1 INFO-VAX	Fri, 25 Oct 2002	Volume 2002 : Issue 589       Contents: RE: (OT) - Parties& Re: ACCVIO with SHOW DEVICE/SIZE=BYTES# Re: Alphaserver 4100 Galaxy Problem " Re: Another strike against Itanium" Re: Another strike against Itanium" Re: Another strike against Itanium" Re: Another strike against Itanium Re: AS1000 4/266 CPU upgrade ?  Re: Availability Manager Problem  Re: Availability Manager Problem backup saveset Re: backup saveset Re: backup saveset Re: backup saveset Re: backup saveset0 Re: backup/image of shadowed volume misses files0 Re: backup/image of shadowed volume misses files0 Re: backup/image of shadowed volume misses files0 Re: backup/image of shadowed volume misses files0 Re: backup/image of shadowed volume misses files0 Re: backup/image of shadowed volume misses files0 Re: batch process won't die after BACKUP aborted Re: Carly on innovation  Re: Carly on innovation  Re: Carly on innovation  Re: Carly on innovation  Re: Carly on innovation  Re: Denial of Service attacks  Denial of Service attacks  Re: Denial of Service attacks + Re: Dual CPU cards vs. dual core processors ( Re: Dual ISA ethernet adapter on VMS 7.2 Re: File Attributes  Re: File Attributes  Re: File Attributes  Re: File Attributes  Re: File Attributes  Re: File Attributes $ Fiorina Gets Feisty at Gartner Event( Re: Fiorina Gets Feisty at Gartner Event Re: Hobbyist LicenseP How can I determine the device of the system disk for another node in a cluster?7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 RE: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! 7 Re: HP tells VMS users "Relax, you are sitting pretty"! A Re: HP-ETS 2002 Presentations are on-line - Pay-per-view ONLY :-(  Incoming !!!! DS10's- It'll be interesting to see what happens here  Re: Monitor System Question  Re: Monitor System Question 4 Re: Need Help really Quick Please: 10/23/02 Noon CST Non-Paged Pool Expansion Re: Non-Paged Pool Expansion Re: OpenVMS 7.3-1 and CSWING Re: OpenVMS 7.3-1 and CSWING Re: OpenVMS 7.3-1 and CSWING Re: OpenVMS 7.3-1 and CSWING Re: OpenVMS 7.3-1 and CSWING Re: page or pagelet  Re: page or pagelet  RE: page or pagelet  Re: page or pagelet  Re: page or pagelet  Re: page or pagelet  Re: Scheduling Question $ Re: TCPIP 5.3 and SMTP authorization$ Re: TCPIP$FTP_SERVER startup problem9 Re: The VMSNET group of forums (was: Re: Acrobat Reader?) 9 Re: The VMSNET group of forums (was: Re: Acrobat Reader?)  Re: TPU port to Linux + Re: Why does this file-spec work only once? + Re: Why does this file-spec work only once? + Re: Why does this file-spec work only once?  XFC  Re: XFC B Re: [OT] - Parties (Re: HP tells VMS users "Relax, you are sittingB Re: [OT] - Parties (Re: HP tells VMS users "Relax, you are sittingB Re: [OT] - Parties (Re: HP tells VMS users "Relax, you are sitting' Re: [OT] - we're back on old cars again   F ----------------------------------------------------------------------    Date: 24 Oct 2002 13:25:52 -0500; From: kaplow_r@eisner.encompasserve.org.TRABoD (Bob Kaplow)  Subject: RE: (OT) - Parties 3 Message-ID: <d6Z9nei61iNx@eisner.encompasserve.org>   o In article <BA52530E3149734A9BAABDBBFA808E4903027B8D@rlghncst964.usps.gov>, VAXVMS <bounce@notmail.com> writes:  > 8 >     And don't forget poison and the infamous petar(d). > @ >     (The use of which in the phrase 'hoist by his own petard' A >      is a pun which involves an archaic term for flatulence...)  > , > Did they have guns in Shakespeare's plays? >  > Mostly daggers as I recall.    IIRC the correct quote is:  A "The first thing we do, let's kill all the lawyers" from Henry VI   K The word is kill, not shoot. Slicing them off a little at a time may be the  best way :-)  1 	26-October, 2001: A day that will live in infamy 4 	Support Freedom: http://www.indefenseoffreedom.org/  J         Guard with jealous attention the public liberty. Suspect every oneJ         who approaches that jewel. Unfortunately, nothing will preserve itE         but downright force. Whenever you give up that force, you are C         inevitably ruined. -- Patrick Henry, speech of June 5 1788     ------------------------------    Date: 25 Oct 2002 03:45:02 +0800, From: Paul Repacholi <prep@prep.synonet.com>/ Subject: Re: ACCVIO with SHOW DEVICE/SIZE=BYTES - Message-ID: <874rbb39kx.fsf@prep.synonet.com>   $ Guy Peleg <guy.peleg@hp.com> writes:  E > I would like to know if you find the SHOW DEVICE/SIZE=BYTES feature E > useful.  For the next version of the O/S we are working on a global = > process cell which will hold the process units. Most of the @ > utilities displaying information in blocks will be modified to > support this. so :   4 > $ SET PROC/UNITS=BYTES will display info in bytes.D > $SET PROC/UNITES=BLOCKS - the default will display info in blocks.   C > The plan is to display information in bytes and not to accept any  > input in bytes.    3 > Do you find this useful? Are we wasting our time?   D How many utilities will be in trouble with the increased size of theE fields? How many digits do you need for a maximum size unit now days?   A Personally, blocks is a MUCH nicer units than bytes. (Or bits...)    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 25 Oct 2002 03:57:28 +0800, From: Paul Repacholi <prep@prep.synonet.com>, Subject: Re: Alphaserver 4100 Galaxy Problem- Message-ID: <87vg3r1ufr.fsf@prep.synonet.com>   4 "Zane H. Healy" <healyzh@shell1.aracnet.com> writes:   > dittman@dittman.net wrote:  B > > I have an Alphaserver 4100 with two 5/533 CPUs and 4GB of RAM.E > > I've followed the Galaxy configuration documentation and have two 8 > > instances.  My lp* variables are defined as follows:   F > Does the 4100 support Galaxy?  I'd thought you needed something like
 > a GS160.  D If you get the Galaxy VCD, you can find a second Galaxy presentationA on it, as well as the one that is in the menu. In that one, Clare D plays with such modern miracles as a DECserver 200, Vaxstation 3100, and a 2100 Alpha.   D A 4100 is a very nice Galaxy box, `enought' for most testing and the like, and no extra bits needed.    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Thu, 24 Oct 2002 15:19:34 -0500 " From: xganon <remailer@xganon.com>+ Subject: Re: Another strike against Itanium 9 Message-ID: <730cac3de8f395a6d422960d2c6a3ce3@xganon.com>   E "Either you're attempting to be humorous, or you're confused.  While   ItanicH will have difficulty surpassing EV7's SPECint performance (and won't be  ableB to touch EV7 in other areas, save for FP) until Alpha development 
 stops, theG new Hammer numbers help prove what I've suspected would be the case all B along:  Hammer will out-perform EV7 pretty much across the board,"    Itanium has and will have higherF transactions with twisty sql and oracle than hammer.  AMD can't afford; the research in branch predictors.  Intel stole the world's C best branch predictor and stuffed it in pentium pro, the subsequent : patent victory by Digital bore that out.  Intel will guardB that BP IP and subsequent home grown BP IP.  Hammer will be better at specint.  Big deal.  D Secondly, EV7 will do better than them both for some time at totallyC random I/O.  EV7 IO port at 3 GB/s.  Granted, you don't hang IO off C every EV7.  But certain folks will be very interested in hanging a  = great many cages off 4/8 CPU "QBBs".  Use your imagination as > to who will be interested in that.  Plowing through very largeD databases EV7 has much higher bandwidth and stream numbers highlight that.    Hammer is a poor EV7 knock-off.    ------------------------------    Date: 25 Oct 2002 04:13:47 +0800, From: Paul Repacholi <prep@prep.synonet.com>+ Subject: Re: Another strike against Itanium - Message-ID: <87r8ef1tok.fsf@prep.synonet.com>   / JF Mezei <jfmezei.spamnot@videotron.ca> writes:   F > At this point in time, one possible exit strategy for Intel would beE > to just plan simple process shrinks for IA64 without any more major D > improvements beyond what has already been committed. It might be a0 > cheao way to bow out without admitting defeat.  ? Other that adding bigger cache, what is the difference from wht  they are planing to do now?    --  < Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------  % Date: Thu, 24 Oct 2002 19:00:56 -0400 * From: "Bill Todd" <billtodd@metrocast.net>+ Subject: Re: Another strike against Itanium 2 Message-ID: <x9ydnT_wbIAF5iWgXTWcpQ@metrocast.net>  9 "Paul Repacholi" <prep@prep.synonet.com> wrote in message ' news:87r8ef1tok.fsf@prep.synonet.com... 1 > JF Mezei <jfmezei.spamnot@videotron.ca> writes:  > H > > At this point in time, one possible exit strategy for Intel would beG > > to just plan simple process shrinks for IA64 without any more major F > > improvements beyond what has already been committed. It might be a2 > > cheao way to bow out without admitting defeat. > A > Other that adding bigger cache, what is the difference from wht  > they are planing to do now?   I According to their public statements, nothing.  But those only go through F Montecito in 2004:  a year or so ago they mentioned a later 'bump' forL Montecito (presumably in 2005), which *could* involve EV7-style on-chip glueI and/or multiple processor cores on the chip (as just mentioned elsewhere: H Montecito is planned for a 90 nm. process; reportedly all the chip spaceJ that frees up will be used for additional cache, but an additional core orJ on-chip glue would be a small enough addition to the total that they mightI squeeze it in later), and no one is talking at all about what might be in F Chivano (possibly 2005, but sounds much more likely to be 2006 or evenH later - in any event, the first Itanic core where the transplanted Alpha- team will likely have significant influence).    - bill   ------------------------------  % Date: Thu, 24 Oct 2002 19:25:32 -0400 * From: "Bill Todd" <billtodd@metrocast.net>+ Subject: Re: Another strike against Itanium 2 Message-ID: <h9GdnXaw_u7BHCWgXTWcpw@metrocast.net>  / "xganon" <remailer@xganon.com> wrote in message 3 news:730cac3de8f395a6d422960d2c6a3ce3@xganon.com... F > "Either you're attempting to be humorous, or you're confused.  While > ItanicI > will have difficulty surpassing EV7's SPECint performance (and won't be  > ableC > to touch EV7 in other areas, save for FP) until Alpha development  > stops, theI > new Hammer numbers help prove what I've suspected would be the case all D > along:  Hammer will out-perform EV7 pretty much across the board," > " > Itanium has and will have higherH > transactions with twisty sql and oracle than hammer.  AMD can't afford$ > the research in branch predictors.  J What makes you think it would have to?  With about 1/3 of Itanic's latencyH to main memory (48 - 57 ns. to a local open page, vs. Itanic's 156 ns.),F mispredicted branches aren't nearly as critical to Hammer (even if itsK improved branch prediction is still not up to Itanic's, which has yet to be I ascertained).  Even 2-hop accesses to remote memory are faster than a zx1 K Itanic access - and since HP seems to be claiming that the zx1 chipset will L outperform other Itanic chipsets, likely including its own Pinnacles chipsetG supporting larger configurations, the average of 1 hop in a 4-processor ( Hammer configuration will be faster yet.     Intel stole the world's E > best branch predictor and stuffed it in pentium pro, the subsequent < > patent victory by Digital bore that out.  Intel will guardD > that BP IP and subsequent home grown BP IP.  Hammer will be better > at specint.  Big deal. > F > Secondly, EV7 will do better than them both for some time at totallyE > random I/O.  EV7 IO port at 3 GB/s.  Granted, you don't hang IO off D > every EV7.  But certain folks will be very interested in hanging a? > great many cages off 4/8 CPU "QBBs".  Use your imagination as @ > to who will be interested in that.  Plowing through very largeF > databases EV7 has much higher bandwidth and stream numbers highlight > that.   I Until such time as Hammer can scale (and scale very effectively) beyond 8 H processors, large EV7 systems will indeed be unbeatable (which I alreadyD said earlier).  For systems in Hammer's range, however, Hammer's I/OD bandwidth, while lower than EV7's, is likely high enough to keep theI processor saturated for any database activity - and the processor will be H faster than EV7.  Furthermore, Hammer has significantly better local and remote memory latency than EV7.    > ! > Hammer is a poor EV7 knock-off.   I No:  Hammer is a *good* EV7 knock-off.  It does make different trade-offs K (1, 2, or 3 HT links instead of EV7's 4-way links), but for anything save a A very large system they're *good* trade-offs, and make it far more K cost-effective as well.  Alpha made the right choices in EV7 for a high-end L product (which by the time EV7 design firmed up it clearly was); Hammer madeG the right choices for a wide-range commodity product - and by virtue of A having a state-of-the-art core (albeit one burdened with the IA32 J instruction set) rather than one designed in the mid-'90s has extended its: reach toward high-end (though size-limited) areas as well.   - bill   ------------------------------  % Date: Thu, 24 Oct 2002 23:03:13 +0200 " From: "t.t" <t.t@poczta.gazeta.pl>' Subject: Re: AS1000 4/266 CPU upgrade ? ) Message-ID: <ap9n54$am8$1@news.gazeta.pl>   J Your 4/266 board is probably 54-23297-04. You can upgrade it to 5/333 with 54-24719-02.2 Here is the list of board you can use with AS1000:   EV4/200  - 54-23297-01 EV4/233  - 54-23297-03 EV4/266  - 54-23297-04 EV5/300  - 54-24719-02 EV5/333  - 54-24799-02 EV5/400   -54-24799-01   Regards - Tomasz    8 "Roland Barmettler" <rob@bbp.remove.ch> wrote in message1 news:20021024151415.1c626eb0.rob@bbp.remove.ch...  > Hello  > F > I've found a 300 MHz EV5 CPU board in a cupboard which was installed6 > in an AlphaServer 1000A before. Its part number is .D > Would any of you have an idea if its possible to upgrade an AS1000> > 4/266 with this board ? I've noticed that there's a firmware/ > upgrade/change for upgrading an AS1000 4/xxx.  >  > Thanks for any help. >  > Greetings, Roland  > H > --------------- bbp - Biveroni Batschelet Partners AG ----------------< >              Bahnhofstrasse 28, CH-5401 Baden, SwitzerlandH > ----------------------------------------------------------------------       --  A Serwis Usenet w portalu Gazeta.pl -> http://www.gazeta.pl/usenet/    ------------------------------  % Date: Thu, 24 Oct 2002 20:18:55 +0200 1 From: "Rene Verhaeghe" <rene.verhaeghe@chello.be> ) Subject: Re: Availability Manager Problem 1 Message-ID: <ATWt9.1045$Yx2.63963@news.chello.be>   K AM on VMS will always try to send its messages to the first defined network E card. If your first netwok card is in a different Lan than your other # systems, that could be the problem.   B You could force the card to be used via a logical name in the file AMDS$LOGICALS.COM   . <rob.buxton@wcc.spam.govt.nz> wrote in message news:3db6f111.518475017@news... 	 > Hi All,  > G > I've had AM running for a while quite successfully. AM analyser is on " > an NT Box. (AM 2.2-1 throughout) > 2 > Last night we went through some Network changes. >  > Background configs. B > Development Cluster, Ethernet only, mixed architecture, VMS 7.2, > 7.2-2, 7.3 & 7.3-1. F > Production Cluster, Ethernet and FDDI, mixed architecture, VMS 7.2 &
 > 7.2-2 only.  > @ > Network changes involved separating the FDDI off to a separateF > network, inaccessible from the rest of the network. It got different > IP Addresses as a result. C > I don't see how IP is an issue as it seems connectivity is at the  > Ethernet level.  > + > Now AM only sees the Development Cluster.  > B > The main difference is that the Prod Cluster has FDDI as well as > Ethernet.  >  > This all worked okay before. > ) > Any reasons why this might be an issue?    ------------------------------  # Date: Fri, 25 Oct 2002 01:00:43 GMT ! From: rob.buxton@wcc.spam.govt.nz ) Subject: Re: Availability Manager Problem % Message-ID: <3db89770.626601705@news>   D On Wed, 23 Oct 2002 19:06:04 GMT, rob.buxton@wcc.spam.govt.nz wrote:   >Hi All, > F >I've had AM running for a while quite successfully. AM analyser is on! >an NT Box. (AM 2.2-1 throughout)  > 1 >Last night we went through some Network changes.  >  >Background configs.A >Development Cluster, Ethernet only, mixed architecture, VMS 7.2,  >7.2-2, 7.3 & 7.3-1.E >Production Cluster, Ethernet and FDDI, mixed architecture, VMS 7.2 &  >7.2-2 only. > ? >Network changes involved separating the FDDI off to a separate E >network, inaccessible from the rest of the network. It got different  >IP Addresses as a result.B >I don't see how IP is an issue as it seems connectivity is at the >Ethernet level. > * >Now AM only sees the Development Cluster. > A >The main difference is that the Prod Cluster has FDDI as well as 
 >Ethernet. >  >This all worked okay before.  > ( >Any reasons why this might be an issue?    # Thanks for the responses - spot on. D I'd looked at the config files and things were getting a bit late...' should've checked the logicals file :-(    Now works as expected    ------------------------------  % Date: Thu, 24 Oct 2002 14:31:07 -0500 - From: Charles Aaron <caaron@ceris.purdue.edu>  Subject: backup saveset 0 Message-ID: <3DB84A7B.F5E0AC42@ceris.purdue.edu>  E I would like to copy a save_set (sy25may02.bck) off of a 40/80 gb dlt F tape to disk and then copy that same save_set to another 40/80, hoping? the save_set size and save_set file attributes remain the same.   0 Can someone email me the commands to use please?   Thanks,  chuck    ------------------------------  % Date: Thu, 24 Oct 2002 15:14:13 -0500 - From: Charles Aaron <caaron@ceris.purdue.edu>  Subject: Re: backup saveset 0 Message-ID: <3DB85495.8324F44F@ceris.purdue.edu>   Antony,   H Thank you for your suggestion. I tried that but the size of the original. save set is approx 303k blocks and when I copy9 it to disk it goes to over 1.8 mill. It does not stay the 
