as2000_power_connectors.png ; what should be connected to connector pointed with ? -sign?   ? What else might it be? What error situations lead to power off?    --   Antti Jrvinen, costello@iki.fi 5             "concerto for two faggots and orchestra"     ------------------------------  % Date: Sat, 10 May 2003 13:41:17 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: Backup question' Message-ID: <3EBD47CD.D2E4163A@fsi.net>    Dirk Munk wrote: >  > Dirk Munk wrote:3 > > I made two image backups to one tape like this:  > > & > > backup/image dka0: mka500:dka0.bck* > > backup/image dka200: mka500:dka200.bck > > % > > Both backup sets are on the tape. 5 > > I was able to verify that with dir MKA400: /size.  > > A > > Now I want to extract some files from the dka200.bck saveset. . > > But somehow I can't get backup to do that. > >  > > I tried: > > 7 > > backup mka500:dka200.bck /select= etc. but no luck.  > > , > > I get the response dka200.bck not found. > > # > > Can somone tell me what to do ?  > >  > > Thanks.... > >  > ' > Thank you all for your contributions.  > 2 > This is the way I was able to extract the files: >  > mount /for mka500: > K > set mag mka500: /skip=files:2   (the tape labels are also seen as a file)  > 7 > now the tape is positioned at the start of dka200.bck  > R > backup mka500:/select=(.....)   (no specification of the name of the backup kit)  ' Hhmmm... Good know. Thanx for that one.   D Does backup then figure out the blocksize based on the length of the first tape I/O? I wonder...    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Sat, 10 May 2003 22:23:39 +0200  From: Dirk Munk <munk@home.nl> Subject: Re: Backup question2 Message-ID: <b9jn6m$qs4$1@news3.tilbu1.nb.home.nl>   David J. Dachtera wrote:  ' >>Thank you all for your contributions.  >>2 >>This is the way I was able to extract the files: >> >>mount /for mka500: >>K >>set mag mka500: /skip=files:2   (the tape labels are also seen as a file)  >>7 >>now the tape is positioned at the start of dka200.bck  >>R >>backup mka500:/select=(.....)   (no specification of the name of the backup kit) >  > ) > Hhmmm... Good know. Thanx for that one.  > F > Does backup then figure out the blocksize based on the length of the > first tape I/O? I wonder...  > Q No, there is a tape label that specifies the the record and block lengths of the  Q file. Somewhere in the VMS documentation you can find a lot of information about  &   the structure of ANSI labeled tapes.   ------------------------------  % Date: Sat, 10 May 2003 17:24:56 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: Backup question& Message-ID: <3EBD7C38.4FA7A61@fsi.net>   Dirk Munk wrote: >  > David J. Dachtera wrote: > ) > >>Thank you all for your contributions.  > >>4 > >>This is the way I was able to extract the files: > >> > >>mount /for mka500: > >>M > >>set mag mka500: /skip=files:2   (the tape labels are also seen as a file)  > >>9 > >>now the tape is positioned at the start of dka200.bck  > >>T > >>backup mka500:/select=(.....)   (no specification of the name of the backup kit) > >  > > + > > Hhmmm... Good know. Thanx for that one.  > > H > > Does backup then figure out the blocksize based on the length of the > > first tape I/O? I wonder...  > > R > No, there is a tape label that specifies the the record and block lengths of theR > file. Somewhere in the VMS documentation you can find a lot of information about( >   the structure of ANSI labeled tapes.  D I misunderstood your statement. I thought you positioned the tape toA beginning of the saveset, after it's label records. You indicated G /skip=file=2, which would leave the tape positioned at the beginning of H the EOF records for the first saveset. So, my original thought was wrong for two reasons.  H The V5 or so VMS doc.'s had a really good layout of ANSI tape labels. It has since disappeared.  H I've done a good amount of work with magtapes, actually. I wrote some DRH DCL that reads the tape labels to try to determine what the correct next@ action should be, rather than using BACKUP to skip over savesets  selectively. Saves LOTS of time!   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  # Date: Sat, 10 May 2003 18:27:25 GMT # From: "John Smith" <a@nonymous.com> 7 Subject: Re: Computerworld: HP continues to support VMS I Message-ID: <hwbva.136181$w7k.92438@news04.bloor.is.net.cable.rogers.com>   ? "Robert Deininger" <rdeininger@mindspring.com> wrote in message F news:rdeininger-1005031227380001@user-uinj56i.dialup.mindspring.com... > In article> <8p8va.134690$w7k.51439@news04.bloor.is.net.cable.rogers.com>,& > "John Smith" <a@nonymous.com> wrote: > = > >"Larry Kilgallen" <Kilgallen@SpamCop.net> wrote in message 0 > >news:38+98L7zEPke@eisner.encompasserve.org...5 > >> In article <3EBCDE91.80B1A1FB@aaa.com>, Jan-Erik 5 > >=?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> writes:  > >> > Keith Parris wrote:/ > >> >> Current roadmaps talk about 8.2 on VAX.  > >> > > >> > Where is that roadmap ? > >> >E > >> > I'v got information (from a HP dealer) that "next" VMS version , > >> > will be the last version for Alpha... > C > Well, HP has already states plans for VMS 7.3-2 and 8.2 on alpha.  The ? > dealer is either confused or a liar.  V7.3-2 will be the last 
