1 INFO-VAX	Tue, 14 Nov 2000	Volume 2000 : Issue 636       Contents: Re: 100BaseT on VAX XMI search& Re: Accessing a OpenVMS server from NT+ Re: Apache and DCL character limit for POST E Re: AS 255/233, which graphics card do I have without opening the box  Re: Beep on shutdown Re: Beep on shutdown$ Can f$getdvi() get ucx socket info ? Re: Change of IP subnet  DE204 under OpenVMS 7.2  Re: DECforms...  disk benchmarking software?  Re: disk benchmarking software?  Re: disk benchmarking software?  Re: Disk performance under VMS( Does CompaQ TCP/IP support secure shell?, Re: Does CompaQ TCP/IP support secure shell? Re: f$search improvement ? Re: f$search improvement ?: Re: Help... Need Software to Logoff users after 20 minutes Re: Malloc + debug question ( Re: NCP commands in the 7.2 Cluster doc? Re: Nested DCL symbol names + Re: open VMS backdrop for motif or xwindows + Re: open VMS backdrop for motif or xwindows % Re: OpenVMS 7.1-2 and cluster forming  Re: Printing blank page  RE: RECALL question/suggestion. 2 Re: Research Systems Inc. dropping support for VMS2 Re: Research Systems Inc. dropping support for VMS5 Re: Resurrecting an Alphastation 200/250: help needed 	 Re: RWAST 	 Re: RWAST - Scarey use of Technology in the US Elections. 1 Re: Scarey use of Technology in the US Elections. . Re: Severity of eBay's ecache problem detailed SSU License Pak  Re: SSU License Pak  RE: Sun "uptime" belly laugher Re: Sun "uptime" belly laugher Re: System disk filling up Re: System disk filling up Re: System disk filling up Re: System disk filling up Re: Technology of US elections Re: Technology of US elections Re: Technology of US ElectionsC RE: Technology of US elections  OR why Palm Beach selected  Solaris C Re: Technology of US elections  OR why Palm Beach selected  Solaris A Re: Technology of US elections OR why Palm Beach selected Solaris 4 Re: Technology of US elections x Brazilian elections4 Re: Technology of US elections x Brazilian elections4 Re: Technology of US elections x Brazilian elections4 Re: Technology of US elections x Brazilian elections: Re: Two Phase Commit (2PC) on VMS - the nature of the fuss! Re: VAXstation 4000-90 Error Code  Wildfire GS160  F ----------------------------------------------------------------------  % Date: Mon, 13 Nov 2000 21:35:31 +0100 " From: "Hans Vlems" <hvlems@iae.nl>' Subject: Re: 100BaseT on VAX XMI search ( Message-ID: <8upimo$95c$1@news.IAEhv.nl>  $ Stick to FDDI, that's all I can say.A The newly formed Digital Network products Group will continue the 8 support for the network components (concentrators etc.).H Compaq FS will support your adapters. We have 6400's and 6300's with theH original DEMFA's without problems. Though we had to check their firmware before a VMS upgrade (7.1).   
 Hans Vlems  * John Nixon heeft geschreven in bericht ...G >Does anyone know of, or have any comments on, any third party 100BaseT K >ethernet adapters that can be used on a VAX 7800 (with XMI)?  My VAXes now K >have a mix of 10BaseT and FDDI.  This is inefficient, and our network guys J >REALLY want to dump all of our FDDI gear.  Only three remaining VAX 7800sH >still require FDDI here, and they will be here at least two more years.1 >(OVMS 7.1, for now and always on these systems).  >  >  >  >    ------------------------------   Date: 13 Nov 2000 22:48:33 GMT From: ka2doug@cs.com (KA2DOUG)/ Subject: Re: Accessing a OpenVMS server from NT > Message-ID: <20001113174833.00868.00001584@ng-cn1.news.cs.com>   ---reply to---0 >Subject: Re: Accessing a OpenVMS server from NT% >From: casinoop2@aol.com  (CasinoOp2) / >Date: 11/09/2000 3:34 PM Central Standard Time ; >Message-id: <20001109163419.10822.00001453@ng-md1.aol.com>  >  >Boy I am feeling real dumb,    M Welcome to the club;-) About the time I get one thing figured out they go and 
 change it.   >what is a "server share"?  A I mean, the share name on the server. ie. \\servername\sharename\    > How do I check permissions?   M If you are using an NT system to administer it's just like NT. If you use the J command line on the OpenVMS system, then LOGON to the domain with whateverO ADMIN user you have, and at the prompt enter SHOW SHARES /PERMISSIONS. Use HELP J to work through the commands to get things set properly if they look wrong (that's what I do).   N The post from sol gongola mentioned setting proper protocols, and that is also a very common problem.  