 same size.   Chuck    ------------------------------  % Date: Fri, 25 Oct 2002 06:11:12 +1000 @ From: "Antony Wardle" <antony.wardle@nospammmmm.optusnet.com.au> Subject: Re: backup saveset < Message-ID: <3db85424$0$18869$afc38c87@news.optusnet.com.au>   wouldn't   mount/over=id tape_drive  5 copy/log tape_drive:[]sy25may02.bck disk_drive:[temp]    dismount tape_drive    init tape_drive new_tape   mount/over=id tape_drive  3 copy/log disk_drive:[temp]sy25may02.bck tape_drive:      work?   = Just pretent its a disk-drive. Can be a bit on the slow side.      kiwi    : "Charles Aaron" <caaron@ceris.purdue.edu> wrote in message* news:3DB84A7B.F5E0AC42@ceris.purdue.edu...E I would like to copy a save_set (sy25may02.bck) off of a 40/80 gb dlt F tape to disk and then copy that same save_set to another 40/80, hoping? the save_set size and save_set file attributes remain the same.   0 Can someone email me the commands to use please?   Thanks,  chuck    ------------------------------  % Date: Fri, 25 Oct 2002 06:36:31 +1000 @ From: "Antony Wardle" <antony.wardle@nospammmmm.optusnet.com.au> Subject: Re: backup saveset < Message-ID: <3db85a14$0$18870$afc38c87@news.optusnet.com.au>  1 Thats because the size of the file on the tape is  cacluated differently.   I can't remember why.   3 Or are you saying that when the saveset was created 7 the size of the files on the disk was only 303k blocks.   3 I would expect that if you backup a 1 gig disk then 7 do a directory of the tape, the size of the file on the ? tape in blocks would be significanly less than 2million blocks. : (about 1 GB). A backup/list of the saveset should show you6 if the saveset is ok, and how big the files in it are.  ? I guess you can always just restore the save set to a disk with 7 the backup command, then put it back to a new tape with  backup.      cheers   ant     : "Charles Aaron" <caaron@ceris.purdue.edu> wrote in message* news:3DB85495.8324F44F@ceris.purdue.edu... Antony,   H Thank you for your suggestion. I tried that but the size of the original. save set is approx 303k blocks and when I copy9 it to disk it goes to over 1.8 mill. It does not stay the 
 same size.   Chuck    ------------------------------   Date: 24 Oct 2002 22:50 CDT ' From: carl@gerg.tamu.edu (Carl Perkins)  Subject: Re: backup saveset - Message-ID: <24OCT200222502909@gerg.tamu.edu>   1 Charles Aaron <caaron@ceris.purdue.edu> writes...  }Antony, } I }Thank you for your suggestion. I tried that but the size of the original / }save set is approx 303k blocks and when I copy : }it to disk it goes to over 1.8 mill. It does not stay the }same size.  }  }Chuck  
 Yes, it does.   K One of your blocks is not the same block as the other block. Or vice-versa.   $ A block on disk is always 512 bytes.  A A block on tape depends on what it is set to, but simply mounting G the tape via "MOUNT TapeDevice: label", no /BLOCKSIZE=n qualifier, will H mount it with 2048 byte blocks. The /BLOCKSIZE=n qualifier on the BACKUPG command has a minimum of 2048 and defaults to 8192 for savesets written I to tape (on disk a backup saveset "block" will span multiple disk blocks, E using a fixed length X byte record format where X defaults to 32256).    --- Carl   ------------------------------    Date: 24 Oct 2002 13:20:59 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones) 9 Subject: Re: backup/image of shadowed volume misses files = Message-ID: <8a646952.0210241220.489823a0@posting.google.com>   n "Stewart, Bill" <wjs-corp@Kaman.com> wrote in message news:<1E4B06029E11D211B47C0000F8207F4D01B9F47D@ESKC2>... > Ken,G > 	Try putting in /ignore=interlock.  If a file is open only the header & > is copied without the ignore switch. >  > Bill >  > ->-----Original Message-----@ > ->From: ken.randell@fortel.com [mailto:ken.randell@fortel.com]- > ->Sent: Thursday, October 24, 2002 10:12 AM  > ->To: Info-VAX@Mvb.Saic.Com 9 > ->Subject: backup/image of shadowed volume misses files  > -> > -> > ->AXP VMS 6.2, 7.2 > ->D > ->I have a weird problem using backup/image on some of my shadowedI > ->volumes where it appears to be missing files...I trust that I'm doing  > ->something wrong. > ->F > ->I have one 'master' system that I make tape images of my disks for8 > ->distribution to other systems.  My scenario is this: > ->? > ->1) create 'master' tape from a shadowed volume (this works)  > ->' > ->$backup/image dsa0: mka600:dsa0.bck  > ->J > ->backup/list of this tape shows all of the files that I would expect to	 > ->have.  > ->E > ->2) take 'master' tape to another system and restore to one volume  > ->(this appears to work) > ->* > ->$backup /image mka600:dsa0.bck dkb100: > ->G > ->3) on the other system, init a second volume to prepare it join the ' > ->shadow set, and create a shadow set  > -> > ->$init dkc100: label 4 > ->$mount dsa0: /shadow=($1$dkb100:,$1$dkc100:) ... > ->D > ->and I get the usual shadow copy operation from dkb100 to dkc100: > ->I > ->4) [problem starts here] Later, if I attempt to back up this shadowed D > ->drive on the new system as follows, I don't get all of the filesJ > ->copied as I would expect -- the files are simply skipped with no errorF > ->message or other (there is no activity or files open on the volume > ->when this backup occurs):  > ->* > ->$backup/image dsa0: mka600:newdsa0.bck > ->G > ->However, backing up by individual directory works.  I can also type ; > ->the files, etc., so the contents do appear to be there.  > ->H > ->Something bizarre that I see is that the FILE_ID for the files whichI > ->are not backed up do NOT match in the 3rd number (RVN I think)...they , > ->are 0 instead of 1 like the other files. > ->E > ->ANA/DISK/LIST DSA0: finds the files, so I think the indexf.sys is 7 > ->okay.  ANA/DISK/REPAIR doesn't find anything wrong.  > ->F > ->What am I missing?  Should I be splitting the shadowset before the > ->backup on the other system?  > -> > ->Thanks.  > ->  E The previous reply may solve your problem any time when the volume is E mounted system or clustered wide. However, I will dismount the volume D and remount the volume local to the backup process. This will insureF all the files are backup from the volume. After the backup, the volume& is remounted system or clustered wide.  F The Relative Volume Number refers to the volume which the file residesB on. If the file resides or found on Volume 2, then the RVN will be "2".   ------------------------------  % Date: Thu, 24 Oct 2002 14:39:34 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> 9 Subject: Re: backup/image of shadowed volume misses files , Message-ID: <3DB83E65.4EAC7CC9@videotron.ca>  @ > ->From: ken.randell@fortel.com [mailto:ken.randell@fortel.com]' > ->$backup/image dsa0: mka600:dsa0.bck   * > ->$backup /image mka600:dsa0.bck dkb100:  N Digital recommends the use of /NOALIAS  with /IMAGE on recent versions of VMS.6 (nothing to do with the TV program). HELP BACKUP/ALIAS  J You should really also have a /LOG and a /JOURNAL=file when you create the$ saveset (and /LOG when you restore).  @ Is it always the same files that get omitted when you restore ?   L Have you tried DIR/FULL on the file ? (file could be marked as nobackup, butE still, its header would be written to disk and an empty file of equal . size/attributes would be created upon restore)  L Another option is that the file isn't a file, but rather just an alias (file entry) pointing to a real file.    ------------------------------  % Date: Thu, 24 Oct 2002 15:52:33 -0400 ) From: "Scott Greig" <jsgreig@geminaq.com> 9 Subject: Re: backup/image of shadowed volume misses files : Message-ID: <y4Yt9.11439$Kf.1314747@news20.bellglobal.com>  A From the hp OpenVMS "Alpha Version 7.3-1 Release Notes" , section   3 4.3 BACKUP Utility-/[NO]ALIAS Qualifier Restriction   F Compaq recommends that you do not use the /ALIAS or /NOALIAS qualifier with the BACKUP/IMAGE command.   FWIW Scott   : "JF Mezei" <jfmezei.spamnot@videotron.ca> wrote in message& news:3DB83E65.4EAC7CC9@videotron.ca...B > > ->From: ken.randell@fortel.com [mailto:ken.randell@fortel.com]) > > ->$backup/image dsa0: mka600:dsa0.bck  > , > > ->$backup /image mka600:dsa0.bck dkb100: > K > Digital recommends the use of /NOALIAS  with /IMAGE on recent versions of  VMS.8 > (nothing to do with the TV program). HELP BACKUP/ALIAS > L > You should really also have a /LOG and a /JOURNAL=file when you create the& > saveset (and /LOG when you restore). > A > Is it always the same files that get omitted when you restore ?  > J > Have you tried DIR/FULL on the file ? (file could be marked as nobackup, but G > still, its header would be written to disk and an empty file of equal 0 > size/attributes would be created upon restore) > H > Another option is that the file isn't a file, but rather just an alias (file ! > entry) pointing to a real file.    ------------------------------  % Date: Thu, 24 Oct 2002 21:39:15 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca> 9 Subject: Re: backup/image of shadowed volume misses files , Message-ID: <3DB8A0C2.A6B2EA21@videotron.ca>   Scott Greig wrote: > C > From the hp OpenVMS "Alpha Version 7.3-1 Release Notes" , section H > Compaq recommends that you do not use the /ALIAS or /NOALIAS qualifier  > with the BACKUP/IMAGE command.  " $HELP BACKUP /IMAGE  (VAX VMS 7.2) ##N Specifying the /IMAGE qualifier WITHOUT also specifying /NOALIAS can result inI incomplete disk or file restoration operations. Therefore Compaq strongly K recommends that you specify /NOALIAS with /IMAGE when performing image mode  backup restoration.  ##  3 So, what is the real scoop on /ALIAS /NOALIAS ?????    ------------------------------  # Date: Fri, 25 Oct 2002 02:16:01 GMT , From: "Ken Randell" <ken.randell@fortel.com>9 Subject: Re: backup/image of shadowed volume misses files 6 Message-ID: <BP1u9.8017$iV1.5981@nwrddc02.gnilink.net>  ' Bad form to follow up on my own post...   6 Upon further investigation, the following falls out --  I Shadowing was not involved in this.  If I split the shadow set, mount the 0 split-off member /nowrite and issue the command:  ) $BACKUP/IMAGE dkcxxx disk$dkcxxx.bck/save   : (there is no activity on this disk, installed files, etc.)  7 I get the same problem; however, the following command:   7 $BACKUP dkcxxx:[*...]*.*.* disk$dkcxxx_noimage.bck/save    gets all of the files.  D So, I can work around my immediate problem & reconstruct this volumeL properly by creating an empty shadowset and restoring from my 'noimage' saveA set to the shadowset above with the appropriate qualifiers, as in J '/owner=original'.  At this point, I don't care about the directory dates,J which are set to the time that I do the restore.  This creates the volume,6 and all files can then be retrieved via $BACKUP/IMAGE.   I'm still perplexed by:   K 1) For my 'missing' files, why does ANA/DISK/LIST show the RVN to be 1 when " DIR/FILE_ID shows the RVN to be 0?J 2) Why does BACKUP/IMAGE not find these files but BACKUP [*...]*.*.* does?E 3) If indexf.sys or other is corrupted, why doesn't ANA/DISK show any 	 problems?    ------------------------------  % Date: Fri, 25 Oct 2002 00:16:56 -0400   From: John Santos <JOHN@egh.com>9 Subject: Re: backup/image of shadowed volume misses files 6 Message-ID: <1021025001159.11075B-100000@Ives.egh.com>  ' On Fri, 25 Oct 2002, Ken Randell wrote:   ) > Bad form to follow up on my own post...  > 8 > Upon further investigation, the following falls out -- > K > Shadowing was not involved in this.  If I split the shadow set, mount the 2 > split-off member /nowrite and issue the command: > + > $BACKUP/IMAGE dkcxxx disk$dkcxxx.bck/save  > < > (there is no activity on this disk, installed files, etc.) > 9 > I get the same problem; however, the following command:  > 9 > $BACKUP dkcxxx:[*...]*.*.* disk$dkcxxx_noimage.bck/save  >  > gets all of the files. > F > So, I can work around my immediate problem & reconstruct this volumeN > properly by creating an empty shadowset and restoring from my 'noimage' saveC > set to the shadowset above with the appropriate qualifiers, as in L > '/owner=original'.  At this point, I don't care about the directory dates,L > which are set to the time that I do the restore.  This creates the volume,8 > and all files can then be retrieved via $BACKUP/IMAGE. >  > I'm still perplexed by:  > M > 1) For my 'missing' files, why does ANA/DISK/LIST show the RVN to be 1 when $ > DIR/FILE_ID shows the RVN to be 0?  B Is the original source disk part of a volume set?  That opens up aD whole new bucket of worms.  I've never used volume sets, but I think? you need to do special things when backing up a volume set.  In D particular, you may be backing up files whose directory is on DSA0:,D but whose contents are on another volume of the set.  Doing an image@ backup of DSA0: may only pick up the directory entry and not theC contents, but doing a non-image backup of DSA0:[directory]*.*;* may @ follow the links and retrieve the data from the other volume(s).    L > 2) Why does BACKUP/IMAGE not find these files but BACKUP [*...]*.*.* does?G > 3) If indexf.sys or other is corrupted, why doesn't ANA/DISK show any  > problems?   @ Have you tried ana/disk on the original system, or were you only* look on the system you had restored it to?   --   John Santos  Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------    Date: 24 Oct 2002 13:00:07 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones) 9 Subject: Re: batch process won't die after BACKUP abortedX= Message-ID: <8a646952.0210241200.3d5b3d2f@posting.google.com>   e Alan Greig <a.greig@virgin.net> wrote in message news:<thpfrug7bni6j0ccnm22a8gpro8l1tu7hp@4ax.com>... . > On Thu, 24 Oct 2002 06:44:16 -0400, JF Mezei' > <jfmezei.spamnot@videotron.ca> wrote:a > 6 > >BACKUP-F-WAITIDLEBCB, attempted wait on idle buffer > H > I'm sure I have seen patches which addressed this error. Maybe nothing- > to do with your problem but worth checking.k >  > >t  E I experience this problem over the years and in the later versions oflF VMS this problem doesn't appear. The backup process uses the paged andE non-paged pool during the backup process. I would look at your virualwE non-paged pool allocation and seen if you haven't max out the virtualsB allocations. You can increase the virtual and physical allocation,C however, be careful, you can allocate more than physical memory you D have. Autogen will not catch it. If you do, the system will crash orF fail to reboot. If can't increase the virtual non-paged allocation dueA lack of memory, then I would decrease the use of process quotas (dC astlm, diolm, biolm, etc.) This will slow down your backup process.i   ------------------------------  % Date: Thu, 24 Oct 2002 08:38:16 -0700u& From: Greg Cagle <gregc@gregcagle.com>  Subject: Re: Carly on innovation/ Message-ID: <urg4v02g5esf1a@corp.supernews.com>E   Bill Todd wrote:8 >  HP gutted its own storage group *before* the merger,     Do you have a citation for this?   -- s
 Greg Cagle gregc at gregcagle dot com   ------------------------------  % Date: Thu, 24 Oct 2002 18:43:21 -0400s* From: "Bill Todd" <billtodd@metrocast.net>  Subject: Re: Carly on innovation2 Message-ID: <PeScnfHh34Xk6iWgXTWcpQ@metrocast.net>  3 "Greg Cagle" <gregc@gregcagle.com> wrote in messagea) news:urg4v02g5esf1a@corp.supernews.com...  > Bill Todd wrote:9 > >  HP gutted its own storage group *before* the merger,o >y" > Do you have a citation for this?  > Sorry - private communication from someone with friends there.   - bill   ------------------------------  % Date: Thu, 24 Oct 2002 16:30:25 -0700e& From: Greg Cagle <gregc@gregcagle.com>  Subject: Re: Carly on innovation/ Message-ID: <urh0k8lri41lca@corp.supernews.com>a   Bill Todd wrote:    > "Greg Cagle"  wrote in message+ > news:urg4v02g5esf1a@corp.supernews.com...  >e > >Bill Todd wrote:e > >e9 > >> HP gutted its own storage group *before* the merger,  > >$# > >Do you have a citation for this?M >T >o@ > Sorry - private communication from someone with friends there.  ? Well, I've got plenty of friends who are *still* there. Perhaps   "gutted" isn't exactly accurate. -- e
 Greg Cagle gregc at gregcagle dot com   ------------------------------  % Date: Thu, 24 Oct 2002 20:06:30 -0400-* From: "Bill Todd" <billtodd@metrocast.net>  Subject: Re: Carly on innovation2 Message-ID: <SeWcnVCSNLdrFyWgXTWcpg@metrocast.net>  3 "Greg Cagle" <gregc@gregcagle.com> wrote in messages) news:urh0k8lri41lca@corp.supernews.com...g > Bill Todd wrote: >-" > > "Greg Cagle"  wrote in message- > > news:urg4v02g5esf1a@corp.supernews.com...2 > >0 > > >Bill Todd wrote:. > > >s; > > >> HP gutted its own storage group *before* the merger,  > > >-% > > >Do you have a citation for this?o > >  > >aB > > Sorry - private communication from someone with friends there. > A > Well, I've got plenty of friends who are *still* there. Perhapsi" > "gutted" isn't exactly accurate.  L And perhaps it is.  I'd be happy to accept any actual numbers you might wishA to provide (though often numbers aren't quite as important as they individuals they represent).   - bill   ------------------------------  % Date: Thu, 24 Oct 2002 21:35:59 -0700t& From: Greg Cagle <gregc@gregcagle.com>  Subject: Re: Carly on innovation. Message-ID: <urhih67nq0o6f@corp.supernews.com>   Bill Todd wrote:    > "Greg Cagle"  wrote in message+ > news:urh0k8lri41lca@corp.supernews.com...t >/ > >Bill Todd wrote:o > >/ > > " > >>"Greg Cagle"  wrote in message- > >>news:urg4v02g5esf1a@corp.supernews.com...h > >> > >> > >>>Bill Todd wrote:e > >>>h > >>>l: > >>>>HP gutted its own storage group *before* the merger, > >>>t% > >>>Do you have a citation for this?i > >> > >>B > >>Sorry - private communication from someone with friends there. > >aB > >Well, I've got plenty of friends who are *still* there. Perhaps# > >"gutted" isn't exactly accurate.O >E >EN > And perhaps it is.  I'd be happy to accept any actual numbers you might wishC > to provide (though often numbers aren't quite as important as theE > individuals they represent).   Sorry, I can't.e --  