 Alpha-only? > release, from what I have heard.  8.0 and 8.1 will be "early"  releases forD > IA64 (similar to 1.0 and 1.5 on alpha), and then starting with 8.2 Alpha F > and IPF VMS will be released together.  None of this is particularly new  > information.  ? So we're talking in what essence amounts to a not distant point C release/minor variant as being the last *Alpha-only* version.  That D Jan-Erik wrote it the way he did may be a verbatim statement of whatC the dealer told him, or a paraphrase, an incomplete statement, or a D difference in language barrier syntax. But it really does not wholly invalidate what was reported.      > < > >> Please report the name of that dealer to Sue Skonetski. > > A > >And thereby removing  a channel for selling VMS, probably in a B > >metropolitan area where nobody else is doing *anything* to sell VMS.B > >Sometimes it's better to deal with the devil you know (the VAR) than4 > >the one that's worse (HP's Marketing Department).  D Ok, so I may have jumped the gun. But given the noted absence of VMS& marketing visible to the public.......      3 > Another brave post from the anonymous John Smith.   9 You want to chat off-line, let me know. Others here have. F I have good and valid reasons, including but not limited to a previousB theft of my identity that took forever to clear up. You don't everA want to go through one of those episodes. Try getting a mortgage,  credit, crossing a border....     E > What do you suppose Sue would do, smite the dealer from the face of  the A > Earth?  She would certainly attempt to contact him and give him A > up-to-date, accurate information regarding VMS.  How would that 	 "remove a  > channel for selling VMS"?   B I don't have an issue with Sue...it's the higher up's who would be  more likely to smite the dealer.    C > Please explain how a "dealer" spreading false information through B > intermediaries is of any use to the VMS community.  Especially a	 dealer as ' > out-of-touch as this one seems to be.   F Again, so we're talking in what essence amounts to a not distant pointD release/minor variant as being the last *Alpha-only* version.  Maybe6 HP has to contact their dealers more about VMS anyway.   ------------------------------  % Date: Sat, 10 May 2003 20:59:40 +0200 9 From: Jan-Erik =?iso-8859-1?Q?S=F6derholm?= <aaa@aaa.com> 7 Subject: Re: Computerworld: HP continues to support VMS ' Message-ID: <3EBD4C1C.23329D3B@aaa.com>    John Smith wrote:  > A > So we're talking in what essence amounts to a not distant point ? > release/minor variant as being the last *Alpha-only* version.   ? Well, the dealar and I was discussing new service contracts for D a couple of new (well, used but "new" to the customer) AlphaServers.@ In that context, the message was that we'd just need the "licensA update service" one year or so, since after the next (major?) VMS E version for Alpha (within a year), there will not be any new versions F released for Alpha. So no need for the "licens update service" then...    = I know this guy very well, and, IMO, he has no reason to make 9 this up. Missinformed, yes, maybe (or even probably !)...   	 Jan-Erik.    ------------------------------  % Date: Sat, 10 May 2003 15:54:26 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>7 Subject: Re: Computerworld: HP continues to support VMS ( Message-ID: <3EBD58E8.7D3CFB2@istop.com>   John Smith wrote: 5 > theft of my identity that took forever to clear up.   J I guess you never got yor name back :-) I could just imagine you at a bar,N meeting a lovely lady who asks "what is your name" ? And you'd have to repond:M Sorry, I don't have one, it was stolen and the police hasn't recovered it yet  " :-) :-) :-) :-) :-)   J (I shoudln't be laughing because of recent attacks against me on usenet by0 anonymous poster may result in the same for me).  D > I don't have an issue with Sue...it's the higher up's who would be" > more likely to smite the dealer.  M Are you kidding ? The higher ups would send the dealer a big gift since he is L helping them implement their long term policy for VMS. As long as the dealerM pushes HP wintel boxes and HP-UX, he has nothing to worry about. But a dealer M that actually markets VMS on behalf of HP, he'd have to worry about his life.    ------------------------------  % Date: Sat, 10 May 2003 13:47:11 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> 2 Subject: Re: DEC TCP/IP Services 5.3 for Hobbyist?' Message-ID: <3EBD492F.ACC70FAA@fsi.net>    "Andreas W. Wylach" wrote: >  > Hello everybody, > E > I am just in the process to build up my private VMS environment und 8 > was asking myself, why the Hobbyist CD just got TCP/IP > Version 5.1 (OpenVMS VAX).N > Does anybody know, when the 5.3 Version is available for the Hobbyist peops?L > Or is there probably a upgrade download somewhere to get the newest TCP/IP