 Good luck.
 Doug Phillips    ------------------------------  % Date: Tue, 14 Nov 2000 07:26:05 +0100 2 From: martin@radiogaga.harz.de (Martin Vorlaender)4 Subject: Re: Apache and DCL character limit for POST; Message-ID: <3a10dafd.524144494f47414741@radiogaga.harz.de>   / Mike Flaherty (mflaherty2@earthlink.net) wrote: M : I am running Compaq Secure Web Server on an Alpha running 7.2 of VMS.  I am E : trying to submit a form via POST to a DCL command procedure but the - : QUERY_STRING symbol keeps being hacked off.  : E : Is there a way to send a large stream of form data to a DCL command ) : procedure without it being chopped off?   C On a POST request, the parameters don't get passed via QUERY_STRING E (unless you provide some in the <FORM ACTION> field directly. Rather, C they are passed via stdin, under CSWS to be read from APACHE$INPUT,  IIRC.    cu,    Martin --J One OS to rule them all       | Martin Vorlaender  |  VMS & WNT programmer7 One OS to find them           | work: mv@pdv-systeme.de N One OS to bring them all      |       http://www.pdv-systeme.de/users/martinv/> And in the Darkness bind them.| home: martin@radiogaga.harz.de   ------------------------------  % Date: Mon, 13 Nov 2000 15:46:33 -0500 5 From: John Johnstone <jj_usenet@onay-amspay.mail.com> N Subject: Re: AS 255/233, which graphics card do I have without opening the box4 Message-ID: <3A100CD9.1EEDE834@onay-amspay.mail.com>  F From $ ANA/SYSTEM a CLUE CONFIG will show any graphics devices, if theI drivers for them have been loaded, as G*: devices (e.g. GYA0:) along with J the type of card in the Device Name field.  If there's nothing there, thenH either there's really a hardware problem or the card that is there isn't6 recognized by the version of DECwindows and/or Open3D.  I From the SRM prompt, a "show config" will list any cards as long as there . aren't any hardware problems.  Something like:  
    PCI Bus  9         Bus 00  Slot 11: Digital TGA2 Graphics Controller   I You may need an upgrade of Open3D or even a downgrade.  I think I've read K that the latest versions have dropped support for some cards.  I don't know J for sure though since there's just too much to be able to remember it all.  K Was this graphics card in this system working with this software previously K and then it stopped working or has something changed in the hardware and/or  software combination?      Tim Llewellyn wrote: >  > Hi > E > I have a AS 255/233 that has recently been reunited with a graphics 
 > console.H > I am sure the monitor worked with DECWindows when I configured the boxF > with VMS 6.2-1 in 1997. Now it is running VMS 7.1 and Motif 1.2-5. I > have a monitorE > connected to the SVGA port. I have a PS/2 mouse and keyboard (Dell)  > connected. > G > The monitor works as a dumb console. However, DECWindows keeps saying E > "No graphics devices are present". Also, and the console mode (>>>)  > there are / > only disks and tapes reported by SHOW DEVICE.  > H > How can I find what graphics card is installed (and in use by VMS as a
 > console)8 > without opening the box? Is this just an Open3D issue? > 	 > Regards  > --8 > Tim Llewellyn, OpenVMS Infrastructure, Remarcs Project2 > MedAS at the BBC, Whiteladies Road, Bristol, UK.C > Email tim.llewellyn@bbc.co.uk. Home tim.llewellyn@cableinet.co.uk  > C > I speak for myself only and my views in no way represent those of  > MedAS or the BBC.    ------------------------------  % Date: Mon, 13 Nov 2000 10:51:54 -0800 0 From: "William S. LaCounte" <vmsmanager@ups.edu> Subject: Re: Beep on shutdown # Message-ID: <3A103849.57D5@ups.edu>   H We had the same problem on some Alphas here. The beeps went away once weD upgraded the Firmware to the latest version and also upgraded to VMS 7.2. This was on a DS20.   Bill   ------------------------------  # Date: Mon, 13 Nov 2000 20:44:55 GMT + From: Craig A. Berry <calepine@my-deja.com>  Subject: Re: Beep on shutdown ) Message-ID: <8upjs4$fq9$1@nnrp1.deja.com>   C Have you removed the /BELL qualifier from all the REPLY commands in  SYS$SYSTEM:SHUTDOWN.COM?  ' In article <3A102AFD.9C210C5C@def.com>,    abc@def.com wrote:
 > Dear all > A > Apologies if this has been asked before, I searched the FAQ and  Dejanews > first. > > > Our (standalone) Alpha machine (OVMS 7.1) repeatedly emits aE > particularly painfully loud beep when I shut it down. I've searched  all B > the .com files called from shutdown.com and remmed out any lines which I E > know to produce a beep, but it's still doing it. There must be some A > commands which beep as a side effect. Has anyone got any ideas?     & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  % Date: Mon, 13 Nov 2000 14:19:02 -0600 * From: Patrick Spinler <pspinler@yahoo.com>- Subject: Can f$getdvi() get ucx socket info ? ) Message-ID: <3A104CB6.CEBED382@yahoo.com>   H Or to ask this another way - can a script running against UCX get device= socket information programatically (e.g., without a 'ucx show  device_socket bgfoo').  G In particular, I'd like to get the associated host information and port  numbers.  E While I appear to be able to get the host name (HOST_NAME item code), E doing a quick dump of all the available f$getdvi() item codes doesn't / show anything promising as far as port numbers.   @ What I'm trying to do is set up a captive login that may only beG accessed by a "set host" or "set host/telnet" from a list of authorized G hosts, and have the captive account trace back the originating account, G and check that the originating account has a specific rights id granted 4 to it.  I'd like to do this in DCL or Perl if I can.  E I can get traceback information now by looking at the TT_ACCPORNAM of H the captive account's terminal, and using that host and port number in aF "ucx show dev".  This works fine, but I'd rather not worry about smallD maintainance issues like ucx show dev's output format changing, etc.   Thanks,  -- Pat   $ ucx show version  @   Digital TCP/IP Services for OpenVMS Alpha Version V4.2 - ECO 24   on a AlphaServer 2100 5/250 running OpenVMS V7.1-2   $ show sys /noproc! OpenVMS V7.1-2  on node <snip>...    --  ?       This message does not represent the policies or positions 1 	     of the Mayo Foundation or its subsidiaries. 3   Patrick Spinler			email:	Spinler.Patrick@Mayo.EDU '   Mayo Foundation			phone:	507/284-9485    ------------------------------  % Date: Mon, 13 Nov 2000 22:36:39 +0100 2 From: "Thomas H. Pauli" <thomas.pauli@t-online.de>  Subject: Re: Change of IP subnet* Message-ID: <3A105EE7.8040401@t-online.de>  
 Hi Rudolf,  A we did something like that a while ago. What we used was UCX SET   INTERFACE commands, I but this could be a nuisance, because the interface names (yes, the node   names) are also storedG in the SYS$SYSTEM:UCX$HOSTS.DAT file and changing of IP addresses will   cause problems. F If you can afford to rebuild this file with UCX SET HOST commands you  might simply rename D this file and create an empty one with UCX CREATE. As far as I know  there should not be a problem F changing the subnet mask alone (see SET INTERFACE) except restartuing  TCP/IP.    Thomas   Rudolf Wingert wrote:    > Hello, > ; > for all how did not read this email and would like to the ? > "I have made some questions, without any answer" thread, here  > the original question: >  > Hello, > K > we must change our IP subnet. The involved product is TCP IP services for F > OpenVMS 4.2 ECO 1. Is there any way to change only the addresses andF > the subnetmask, without running the configuration procedure on everyH > node? We do have only one one bootnode cluster with a common UCX root.; > What are the hints, which I have to do before the change?  >   > TIA and regards Rudolf Wingert >  >    ------------------------------  % Date: Mon, 13 Nov 2000 17:31:27 -0500 / From: "Howard E. Arnold" <arnoldh@celerent.com>   Subject: DE204 under OpenVMS 7.2? Message-ID: <NEBBKBPAMLAEPFNHOPFBOEIDCBAA.arnoldh@celerent.com>   E Are the DE204 network cards supported under OpenVMS 7.2?  