 Greg Cagle gregc at gregcagle dot com   ------------------------------    Date: 24 Oct 2002 13:43:59 -0500+ From: young_r@encompasserve.org (Rob Young)e& Subject: Re: Denial of Service attacks3 Message-ID: <461ZOT1$owbv@eisner.encompasserve.org>   M In article <3DB83C70.3C1770C7@mist.com>, GreyCloud <cumulus@mist.com> writes:l; > Does anybody know if the recent denial of service attacks = > that were mentioned in the news know what type of O/S thesen > were running on???     	Root Servers?  Yes.  AIX.  K http://www-1.ibm.com/servers/eserver/pseries/solutions/success/network.htmlu   				Rob    ------------------------------  % Date: Thu, 24 Oct 2002 11:31:12 -0700." From: GreyCloud <cumulus@mist.com>" Subject: Denial of Service attacks( Message-ID: <3DB83C70.3C1770C7@mist.com>  9 Does anybody know if the recent denial of service attacks ; that were mentioned in the news know what type of O/S thesel were running on???   ------------------------------  + Date: Thu, 24 Oct 2002 16:54:43 -0600 (MDT)-" From: John Nebel <nebel@csdco.com>& Subject: Re: Denial of Service attacksF Message-ID: <Pine.OSF.4.21.0210241652020.1794-100000@athena.csdco.com>  H F which is one of the busiest uses TRU64 on ES40s was one of the 4 which
 stayed up.   www.isc.orgi  
 John Nebel      On 24 Oct 2002, Rob Young wrote:  O > In article <3DB83C70.3C1770C7@mist.com>, GreyCloud <cumulus@mist.com> writes:i= > > Does anybody know if the recent denial of service attackss? > > that were mentioned in the news know what type of O/S thesed > > were running on??? >  >  > 	Root Servers?  Yes.  AIX. > M > http://www-1.ibm.com/servers/eserver/pseries/solutions/success/network.html  > 	 > 				Robm >  >    ------------------------------  % Date: Thu, 24 Oct 2002 18:31:06 -0400a* From: "Bill Todd" <billtodd@metrocast.net>4 Subject: Re: Dual CPU cards vs. dual core processors2 Message-ID: <3kKdnVwt3LIG6SWgXTWcpw@metrocast.net>  . "John Smith" <a@nonymous.com> wrote in messageC news:EETt9.60058$mxk1.47240@news04.bloor.is.net.cable.rogers.com...  > L http://story.news.yahoo.com/news?tmpl=story2&ncid=1208&e=2&u=/pcworld/200210# > 23/tc_pcworld/106278&sid=95612453  >h >c& > HP Doubles Up on Itaniums in Servers > Wed Oct 23, 5:00 PM ET  > Ashlee Vance, IDG News Service > K > Hewlett-Packard will double the processor count in its Itanium servers inoJ > 2004, as the company attempts to fill a technology gap caused by Intel'sK > decision not to make multicore Itanium processors until the middle of the 	 > decade.t  H Ahah!  More confirmation that nothing new will be coming down the ItanicK pike until 2005 at the earliest.  Plus an indication that HP is starting to J recognize that Itanic may be the SPARC of this decade ("Lower performance?D No problem:  just plug in some more processors" - Rob will love it).   >"F > HP plans to make it possible to upgrade current servers based on the 64-bitJ > Itanium 2 chip by allowing customers to replace existing processor cardsL > with new ones that contain twice as many chips, said Shane Robison, seniorJ > vice president of corporate strategy and technology and chief technology > officer at HP.  E Ah, yes:  the architect of Alpha's (and one might suspect later HP's)d demise.i  >  This technology will also work with future generations of theL > Itanium processor and should provide a performance boost for HP's systems.I > It will also work as a stopgap measure to help HP compete in the serverCH > market as Intel does not plan to release similar technology known as a) > multicore processor until 2005 or 2006.a  L First actual date I've seen associated with this feature.  The question thenG becomes whether it's the 'bump' to Montecito mentioned a year or so ago4G (which I had suspected would be EV7-style on-chip glue, but it could beAI either or both - the on-chip glue being the more important of the two) ormD will wait until Chivano (whatever that may be, likely in 2006 at the
 earliest).  I > "We have developed our own dual processor cache daughter card," Robison K > said. "Customers can double the number of CPUs (central processing units)aJ > within most HP systems--anything that has a McKinley, which is Itanium 2 inK > it, or the Madison processor--simply by upgrading to this daughter card."a  K HP 'innovation' at its finest - what a guy!  Of course, this does mean thatCJ whatever bus/memory bottlenecking the system was already experiencing willI suddenly become twice as bad:  that could be why other manufacturers haverJ tended to coordinate increases in processor count with surrounding chipsetJ enhancements.  And of course an additional layer of arbitration in the newK package (between its two processors and the chipset that's expecting to seejH only one) won't help memory latency either (did anyone notice the recentL change in the latency figures in the vaunted zx1 chipset from 112 ns. to 156L ns?  it turns out that the lmbench version used to calculate the former used3 a stride that allowed the chipset to pre-fetch...).e   >i >  > Matching RivalsnL > HP will make the dual processor cards available in the first half of 2004.I > The card will contain the third-generation Madison chips, which are due  out F > in the first half of 2003. Customers with either McKinley servers orH > Madison-based servers will be able upgrade two to 64 processor systems when$ > the cards arrive in 2004, HP said. >sL > Doubling the processor count in its high-end servers would help HP competeD > against IBM and Sun. IBM has already rolled out a dual-core Power4	 processorn. > for its Unix (news - web sites) server line.  I Actually, the dual-core POWER4 appeared about a year ago, each individualdJ POWER4 core providing close to Itanic-level performance at only about halfJ the power consumption.  And with a couple more process-shrinks in store inI the 2004 time-frame, one might expect more than 2 cores on POWER chips by E the time Shane's 10-pounds-of-you-know-what-in-a-5-pound-bag hits the  streets.     Sun and HP also plan to releaseJ > dual-core chips for their servers based on RISC (reduced instruction set" > computing) technology next year.  K So Itanic will only be 2 - 3 years behind them in this respect, as comparedcI with 4 - 5 years behind IBM.  And only a couple of years behind Alpha and:I Hammer in the area of on-chip glue (whoops:  USIII got its on-chip memoryeL controller last year, so I guess they'll again be 4 - 5 years behind in thatL instance).  Clever, these Intel chaps:  no wonder Itanic is destined to take over the world.   *  HP, however, had not previously announcedI > similar technology for its Itanium line of servers that will eventuallyh > replace its RISC systems.   F And, of course, isn't announcing it now either:  dual-chip modules areI really not all that 'similar' to dual-core chips - consider, for example,fB issues of power consumption and cooling (leaving aside the latency5 advantages in on-chip inter-processor communication).h   >mF > Packing more processors into existing servers helps customers make a7 > significant upgrade without paying for a new chassis.s  G Or, if you wished to be cynical, helps keep customers in the fold (with6K incremental additional purchases) while Intel keeps trying to catch up withsG the competition those customers might otherwise be inclined to move to,o@ while still leaving open the possibility of selling them the newE latest-and-greatest Itanic box when it finally appears.  Sort of liketH changing automobile tail-fins in the '50s, and not necessarily much more functional.l    HP users will be ableI > to boost the performance of their systems by ripping out daughter cardsa withL > the old processors and replacing them with the new double processor cards,J > said Dean McCarron, principal analyst at market research company Mercury > Research.nK > "In general, with an SMP (symmetric multiprocessing) system, when you add < > more processors, you get more performance," McCarron said.  K Because in general, the box you add them to was designed to accommodate the K additional processing elements, rather than designed to handle only half asb many.i    "It depends ong: > the applications, though, as to how well it will scale."   Well, that too, of course.   ...   L > HP's decision to make the double processor cards will buy the company someL > time before Intel ships full fledged multi-core chips later in the decade.  G Buying time seems to be the watch-phrase of this architecture.  Itanic1iH (which originally was supposed to be a real product - in 1997...) boughtH time until Itanic2 could *really* blow the competition away (though onlyE about a year ago, articles were starting to caution that the McKinleyvJ ramp-up would still be slow:  they might only be shipping about 100K unitsE per quarter by now...).  Now McKinley is buying time until the *real* J ramp-up can begin with Madison.  And in 2004 the two-headed HP mutant will( buy time until the real thing arrives...  L > HP, which co-developed Itanium with Intel, said this move is an example ofE > its capability to add technology to a chip it does not manufacture.c >eG > "This gives us another example where we invest and innovate on top ofw. > industry standard components," Robison said.  J And what a fine example of innovation it is.  At the very least comparableJ to Compaq's addition of the 'Internet key' - though one expects they'll beF charging a bit more for it (two Itanics don't come cheap, nor does theL non-industry-standard packaging - oh, dear! - that it will take to make them appear to be one).   - bill   ------------------------------  % Date: Thu, 24 Oct 2002 18:51:49 -0700a& From: Tom Crabtree <tccrab@sunset.net>1 Subject: Re: Dual ISA ethernet adapter on VMS 7.2l* Message-ID: <3DB8A3B5.A71ED8D6@sunset.net>   TR:   M > The board is officially not supported for OpenVMS, but I found descriptions - > of how to get it running with a Multia Kit. G Yes.  It works.  I've got such an animal setting in my 'data dungeon'. -E Admittedly I've not used it in a while, but it ran just fine for manya years.  rI > This Multia Kit, what exactly is that? Will I get it when I buy OpenVMSm > 7.2/Hobby License?G You will receive the latest release of OpenVMS Hobbyist (7.3 ? guys..?)dH once Montgar releases it.  Don't know for sure if they've caught up with orders yet.e  J > I have two DEC Etherworks 3 DE204 ISA cards. I found out these cards areH > supported by OpenVMS. However, I read that OpenVMS 6.2 doesn't supportB > multiple ISA ethernet adapters. Is this still the case with 7.2?D Dump the ISAs, nothing but trouble. Recommend DE500's.  You can pickD them up on eBay for cheap. DE435 and 450's will work too.  I've seen them for $5.95 on eBay.-  -L > Also, I read that OpenVMS only supports a single PCI device. Would puttingL > in a PCI nic and an ISA nic work? Is there a list of PCI/ISA hardware that0 > is supported? Any non-DEC network cards maybe?B Wrong. Where did you read this?  Put as many PCI cards in as slots9 allow. I seem to recall the the AXPCI33 has 5 PCI slots.     L > The reason why I ask is because ultimately I'd like to perform NAT routingM > on the system to share my cable internet connection. Is this even possible?,= > Has anyone ever done this and is there software to do this?uB OK, but routers are cheaper (and easier to set up).  You can buy a/ Linksys for $60-80.  Uses less electricity too.1  1L > Finally, the board has an IDE port and a SCSI-2 port. Can I simply connectL > any IDE or SCSI harddrive on the system, no limitations? And what about CDH > ROM drives? Will an IDE ATAPI CD ROM drive on the IDE port do the job?  B OpenVMS does not like IDE drives.  They will cause you nothing butH trouble, especially in a system of this age.  Find a couple of RZ28 HD'sF and a DEC RRD42 or 43 CDRom (they're slow, but you will only use it toH load software).  You will have enough fun learning OpenVMS, don't hamper$ yourself with hardware problems too.  G Graphics cards will be a bit of a problem, OpenVMS is quite fussy aboutv& them too.  What do you intend to use? B I've used S3 Trio, Matrox Millennium, Powerstorm 3D30 & 3D40, Elsa$ Gloria (with the Permedia chip set).   Tom Crabtree OpenVMS Neophyte   ------------------------------    Date: 24 Oct 2002 13:37:21 -0500 From: briggs@encompasserve.org Subject: Re: File Attributes3 Message-ID: <dbcBHoC0KSMf@eisner.encompasserve.org>   c In article <3DB8134A.B784960B@mindspring.com>, Atlant Schmidt <atlantnospam@mindspring.com> writes:0! > briggs@encompasserve.org wrote:. > D >> Yes, we have the equivalent of system().  And like system(), thatD >> call is limited by the ability of the caller to access the output- >> stream of the resulting forked subprocess.s >>H >>  call lib$spawn ( 'WRITE SYS$OUTPUT F$FILE("your.file.name","RDT")' ) >>H >> Now all you need to do is harvest the revision date and time from the0 >> output stream of that created subprocess. ... > 0 > So there's no equivalent of system()'s brother. > (wrapper routine) popen(), which makes quick* > work of the output processing (and makes > `backticks` possible)?  ? If you want to deal with null terminated strings, I don't thinkEA that there is anything that prevents a Fortran program from using  decc$popen and decc$fgets.   CC     Run-time_functions  	     popenh  %        Initiates a pipe to a process.           Syntax:                #include <stdio.h>h  A              FILE *popen (const char *command, const char *type);d   	John Briggs   ------------------------------  % Date: Thu, 24 Oct 2002 11:35:38 -0400s2 From: Atlant Schmidt <atlantnospam@mindspring.com> Subject: Re: File Attributes. Message-ID: <3DB8134A.B784960B@mindspring.com>   briggs@encompasserve.org wrote:s  C > Yes, we have the equivalent of system().  And like system(), thatsC > call is limited by the ability of the caller to access the output , > stream of the resulting forked subprocess. > G >  call lib$spawn ( 'WRITE SYS$OUTPUT F$FILE("your.file.name","RDT")' )o >nG > Now all you need to do is harvest the revision date and time from the-/ > output stream of that created subprocess. ...4  . So there's no equivalent of system()'s brother, (wrapper routine) popen(), which makes quick( work of the output processing (and makes `backticks` possible)?   Atlant   ------------------------------  # Date: Thu, 24 Oct 2002 17:44:15 GMTx9 From: Hein van den Heuvel <hein_netscape@eps.zko.dec.com>8 Subject: Re: File Attributes/ Message-ID: <3DB830A4.F995EF15@eps.zko.dec.com>-  > > It might be nice to have a lib$file_attributes function thatI > corresponds to the lexical, but on the other hand it's not particularly 4 > difficult to write such a utility routine oneself.  A Theo de Klerk wrote one in Pascal back in 1986. Is it in his boo?v  0 The first questions for this started in 1984.... For your amuzement:_   Notes>H ------------------------------------------------------------------------H Note 438.3         Callable lexical functions                     3 of 5H xxxxxx::LIONEL                                7 lines  22-OCT-1984 16:11H ------------------------------------------------------------------------G The RTL in V4 has callable versions of F$GETDVI, F$GETJPI and F$GETSYI.kK These are called LIB$GETDVI, etc.  Note that these are implemented entirely > inside the RTL; they do not call into DCL so you need not have9 a CLI around to use them.  I wanted to also do a callable-A F$FILE_ATTRIBUTES, but ran out of time.  There is a bit of latent02 support for such a thing in the existing routines.%                                 Stevea    = The real thing never happened.  Often requested, never deemed>; important enough as the XAB route is not all that hard oncea1 you have made your first one using a few examplese  1 Two fortran examples from yours truly appended...i= One specifically for RDT (to SET actually) as this is part ofx4 the generic XAB and one more to get CDT using XABDAT   hth, Hein.e  H ------------------------------------------------------------------------H Note 588.3     How do I change revision date and time (XABRDT)    3 of 6H CASEE::VANDENHEUVEL "Hein, Valbonne."_mm_/\_mm_"  69 lines 2-MAY-1988B                -< working sample code to set RDT from useropen. >-H ------------------------------------------------------------------------  C     You are being caught by some idiosyncrasies for the fortran XABs definitions.  G     Every XAB has a common part and a specific part. It just so happensd?     that *all* (2) of the xabrdt fields are in the common part!sC     The solution at the end of this note makes naughty use of this.aH     Note: I tend to use subroutines for the common parts (also included)  8     RMS programming: all XAB's must be chained together.  <     USEROPEN programming: leave XAB chains as you find them.  D         It is a good (required) practice to hookup your XAB allowingC         for XABs provided by the language support RTL and to removez!         your XAB before returningp  @     USEROPEN programming: Do not lie too much about file status.  5         Indicate succes if opened, failure if closed.,  ?     Giving the above, the quickest (not best) solution becomes:              : as before 	         :d c      initialize the xabrdt  )        useropen_xab.xab$b_cod = xab$c_rdtu,        useropen_xab.xab$b_bln = xab$c_rdtlen9        useropen_xab.xab$l_nxt = 0               !No next.t/        useropen_xab.xab$l_rdt0    = new_date(1) /        useropen_xab.xab$l_rdt4    = new_date(2)a  5        fortran_xab = fab.fab$l_xab              !SavecA        fab.fab$l_xab = %loc(useropen_xabdat)    !Point to DAT XAB          fab.fab$b_fac = fab$v_upd'        set_date = sys$create(%ref(fab))-A        fab.fab$l_xab = %loc(useropen_xab)       !Point to RDT XABm8        set_date = sys$close(%ref(fab))          !Set RDT8        fab.fab$l_xab = fortran_xab              !RestoreI        set_date = 0                             !Fake failure to indicatei close.