 > Version?  H hp remains rather set against downloadables. Guess LMF hacks have become too well known.   G I posted some time ago about a bunch of DCL that I cooked up to build a G hobbyists' LP distro. from the standard-issue SPL. Any interest? E-mail H me privately. How to demung the reply-to should be obvious. If you don'tD get a reply, try using dls.net instead of fsi.net. Somewhere betweenE forwarding and spam filtering, I think I may be losing some messages.    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------   Date: 11 May 03 06:09:42 +0200) From: p_sture@elias.decus.ch (Paul Sture) J Subject: Re: DECnet-Plus (DECnet PhaseV) (DECnet/OSI) migration necessary?) Message-ID: <hZW2+sck7vS6@elias.decus.ch>   h In article <d7791aa1.0305100429.7c139309@posting.google.com>, bob@instantwhip.com (Bob Ceculski) writes:\ > p_sture@elias.decus.ch (Paul Sture) wrote in message news:<D6FLjSctJlPC@elias.decus.ch>... >> >  P >> > If 5 doesn't give you anything of value in your environment, can you reallyH >> > justify spending time/money learning/installing/fiddling with  it ? >>  % >> The cost of prior version support?  > : > about $300 to $400 per year ... very inexpensive ... and6 > well worth the money instead of spending 5x the time > on phase V ...  = The only way I can believe that is if it's  a per-cpu figure.   = Or your sales droid has hidden the figures elsewhere perhaps.    Sorry, got lotsa cpus here.   $ That 5x figure is total garbage too.  A Whilst I am at it, your company's website won't let me in without I Javascript enabled - just a blank page, and when I go in with JS enabled, D 2 out of the 3 history links come up with "Page under construction".   Not exactly awe inspiring.     --  
 Paul StureD 'scuse me while I go off and study whatever I can find on those fine/ AS400 systems that Bob has difficulty with :-)     ------------------------------  % Date: Sat, 10 May 2003 13:26:33 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> * Subject: Re: Go to an AS/400 from VAX/VMS?' Message-ID: <3EBD4459.4AA3B26A@fsi.net>    Paul Sture wrote:  > X > In article <3EB953F8.638371CB@istop.com>, JF Mezei <jfmezei.spamnot@istop.com> writes: > > RC Bryan wrote: J > >> If I were in your position, I would struggle to find any way to avoidA > >> the AS400.  You can get a nice Alpha on EBAY that would be a J > >> substantial step up from your 3100 for under $1000 and get some up to > >> date software from HPQ  > > O > > But if you APPLICATION vendor has abandonned VMS, not even a marvel machine ' > > will make that software run on VMS.  > > P > > The OS is neat, but for business, it does very little. It is the application > > that drives the business.  > J > You have it in one JF. The business application is everything. Sometimes9 > our bias towards VMS can get in the way of seeing that.   F ...and hp/Q's bias towards non-VMS keeps them from seeing it, as well.   --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Sat, 10 May 2003 13:24:11 -0500 1 From: "David J. Dachtera" <djesys.nospam@fsi.net> E Subject: Re: How Alpha will save Itanium - must reading for Bill Todd ' Message-ID: <3EBD43CB.4FCC0E25@fsi.net>    Bob Ceculski wrote:  > f > "Bill Todd" <billtodd@metrocast.net> wrote in message news:<ovqcndxS1b2GNCGjXTWcog@metrocast.net>... > > > ' > > > 121000 tpmC.   At $4.97 per tpmC.  > > > C > > > So all you have left is "$ per metric" and of course cheering = > > > on IBM and Opteron - never acknowledging your circle is  > > > getting tighter. > > > @ > > > Oh... at $4.97 per tpmC for that 4 processor Madison, your< > > > "$ per metric" argument is looking a bit weak at that. > > P > > Not really, Rob:  if you're looking at small systems and are cost-conscious,> > > then the 4-processor Opteron at $2.76 per tpmC is for you. > > 
 > > - bill > < > except oopsteron does/will not run OpenVMS ... add that in= > to the equation (tco) and Itanium will blow oopsteron away!   H Unless, of course, HP/Q sees a good enough reason to do an AMD64 port...  1 ...then again, vision is not their strong suit...    --   David J. Dachtera  dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/    ------------------------------  % Date: Sat, 10 May 2003 14:25:04 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>E Subject: Re: How Alpha will save Itanium - must reading for Bill Todd ) Message-ID: <3EBD43FD.812CB949@istop.com>    Bob Ceculski wrote: < > except oopsteron does/will not run OpenVMS ... add that in= > to the equation (tco) and Itanium will blow oopsteron away!   J If Opteron is 8086 compatible, then I wil lbe able to run charron-vax. andL thus run VMS. And if charron make an alpha emulator that runs on opteron, it would be able to run Alpha-VMS.   M And wouldn't it be a hoot if that generated performance that was more or less   on par with IA64 based systems ?   ------------------------------    Date: 11 May 2003 00:32:48 -0500+ From: young_r@encompasserve.org (Rob Young) E Subject: Re: How Alpha will save Itanium - must reading for Bill Todd 3 Message-ID: <Jj+dR7teY0eX@eisner.encompasserve.org>   _ In article <ovqcndxS1b2GNCGjXTWcog@metrocast.net>, "Bill Todd" <billtodd@metrocast.net> writes:     : > "Rob Young" <young_r@encompasserve.org> wrote in message   >  >>9 >> Additionally, HP has the highest 4-CPU result on tpmC:  > L > Quite likely only because IBM doesn't bother submitting scores for systemsL > that small.  If their 32-processor system scaled perfectly linearly then aK > 4-processor version would yield a score of over 85,000 tpmC, so take your 	 > choice:     C http://www.tpc.org/tpcc/results/tpcc_result_detail.asp?id=103042401   6 	85000 tpmC is a lot less than 121000.  So even if IBM9 	were to submit a 4 CPU result (not sure - the IBM config @ 	describes 8-way MCM.  Not sure if the p690 goes down to 4-way).    > >  either close to linear scaling to large processor counts isN > possible in this benchmark, which means that SuperDome's scalability is justM > as lousy as a cursory glance makes it appear to be, or the new POWER4+s are ? > likely faster in a 4-processor configuration than Madison is.  >   9 	Not at all.  You would like to believe that.  Power4+ is ; 	obviously less powerful than Madison in 4 CPU configs (if  F 	a comparison could be made - not sure) and more expensive.  An 8-way 9 	MCM lists for $350000 in the p690 config at tpmC.   Each B 	additional Madison in HP's listing goes for $9900.  Now obviously9 	this is high-end versus low-end and that is a recipe for < 	disaster when comparing.  That said, I don't see IBM's p690; 	prices getting knocked back as much when shifting to 4-way 	 	configs.     >>F >> http://www.tpc.org/tpcc/results/tpcc_result_detail.asp?id=103042401 >>$ >> 121000 tpmC.   At $4.97 per tpmC. >>@ >> So all you have left is "$ per metric" and of course cheering: >> on IBM and Opteron - never acknowledging your circle is >> getting tighter.  >>= >> Oh... at $4.97 per tpmC for that 4 processor Madison, your 9 >> "$ per metric" argument is looking a bit weak at that.  > N > Not really, Rob:  if you're looking at small systems and are cost-conscious,< > then the 4-processor Opteron at $2.76 per tpmC is for you. >   < 	Well that's a nice whitebox number with decent performance.  > 	So you do still have "$ per metric" argument - Opteron versus 	Madison.  Reviewing:   , 	Madison is considerably more powerful than:  " 		Opteron in TPC-C, SAP SD, SpecFp> 		Power4 same but add SpecInt - for low CPU count comparisons.  : 	Who is to say a Madison box isn't lurking in the wings to> 	make 32-64 CPU results higher?  Hein basically hints at that:  _ http://groups.google.com/groups?selm=3E63CCD2.9A6E9124%40eps.zko.dec.com&oe=UTF-8&output=gplain   I "What the Itanium and it's follow ups needs is a good large SMP SYSTEM to G be put in and flourish and a good OS and DB to get everything out of it G that is potentially in there. Obviously HP is working on such solutions + but I can not say when/where/how-many-CPUs.   ? The NEC benchmark is just a stake in the ground. A first step."     8 	The inference of course is there is higher numbers than9 	what NEC demonstrated coming out of HP.  Wonder how soon , 	the IBM folks will have to eat these words:  ( http://www.theinquirer.net/?article=9416  M It took a further dig at HP and Intel. Adalio Sanchez, general manager of the F Eserver division, said: "We don't just assemble boxes with third party
 components".    6 	I'm sure that fired up the troops at HP.  Surely they: 	are pushing hard to hit 681000 tpmC with 32 Madisons ;-).; 	Looking at some examples, I'm seeing 75-80 percent scaling E 	factor across architectures.  Ideally, an HP 32 next-gen box should   	be able to ? 	get 720K to 760K tpmC *if* their 4-way result is any indicator 9 	(the CPU is certainly capable of it - let's see how they : 	did on their next gen box.  Bet it has higher bandwidth).  > 	Madison is dominating the 4-CPU space for performance.  Sure,C 	Opteron is in a nice niche there.  Value 4-CPU space.  But Opteron > 	will get its competition from 3.0 GHz Xeons coming as Madison> 	is the high performer in that space (running real benchmarks,C 	i.e. SAP, TPC - non-clustered, SpecFp - neck and neck with Opteron 
 	on SpecInt).   < 	Interesting to see IBM being squeezed terribly in the 4-CPU- 	space by Xeon and now Opteron and Madison.     @ 	Want to get a feel for Alphas fate hadn't it been knifed?  Look> 	at Sun and IBM over the next 2-3 years.  Seriously retreatingC 	to the high-end.  Madison and follow-ons going to meet them there.    				Rob    ------------------------------  % Date: Sat, 10 May 2003 20:02:29 +0200 , From: Albrecht Schlosser <ajs856@tiscali.de>3 Subject: Re: How to determine boot device from DCL? , Message-ID: <irej9b.m54.ln@news.hus-soft.de>   sms@antinode.org wrote:  > + > From: p_sture@elias.decus.ch (Paul Sture)  > X-Gateway-Source-Info: USENET  > F > > In article <03050711214800@antinode.org>, sms@antinode.org writes: > > > 5 > > > ALP $ write sys$output f$getenv( "bootdef_dev")  > > > SCSI 0 6 0 0 0 0 0 > > >  > > > Not exactly "dka0".  > >  > > Or even: > > / > > $ write sys$output f$getenv( "bootdef_dev")  > > RAID 1 8 0 0 0 0 0 > > ( > > which presents itself to VMS as DRA0 > = >    On my AlpSta 200 4/233 systems, f$getenv( "booted_dev"):  >  > DKA0:    SCSI 0 6 0 0 0 0 0  > DKA400:  SCSI 0 6 0 4 400 0 0   > DKB500:  SCSI 0 12 0 5 500 0 0 >  > Where: >  > >>>show device > dka0.0.0.6.0 > dka400.4.0.6.0 > dkb500.5.0.12.0  > [...]  > pka0.7.0.6.0 > pkb0.7.0.12.0  > ' >    So, I'd guess it's something like:  >  > SCSI 0 12 0 5 500 0 0  >      ?  | ? |   | ? ? C >         |   |   +------- Composite device number (100 * ID + LUN) " >         |   +----------- SCSI id& >         +--------------- PCI slot id > J > I suspect that one mystery value (the first one?) is the PCI bus number.I > (I have only bus "00".)  Another should be the SCSI LUN,  (Mine are all  > zero.) > H >    For one stable system, where the PK (or whatever) adapter names areI > known, and can be correlated with their bus slot (or whatever) numbers, H > the rest of the conversion is easy enough.  It's not obvious to me howE > the required info can be gathered without resort to a console "show  > device" report, however.  + As someone (not me) suggested recently  :-)   & $ diag /trans/inc=config/rev/out=x.tmp  