        return>
        end  H C-----------------------------------------------------------------------+         SUBROUTINE SET_XAB(XAB,COD,BLN,NXT)s         IMPLICIT NONE          INCLUDE '($XABDEF)'e         RECORD  /XABDEF/ XAB         BYTE    COD, BLN         INTEGER NXTl         XAB.XAB$B_COD=CODn         XAB.XAB$B_BLN=BLN          XAB.XAB$L_NXT=NXT1         END>H C-----------------------------------------------------------------------+         SUBROUTINE GET_XAB(XAB,COD,BLN,NXT)2         IMPLICIT NONEi         INCLUDE '($XABDEF)'          RECORD  /XABDEF/ XAB         BYTE    COD, BLN         INTEGER NXTt         COD=XAB.XAB$B_CODo         BLN=XAB.XAB$B_BLN          NXT=XAB.XAB$L_NXTa         END   6 That reply was in response to a failing program beloe:          program set_new_date           implicit none          external set_date(        common /expiration_date/ new_date  (        integer*4 new_date(2), sys$bintim#        integer*4 return_status, iosI        character*23 ascii_date  -        ascii_date = '31-DEC-1989 00:00:00.00'n  , c      convert the ascii date to system date c      & time format  6        return_status = sys$bintim (%descr(ascii_date),&      1                 %ref(new_date))#        if (.not.return_status) thenr0                type *,'Error calling SYS$BINTIM'2                call lib$stop (%val(return_status))
        end ifn   c      create the file  ;        open (unit=1,status='old',organization='sequential',r9      1         file='set_new_date.seq',useropen=set_date,>      2         iostat=ios)  "        write(6,*)'open_ios : ',ios          close(unit=1)  
        end   c      useropen routine   0        integer*4 function set_date (fab,rab,lun)          include '($fabdef)'        record /fabdef/ fab        record /fabdef/ fab          include '($rabdef)'        record /rabdef/ rab          include '($xabdef)'#        record /xabdef/ useropen_xab,          include '($xabdatdef)'t)        record /xabdatdef/ useropen_xabdats          include '($xabrdtdef)'l)        record /xabrdtdef/ useropen_xabrdth  (        common /expiration_date/ new_date  $        integer*4 new_date(2), status=        integer*4 sys$create, sys$connect, sys$open, sys$closeh   c      initialize the xabdat  5        useropen_xabdat.xabdatdef$$_fill_1 = xab$c_datt8        useropen_xabdat.xabdatdef$$_fill_2 = xab$c_datlen  /        useropen_xabdat.xab$l_edt0 = new_date(1)h/        useropen_xabdat.xab$l_edt4 = new_date(2)   /        useropen_xabdat.xab$l_bdt0 = new_date(1)m/        useropen_xabdat.xab$l_bdt4 = new_date(2)v/        useropen_xabdat.xab$l_cdt0 = new_date(1)p/        useropen_xabdat.xab$l_cdt4 = new_date(2)k  /        useropen_xab.xab$l_rdt0    = new_date(1)s/        useropen_xab.xab$l_rdt4    = new_date(2)i   c      initialize the xabrdt  5        useropen_xabrdt.xabrdtdef$$_fill_1 = xab$c_rdtf8        useropen_xabrdt.xabrdtdef$$_fill_2 = xab$c_rdtlen  ,        fab.fab$l_xab = %loc(useropen_xabdat)         fab.fab$b_fac = fab$v_upd  '        set_date = sys$create(%ref(fab))b  % c      set_date = sys$open(%ref(fab))e           if (.not.set_date) return  )        set_date = sys$connect (%ref(rab))c         if (.not.set_date) return  
        returng
        end  2                                -< ** Hackers ** >-G =======================================================================sG Note 587.5  Getting the creation date of a file from a FORTRAN program.oG CASEE::VANDENHEUVEL "Why is Software so Hard?"   102 lines  21-APR-1989 -                                    -< SMOP >-lF ----------------------------------------------------------------------  D     It is merely a matter of using a useropen, hooking up a DAT XAB.D     Should be a SMOP. Xcept that dealing with XABs is a problem withF     Fortran because following a pointer chain is relativly hard to do.  D     I am not really a Fortran programmer so I was a bit reluctant toE     publish the way I dealt with this problem previously. Perhaps the2=     following sample code may be of use to someone after all.u  G     The critical parts of the code are the XAB_SET and XAB_GET routines9I     the deal with the 'generic' XAB data fields and which are fooled intoi8     using passed values for addresses when appropriated.5     (It is Fortran-only, but is it clean Fortran? :-)h  G     The particular example inits a DAT and a PRO XAB and hooks them up,oK     opens an (existing) file, and unhooks the added Xab to prevent problemsm  ?     with potential RTL XAB controlblock deallocation functions.d  	     Hein.m           IMPLICIT        NONE!         CHARACTER*80    FILE_NAMEd&         INTEGER*2       FILE_NAME_SIZE%         INTEGER*4       UFO_OPEN, LUNm          EXTERNAL        UFO_OPENE         CALL LIB$GET_INPUT (FILE_NAME, 'File name? ', FILE_NAME_SIZE).<         OPEN    (UNIT=LUN, FILE=FILE_NAME(1:FILE_NAME_SIZE),2         1       STATUS='OLD', USEROPEN = UFO_OPEN)                 :s                 :          ENDm  H C-----------------------------------------------------------------------.         INTEGER FUNCTION UFO_OPEN(FAB,RAB,LUN)         IMPLICIT        NONE2         INTEGER*4       NXT, LAST_XAB, LUN, STATUS         INTEGER*2       BUFLEN-         INTEGER         SYS$OPEN, SYS$CONNECTt%         CHARACTER*200   ACLBUF,OUTBUFn          BYTE            COD, BLN         INCLUDE '($FABDEF)'-         INCLUDE '($RABDEF)'          INCLUDE '($XABDATDEF)'         INCLUDE '($XABPRODEF)'I         RECORD  /FABDEF/FAB, /RABDEF/RAB, /XABPRODEF1/PRO, /XABDATDEF/DAT     B C       Make this PRO a real XABPRO by Filling in the code fields.6 C       This XAB will be the last in the chain (NXT=0)  :         CALL SET_XAB(DAT,XAB$C_DAT,XAB$C_DATLEN,%LOC(PRO))I         CALL SET_XAB(PRO,XAB$C_PRO,XAB$C_PROLEN,0)      !Init COD and BLNPH         PRO.XAB$L_ACLBUF = %LOC(ACLBUF)                 !Init ACL buffer pointer H         PRO.XAB$W_ACLSIZ = 200                          !Init ACL buffer sizeG         NXT = FAB.FAB$L_XAB                             !At least 1 XAB  (XABFHC)G         DO WHILE (NXT.NE.0)                             !Walk the chainAH             LAST_XAB = NXT                              !Pointer to last XAB2K             CALL GET_XAB(%VAL(LAST_XAB),COD,BLN,NXT)    !Should check COD..c  B         END DO                                          !..want no duplicates.0C         CALL SET_XAB(%VAL(LAST_XAB),COD,BLN,%LOC(DAT))  !Hook'm up.r           STATUS = SYS$OPEN(FAB)-         IF (STATUS) STATUS = SYS$CONNECT(RAB)h         UFO_OPEN = STATUSsE         CALL SET_XAB(%VAL(LAST_XAB),COD,BLN,0)          !And Unhook'mC  5         CALL SYS$ASCTIM (BUFLEN,OUTBUF,DAT.XAB$Q_CDT)t,         TYPE *,' Created: ',OUTBUF(0:BUFLEN)H         CALL SYS$FORMAT_ACL (ACLBUF(0:PRO.XAB$W_ACLLEN), BUFLEN, OUTBUF,         1       50, '**', 3)  @         WRITE (5,1) PRO.XAB$W_GRP, PRO.XAB$W_MBM, PRO.XAB$W_PRO,B         1               PRO.XAB$W_ACLLEN, PRO.XAB$L_ACLSTS, BUFLEN>  1      FORMAT (' Owner=[',O3,',',O3,'], Protection mask=',O6,8         1       ' BUFLEN=',I3,', ACLSTS=',Z8,' OUT=',I3)         TYPE *,OUTBUF(0:BUFLEN),         END   H C-----------------------------------------------------------------------+         SUBROUTINE SET_XAB(XAB,COD,BLN,NXT)          IMPLICIT NONEB         INCLUDE '($XABDEF)'p         RECORD  /XABDEF/ XAB         BYTE    COD, BLN         INTEGER NXTt         XAB.XAB$B_COD=CODh         XAB.XAB$B_BLN=BLNy         XAB.XAB$L_NXT=NXT          ENDu  H C-----------------------------------------------------------------------+         SUBROUTINE GET_XAB(XAB,COD,BLN,NXT)-         IMPLICIT NONE          INCLUDE '($XABDEF)'6         RECORD  /XABDEF/ XAB         BYTE    COD, BLN         INTEGER NXTa         COD=XAB.XAB$B_COD          BLN=XAB.XAB$B_BLN1         NXT=XAB.XAB$L_NXT          ENDO   ------------------------------  % Date: Thu, 24 Oct 2002 15:03:18 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>  Subject: Re: File Attributes, Message-ID: <3DB843F3.13674BBB@videotron.ca>   "Craig A. Berry" wrote:-> > It might be nice to have a lib$file_attributes function thatI > corresponds to the lexical, but on the other hand it's not particularlyt4 > difficult to write such a utility routine oneself.   #include rms   struct FAB myfab ; struct NAM mynam ; struct XABDAT mydates;  L myfab = cc$rms_fab ;  /* initialise the blocks to lost fo default values */  mynam = cc$rms_nam ; t mydates = cc$rms_xabdat ;   ( myfab.fab$l_fna = "Chocolate_cake.txt" ;+ myfab.fab$w_fns = strlen(myfab.fab$l_fna) ;a myfab.fab$l_nam = &mynam ;K myfab.fab$l_xab = &mydates ; /* point the FAB to the first XAB in a list */   = mydates.xab$l_nxt = 0 ;  /* end of the linked list of XABs */r  # status = SYS$DISPLAY(&myfab, 0,0 );   < After this, you have the creation date in: mydates.xab$q_cdt  K It is in a VMS quadword date-time format. You can use some LIB$ routines toe format it into text date/time).   R The NAM block will get filled with the parsed filename with device, directory etc.  G There are various XAB types you can link to the chain to obtain variouskE information such as allocation, file attributes (record format etc ).t  N The RMS manual may seem daunting, but once you get the hang of it, it is quite simple to get done.u   ------------------------------   Date: 24 Oct 2002 22:03 CDTt' From: carl@gerg.tamu.edu (Carl Perkins)  Subject: Re: File Attributes- Message-ID: <24OCT200222032129@gerg.tamu.edu>a  e In article <3DB7EEE3.BBE91ABD@mindspring.com>, Atlant Schmidt <atlantnospam@mindspring.com> writes...A }JF Mezei wrote: }  }> Bill wrote:H }> > I can get to the file attributes (i.e. "Revision Date") through theH }> > Lexical f$file_attributes within a command procedure.  But how do I; }> > get to the same information through a Fortran program?v }> }> HELP RMSa }>N }> Look at FAB and various XABs that get populated with a SYS$OPEN (as well as }> other calls). } ' }Look at the FABs and RABs? Good golly!r0 }No wonder you VMS folks are paid the big bucks!. }You work very, very hard for simple data! :-) } / }If this question of file attributes isn't on a.) }performance-critical path, maybe it'd bep' }better to spawn something simple (DCL,.$ }Perl, whatever) and use *IT* to get }the data for you. } - }I ask the following question is seriousness,-% }and *NOT* as a one-upping technique:f } , }Has VMS got the equivalent of system() yet?/ }Creating the code to get file attributes would $ }take about a half minute with that. }  }Atlantn  @ FABs, RABs, and XABs are trivial. They are just data structures,? the definitions of which are well known. If you can't examine a A data structure for the values you want, you should step back fromaB the keyboard and stop doing anything resembing attempting to write
 a program.  D VMS has has system() type calls practically forever. Doing this with? such a call is *more* complex than using the RMS calls and datae structures (and slower, too).   E RMS can get complicated, but not in this case. All you do for this isiB define the data structures, initialize them, and then do somethingF that loads them like a Sys$Open( fab) (followed by a Sys$Close( fab)).E What could be easier? Certainly not spawning a subprocess and parsing  it's text output.:   --- Carl   ------------------------------   Date: 24 Oct 2002 22:16 CDTn' From: carl@gerg.tamu.edu (Carl Perkins)p Subject: Re: File Attributes- Message-ID: <24OCT200222161038@gerg.tamu.edu>o  e In article <3DB8134A.B784960B@mindspring.com>, Atlant Schmidt <atlantnospam@mindspring.com> writes...-  }briggs@encompasserve.org wrote: } D }> Yes, we have the equivalent of system().  And like system(), thatD }> call is limited by the ability of the caller to access the output- }> stream of the resulting forked subprocess.r }>H }>  call lib$spawn ( 'WRITE SYS$OUTPUT F$FILE("your.file.name","RDT")' ) }>H }> Now all you need to do is harvest the revision date and time from the0 }> output stream of that created subprocess. ... } / }So there's no equivalent of system()'s brother.- }(wrapper routine) popen(), which makes quick ) }work of the output processing (and makesm }`backticks` possible)?a }  }Atlantj   $ help cc run popena CC     Run-time_functions  	     popenp  %        Initiates a pipe to a process.e          Syntax:                #include <stdio.h>p  A              FILE *popen (const char *command, const char *type);s  K It's in there. It may, or may not, do everything you might expect it to do.h- (This is a fairly new addition to the C RTL.)r  @ Have you ever considered checking these things out for yourself?   --- Carl   ------------------------------  # Date: Thu, 24 Oct 2002 21:53:53 GMTr# From: "John Smith" <a@nonymous.com>-- Subject: Fiorina Gets Feisty at Gartner Event2H Message-ID: <RZZt9.62510$Q3S.43404@news01.bloor.is.net.cable.rogers.com>  = Somewhat stale dated now, but a few comments are interesting..    8 http://www.baselinemag.com/article2/0,3959,605377,00.asp       October 8, 2002p$ Fiorina Gets Feisty at Gartner Event   By  Tom Steinert-Threlkeld    G LAKE BUENA VISTA, Fla.-Merger accomplished, Carly Fiorina is coming outo feisty..  G In one 45-minute session at a gathering of 6,000 information-technology E executives and potential customers, the chairman and CEO of the newlyvL combined Hewlett-Packard and Compaq Computer corporations dissed rivals suchK as Dell Computer as a mere distribution mechanism, Sun Microsystems and EMCtL as "stand-alone hot box companies'' and IBM as a company that cares more for* its own needs than those of its customers.  H Affording her management team a B-plus grade for the way it has absorbedJ Compaq into Hewlett-Packard, Fiorina showed no fear at the prospect of theI industry's star entrepreneur, Michael Dell, now taking aim on the printer-D market, where HP generates a disproportionate share of its corporate profits.  H "The fact that Dell makes an announcement doesn't send us shaking in ourL boots," she said at the Gartner Symposium and ITxpo here. "Dell has become aF channel. They're not a technology company. They're not an R&D company.0 They're just trying to expand their portfolio.''  I In the first nine months of this year, HP's imaging and printing business>H accounted for $14.7 billion of the company's $54.9 billion in sales. ButF it's the only really profitable part of the business. As that unit hasL recorded $2.32 billion of earnings from operations in the first nine months,. HP as a whole has recorded just $2.26 billion.  H Sensing a possible vulnerability as HP merges with Compaq, Dell recentlyH reached an agreement with Lexmark International to have printers and ink, cartridges manufactured under the Dell name.  L But Fiorina dismissed the pact as a marketing move, since Dell already sells* Lexmark printers, under the Lexmark brand.  I "'Dude, you're getting a Lexmark' just doesn't have the same ring,'' said0F Fiorina, "which is maybe why Michael Dell has decided to private-label printers.''e  G As HP rolls out 50 new printing products, the company is also investingwH about $900 million a year in R&D and has, for instance, 100 patents on aI sub-$50 printer in its lineup. By contrast, Fiorina asserted, the Lexmarkk: deal shows that Dell is mainly a means of moving products.  J "Michael Dell can be a channel of distribution, but that's not the same as$ being a printer provider," she said.       HP's Corporate Focus  J She was similarly disparaging about other rivals, who she said lacked HP'sL focus on the needs of corporate customers-everything from the desktop to theL server, and from the network to the consulting services needed to weave wide( varieties of equipment options together.  I Even though HP lost ground in the second quarter of this year to IBM, Sun-K and Dell in the sales of server computers to corporations, Fiorina remained I unfazed. She attributed the drop in share to customers examining HP's new G product road maps, fresh from the consolidation of HP and Compaq serverG lines. *(1)u  K In its summary report for the second quarter of this year, Gartner anointedtL the newly united HP as the world leader in servers, with 30.5 percent of theK market. But sales actually fell 9.2 percent from the year-ago quarter, wheniK it was two separate companies. IBM also lost ground, with its share fallingm" from 15.0 percent to 14.5 percent.  K By contrast, Sun sales in an overall flat market gained 10.8 percent to 6.5eL percent of the market. Dell's unit sales gained 12.9 percent, giving it 18.08 percent of the 1 million servers shipped in the quarter.  G In the long run, according to Fiorina, Sun could not keep up with HP byiA focusing on fast machinery, using a proprietary operating system,pJ Solaris.*(2)  She said Sun's chief executive, Scott McNealy, was running aL "stand-alone hot box company," as was EMC, the storage server market leader.I Customers are no longer won over by speed of servers or desktop machines,c	 she said.   : "In our judgment, those days are gone forever,'' she said.  I McNealy, taking the stage 45 minutes after Fiorina's presentation, seemedaI unfazed, noting that Sun has gained more share from HP over the last nineaK months than during any time in its history. He noted that Sun now dominateseK the market for servers that crunch information 64 bits at a time. HP's playaG in that arena, based on a joint effort with Intel on a 64-bit processorgL series called Itanium, in development for roughly seven years, has been slowK to get off the ground. Intel's chief executive, Craig Barrett, who followedlK Fiorina on stage, predicted that sales of Itanium products would pick up inn+ the second half of this year and into next.   K After citing enterprise computing as the motive behind the merger of HP andoH Compaq *(3), Fiorina also decried her main rival in providing consultingE services in that field. IBM, whose Global Services unit is the marketf2 leader, "has a very different approach," she said.  G IBM has its own way of doing business, she said, focused on proprietaryaK services and hardware *(4). Even though HP, too, sells its own systems, thetE company offers a wider variety of choices, including Unix and Windowsn systems.  G She also said that IBM's acquisition of PricewaterhouseCoopers will notqK help. HP had considered buying PwC before IBM. "We walked away from the PwCrI opportunity not once but five times. And wound up buying Compaq instead."u  I Fiorina, apparently, doesn't seem to worry about antagonizing her rivals.)/ "We are going to go on the offense,'' she said.b   ------    E (1) "She attributed the drop in share to customers examining HP's newaG product road maps, fresh from the consolidation of HP and Compaq server  lines."tA        In the enterprise space she seems to hold near and dear, a J significant part of the drop is attributable to factors all in c.o.v. know