 DECevent V2.9  $ sea x.tmp "device dk" 9     device dka0.0.0.5.0 (RZ1CF-CF) found on pka0.0.0.5.0  8     device dka400.4.0.5.0 (RRD47) found on pka0.4.0.5.0 =     device dkc300.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc301.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc302.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc303.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc304.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc305.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc306.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc307.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  9     device dka0.0.0.5.0 (RZ1CF-CF) found on pka0.0.0.5.0       ... (everything twice) ...   Albrecht Schloer    ------------------------------  % Date: Sat, 10 May 2003 20:16:06 +0200 , From: Albrecht Schlosser <ajs856@tiscali.de>3 Subject: Re: How to determine boot device from DCL? , Message-ID: <2lfj9b.a84.ln@news.hus-soft.de>   Albrecht Schlosser wrote:  > - > As someone (not me) suggested recently  :-)  > ( > $ diag /trans/inc=config/rev/out=x.tmp >  > DECevent V2.9  > $ sea x.tmp "device dk" : >     device dka0.0.0.5.0 (RZ1CF-CF) found on pka0.0.0.5.09 >     device dka400.4.0.5.0 (RRD47) found on pka0.4.0.5.0 > >     device dkc300.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc301.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc302.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc303.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc304.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc305.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc306.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc307.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0: >     device dka0.0.0.5.0 (RZ1CF-CF) found on pka0.0.0.5.0  >     ... (everything twice) ...  B You might get more such lines, because diag may find more than one@ config entry (I had created a new error log file some time ago).   So this is better:  , $ boottime=f$extr(0,11,f$getsyi("boottime"))- $ diag /since='boottime'/inc=config/out=x.tmp   
 DECevent V2.9  $ sea x.tmp "device dk" 9     device dka0.0.0.5.0 (RZ1CF-CF) found on pka0.0.0.5.0  8     device dka400.4.0.5.0 (RRD47) found on pka0.4.0.5.0 =     device dkc300.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc301.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc302.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc303.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc304.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc305.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc306.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0  =     device dkc307.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0     Albrecht Schloer    ------------------------------  + Date: Sat, 10 May 2003 15:37:14 -0500 (CDT)  From: sms@antinode.org3 Subject: Re: How to determine boot device from DCL? ) Message-ID: <03051015371473@antinode.org>   , From: Albrecht Schlosser <ajs856@tiscali.de>  . > $ boottime=f$extr(0,11,f$getsyi("boottime"))/ > $ diag /since='boottime'/inc=config/out=x.tmp  >  > DECevent V2.9  > $ sea x.tmp "device dk" : >     device dka0.0.0.5.0 (RZ1CF-CF) found on pka0.0.0.5.09 >     device dka400.4.0.5.0 (RRD47) found on pka0.4.0.5.0 > >     device dkc300.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0> >     device dkc301.3.0.14.0 (IFT 3102) found on pkc0.3.0.14.0 > [...]   /    Interesting, but not universally successful:   3 ALP $ diagnose /translate /include = configurations   
 DECevent V3.3 + ALP $ write sys$output f$getsyi( "VERSION")  V7.2-1  H ------------------------------------------------------------------------  4    Steven M. Schweda               (+1) 651-699-98183    382 South Warwick Street        sms@antinode.org     Saint Paul  MN  55105-2547    ------------------------------  % Date: Sat, 10 May 2003 22:07:37 +0100 * From: "John Travell" <john@travell.uk.net>+ Subject: Re: INDEX.SYS size and performance 5 Message-ID: <b9jq61$k0v26$1@ID-120847.news.dfncis.de>   < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3EBC4C8D.E56078C2@fsi.net...   J > > > Might not work anyway, there is at least one block in the index file for J > > > each file header. The minimum amount of disk space consumed per fileE > > > name is one block, even for a file with zero blocks of content.  > > F > > That one block is, however, not allocated to the zero-length file,J > > rather to INDEXF.SYS. If INDEXF is pre-allocated using /HEADERS on theK > > INITIALIZE command, the net change in freespace is still zero until the G > > directory where the file is entered needs to be extended, and then, D > > again, the increased allocation belongs to a file other than the > > zero-length file.  > >   K Your comment about 'allocated' may not be referring to the way I understand  things, but...I Enable disk quotas, then account for every block allocated to your quota.   J 1. The allocated size of all of the files and directories inside and below your top level directory. 7 2. The allocated size of your top level directory file. K 3. All of the blocks in INDEXF.SYS containing headers for all of the above.   L So, if disk quotas are enabled, creating a zero block file will add at leastK 1 to your quota, and maybe more if the fact of creating the file means that I the directory file needs to be extended to accomodate the extra directory  entry.     -- John Travell  VMS crashdump expertise for hire john@travell.uk.net  http://www.travell.uk.net/       --- & Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.478 / Virus Database: 275 - Release Date: 06/05/2003    ------------------------------  % Date: Sat, 10 May 2003 21:36:51 +0100a* From: "John Travell" <john@travell.uk.net>+ Subject: Re: INDEX.SYS size and performancee5 Message-ID: <b9jnt5$jhb3s$1@ID-120847.news.dfncis.de>p  < "David J. Dachtera" <djesys.nospam@fsi.net> wrote in message! news:3EBC4C8D.E56078C2@fsi.net...  > John Travell wrote:e > >s: > > "Jim Brankin" <jbrankin@ntlworld.com> wrote in message; > > news:863f19d6.0305090017.133baf81@posting.google.com...g< > > > John Laird <john@laird-towers.org.uk> wrote in message8 > > news:<efklbv8hh3ett5i0tdj75l6gtms6tmsgre@4ax.com>... > > > .B > > > > ? > > > > Let's see you create more files than clusters, then :-)b > > > >e > > >C. > > > No problem. Make them zero length files. > >NL > > Might not work anyway, there is at least one block in the index file forH > > each file header. The minimum amount of disk space consumed per file name isp; > > one block, even for a file with zero blocks of content.g > D > That one block is, however, not allocated to the zero-length file,H > rather to INDEXF.SYS. If INDEXF is pre-allocated using /HEADERS on theI > INITIALIZE command, the net change in freespace is still zero until the-E > directory where the file is entered needs to be extended, and then, B > again, the increased allocation belongs to a file other than the > zero-length file.R >AG > I forget when it appeared, but CREATE/DIRECTORY now has a /ALLOCATION41 > qualifier; so, you can pre-extend that as well.@ >tI > So, it *IS* possible to use up the entire /MAXFILES allotment with zeroo > change in freespace. >n" But... Thats NOT the challenge...!? It was:  >>>Let's see you create more files than **_clusters_**o  D Even if the file is zero blocks long, it still has a header block in8 INDEXF.SYS. You would need to set /MAXFILES greater than# the number of clusters on the disk.s> In other words, the indexfile bitmap has to be larger than theA volume bitmap. I doubt that INIT allows you to do this, as all ofa> the native filesystems on VMS assume that you cannot have more! than 1 file in each disk cluster.m   -- John Travell  VMS crashdump expertise for hire john@travell.uk.netv http://www.travell.uk.net/           ---s& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.478 / Virus Database: 275 - Release Date: 06/05/2003e   ------------------------------  % Date: Sat, 10 May 2003 22:11:42 +0100 + From: John Laird <john@laird-towers.org.uk>0+ Subject: Re: INDEX.SYS size and performancei8 Message-ID: <eeqqbvs67k1l7r8pmond7lfref5dggunak@4ax.com>  H On Sat, 10 May 2003 21:36:51 +0100, "John Travell" <john@travell.uk.net> wrote:  # >But... Thats NOT the challenge...!m@ >It was:  >>>Let's see you create more files than **_clusters_** > E >Even if the file is zero blocks long, it still has a header block in.9 >INDEXF.SYS. You would need to set /MAXFILES greater thann$ >the number of clusters on the disk.? >In other words, the indexfile bitmap has to be larger than thetB >volume bitmap. I doubt that INIT allows you to do this, as all of? >the native filesystems on VMS assume that you cannot have moreb" >than 1 file in each disk cluster.  F All quite correct.  I did say the answer was a cheat, and have already indicated what it is ;-)  G And actually it wasn't a challenge to prove anything per se, as much as-  pointing out that the statement:  M >The size of BITMAP.SYS is defined by the number of clusters on the disk, and A >has nothing to do with the number of files the disk may contain.7  K is not totally accurate.  There are other limiting factors, but clearly yous< cannot have more files than clusters, so they *are* related.     	John1   ------------------------------  % Date: Sat, 10 May 2003 17:18:57 -0500y1 From: "David J. Dachtera" <djesys.nospam@fsi.net>r+ Subject: Re: INDEX.SYS size and performancee' Message-ID: <3EBD7AD1.D01E4BEA@fsi.net>d   John Laird wrote:  > J > On Sat, 10 May 2003 21:36:51 +0100, "John Travell" <john@travell.uk.net> > wrote: > % > >But... Thats NOT the challenge...!tB > >It was:  >>>Let's see you create more files than **_clusters_** > > G > >Even if the file is zero blocks long, it still has a header block in ; > >INDEXF.SYS. You would need to set /MAXFILES greater thanh& > >the number of clusters on the disk.A > >In other words, the indexfile bitmap has to be larger than theiD > >volume bitmap. I doubt that INIT allows you to do this, as all ofA > >the native filesystems on VMS assume that you cannot have moret$ > >than 1 file in each disk cluster. > H > All quite correct.  