 all too well.t  L (2) "In the long run, according to Fiorina, Sun could not keep up with HP byK focusing on fast machinery, using a proprietary operating system, Solaris."yK       The corollary is that HP's own proprietary operating systems, OpenVMSnK and PH-UX (not counting Tru64), are in jeapordy as well from something too.oG But what is that something? Linux, or that famous non-proprietary, openr  operating system from Microsoft?  L (3) "After citing enterprise computing as the motive behind the merger of HP and Compaq ..." K       I'm pretty sure that OpenVMS customers sure don't feel that way afteruB Stallard's comments, and the lack of marketing effort for the most) enterprise ready operating system HP has.s  L (4) "IBM has its own way of doing business, she said, focused on proprietary services and hardware."eL       HP's hardware isn't proprietary either? What would you call Superdome,K Tandem, or Alphaservers? In the end, even if HP uses a low-volume chip liketI IA-64, it's just as proprietary as any other chip from any other fab. IBMeI sells Power to others, Intel sells chips to others, Sun's fab-of-the-weekrK sells chips to others. Think a HP IA-64 server with their own logic chipsetr isn't proprietary?  4       And just what are proprietary services anyway?   ------------------------------  % Date: Thu, 24 Oct 2002 20:36:00 -0400l* From: "Bill Todd" <billtodd@metrocast.net>1 Subject: Re: Fiorina Gets Feisty at Gartner Event 2 Message-ID: <a-udnaJWno18DCWgXTWcpg@metrocast.net>  . "John Smith" <a@nonymous.com> wrote in messageB news:RZZt9.62510$Q3S.43404@news01.bloor.is.net.cable.rogers.com...? > Somewhat stale dated now, but a few comments are interesting.-  L Indeed they are.  This is the third recent interview in which Carly has comeK off sounding more and more like a Luddite:  "Technology bad (at least if HPiG has to own it), service good!"  As I noted elsewhere, kind of a strange K attitude for the CEO of a company with 'invent' prominently featured in itshJ logo, but I guess Carly's interpretation is that it means *organizational*F (re-)invention (as in "The *New* HP [Way]") rather than something more product-related.   >l >a: > http://www.baselinemag.com/article2/0,3959,605377,00.asp   ...h  J > "The fact that Dell makes an announcement doesn't send us shaking in ourL > boots," she said at the Gartner Symposium and ITxpo here. "Dell has become asH > channel. They're not a technology company. They're not an R&D company.  ? And HP is?  Only because she hasn't yet finished redefining it.l   ...s   > HP's Corporate Focus >cL > She was similarly disparaging about other rivals, who she said lacked HP'sJ > focus on the needs of corporate customers-everything from the desktop to thekI > server, and from the network to the consulting services needed to weavec wide* > varieties of equipment options together.  J Watch out, Office Max:  she's got your business in her sights next (unless. she gets into 'temp' service personnel first).  D For a clearly intelligent woman, her understanding of business seemsA remarkably simplistic.  "Horizontal integration - GOOD.  VerticalyL integration - BAD" about seems to sum it up, whereas - perhaps especially inG technology-intensive industries - it's the right *balance* between bothc4 types (and the details of each) that spells success.   ...   I > In the long run, according to Fiorina, Sun could not keep up with HP byhC > focusing on fast machinery, using a proprietary operating system,oL > Solaris.*(2)  She said Sun's chief executive, Scott McNealy, was running aF > "stand-alone hot box company," as was EMC, the storage server market leader.eK > Customers are no longer won over by speed of servers or desktop machines,. > she said.n >n< > "In our judgment, those days are gone forever,'' she said.  I In her judgement, HP was going to blossom rather than flounder during herbF oh-so-astute stewardship.  In her judgement, merging with Compaq was aK marriage made in heaven and the merger benefits were already supposed to bemJ pouring in.  In her judgement, the increasingly pessimistic adjustments inL internal projections that occurred during the pre-merger months were invalidK and definitely not something to let the rest of her Board of Directors knowa# about (let alone the shareholders).1  I Perhaps it's her judgement that needs reinventing rather than the rest oft HPaq.n   ...u  K > Fiorina, apparently, doesn't seem to worry about antagonizing her rivals.l   Or her customers.y  1 > "We are going to go on the offense,'' she said.   % When exactly would that be?  And how?u   - bill   ------------------------------  % Date: Thu, 24 Oct 2002 20:07:44 -0500n1 From: "David J. Dachtera" <djesys.nospam@fsi.net>  Subject: Re: Hobbyist Licensem' Message-ID: <3DB89960.527C0DAF@fsi.net>    Martin Vorlaender wrote: > 5 > Mark Buda (buda@tabasco.zko.dec.no.spam.com) wrote: 6 > > "Larry Kilgallen" <Kilgallen@SpamCop.net> wrote... > > > Doc.Cypher writes:K > > > > The extra ones I use are Multinet and PerfectDisk. IIRC PerfectDisk.+ > > > > doesn't use standard VMS licensing.d > > > > J > > > > However, I do think the generic hobbyist license is a good idea asK > > > > companies which use it no longer need to muck about issuing keys ton > > > > Hobbyists. > > >lI > > > Some vendors may feel they need to know who is using their Hobbyist- > > > license.@ > > > Others are more interested in reducing their own effort to > > > participate.2 > > > Some probably don't understand the question. > >A  > > What is "THE" question?  :-) > H > Well, y'know, the question about life, the universe, and everything... > to which the answer is 42g  % It is? Damn this dyslexia - I got 24!v   -- u David J. Dachterag dba DJE Systemse http://www.djesys.com/  ( Unofficial Affordable OpenVMS Home Page: http://www.djesys.com/vms/soho/l   ------------------------------  % Date: Thu, 24 Oct 2002 16:13:31 -0400b From: norm.raphael@metso.comY Subject: How can I determine the device of the system disk for another node in a cluster?o? Message-ID: <OF641887E1.62603D0E-ON85256C5C.006EB209@metso.com>o  5 Is there any easier way than capturing the output of:s $ RUN SYS$SYSTEM:SYSMANV SYSMAN> SET ENVIRONMENT/CLUSTER6+ %SYSMAN-I-ENV, current command environment:d!      Clusterwide on local cluster 9      Username MASTER       will be used on nonlocal nodes.  % SYSMAN> DO SHOW LOGICAL SYS$SYSDEVICE.1 %SYSMAN-I-OUTPUT, command execution on node NODEAr1    "SYS$SYSDEVICE" = "DSA100:" (LNM$SYSTEM_TABLE)n1 %SYSMAN-I-OUTPUT, command execution on node NODEBn0    "SYS$SYSDEVICE" = "DSA11:" (LNM$SYSTEM_TABLE)1 %SYSMAN-I-OUTPUT, command execution on node NODEC 1    "SYS$SYSDEVICE" = "DSA150:" (LNM$SYSTEM_TABLE)s SYSMAN> EXIT $I  6 so that I know for each node the system disk device on   each of the other nodes?   ------------------------------  % Date: Thu, 24 Oct 2002 15:12:35 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>e@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!, Message-ID: <3DB84621.1DC2A6D2@videotron.ca>   Jerry Leslie wrote: J >         The United States is not trading with China in the normal sense.K >    The Chinese have access to U.S. markets for products made with ChinesenI >    labor. In exchange U.S. companies have access to Chinese markets andbI >    U.S. markets with products made by Chinese labor. In this "exchange" 3 >    where lies the advantage for the U.S. economy?c  L A good example: At one point, some US politician was complaining about Kraft# not exporting any cheese to Japan.    J While t was true that Kraft-USA wasn't exporting to Japan, Kraft-AustraliaK was.  When you have a multinational corporation it will use facilities that J are most cost effective to do the job. And if due to transportand currencyL etc, it is more profitable for Kraft to take australian-made cheese and shipM it to Japan compared to taking USA made cheese and shipping it to Japan, then 	 why not ?p  D In the end, the profits made by Kraft Australia will return to KraftN headquarters in the USA, and that money will benefit the shareholders of KraftN in the form of dividends, and benefit the USA government because Kraft will beJ paying some corporate taxes on those profits that are returned to the USA.  J Why do you think that the USA never complains about the HUGE trade deficitM with Canada ? (orders magnitudes bigger than the US trade deficit with Japan)oL That is because even though the USA may import much more from canada that itK exports, it usually imports goods made in Canada by a USA owned corporationwL and the profits go back to the USA, so in total, the USA is a winner because9 more profits go from Canada-USA than goods go Canada-USA.    ------------------------------  % Date: Thu, 24 Oct 2002 14:56:30 -0700q$ From: Shane Smith <ssmith@icius.com>@ Subject: RE: HP tells VMS users "Relax, you are sitting pretty"!0 Message-ID: <01C27B6D.9A134100@sulfer.icius.com>  G Not sending American cheese to China could be considered a humanitarianc act...   -----Original Message-----4 From: JF Mezei [mailto:jfmezei.spamnot@videotron.ca]) Sent: Thursday, October 24, 2002 12:13 PMd To: Info-VAX@Mvb.Saic.Com @ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!     Jerry Leslie wrote:-J >         The United States is not trading with China in the normal sense.K >    The Chinese have access to U.S. markets for products made with ChineseBI >    labor. In exchange U.S. companies have access to Chinese markets and,I >    U.S. markets with products made by Chinese labor. In this "exchange"i3 >    where lies the advantage for the U.S. economy?u  F A good example: At one point, some US politician was complaining about KraftS# not exporting any cheese to Japan. T  : While t was true that Kraft-USA wasn't exporting to Japan, Kraft-Australia F was.  When you have a multinational corporation it will use facilities thatA are most cost effective to do the job. And if due to transportandq currencyG etc, it is more profitable for Kraft to take australian-made cheese and  shipH it to Japan compared to taking USA made cheese and shipping it to Japan, then	 why not ?   D In the end, the profits made by Kraft Australia will return to KraftH headquarters in the USA, and that money will benefit the shareholders of Kraft_F in the form of dividends, and benefit the USA government because Kraft will besE paying some corporate taxes on those profits that are returned to the. USA.  B Why do you think that the USA never complains about the HUGE trade deficituF with Canada ? (orders magnitudes bigger than the US trade deficit with Japan)D That is because even though the USA may import much more from canada that it ? exports, it usually imports goods made in Canada by a USA owned: corporation D and the profits go back to the USA, so in total, the USA is a winner because-9 more profits go from Canada-USA than goods go Canada-USA.    ------------------------------  % Date: Thu, 24 Oct 2002 21:46:38 -0400d( From: David Froble <davef@tsoft-inc.com>@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!, Message-ID: <3DB8A27E.7070703@tsoft-inc.com>   Alan Greig wrote:c  4 > On Wed, 23 Oct 2002 09:40:57 -0400, Atlant Schmidt& > <atlantnospam@mindspring.com> wrote: >  > - >>The corporations have already exported many - >>(most?) of the high-paying blue-color jobs.e. >>Now, they're going to do the same thing with# >>the high-paying white-color jobs.a >> > H > While not a great defender of Capitalism, I have to point out that theH > logic here is supposed to be that jobs are slowly transferred to lowerF > GDP economies, paying high salaries relative to their own economies. > C > Over time the injection of cash brings these countries up the GDP-G > table and wages slowly increase towards those of the dominant economya' > from which the jobs were transferred.H > G > This is self balancing because the poorer countries are servicing the D > wealthier countries. Take too much out of the richer economies and > they have nowhere to sell to.- > H > What would you rather have? A few million more dead of malnutrition inF > the third world or more fat Americans and Europeans. Ok, I'm playing= > Devil's Advocate here. But that;'s supposed to be the idea.t  N Well, there is the need to have all people have some chance at a decent life. L Yes, when something can be packaged as a commodity, then it can be produced 	 anywhere.u  P But #1.  I seriously doubt that someone in India can do much for a project that O takes local hands on work and experience.  The companies that have these needs rM and still get stupid and try to save a few bucks won't be around to make the u same mistake too many times.  
 Butt #2.  :-)r  N The fattest part of those Americans and Europeans is indeed their butt, which Q they want to sit on and continue to reap the same rewards that they did 20 years hP ago for doing the same job that they did 20 years ago.  They don't want to push  the frontiers.  M When did the first man walk on the moon?  Nineteen f**king sixty nine!!!  33 $O years have passed.  Can we get a man to the moon today?  What about beyond the  P moon?  There is no resolve to push at frontiers.  Just the wish to sit on those O fat butts and continue to get pats on the back for old accomplishments.  Well,  P the pat on the back is turning into a shove out of the way.  When the people of J the USA again want to lead, then there will be followers.  Not until then.  Q Note that a small fraction of the US defense budget would be more than enough to )P get innovation and adventure going again.  It's probably not even possible (Ok, P I'm exaggerating) to measure the fraction of the entitlements budget that would  get things going again.   Q And I really don't think that the current powers that be can even imagine how to UQ use the technology that they have.  It doesn't seem that the spy satellites were EN used to record all activity in the Washing DC area.  After the next shooting, 2 you just watch the film and follow the moron home.   Dave   ------------------------------  # Date: Fri, 25 Oct 2002 04:26:43 GMT-1 From: LESLIE@JRLVAX.HOUSTON.RR.COM (Jerry Leslie)U@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!< Message-ID: <7K3u9.193779$8o3.5677124@twister.austin.rr.com>  ) David Froble (davef@tsoft-inc.com) wrote:  : L : When did the first man walk on the moon?  Nineteen f**king sixty nine!!!  I : 33 years have passed.  Can we get a man to the moon today?  What about  M : beyond the moon?  There is no resolve to push at frontiers.  Just the wish  I : to sit on those fat butts and continue to get pats on the back for old iJ : accomplishments.  Well, the pat on the back is turning into a shove out I : of the way.  When the people of the USA again want to lead, then there i% : will be followers.  Not until then.m  H One of NASA's biggest shortcomings is its failure to market the value of/ spinoffs of aerospace technolgy to the economy.s  A I heard a statement once that for every billion dollars spent on  D the space program, the economy's GNP went up a certain percent from > the spinoffs. I'll see if I can find the statement via google.  G A marketing failure that cost most of the enterprise - sounds familiar.>  I Another statement was that if NASA's funding had continued at a constant c8 and sustainable level, we'd already have people on Mars.  F NASA has been killing projects recently, such as the X-38 project, andH more recently (and probably correctly) the Shuttle Launch Control System upgrade:  ;    http://www.space.com/missionlaunches/fl_clcs_020918.html'4    NASA Kills 'Wounded' Launch System Upgrade at KSC    2 --Jerry Leslie   (my opinions are strictly my own)9   Note: leslie@jrlvax.houston.rr.com is invalid for emailb   ------------------------------  % Date: Fri, 25 Oct 2002 00:48:59 -0400 * From: "Bill Todd" <billtodd@metrocast.net>@ Subject: Re: HP tells VMS users "Relax, you are sitting pretty"!2 Message-ID: <z4idndtWyrK2UCWgXTWcpQ@metrocast.net>  > "Jerry Leslie" <LESLIE@JRLVAX.HOUSTON.RR.COM> wrote in message6 news:7K3u9.193779$8o3.5677124@twister.austin.rr.com...+ > David Froble (davef@tsoft-inc.com) wrote:t > :eL > : When did the first man walk on the moon?  Nineteen f**king sixty nine!!!J > : 33 years have passed.  Can we get a man to the moon today?  What aboutI > : beyond the moon?  There is no resolve to push at frontiers.  Just thee wishJ > : to sit on those fat butts and continue to get pats on the back for oldK > : accomplishments.  Well, the pat on the back is turning into a shove out_J > : of the way.  When the people of the USA again want to lead, then there' > : will be followers.  Not until then.a > J > One of NASA's biggest shortcomings is its failure to market the value of1 > spinoffs of aerospace technolgy to the economy.   H Possibly.  But if we had the right attitude, such marketing would not beK necessary - while with our current attitude, it's very questionable whether  it would be effective.  