I did say the answer was a cheat, and have already > indicated what it is ;-) > I > And actually it wasn't a challenge to prove anything per se, as much as?" > pointing out that the statement: > O > >The size of BITMAP.SYS is defined by the number of clusters on the disk, andtC > >has nothing to do with the number of files the disk may contain.  > M > is not totally accurate.  There are other limiting factors, but clearly youc> > cannot have more files than clusters, so they *are* related.  @ Well, I showed how to create more files on a disk than there areG clusters. Remember, a file header is one *BLOCK* in a cluster, *NOT* anbF entire cluster! Therefore, the number of zero-length files that can beF created is roughly the block size of INDEXF.SYS, minus the size of theE INDEXF bitmap and the other data the file contains, times the clusterv size.   E Of course, if you do SET FILE/ENTER repeatedly for one file, that one G file can exist in multiple paths, and so can appear multiple times in a 9 recursive directory listing. Anyone who has ever done DIRl( SYS$SYSDEVICE:[000000...] has seen that.  E So, unless you're not explaining yourself well, yes, you *CAN* create 1 more files than the number of available clusters.m   --   David J. Dachtera0 dba DJE Systems  http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/c   ------------------------------  % Date: Sat, 10 May 2003 20:15:47 -0400t* From: JF Mezei <jfmezei.spamnot@istop.com>+ Subject: Re: INDEX.SYS size and performance ) Message-ID: <3EBD9623.CB7B81D3@istop.com>i  ; > >INDEXF.SYS. You would need to set /MAXFILES greater than & > >the number of clusters on the disk.  H If empty one file takes one block in indexf.sys, then you'd have to have8 maxfiles greater than the number of blocks on the drive.  K If you have a cluster factor of 3, you can put 3 (empty) files per cluster.s   ------------------------------  % Date: Sat, 10 May 2003 14:28:10 -0400 * From: JF Mezei <jfmezei.spamnot@istop.com>Y Subject: Re: Not fixed yet but still trying: Problems changing from serial port to DecSer ) Message-ID: <3EBD44B8.21C0CC0E@istop.com>s   Another thing to check:   < process A on the vax/alpha has connected to the serial port.2 process B on the same node does a SHOW PROC A/CONT  L While you type stuff on A, you look at the screen of SHOW PROC to see if the IO/CPU changes.bK Then, while you type stuff from the terminal attached on the decserver, youIG also monitor the output of show proce/cont to see if data sent from tehc/ decserver causes the A process to do something.n   ------------------------------  % Date: Sat, 10 May 2003 21:05:13 +0200u6 From: Adolf Sonderegger <adolf.sonderegger@bluewin.ch>Y Subject: Re: Not fixed yet but still trying: Problems changing from serial port to DecSerf* Message-ID: <3EBD4D69.15C39B0E@bluewin.ch>  
 Roscoe wrote:.  E > Thanks for all the responses thus far.  I'm not sure if this is the H > proper way to post a follow-up so I appologize if there's another way. >bD > I've tried several things as suggested from your guidance however,; > it's still not working.  Here's what's happened thus far.  >iA > 1) I tried changing the terminal characteristics to Hostsync ascG > suggested(instead of No Hostsync).  However, this did not make it anym	 > better.c > H > 2) To answer Mr. Sonderegger's question - yes, this is how my LTA port > was setup: >f > $ RUN SYS$SYSTEM:LATCP$ >   CREATE PORT LTA2312 /APPLICATIONJ >   SET PORT LTA2312 /NODE=3Ddecserver/PORT=3Dprot /APPLICATION /NOQUEUED=   >cE > 3) To answer Mr. Brown's question - I do not have a service definednE > for the port.  Just an LTA (LTA2312:).  And yes, I can SET HOST/DTEe0 > LTA2312: and bidirectional comms are possible. >hD > 4) On a positive note I was able to use the following from DCL andC > receive data from the VT340 (and BTW that's in REMOTE mode on thep > terminal server).n >g > $ ALLOCATE LTA2312: TERM5 > $ DIR /OUT=3DTERM (directory listing goes to VT340) A > $ DUMP TERM (keyboard entry is returned after carriage returns)  >ID > 5) The most prevalent information given on your replies was that aB > "write" must be issued to the port first in order to establish aG > connection.  So, I headed off in that direction.  I also found an oldaC > post entitled "Need help reading data from a decserver port".  It = > probably would have helped if I would have told you that mywG > application was written in Fortran.  It was great news when I saw Mr.eB > Moulton's Fortran code describing this exact problem and his fixD > (shown directly below).  However, I tried it and I can't get it toF > work either.  This is of course placed right before my SYS$QIOW call  > that reads data from the port. >a; >     FUNCTION_CODE =3D JIOR(IO$_TTY_PORT, IO$M_LT_CONNECT). >>$ >     RETURN=3DSYS$QIOW(,%VAL(CHAN), > $   %VAL(FUNCTION_CODE), > $   IOSB,r. > $   ,,              ! AST address, parameter/ > $   ,,,,,)  ! This function has no parameters  >t, >  A read QIO to the channel will then work. >eC > So, I'm one step closer in that I can read data from the VT340 ins@ > order to test the serial port connection.  However, my Fortran) > application does not want to cooperate.l > A > No doubt I'm still doing something wrong.  I'm just hoping that 2 > someone much brighter than me can figure it out. >i > Thanks again,  >> > Roscoe at ISG Steel Company( >m5 > PS - Here's the terminal settings in case it helps:t >lB > Terminal: _LTA2312:   Device_Type: Unknown       Owner: No Owner >MD >    Input:    9600     LFfill:  0      Width:  80      Parity: None2 >    Output:   9600     CRfill:  0      Page:   24 >o > Terminal Characteristics: G >    Interactive        Echo               Type_ahead         No EscapehD >    Hostsync           TTsync             Lowercase          No Tab@ >    Wrap               Scope              No Remote          No
 > EightbitE >    Broadcast          No Readsync        No Form            FulldupiD >    No Modem           No Local_echo      No Autobaud        HangupG >    No Brdcstmbx       No DMA             No Altypeahd       Set_speedo@ >    No Commsync        Line Editing       Overstrike editing No
 > FallbackH >    No Dialup          No Secure server   No Disconnect      No PasthruH >    No Syspassword     No SIXEL Graphics  No Soft Characters No Printer > Port@ >    Numeric Keypad     No ANSI_CRT        No Regis           No > Block_mode@ >    No Advanced_video  No Edit_mode       No DEC_CRT         No
 > DEC_CRT2@ >    No DEC_CRT3        No DEC_CRT4        No DEC_CRT5        No > Ansi_Color >    VMS Style Input   Helloo  6 There is the Book OpenVMSI/O User=92s Reference Manual on URL:kJ http://h71000.www7.hp.com/doc/731FINAL/DOCUMENTATION/PDF/OVMS_731_IO_USER= =2EPDF =2Ei  @ It has a LT Example on 5.6 Terminal Driver Programming Examples.   regards  Adolf Sondereggeri   ------------------------------  % Date: Sat, 10 May 2003 14:23:15 -0400v* From: JF Mezei <jfmezei.spamnot@istop.com>Y Subject: Re: OT - PVS (was Re: DECnet-Plus (DECnet PhaseV) (DECnet/OSI) migration necessa ) Message-ID: <3EBD4391.997F1B65@istop.com>h   "Bradford J. Hamilton" wrote: P > Not what COMPAQ (at the time) quoted us for PVS - their figure was $11K/month.E > Made it very easy to justify the migration to our customer.     :-)n  ' OK, I have to take issue with this now.s  M This seems to be some remnant of the big bad Digital. We'll raise the support.> costs of Decnet 4 in order to force customers to use decnet 5.  ? To Digital/HP: which actually costs less to maintain/support ? s  - Which is first/easiest to be ported to IA64 ?e  R Seems to me that it is 5 that shoudl cost more to support and 4 should be cheaper.  H You need the extra functionality of 5, then pay extra for it. You want aH stable, "mature" product, then you pay less because you can't expect new features etc on it.   H This reminds me of Digital still selling RD54 drives (154 meg) for $9000T (canadian) when you could have drives in the 1 gig range for a couple hundred bucks.   ------------------------------  % Date: Sat, 10 May 2003 13:49:17 -0500g1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: Splitting cluster' Message-ID: <3EBD49AD.A9ED3977@fsi.net>h   Paddy O'Brien wrote: > A > I'm afraid that I currently have no access to any fine manuals.  > G > I have the remains of a heterogeneous cluster: one VAX (was four) ands > one Alpha. > I > In order to get them to boot separately and not as a cluster, do I onlyoN > need to change VAXCLUSTER to 0 in each of the MODPARAMS.DAT and run Autogen? > G > Do I need to worry about other system parameters, e.g. quorum values?   E I should think that with clustering turned off, things like votes ands quorum would be meaningless.  = Caveat re: shared interconnects, as another poster mentioned.1   -- 0 David J. Dachtera  dba DJE Systemsf http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/5   ------------------------------   End of INFO-VAX 2003.259 ************************clear up.   J I guess you never got yor name back :-) I could just imagine you at a bar,N meeting a lovely lady who asks "what is your name" ? And you'd have to repond:M Sorry, I don't have one, it was stolen and the police hasn't recovered it yet  " :-) :-) :-) :-) :-)   J (I shoudln't be laughing because of recent attacks against me on usenet by0 anonymous poster may result in the same for me).  D > I don't have an issue wi5ᴏ < uyqp"W)WN]?9p 7(_g9!В^aAr>+MMe5b$$ZcEڍgi+VPkZ*DBp|5ّh,[XE3guaqA]*|^uN\;ElGHL|}&@k:xP5L<,+>m$KcU1[Sr:vLDHX=g哴e03d[)n#!y~9m_~>U4U_ĭBn
Ќ0f=wfYP3#Ff?- ]%f(xФK=Xt A?+ĭD$\*t,H8 A?Cfi&V=]jQd*=TMjn^CcBNbɇ'A2C=yl.2ױAJ=~jgcKTါ&i.1JOqČD74O_b.)Ufd (	D>:nWY{#K=1+!˄[~u-x͸*㫡P^.>&	ǉeTװeLռ֗z^[xM(xIuU9Qլܹimo8VJ F^z6#La~B,Qc<&TCކB|P?[E;	~K~R1Gf%˙락\[2o?sҿqGYm:<9N0˓NqUFgL"hF>2d<b0@,PK    :H,ԖE  A     gawk-3_1_1/intl/ref-add.sinS]k@|ppPJR;1M!N:*ߙ]S߻:9$y^AÕb+bMp,ӆrHaG
YQtL`t8|8bf!trV=UF؀?&8PF
֎ `sYsq2%ئfĝL0ܐ"#*uZɌ3R Ė{Sߒ0IXyΉpMz۠N=PT½ߒ_-)5eU!%Ԗ}a XDQHVcFRsvrͶLRPhr9ny1fA'˻ Bxbz\xx99!+<k*G)v>g$wܥ@G+^q!(zgy=|xpqP\\{|%z`x>F//8h%>rEB]Zi
2ݷҿ2^m?YoO]]ldI^Ms.1``F8Kᱸ
ɏfzX">\wutPK    :H,#7       gawk-3_1_1/intl/ref-del.sinS]k@|K?Ji]ec'-$YZYG;swr*J{W'@K'vggvOWHH=!7 RZPNTJiCkGrYRt6r_8\.q=0Du!,t2V=,URF8ȞƨuT(1!#w#H68L5<7 jXm7rgqK(VRNIY`My]Kߒ0I78|uLB>6YJpoIU/aY,K*{-qܭ7	Ad;lWhXp,%C#y<a&.ɶѱX&yc d9My^Q>pf,eeB*JI@gK^y!s(zx68h !!,EJB\577