L We didn't venture into space so that we'd have better frying pans:  that was just a side benefit.   ...h  J > Another statement was that if NASA's funding had continued at a constant: > and sustainable level, we'd already have people on Mars.  F To be specific, I think the meaning was that we'd have (or at least beJ capable of having) a sustained presence there, not just have been and leftF already.  The distance isn't that much more of a problem than having aG sustained presence in low Earth orbit, once you've established a supply = schedule that can tolerate several-years'-long transit times.a  L One of the few possible justifications for having the immense wealth that weI are so lucky to have is the willingness to spend it on ventures that willeH expand the horizons and potential of humanity rather than squander it onJ personal luxuries.  That's why the killing of Alpha was so disgusting:  anF attempt (and a monumentally stupid one at that) to maximize short-term. profit at the expense of long-term excellence.   - bill   ------------------------------   Date: 25 Oct 2002 03:32:05 GMT& From: peter@abbnm.com (Peter da Silva)J Subject: Re: HP-ETS 2002 Presentations are on-line - Pay-per-view ONLY :-(- Message-ID: <apadvl$t7v@web.eng.baileynm.com>n  / In article <urea0cnsrmeieb@corp.supernews.com>,E( Jeff Killeen <KilleenJ@toast.net> wrote:F > To the CD issue - we sold so few of them that we lost money on them.  K I have a little experience with small volume CDs. Even a few years ago, youNL could get a run of 1000 CDs cut from a master for under $1000, with printingL and sleeves. If you charge $20/CD you'll make back the production costs with 50 CDs.   I If ENCOMPASS is too big to gear down to that small a volume, contract the-L whole thing out. Let someone else take the risk. But if ENCOMPASS is too big) to do that, then it's just plain too big.S  $ > There is a magic point at which isE > less expensive to ship a CD to everyone versus the overhead cost ofR# > maintaining a order entry system.Y  6 Don't maintain one. Let Amazon or Yahoo do it for you.   -- LO I've seen things you people can't imagine. Chimneysweeps on fire over the roofs O of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All L these things will be lost in time, like chalk-paintings in the rain.   `-_-'K Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`D   ------------------------------  % Date: Thu, 24 Oct 2002 20:31:12 -0400, From: "ICUSC" <sales@hpaq.net> Subject: Incoming !!!! DS10'sB/ Message-ID: <urh427jlgge643@news.supernews.com>   " More DS10's are in their way in...     DS10 466Mhz Systems  DS10 600Mhz Systems     Custom configurations no problem  & Save up to 70% on Manufacturer pricing  G Yes... We're trying to save you money and keep you using VMS; after all  someone needs to !!!     -- Island Computers US Corp.t 2700 Gregory St, Ste 180 Savannah GA 31404, USA Tel: (00) 1 912 447 662S Fax: (00) 1 912 201 0096 sales@hpaq.net www.islandco.com   ------------------------------  # Date: Thu, 24 Oct 2002 20:32:24 GMTT# From: "John Smith" <a@nonymous.com>r6 Subject: It'll be interesting to see what happens hereH Message-ID: <sNYt9.61917$Q3S.58334@news01.bloor.is.net.cable.rogers.com>  2 http://www.eweek.com/article2/0,3959,644757,00.asp   http://www.openhack.com/  2 http://www.eweek.com/article2/0,3959,643205,00.asp   ------------------------------    Date: 24 Oct 2002 14:05:23 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones) $ Subject: Re: Monitor System Question= Message-ID: <8a646952.0210241305.33c1ccc5@posting.google.com>C  ] "Alder" <MUNDDGNTDYTV@spammotel.com> wrote in message news:<3db580d8$1@obsidian.gov.bc.ca>... M > I think it's simply the result of trying to keep the data entirely "within":G > the outline of the scale.  A 100% reading starts one cell in from the-H > leftmost graph boundary, and ends one cell in from the rightmost graphK > boundary.  The designers allow the bar to cross the percentile lines, butXE > not those of 0 and 100 (see below).  That's what it looks like on aOM > character-cell terminal anyway.  I can't get this stupid Outlook Express to H > show fixed-width fonts, so hopefully it hasn't been mangled too badly. > K > As for "resolution", it would be more useful to have the actual CPU usageLI > displayed numerically "on top of" or at the end of the individual bars, & > perhaps with a decimal place or two. >  >  > 5 >                             OpenVMS Monitor Utility 5 >                              TOP CPU TIME PROCESSES,0 >                                  on node HOBBY5 >                             22-OCT-2002 09:32:35.11s > G >                                      0         25        50        75  > 100tD >                                      + - - - - + - - - - + - - - - > + - - - - -+& >  00000117  SETI@home 54%         100* > ########################################F >                                      |         |         |         | > |t >  00000154  SYSTEMyF >                                      |         |         |         | > |y > F >                                      |         |         |         | > |c > F >                                      |         |         |         | > |m > F >                                      |         |         |         | > |  > F >                                      |         |         |         | > |D > F >                                      |         |         |         | > |q > D >                                      + - - - - + - - - - + - - - - > + - - - - -+ >  >  > Aldera >  > H > "Phillip Helbig" <HELBPHI@sysdev.deutsche-boerse.com> wrote in message7 > news:01KNYOI3V5ZM9QXRDQ@sysdev.deutsche-boerse.com...kF > > > One thing that has always bothered me is the MONITOR PROC/TOPCPU >  display.c > > >mL > > > The first three quartiles (0%-25%, 25%-50%, 50%-75%) are ten character= > > > cells wide. But the fourth is eleven character cells!!!- > > > ' > > > What kind of scale is that??  ;-)l > >. > > Interesting. > > = > > Definitely, someone In The Know needs to comment on this.g > >aL > > Perhaps the idea is to give a bit more resolution at the high end, sinceI > > if two processes are using about the same amount of CPU, you are more J > > interested in which is using more at the high end than at the low end.H > > But in that case, the scale should be even more non-linear, like theI > > fuel gauge on a car (but in reverse, with more resolution at the higha9 > > end, not at the low end which is important for fuel).t  A Outlook Express uses Arial as its default font. You must copy the D screen or file to MS Word and use a Courier New or r-ansi font whichE use by MS Outlook Express. You can copy-paste or attached the MS Wordf doc. to the email.   ------------------------------    Date: 25 Oct 2002 03:50:42 +0800, From: Paul Repacholi <prep@prep.synonet.com>$ Subject: Re: Monitor System Question- Message-ID: <87znt31ur1.fsf@prep.synonet.com>s  * per@nospam.mimer.se (Per Schrder) writes:  ? > Phillip Helbig <HELBPHI@sysdev.deutsche-boerse.com> wrote in t2 > <01KNXIP5KJCQ9QWKOO@sysdev.deutsche-boerse.com>:   A > >The memory display has units;; the others don't.  The units inaA > >these cases are all pure numbers, except for CPU, which is perID > >cent.  Note that the maximum scales with the number of CPUs.  :-)  dB > One thing that has always bothered me is the MONITOR PROC/TOPCPU
 > display.  e> > The first three quartiles (0%-25%, 25%-50%, 50%-75%) are tenC > character cells wide. But the fourth is eleven character cells!!!S  l# > What kind of scale is that??  ;-)    A fencepost one...   -- a< Paul Repacholi                               1 Crescent Rd.,7 +61 (08) 9257-1001                           Kalamunda. @                                              West Australia 6076. Raw, Cooked or Well-done, it's all half baked.F EPIC, The Architecture of the future, always has been, always will be.   ------------------------------    Date: 24 Oct 2002 12:17:55 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones) = Subject: Re: Need Help really Quick Please: 10/23/02 Noon CSTD= Message-ID: <8a646952.0210241117.45b043ef@posting.google.com>i  P Jim Agnew <jpagnew@vcu.edu> wrote in message news:<3DB7F098.28B102F8@vcu.edu>...* > hhmm... I like that "dead sergeant"..... >  > Shane Smith wrote: > > J > > I believe it's "SET CONSOLE SERIAL" issued at the dead sergeant prompt
 > > ">>>". > > 	 > > Shanel > >  > > -----Original Message-----2 > > From: Shawn [mailto:shawnm@cotepe.carenet.org]. > > Sent: Wednesday, October 23, 2002 10:14 AM > > To: Info-VAX@Mvb.Saic.Come= > > Subject: Need Help really Quick Please: 10/23/02 Noon CST' > > H > > I know there is a command to re-direct the output of an Alpha ServerH > > running OpenVMS 7.2-1 to a VT Terminal on the Com Ports, rather than. > > the monitor attached to the Graphics Card. > >   H > > Can someone please tell me this command and can it be done while theA > > system is running or do I need to do this at the Boot Prompt.n > >  > > Thanks,) > > 	 > > Shawnr > >  > > sfm1115@bjc.orgv  E If you want to change the characterics of the graphic port, you enterhE the command, "set console serial", at the boot prompt (>>>) which canuF only be access when the system is down. However, you can turn on OPA0:> port and attached a VT terminal to it and turn on the port forF broadcast using the set term command. This is the old operator console6 that has been turned off in lieu of the graphics port.   ------------------------------  % Date: Thu, 24 Oct 2002 22:55:32 +0200l) From: "Steven Thompson" <steven@omga.biz> ! Subject: Non-Paged Pool ExpansionFB Message-ID: <ap9mo9$kds$1@nsnmrro2-gest.nuria.telefonica-data.net>   Hi Gurus  H We've just upgraded a 2 node Alpha cluster from 7.2-1 to 7.3, patched it3 with all but 1 new patch that has just been issued.n* It's got the TCPIP stack version 5.1 Eco4.  G I suspect the Packet Type FRK as being the problem, ie. growing and notp releasing memory.i= Seen by analyzing the system with the SHOW POOL/NONPAGED/SUMMn  0 What affects this packet, (what is this packet)?   Thanks Steven   ------------------------------  % Date: Thu, 24 Oct 2002 19:51:28 -0400t, From: Mark Jilson <jilly@clarityconnect.com>% Subject: Re: Non-Paged Pool Expansiono3 Message-ID: <wt%t9.143$g4.64820@newsfeed.slurp.net>t   SDA> SET OUT {file.ext}y SDA> SHOW POOL/NONPAG/TYPE=FRK	 SDA> EXITi  M Now look at the file and see if any of the data in it looks repetitive.  You sL could also pull out all of the address' of the FRK blocks and put them into 	 a file as  FORMAT {address}/TYPE=FKB  FORMAT {address}/TYPE=FKB etc.   Then in SDA 	 READ/EXECh READ SYS$LOADABLE_IMAGES:SYSDEFo SET OUT FORMATTED.LOG&. @{file.ext} ! file with allthe FORMAT commands EXIT  L Then Sort the output on the FKB$L_FPC and see if there are a number of fork $ blocks pointing to the same address.       Steven Thompson wrote:  
 > Hi Gurus > J > We've just upgraded a 2 node Alpha cluster from 7.2-1 to 7.3, patched it5 > with all but 1 new patch that has just been issued.i, > It's got the TCPIP stack version 5.1 Eco4. > I > I suspect the Packet Type FRK as being the problem, ie. growing and not  > releasing memory.a? > Seen by analyzing the system with the SHOW POOL/NONPAGED/SUMMt > 2 > What affects this packet, (what is this packet)? >  > Thanks > Steve    -- nC Jilly - Working from Home in the Chemung River Valley - Waverly, NY H       - jilly@clarityconnect.com                      - Brett Bodine fanH       - Mark.Jilson@hp.com                            - since 1975 or soH       - http://www.jilly.baka.com           - http://www.brettbodine.com   ------------------------------  + Date: Thu, 24 Oct 2002 21:55:06 +0000 (UTC) , From: lewis@mazda.mitre.org (Keith A. Lewis)% Subject: Re: OpenVMS 7.3-1 and CSWINGe. Message-ID: <ap9q7q$kma$1@newslocal.mitre.org>   carlos.costa@datawest.ca (Carlos Costa) writes in article <7edc7395.0210220742.6bd0ce68@posting.google.com> dated 22 Oct 2002 08:42:57 -0700:s >Hello, everyone.  > B >I recently upgraded VMS to 7.3-1 (from VMS 7.3-0) on two separateE >clusters. I have also been using CSWING for years now on various VMSo >versions and I love it. >eE >However, in 7.3-1 CSWING no longer works (it does under 7.3-0). WheneB >it goes to search a new directory tree it goes off to never-never >land.  
 Confirmed.   		7.3		7.3-1 		---		-----! old image	works		never-never lando& relinked image	works		access violation  K Recompiling the source fails because my relatively recent C compiler whinesp about getenv() being an int...  + --Keith Lewis              klewis$mitre.orgs> The above may not (yet) represent the opinions of my employer.   ------------------------------  + Date: Thu, 24 Oct 2002 22:14:14 +0000 (UTC)t, From: lewis@mazda.mitre.org (Keith A. Lewis)% Subject: Re: OpenVMS 7.3-1 and CSWINGw. Message-ID: <ap9r7u$kma$2@newslocal.mitre.org>   carlos.costa@datawest.ca (Carlos Costa) writes in article <7edc7395.0210230744.2f4fdb88@posting.google.com> dated 23 Oct 2002 08:44:20 -0700:cF >Harry did wonder, though, if I was using the now filesystem, ODS5, asE >he was pretty certain that CSWING would not work with it. I am *not*nD >using ODS5, but his comment got me to thinking that perhaps that isC >where the problem in VMS lies. Whatever they put into VMS for ODS5.A >perhaps is not 100% binary compatible with 7.3-0. This has wider > >implications than just CSWING. What else isn't going to work?  K I am using ODS-5 on my 7.2-1, 7.3 and 7.3-1 cluster nodes.  The old systemsaE work, 7.3-1 doesn't.  As you said, it goes off into never-never land.'  G When I relink the old .OBJ files on 7.3-1 and then run it bombs with ant7 access violation regardless of the runtime VMS version.a  H I couldn't get the C source to recompile on the Compaq C 6.4.  It thinks# getenv() is an int for some reason.   E >Harry also said that VMS has speeded up the file internals that made E >reading the filesystem directory necessary in CSWING, and he thoughtg? >that it would be an idea for someone to re-write the directoryhC >searching code (in get_dirs() on SWING.C) to use the improved fileaE >lookup calls. Any willing VMS C programmers out there? It would be a $ >great service to the VMS community.  8 Send me an Elsa Gloria Synergy card and I'll do it!  :^)  + --Keith Lewis              klewis$mitre.orgH> The above may not (yet) represent the opinions of my employer.   ------------------------------  % Date: Fri, 25 Oct 2002 12:50:53 +1000s* From: "Dale King" <dalek@forpresident.com>% Subject: Re: OpenVMS 7.3-1 and CSWING ' Message-ID: <apabid$17q$1@lore.csc.com>o   Paul Repacholi wrote:a1 > carlos.costa@datawest.ca (Carlos Costa) writes:n >sD >> I recently upgraded VMS to 7.3-1 (from VMS 7.3-0) on two separateG >> clusters. I have also been using CSWING for years now on various VMSt >> versions and I love it. > G >> However, in 7.3-1 CSWING no longer works (it does under 7.3-0). WhenpD >> it goes to search a new directory tree it goes off to never-never >> land. >,E >> Has anyone else experienced this, and more importantly, is there a  >> fix out?g >n. > It has done this `forever' with ISO9660 CDs. >aG >> This problem is also disconcerting because 7.3-1 has, in HP/Compaq'sdA >> own words, "100% binary compatibility with previous versions".. > H > I think you missunderstand the statment, and some of the moderatly susB > things CSWING does. But I too am curious as to what has changed.  O One thing I've noticed between V7.3 and V7.3-1 that I can't find in the releaseu> notes is that the output of the DIRECTORY command has changed.   Where the command:8     $ DIR/COLUMNS=1/SIZE=ALL/WIDTH=FILENAME=40/DATE=MODIO used to look "nice" under the former, it now wraps under the latter.  I presume 2 this is to allow larger filesizes to be displayed.   I wonder if this is related?   ------------------------------  % Date: Fri, 25 Oct 2002 04:27:40 +0200e2 From: martin@radiogaga.harz.de (Martin Vorlaender)% Subject: Re: OpenVMS 7.3-1 and CSWINGh; Message-ID: <3db8ac1c.524144494f47414741@radiogaga.harz.de>d  - Keith A. Lewis (lewis@mazda.mitre.org) wrote: M > Recompiling the source fails because my relatively recent C compiler whinesy  > about getenv() being an int...   Easy: just add   #include <stdlib.h>.  5 at the top of the SWING.C. And while we're at it: addu   #include <string.h>   I as well to get rid of the %CC-I-INTRINSICINT messages about the intrinsic 5 strlen()'s return being converted from size_t to int.a  K But there's a *lot* more to clean up to get a clean compile with DEC C evenaG in /STANDARD=VAXC mode, e.g. many routines returning char* are declared K with an implicit int return type. And no prototypes. But I might be tempted  to give it a try...-   cu,-   Martin -- -G  Your mouse has moved.     | Martin Vorlaender  |  VMS & WNT programmerd4  Windows must be restarted | work: mv@pdv-systeme.deH  for the change to take    |    http://www.pdv-systeme.de/users/martinv/;  effect. Reboot now? [OK]  | home: martin@radiogaga.harz.deQ   ------------------------------   Date: 25 Oct 2002 04:50:55 GMT- From: djweath@attglobal.net (Dave Weatherall) % Subject: Re: OpenVMS 7.3-1 and CSWINGe5 Message-ID: <DTiotGxQ0bj6-pn2-SWTEopChNI5z@localhost>r  A On Thu, 24 Oct 2002 06:05:20 UTC, Arne =?ISO-8859-1?Q?Vajh=F8j?= o <arne@vajhoej.dk> wrote:   > Carlos Costa wrote:o > l > > "DAniel Allen" <dallen@nist.gov> wrote in message news:<BCEGLBGJDODLELBJIADKEEOKCBAA.dallen@nist.gov>...< > >>The status checking in this code snippet is not correct: > >>K > >>	1)  The QIO status should check not for SS$_NORMAL but any error value  > >> > > I > > I am missing something here. The code doesn't check for SS$_NORMAL itcD > > checks for not-equal-to SS$_NORMAL. Wouldn't that be the same asF > > checking for any error value? In other words, if the statis is NOTF > > NORMAL then an error occurs. If there was an error then SS$_NORMAL > > would not return, would it?e >  > B > I think one of the points is that there are other good status'es > than SS$_NORMAL.  F True, one would normally test for 'odd'ness. Odd being good, even bad.E The point being that bit 0 signals success(1) or fail(0), other bits  D in the status return qualify the 'goodness' or describe the failure.  C In DEC-Fortran you can test the INTEGER return as though it were a c) LOGICAL ,i.e. TRUE (good) or FALSE (bad).t  E In C I've used a macro vms_ok(status) which tests the state of bit 0.l  0 What I do in PASCAL .... ISTR    if odd(status).  % In macro, use BLBS/BLBC instructions.n   That's my languages exhausted.   -- r Cheers - Dave.   ------------------------------    Date: 24 Oct 2002 13:56:36 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones)c Subject: Re: page or pagelet< Message-ID: <8a646952.0210241256.296d1d1@posting.google.com>  X VAXman-  @SendSpamHere.ORG wrote in message news:<00A15D65.98CDBFDF@SendSpamHere.ORG>...Z > In article <ap41pv$t0g1@rain.i-cable.com>, "Kenneth" <yeung_kenneth@hotmail.com> writes:L > >In VMS 7.2-1, Alpha 8400. Does any on know when I do a mon page, the freeM > >page list and the modified page list shown is in the unit of page (8KB) ork > >pagelet (512K)? > >d > >n >  > Hmm.  Let's see. > # > When I do it on my system, I see:  >  > $ MONITOR PAGE5 >                             OpenVMS Monitor Utilityh8 >                            PAGE MANAGEMENT STATISTIICS >   :sJ >                                     CUR        AVE        MIN        MAX >   :oM >     Free List Size               237273.00  237273.59  237273.00  237276.00nM >     Modified List Size              350.00     350.00     350.00     350.00i >  >  > Then, if I issue:h >  > $ SHOW MEMORY/PHYSICALB >               System Memory Resources on 22-OCT-2002 15:05:49.40 > N > Physical Memory Usage (pages):     Total        Free      In Use    ModifiedN >   Main Memory (2048.00Mb)         262144      237442       24353         349 > P > Of the physical pages in use, 5523 pages are permanently allocated to OpenVMS. >  >  > 3 > It would look to me to be a page of 8K.  (2GB/8K)  >  > ' > FYI, a pagelet is 512 bytes not 512K.D  B A page on an Alpha VMS machine is 8192 bytes per page and on a VAXB machine it is 512 bytes per page. This is due to RISC vs CISC chipB technology where the image sizes grew to be 5-6 time larger on theE RISC chip. A pagelet on an Alpha VMS processor is 512 bytes. It takesrD 16 pagelets (16*512=8192) to equal to a page on a Alpha VMS machine.   ------------------------------  # Date: Thu, 24 Oct 2002 21:21:01 GMT " From:   VAXman-  @SendSpamHere.ORG Subject: Re: page or pagelet0 Message-ID: <00A15F0A.315610CA@SendSpamHere.ORG>  v In article <8a646952.0210241256.296d1d1@posting.google.com>, jones.computer.srv@worldnet.att.net (Daryl Jones) writes:Y >VAXman-  @SendSpamHere.ORG wrote in message news:<00A15D65.98CDBFDF@SendSpamHere.ORG>...e[ >> In article <ap41pv$t0g1@rain.i-cable.com>, "Kenneth" <yeung_kenneth@hotmail.com> writes:hM >> >In VMS 7.2-1, Alpha 8400. Does any on know when I do a mon page, the freeiN >> >page list and the modified page list shown is in the unit of page (8KB) or >> >pagelet (512K)?- >> > >> > >> t >> Hmm.  Let's see.e >> p$ >> When I do it on my system, I see: >>   >> $ MONITOR PAGE 6 >>                             OpenVMS Monitor Utility9 >>                            PAGE MANAGEMENT STATISTIICS  >>   :K >>                                     CUR        AVE        MIN        MAXr >>   :N >>     Free List Size               237273.00  237273.59  237273.00  237276.00N >>     Modified List Size              350.00     350.00     350.00     350.00 >> S >>   >> Then, if I issue: >> S >> $ SHOW MEMORY/PHYSICALtC >>               System Memory Resources on 22-OCT-2002 15:05:49.40n >> AO >> Physical Memory Usage (pages):     Total        Free      In Use    ModifiedCO >>   Main Memory (2048.00Mb)         262144      237442       24353         3491 >> (Q >> Of the physical pages in use, 5523 pages are permanently allocated to OpenVMS.D >> " >> S >>  4 >> It would look to me to be a page of 8K.  (2GB/8K) >> n >> O( >> FYI, a pagelet is 512 bytes not 512K. >TC >A page on an Alpha VMS machine is 8192 bytes per page and on a VAXeC >machine it is 512 bytes per page. This is due to RISC vs CISC chip-C >technology where the image sizes grew to be 5-6 time larger on thef >RISC chip.@    No it's not!  Who told you this?    < > A pagelet on an Alpha VMS processor is 512 bytes. It takesE >16 pagelets (16*512=8192) to equal to a page on a Alpha VMS machine.S  7 Unless your particular Alpha has 16K, 32K or 64K pages.T --O VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COMn            .5   "Well my son, life is like a beanstalk, isn't it?" S   ------------------------------  % Date: Thu, 24 Oct 2002 15:39:40 -0700e# From: "Tom Linden" <tom@kednos.com>o Subject: RE: page or pagelet9 Message-ID: <CIEJLCMNHNNDLLOOGNJIKEIAFPAA.tom@kednos.com>o   >-----Original Message-----.C >From: VAXman- @SendSpamHere.ORG [mailto:VAXman- @SendSpamHere.ORG]t) >Sent: Thursday, October 24, 2002 2:21 PMo >To: Info-VAX@Mvb.Saic.Com >Subject: Re: page or pagelet  >e >v= >In article <8a646952.0210241256.296d1d1@posting.google.com>,s: >jones.computer.srv@worldnet.att.net (Daryl Jones) writes:- >>VAXman-  @SendSpamHere.ORG wrote in message - >news:<00A15D65.98CDBFDF@SendSpamHere.ORG>... 8 >>> In article <ap41pv$t0g1@rain.i-cable.com>, "Kenneth"$ ><yeung_kenneth@hotmail.com> writes:? >>> >In VMS 7.2-1, Alpha 8400. Does any on know when I do a mone >page, the freesA >>> >page list and the modified page list shown is in the unit of  >page (8KB) or >>> >pagelet (512K)? >>> >u >>> >  >>>S >>> Hmm.  Let's see. >>> % >>> When I do it on my system, I see:t >>>d >>> $ MONITOR PAGE7 >>>                             OpenVMS Monitor Utilityh: >>>                            PAGE MANAGEMENT STATISTIICS >>>   :oL >>>                                     CUR        AVE        MIN        MAX >>>   :r9 >>>     Free List Size               237273.00  237273.59e >237273.00  237276.00e9 >>>     Modified List Size              350.00     350.00  >350.00     350.00 >>>- >>>- >>> Then, if I issue:  >>>1 >>> $ SHOW MEMORY/PHYSICALD >>>               System Memory Resources on 22-OCT-2002 15:05:49.40 >>>i@ >>> Physical Memory Usage (pages):     Total        Free      In >Use    Modified8 >>>   Main Memory (2048.00Mb)         262144      237442 >24353         349 >>>e< >>> Of the physical pages in use, 5523 pages are permanently >allocated to OpenVMS. >>>4 >>>2 >>>35 >>> It would look to me to be a page of 8K.  (2GB/8K)  >>>s >>>R) >>> FYI, a pagelet is 512 bytes not 512K.e >>D >>A page on an Alpha VMS machine is 8192 bytes per page and on a VAXD >>machine it is 512 bytes per page. This is due to RISC vs CISC chipD >>technology where the image sizes grew to be 5-6 time larger on the >>RISC chip. >a! >No it's not!  Who told you this?p  C I think it is probably close to true.  When you write compilers you G have occassion to to look at things like this and I think it is obvious E to anyone who has coded assembly on VAX and Alpha that it takes a lotaB more instructions to accomplish the same task.  In the case of ourH PL/I compiler, on Alpha it is more than 7 times larger.  Of course, thisF uses the GEM backend versus VCG, so I think the 5-6 figure is probablyH fairly accurate.  If you compare the sizes of DECC you will find about aI 9:1 ratio, but again, part of that may be due to the much larger backend.    >e >d= >> A pagelet on an Alpha VMS processor is 512 bytes. It takesfF >>16 pagelets (16*512=8192) to equal to a page on a Alpha VMS machine. >a8 >Unless your particular Alpha has 16K, 32K or 64K pages. >--f3 >VAXman- OpenVMS APE certification number: AAA-0001u >VAXman(at)TMESIS(dot)COMs > 5 >  "Well my son, life is like a beanstalk, isn't it?"h >  >---' >Incoming mail is certified Virus Free.p; >Checked by AVG anti-virus system (http://www.grisoft.com).aB >Version: 6.0.404 / Virus Database: 228 - Release Date: 10/15/2002 >  ---t& Outgoing mail is certified Virus Free.: Checked by AVG anti-virus system (http://www.grisoft.com).A Version: 6.0.404 / Virus Database: 228 - Release Date: 10/15/2002e   ------------------------------  % Date: Thu, 24 Oct 2002 19:57:41 -0400d* From: "Bill Todd" <billtodd@metrocast.net> Subject: Re: page or pagelet2 Message-ID: <leednY78xbt4FSWgXTWc3Q@metrocast.net>  D "Daryl Jones" <jones.computer.srv@worldnet.att.net> wrote in message6 news:8a646952.0210241256.296d1d1@posting.google.com...   ...>  D > A page on an Alpha VMS machine is 8192 bytes per page and on a VAXD > machine it is 512 bytes per page. This is due to RISC vs CISC chipD > technology where the image sizes grew to be 5-6 time larger on the > RISC chip.  H That sounds like far too high a ratio for code that's exactly equivalentH (I'd believe something more like 2, possibly as much as 3), but I'll letH people in a position to measure equivalent compiled images address that.  I Regardless, that's almost certainly not the reason for the increased pagecL size:  after all, *data* didn't increase in size due to the change from CISCK to RISC.  What happened was mostly the passage of time:  a few MB of memoryaJ constituted a significant cash outlay when VAX was designed, and pages hadJ to be small (a disk sector being a convenient size) in order to be able toJ accommodate lots of processes and pieces of the system.  By the time AlphaK came around 15 years later, memory prices had plummeted enough to make much I larger memory sizes - and larger pages - reasonable, and all other thingsiI being equal, the larger the page size, the better, as long as it's not sotK large that the actual streaming transfer time to/from disk starts to becomeeD significant (with current disk characteristics - and memory prices -J anything up to 64 KB has become fairly reasonable for a minimum page size;L the option to use larger sizes than the minimum is also often useful - e.g., to avoid pressure on the TLB).   - bill   ------------------------------  % Date: Thu, 24 Oct 2002 20:50:02 -0400n) From: "Neil Rieck" <n.rieck@sympatico.ca>a Subject: Re: page or pagelet: Message-ID: <Zy0u9.11592$Kf.1396056@news20.bellglobal.com>  D > A page on an Alpha VMS machine is 8192 bytes per page and on a VAXD > machine it is 512 bytes per page. This is due to RISC vs CISC chipD > technology where the image sizes grew to be 5-6 time larger on theG > RISC chip. A pagelet on an Alpha VMS processor is 512 bytes. It takesoF > 16 pagelets (16*512=8192) to equal to a page on a Alpha VMS machine.  K Page size has nothing to do with RISC/CISC but it has everything to do withsK what is a practical minimum amount of bytes to page (or swap) to/from disk.gL Further more, Alpha doesn't always have a page size of 8192 as the followingF quote from the SYS$GETSYI portion of the OpenVMS System Service Manual
 testifies.   SYI$_PAGE_SIZE  @ Returns the number of CPU-specific bytes per page in the system.  + On VAX systems, $GETSYI always returns 512.   = On Alpha systems, CPU page size varies from system to system.g  G On Alpha and VAX systems, because this number is a longword, the bufferh= length field in the item descriptor should specify 4 (bytes).h  A http://www.openvms.compaq.com:8000/731FINAL/4527/4527pro_065.htmlE   ~~~~~   F For those people who don't think this info is important, check out the1 following system services: sys$crmpsc, sys$mgblsc    ~~~~~-  I For those people who aren't familiar with system calls, try the followingE lexical example in DCL:M  ' $write sys$output f$getsyi("PAGE_SIZE")       
 Neil Rieck Kitchener/Waterloo/Cambridge,D Ontario, Canada.! http://www3.sympatico.ca/n.rieck/ 8 http://www3.sympatico.ca/n.rieck/links/cool_vax_vms.html   ------------------------------   Date: 24 Oct 2002 23:03 CDT ' From: carl@gerg.tamu.edu (Carl Perkins)a Subject: Re: page or pagelet- Message-ID: <24OCT200223032908@gerg.tamu.edu>o  $ VAXman-  @SendSpamHere.ORG writes...= }> A pagelet on an Alpha VMS processor is 512 bytes. It takesfF }>16 pagelets (16*512=8192) to equal to a page on a Alpha VMS machine. } 8 }Unless your particular Alpha has 16K, 32K or 64K pages. }--tP }VAXman- OpenVMS APE certification number: AAA-0001     VAXman(at)TMESIS(dot)COM  ? Unless he's got either a new Marvel or one of the Cray designs, @ the odds of the page size not being 8K is, as far as I can tell,
 exactly zero.i  > Every Alpha that runs VMS that you can actually buy up throughB today has had a page size of 8KB. Marvel may, or may not, (I don't? actually know) change this when it becomes possible to actually A buy one without being on the privileged "can get them early" list  (who have some now).   --- Carl   ------------------------------    Date: 24 Oct 2002 12:28:25 -07007 From: jones.computer.srv@worldnet.att.net (Daryl Jones)3  Subject: Re: Scheduling Question= Message-ID: <8a646952.0210241128.1b8e7918@posting.google.com>:  \ carl@gerg.tamu.edu (Carl Perkins) wrote in message news:<23OCT200222285878@gerg.tamu.edu>...1 > "Barry Treahy, Jr." <Treahy@MMaz.com> writes...i > }Herbert Stoeri wrote: > }> Hi, > }>D > }> I stumbled on a problem, most likely connected to schduling of J > }> processes on openVMS ALPHA 7.3. User A  runs an interactive programm H > }> (gnuplot) doing a lot of I/O and using a neglegible amount of CPU, K > }> user B runs interactively a CPU-bound program, doing only computation iG > }> and no I/O. Everything works fine, A's program gets it's priority .# > }> boost and is fully responsive.a > }>F > }> If however A runs both the CPU-bound program and the interactive N > }> program under his (A's) username, the interactive program gets very slow. > } J > }You are unclear as to whether the CPU task is non-interactive, as well K > }as what the base priority is, and the second task which A is also doing  H > }that is interactive, you provide no details;  Is it CPU, I/O, Memory G > }intensive?  Is it the same program?  If the same program, are there eI > }large memory demands that you are overlooking?  If both are producing  G > }significant I/O, is it being bottlenecked on the same file (ie lock m& > }conflicts), same drive, or adapter? > } > > }You are very generic and have provided too little detail... > }  > }Barry > C > There is also the question of how user A is running both of them.e > @ > Are the completely separate processes, or is one (or both) runE > as a subprocess? The subprocess way can kill you via quota sharing.  > 
 > --- Carl  A I would check the processes quotas and memory allocations between A username A and B. If username A quotas and memory allocations are.C larger than username B, then the CPU bound process may suck the CPUsC and leave the interactive process having a hard time getting enougheC time on the CPU to do its work in a timely fashion. This is why CPUt? bound processes are run at a lower priority than an interactivee process.   ------------------------------  % Date: Fri, 25 Oct 2002 05:58:48 +0200 2 From: martin@radiogaga.harz.de (Martin Vorlaender)- Subject: Re: TCPIP 5.3 and SMTP authorizatione; Message-ID: <3db8c178.524144494f47414741@radiogaga.harz.de>d  , David Webb (david20@alpha1.mdx.ac.uk) wrote: > 2) SMTP AUTH >  >    This is an RFC. >sH >    This is supported by PMDF on VMS and possibly other mail products - >    MX, Multinet, TCPWARE ?  3 MX 5.2 has it as well. Don't know about the others.c   cu,.   Martin -- -F                           | Martin Vorlaender  |  VMS & WNT programmer3  Cetero censeo            | work: mv@pdv-systeme.de0F  Redmondem delendam esse. |   http://www.pdv-systeme.de/users/martinv/:                           | home: martin@radiogaga.harz.de   ------------------------------  # Date: Fri, 25 Oct 2002 02:35:33 GMTC( From: Alder <PGDEHMKOKIMD@spammotel.com>- Subject: Re: TCPIP$FTP_SERVER startup probleme, Message-ID: <3DB8ADED.9050803@spammotel.com>  D I'm having a similar problem trying to configure Anonymous FTP with I TCP/IP Services for OpenVMS 5.0 on OpenVMS Alpha 7.2 (the hobbyist kit). oF    Anonymous logins are refused because of a privilege issue, I think.  I I configured it from the menus in TCPIP$CONFIG and start the server from . the SYSTEM account with:   $ @TCPIP$SYSTEM:TCPIP$FTP_RUNa  @ Here's what pops out when I connect from another host on my LAN:  B ******************************************************************    &                  FTP server started on@                  SZEGED.EREBUS.CA        24-OCT-2002 11:15:05.86    B ******************************************************************    C %TCPIP-I-FTP_SESCON, FTP SERVER: session connection from KIRALY at  
 24-OCT-2002 1l
 1:17:28.75  + %TCPIP-I-FTP_NODE, client host name: KIRALY'' %TCPIP-I-FTP_USER, user name: ANONYMOUSH %TCPIP-I-FTP_OBJ, object: - ? %TCPIP-I-FTP_CHINFO, ANONYMOUS: Error locking pages into memory2G %SYSTEM-F-NOPRIV, insufficient privilege or object protection violationD? %TCPIP-E-FTP_LOGFAL, remote interactive login failure anonymousD  + -TCPIP-I-FTP_NODE, client host name: KIRALYyG -SYSTEM-F-NOPRIV, insufficient privilege or object protection violationu  J Can anyone suggest which permissions have to be adjusted and how?  Thanks.   Alderu   ------------------------------  % Date: Thu, 24 Oct 2002 12:44:29 -0700p0 From: Mark Berryman <Mark.Berryman@Mvb.Saic.Com>B Subject: Re: The VMSNET group of forums (was: Re: Acrobat Reader?), Message-ID: <3DB7EB2D.61AE887C@Mvb.Saic.Com>   Bill Gunshannon wrote: > E > In article <1fkjrte.1xzgh9g1weeh98N%Didier.Morandi.nospam@free.fr>,l@ >         Didier.Morandi.nospam@Free.fr (Didier Morandi) writes:K > > Thank you Hoff, I stand accused. But I'm happy to hear about the VMSNETcL > > group of forums again. There is nearly no activity in there. And the VMS > > FAQ sez: > >0I > > "Another source of free software is the vmsnet.sources newsgroup (and7J > > the corresponding vmsnet.sources.d discussion group).  See the monthlyG > > posting "vmsnet.sources archives" for a list of sites which archive.# > > submissions to vmsnet.sources".@ > >o= > > There is as of today no messages in vmsnet.sources nor in.L > > vmsnet.sources.d. And I did not find the monthly posting "vmsnet.sources. > > archives". Maybe the FAQ could be updated. > >rG > > Also, It looks like these forums are mostly a public repository for C > > japanese cartoons posted there by the www.frozen-layer.com site G > > WebMaster, and also some DECUS members activity, which is generallyT< > > followed-up in our Group (but for pdp and a few others). > >I > C > Becuase of the ease with which code sources can be made availablehB > through the web the use of USENET as a source distribution mediaB > has all but ceased.  A quick glance at the various archives showB > no new sources in any of the major groups since at best 1998 andC > in most cases since 1996.  vmsnet.sources was a much lower volume < > group and likely pretty much died out about the same time. > 8 > They were a tremendous resource whose time has passed. >   E That is correct.  Submissions to vmsnet.sources dried up at about the E time the internet went commercial and access to data via web browsers  became available to the masses.   > The VMSNET community is still thriving.  While general purposeE newsgroups within vmsnet are very low volume (mainly because everyone#H makes such postings here), those newsgroups dedicated to specific topics) (mail, IP stacks, etc.) are quite active.   G Vmsnet.sources and vmsnet.sources.d should probably be formally removed C from usenet as they get nothing but spam and have been that way for > several years (the single exception: Hunter's announcements in? vmsnet.sources.d when something is updated on his web archive).|  
 Mark Berryman  Mark.Berryman@Mvb.Saic.Com vmsnet.sources moderator   ------------------------------   Date: 24 Oct 2002 21:32 CDT ' From: carl@gerg.tamu.edu (Carl Perkins) B Subject: Re: The VMSNET group of forums (was: Re: Acrobat Reader?)- Message-ID: <24OCT200221320304@gerg.tamu.edu>   8 Didier.Morandi.nospam@Free.fr (Didier Morandi) writes...H }Thank you Hoff, I stand accused. But I'm happy to hear about the VMSNET= }group of forums again. There is nearly no activity in there.  }D.e  ; Some of them have activity. I know the vmsnet.mail.pmdf ande> vmsnet.networks.tcp-ip.multinet groups (the only ones I check)6 receive some traffice - a few messages a day, usually.   --- Carl   ------------------------------  + Date: Thu, 24 Oct 2002 20:32:19 +0000 (UTC)e, From: lewis@mazda.mitre.org (Keith A. Lewis) Subject: Re: TPU port to Linux. Message-ID: <ap9lcj$kj5$1@newslocal.mitre.org>   "Tony Scandora" <Scandora@cmt.anl.gov> writes in article <ap6eup$bde$1@milo.mcs.anl.gov> dated Wed, 23 Oct 2002 10:22:36 -0500:hJ >Years ago I saw a product called nu/TPU that impressed me at the time.  IF >don't remember if it could read a section file, but it could at leastG >compile and run TPU source, including EVE.  They're still in business:r' > http://www.asoft-dev.com/tpu_info.htmA  J I used to stop by that guy's booth at Decus regularly and explain that youL can't charge $500 for a piece of software designed to run on a PC.  It looks< like he's dropped the price to $250 now (Windows version).    J The cheapest Linux license is still $500 -- I bought a used Alpha for lessL than that, and I can run the *real* TPU on my Linux display using X-windows!  B Compaq TCPIP 5.3's NFS lets you mount Unix volumes as ODS-5, whichH eliminates most of the annoying filename translation.  I've been editing Sun-hosted files using TPU.y  + --Keith Lewis              klewis$mitre.orgi> The above may not (yet) represent the opinions of my employer.   ------------------------------  # Date: Thu, 24 Oct 2002 18:02:42 GMTa9 From: Hein van den Heuvel <hein_netscape@eps.zko.dec.com>n4 Subject: Re: Why does this file-spec work only once?/ Message-ID: <3DB834F7.FCC8C955@eps.zko.dec.com>i   Paul Repacholi wrote:-  2 > spamsink2001@yahoo.com (Alan E. Feldman) writes: >lG > > Why does a file-spec containing "password" as shown below cause theeF > > second (and all subsequent) lexical functions that use it to fail?E > > If I make any change at all to "password", even just capitalizing  > > it, it works fine. Example:  > 
 > [sniped] >tB > Yep, same as in 87. It is an RMS bug, it looks inside the accessE > control string and thinks it knows what is or isn't a valid string.-C > *EVEN THOUGHT THE SPEC CLEARLY STATES THAT IT IS NOT INTERPRETED*o  J I guess we should have changed to doc. RMS does look, for one and only one word: password  < It does does in module RM0SCAN where it sets flag FSCB$V_PWDG This flag is later used by RM0XPFN. The relevant comment in that module3 reads.   ;2" ; Logical node name translation... ;dI ; Attempt to translate the node name string of the (left-most) node spec. G ; For the translation to succeed, the equivalence string must represent I ; one or more concatenated node specs, i.e., each field must be delimitedm ; by a double colon. ; E ; Note: If node name is prefixed by underscore, no translation occurso& ;       but the underscore is removed. ;aH ; Note: If node name contains an access control string with the password maskedJ ;       out, then an attempt is made to translate the string consisting ofK ;       <node name + access control string> as a special logical node name.   F Hmmm,... would pre-pending an underscore work? Frankly my dear I don't care!m  I > Some one will no doubt say you should not use `password' as a password. B > Ask them why, as Password or pasSword are OK. PLease explain....  G Oh yeah. Did so in the very first reply in the string. Since explained.l   Grins, Hein   ------------------------------    Date: 24 Oct 2002 11:19:35 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)4 Subject: Re: Why does this file-spec work only once?= Message-ID: <b096a4ee.0210241019.2032e311@posting.google.com>o  Y briggs@encompasserve.org wrote in message news:<nsQoMZfr3i0X@eisner.encompasserve.org>... o > In article <b096a4ee.0210220908.34ec037@posting.google.com>, spamsink2001@yahoo.com (Alan E. Feldman) writes:at > > Hein van den Heuvel <hein_netscape@eps.zko.dec.com> wrote in message news:<3DB456F0.DA6C4AAE@eps.zko.dec.com>... > >> "Alan E. Feldman" wrote:3 > >> e
 > >> > Hello,b > >> >J > >> > Why does a file-spec containing "password" as shown below cause theI > >> > second (and all subsequent) lexical functions that use it to fail?o > >> e [...]e > > Check out this example:  > >  > > ABCDE$ TYPE WB2.COMG > > $    ON WARNING THEN EXITa > > $    NODE_INT = 1k > > $    NODE_NAME = "ABCDE" > > $    CODE = A !- > > $    UPDATE_LOG_FILE_SPEC =:3 > > NODE_NAME+CODE+"::FTTOP:UPDATE_''NODE_INT'.LOG".9 > > $    UPDATE_LOG_FILE = F$SEARCH(UPDATE_LOG_FILE_SPEC)w9 > > $    SEARCH 'UPDATE_LOG_FILE' BEGIN,MANAGER/MATCH=ANDe7 > > $    SEARCH 'UPDATE_LOG_FILE' "BEGIN NODE FUNCTION" 
 > > ABCDE$ > > ABCDE$ @WB2n > > $    ON WARNING THEN EXITi > > $    NODE_INT = 1o > > $    NODE_NAME = "ABCDE" > > $    CODE = A ! E > > $    UPDATE_LOG_FILE_SPEC = NODE_NAME+CODE+"::FTTOP:UPDATE_1.LOG"}9 > > $    UPDATE_LOG_FILE = F$SEARCH(UPDATE_LOG_FILE_SPEC)L= > > $    SEARCH ABCDE"FT password"::_DSA2:[FT]UPDATE_1.LOG;19t > > BEGIN,MANAGER/MATCH=ANDT= > >     "GVT" Update Data Base Manager Begin Node Function...aI > > $    SEARCH ABCDE"FT password"::_DSA2:[FT]UPDATE_1.LOG;19 "BEGIN NODEo
 > > FUNCTION"i, > > %SEARCH-W-OPENIN, error opening ABCDE"FT1 > > password"::_DSA2:[FT]UPDATE_1.LOG;19 as inputl3 > > -RMS-E-FND, ACP file or directory lookup failedg@ > > -SYSTEM-F-INVLOGIN, login information invalid at remote node# > > %SEARCH-E-NOFILE, no file foundw
 > > ABCDE$ > >  e8 > > OK. WHY does it fail for the second SEARCH command?  > ? > Because you have discovered a security hole in DECnet on VMS.  > D > Since you have specified incorrect access control information, the< > first SEARCH should fail as well.  But instead, it uses an> > existing DECnet FAL listener process on the remote node that? > that should have reverted to the idle process state after the  > F$SEARCH terminated. [...] < > You're not refreshing the symbol.  You're creating another< > remote DECnet process on the far end node.  Or pushing oneF > into the "will accept improperly authenticated connections" state at > any rate.g  DC OK. I thought that because the first SEARCH command worked that thel7 symbol somehow still contained the correct password and@C VMS/DCL/DECnet/whatever was still hiding it behind "password". So Io- believe you are right or at least very close.o  F Strangely, the following code works most of the time but every once in? a while the second SEARCH command fails with "login informationr invalid at remote node"e   $ TYPE WB3.COM $    ON WARNING THEN EXITs $    NODE_INT = 1r $    NODE_NAME = "IDS15" $    CODE = A !n	 $_EXISTS:rF $    UPDATE_LOG_FILE = NODE_NAME+CODE+"::FTTOP:UPDATE_''NODE_INT'.LOG"5 $    SEARCH 'UPDATE_LOG_FILE' BEGIN,MANAGER/MATCH=ANDh3 $    SEARCH 'UPDATE_LOG_FILE' "BEGIN NODE FUNCTION"r $i  ( BTW, this was VMS 6.2, DECnet Phase IV.   ' Strangely, a NCP SHOW EXEC CHAR yields h  6 Identification           = DECnet for OpenVMS VAX V6.1  
 Thanks again.-   Disclaimer: JMHO Alan E. Feldmanc   ------------------------------    Date: 24 Oct 2002 11:23:16 -0700. From: spamsink2001@yahoo.com (Alan E. Feldman)4 Subject: Re: Why does this file-spec work only once?= Message-ID: <b096a4ee.0210241023.2b9b67d1@posting.google.com>2  j whohe@whoever.com (DL Phillips) wrote in message news:<af0dc2ea.0210231145.6b3bfe93@posting.google.com>... > Alan,  > * > Do the NETSERVER.LOG files yield a clue? >  >  DL Phillips t   No. Thanks anyway.   Disclaimer: JMHO Alan E. Feldmant   ------------------------------  % Date: Fri, 25 Oct 2002 06:15:22 +1000v@ From: "Antony Wardle" <antony.wardle@nospammmmm.optusnet.com.au> Subject: XFC< Message-ID: <3db8551e$0$18873$afc38c87@news.optusnet.com.au>   Hi  > As I understand it, VCC_FLAGS = 1 means that VMS will use XFC.K On 7.3 VCC_FLAGS set to 1 means that it will use the pre 7.3 version of XFCiK and you set it to 2(?) for the newer way, but you should have the xfc patchr on.   B If I type show memory/reserve it should show me how much memory isF allocated for XFC. Is this correct? I have VCC_FLAGS set but no memoryD reserved, so is the xfc working, or is vms just using memory without
 reserving it?   ' I am finding it all a little confusing.t   kiwi   ------------------------------  % Date: Thu, 24 Oct 2002 22:54:44 +0200i" From: "t.t" <t.t@poczta.gazeta.pl> Subject: Re: XFC) Message-ID: <ap9ml8$91r$1@news.gazeta.pl>    Hi,dJ VCC_FLAGS set to 1 means, that you use old cache mechanism (VCC Cache). ToJ use new one (XFC) you have to set VCC_FLAGS to 2, which is default in 7.3.' Of course you have to apply XFC patch:) I When XFC is active you can see memory reservation with show mem/cache andoC cache utilisation for particular disk with sho mem/cache=vol=dkaXXX    Regards - Tomasz    K "Antony Wardle" <antony.wardle@nospammmmm.optusnet.com.au> wrote in messageC6 news:3db8551e$0$18873$afc38c87@news.optusnet.com.au... > Hi >i@ > As I understand it, VCC_FLAGS = 1 means that VMS will use XFC.I > On 7.3 VCC_FLAGS set to 1 means that it will use the pre 7.3 version ofu XFC.G > and you set it to 2(?) for the newer way, but you should have the xfc  patcho > on.  > D > If I type show memory/reserve it should show me how much memory isH > allocated for XFC. Is this correct? I have VCC_FLAGS set but no memoryF > reserved, so is the xfc working, or is vms just using memory without > reserving it?f >o) > I am finding it all a little confusing.  >9 > kiwi >a >>       -- eA Serwis Usenet w portalu Gazeta.pl -> http://www.gazeta.pl/usenet/w   ------------------------------    Date: 24 Oct 2002 13:29:36 -0500; From: kaplow_r@eisner.encompasserve.org.TRABoD (Bob Kaplow)yK Subject: Re: [OT] - Parties (Re: HP tells VMS users "Relax, you are sittingn3 Message-ID: <EvigJuvrPTfW@eisner.encompasserve.org>a  c In article <3DB7F73B.71E424E6@mindspring.com>, Atlant Schmidt <atlantnospam@mindspring.com> writes:o- > Sometimes, it doesn't matter *WHO* you voteN$ > for; Republicans win regardless... > - > Have you been following the news around the ( > "Early Voting" going on down in Texas?  J Texas? Amateurs. I was born and raised in Chicago, where the Democrats won before the poles opened.    J Before his death JFKjr was making a speech in Chicago. He talked about hisB family ties to the city. He mentioned how they'd bought buildings,% companies, and even an election here..  H Now that he's dead, he's still voting a straight democratic ticket here.  1 	Bob Kaplow	NAR # 18L	TRA # "Impeach the TRA BoD"e& 		>>> To reply, remove the TRABoD! <<<B Kaplow Klips & Baffle:	http://www.pleimling.org/le/Phantom4000.pdf: NIRA:	http://www.nira-rocketry.org	NAR:	http://www.nar.org  1 	26-October, 2001: A day that will live in infamyi4 	Support Freedom: http://www.indefenseoffreedom.org/  J         Guard with jealous attention the public liberty. Suspect every oneJ         who approaches that jewel. Unfortunately, nothing will preserve itE         but downright force. Whenever you give up that force, you arebC         inevitably ruined. -- Patrick Henry, speech of June 5 1788     ------------------------------    Date: 24 Oct 2002 13:31:07 -0500; From: kaplow_r@eisner.encompasserve.org.TRABoD (Bob Kaplow) K Subject: Re: [OT] - Parties (Re: HP tells VMS users "Relax, you are sitting?3 Message-ID: <hYPJgoXGH8VV@eisner.encompasserve.org>-  _ In article <m9mdnckWgN9cniWgXTWcog@metrocast.net>, "Bill Todd" <billtodd@metrocast.net> writes:lI >> Where I live half the candidates on next months ballot will be runningWH >> unopposed, even by the other "major" party. Sort of reminds me of theJ >> election recently won in Iraq by Sadaam. I refuse to vote for unopposedB >> candidates, especially since they never represemt MY positions. > @ > That's a perfect opportunity to voice your support for change.  I And how do I vote for change when there is only one name on the ballot???n  1 	26-October, 2001: A day that will live in infamy 4 	Support Freedom: http://www.indefenseoffreedom.org/  J         Guard with jealous attention the public liberty. Suspect every oneJ         who approaches that jewel. Unfortunately, nothing will preserve itE         but downright force. Whenever you give up that force, you are_C         inevitably ruined. -- Patrick Henry, speech of June 5 1788 e   ------------------------------  % Date: Thu, 24 Oct 2002 18:49:55 -0400 * From: "Bill Todd" <billtodd@metrocast.net>K Subject: Re: [OT] - Parties (Re: HP tells VMS users "Relax, you are sitting)2 Message-ID: <c4WdnWN3W8Se5CWgXTWcog@metrocast.net>  H "Bob Kaplow" <kaplow_r@eisner.encompasserve.org.TRABoD> wrote in message- news:hYPJgoXGH8VV@eisner.encompasserve.org...E@ > In article <m9mdnckWgN9cniWgXTWcog@metrocast.net>, "Bill Todd"  <billtodd@metrocast.net> writes:K > >> Where I live half the candidates on next months ballot will be running J > >> unopposed, even by the other "major" party. Sort of reminds me of theL > >> election recently won in Iraq by Sadaam. I refuse to vote for unopposedD > >> candidates, especially since they never represemt MY positions. > >tB > > That's a perfect opportunity to voice your support for change. >nK > And how do I vote for change when there is only one name on the ballot???G  I Write-ins are usually allowed.  And if you know in advance there's no oneoF you wish to vote for for a particular office, you can even organize an, actual write-in campaign to make your point.   - bill   ------------------------------    Date: 24 Oct 2002 13:31:52 -0500; From: kaplow_r@eisner.encompasserve.org.TRABoD (Bob Kaplow)s0 Subject: Re: [OT] - we're back on old cars again3 Message-ID: <5t2cc6Y0SUAX@eisner.encompasserve.org>   c In article <VoJuIkeW+M0r@eisner.encompasserve.org>, koehler@encompasserve.org (Bob Koehler) writes: k > In article <20021023182149.I15801@eisenschmidt.org>, John Eisenschmidt <jweisen@eisenschmidt.org> writes:6 >>  H >> To actually drift back on topic (the topic itself being offtopic), if. >> Microsoft were a car, I'm thinking Corvair. > J >    Nope.  GM actually fixed a lot of the design problems in the Corvair.   Yugo?   1 	26-October, 2001: A day that will live in infamye4 	Support Freedom: http://www.indefenseoffreedom.org/  J         Guard with jealous attention the public liberty. Suspect every oneJ         who approaches that jewel. Unfortunately, nothing will preserve itE         but downright force. Whenever you give up that force, you arelC         inevitably ruined. -- Patrick Henry, speech of June 5 1788 b   ------------------------------   End of INFO-VAX 2002.589 ************************chipB technology where the image sizes grew to be 5-6 time larger on theE RISC chip. A pagelet on anڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ    ڽ     ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    	۽    
۽    ۽    ۽    
۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽     ۽    !۽    "۽    #۽    $۽    %۽    &۽    '۽    (۽    )۽    *۽    +۽    ,۽    -۽    .۽    /۽    0۽    1۽    2۽    3۽    4۽    5۽    6۽    7۽    8۽    9۽    :۽    ;۽    <۽    =۽    >۽    ?۽    @۽    A۽    B۽    C۽    D۽    E۽    F۽    G۽    H۽    I۽    J۽    K۽    L۽    M۽    N۽    O۽    P۽    Q۽    R۽    S۽    T۽    U۽    V۽    W۽    X۽    Y۽    Z۽    [۽    \۽    ]۽    ^۽    _۽    `۽    a۽    b۽    c۽    d۽    e۽    f۽    g۽    h۽    i۽    j۽    k۽    l۽    m۽    n۽    o۽    p۽    q۽    r۽    s۽    t۽    u۽    v۽    w۽    x۽    y۽    z۽    {۽    |۽    }۽    ~۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    ۽    