We have one K installed in a Alphastation 400 4/233 and when we try to configure TCPIP it  fails with no such interface.    Howard   ------------------------------    Date: 13 Nov 2000 17:36:18 -0500* From: kuhrt@eisner.decus.org (Marty Kuhrt) Subject: Re: DECforms...+ Message-ID: <g4qc$Z+9od4t@eisner.decus.org>   ] In article <3A0A08EE.58626E87@triton.com.no.spam>, Jay Olson <jjo@triton.com.no.spam> writes:  > Kevin Iliff wrote: >>  	 >> Hello,  >>  : >> Are there any DECforms users, or former DECforms users,9 >> monitoring this newsgroup?  Our company is moving from 8 >> VTs to PCs, so we're looking to replace DECforms with6 >> something more PC compatible.  Does anyone have any9 >> insights as to what our options might be?  Does anyone 3 >> know of another newsgroup that might be helpful?  > G > Three of my major clients use DECforms, and have been doing so for 10 J > years. They haven't replaced it, even though VT-style terminals are hardE > to come by, because it still works fine and nothing better has come  > along :).  > J > If you are comfortable with DECforms and you have an application writtenF > using DECforms and your programmers know DECforms, then the sensibleG > thing to do is to get some terminal emulator software for the PCs and H > leave the application alone. This applies to any "legacy" software. If > it ain't broke, don't fix it.  > J > There is an animal called the DECforms Web Connector or some such thing,G > and I believe it is included with DECforms. I haven't used it before, E > but it seems rather strange to try grafting DECforms into a browser > > environment. In my opinion, if you want to go with somethingH > browser-based, you need to change your whole user interaction paradigm" > (i.e. rewrite your application). > * > 	- Jay Olson (jjo "at" triton "dot" com) > 	Triton Software Group LLC  ; DECForms Web Connector allows a DECForm application to used ? unchanged from a browser.  As I undersand it they separated the > form management from the display functions, which allows for a> "form gui server" to handle how the stuff is displayed.  Check6 out the DECForms link on the Application Compilers and Environments page A http://www.openvms.compaq.com/commercial/cace.html for more info.    ------------------------------  # Date: Tue, 14 Nov 2000 02:17:51 GMT / From: "Thomas Williams" <tawilli@earthlink.net> $ Subject: disk benchmarking software?E Message-ID: <jh1Q5.15592$Gd7.877230@newsread2.prod.itd.earthlink.net>   K Does anyone know where I can get a (preferably free) program to do some I/O G benchmarking for VMS? I'd like to compare VMS volume shadowing vs. ASCI A shadowing (3rd party) vs HSZ-70 disk mirroring and different RAID J configurations.  I know HW shadowing offload the CPU and should be faster,G but I'd like to see some numbers for my specific configuration.  Thanks    Tom Williams   ------------------------------  % Date: Mon, 13 Nov 2000 21:32:13 -0600 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> ( Subject: Re: disk benchmarking software?- Message-ID: <3A10B23D.89F05EE4@earthlink.net>    Thomas Williams wrote: > M > Does anyone know where I can get a (preferably free) program to do some I/O I > benchmarking for VMS? I'd like to compare VMS volume shadowing vs. ASCI C > shadowing (3rd party) vs HSZ-70 disk mirroring and different RAID L > configurations.  I know HW shadowing offload the CPU and should be faster,I > but I'd like to see some numbers for my specific configuration.  Thanks   : Well, I would expect the software to be its own benchmark.  + Here's some numbers from my own experience:   F Host based shadow-copy, 4GB disks: 2 hours to shadow-copy (OVMS V5.5-2+ (VAX), V6.2-1H3 (Alpha) and V7.1-2 (Alpha))   : HSJ based mirroring, 9GB disks: 50 minutes to mirror copy.  G Third-party shadowing software is not recommendable, IMO. The "shadows"ME are usually MOUNTed /FOREIGN - there is no "hot copy". If the primaryaH fails, that volume becomes unavailable until manual intervention occurs.   -- n David J. Dachterae dba DJE Systemse http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/r  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.$   ------------------------------    Date: 13 Nov 2000 23:57:57 -0500* From: young_r@eisner.decus.org (Rob Young)( Subject: Re: disk benchmarking software?+ Message-ID: <PbNJp6866UfO@eisner.decus.org>d  w In article <jh1Q5.15592$Gd7.877230@newsread2.prod.itd.earthlink.net>, "Thomas Williams" <tawilli@earthlink.net> writes:iM > Does anyone know where I can get a (preferably free) program to do some I/OpI > benchmarking for VMS? I'd like to compare VMS volume shadowing vs. ASCIeC > shadowing (3rd party) vs HSZ-70 disk mirroring and different RAID L > configurations.  I know HW shadowing offload the CPU and should be faster,I > but I'd like to see some numbers for my specific configuration.  Thanksd >  > Tom Williams >   	 	TESTDEV:a  G http://evaxsw.tuwien.ac.at/info/html/freewarev40-directory.html#Storaget  > 	Works for me and you can specify I/O size and number of I/Os, 	etc.p  ? 	When you say different RAID configurations.  How deep are youre? 	pockets?  How many controllers do you have?  Since you mentionh@ 	HSZ70, I'm assuming UltraScsi Hub so you have 70 MB/sec to playA 	with (I believe).  Do you have a lot of I/O opposed to bandwidthe 	needs?o  @ 	What you could do is buy RAID Software for OpenVMS, it is about> 	$1200 I suppose.  This will allow you to create software 0+1.  ; 	Tell us about your I/O condition.  Are you seeing a lot ofx< 	"wait I/O"  <== for the Unix folks.  i.e.  how do  you look   			$ MONITOR DISK/ITEM=QUEUE  1 	How bad off are you?  What would you like to do?i   				RobE   ------------------------------  % Date: Mon, 13 Nov 2000 21:47:38 +0000-' From: Brian@pickroad-demon.nospam.co.uk ' Subject: Re: Disk performance under VMSm8 Message-ID: <tmn01ts3il4p91hlfsk2nkjrsgt223ikiq@4ax.com>  E Have you got/turned on the writeback cache? We run a 4GL to RMS files E which is IO intensive on a Multisite cluster. Without writeback Cacheu@ at the controller level we would not get any work done. Also tryB upgrading the disks from 7200 RPM Ultra SCSI-2 disks to controller@ stripesets of 2 x 18Gb disks at 10,000 RPM or 4 x 9.1Gb 10,00RPM4 drives as this will split the load across splindles.  C I assume that both members of your shadow set are on different SCSI  busses at the contoller level. d  E Are you HSZ80 dual controllers as you could prefer each member of the E shadow set to a different controller but this will slow down a shadowi copy if you do them.  = On Thu, 26 Oct 2000 06:54:18 GMT, a_haines@my-deja.com wrote:i  E >We have some very performance critical systems that also need to logtH >information to disk. So, I have been doing some investigation into diskA >I/O performance (such as comparing RMS block I/O with Fast I/O).4 >3B >The configuration is VMS V7.1-2, 4 CPU ES40, 8192MB memory, HSZ80B >controller, OpenVMS shadow set of two Compaq BA03611C9B (17769177H >block) disks. The maximum_cached_transfer_size on the controller is set >to 127 for all disks. >aC >One test is to create a 500,000 block file at the beginning of the E >empty disk and fill it with zeros using 127 block writes (either RMSyD >block I/O or $IO_PERFORMs). I record the elapsed & CPU time for theF >write operations (not for file creation/deletion), then move the fileG >1,000,000 LBN's down the disk and repeat the test, repeating until thea >end of the disk.c >kG >It takes twice as long (35 seconds) to write the file at the end (highiE >LBN) of the disk than it does at the beginning (low LBN) of the diskt >(18 seconds). >h >Compaq have said: >hF >Data transfer at the beginning of a disk is faster then at the end of( >the disk. There are 2 reasons for this:G > 1- at the start there are more sectors per rotation, so in 1 rotationg# >more sectors/data can be accessed.sF > 2- at the end of the disk has less sectors per rotation resulting in >more headmovements. >-A >This is not what I expected. Does anyone have any views on this?d >a >h' >Sent via Deja.com http://www.deja.com/A >Before you buy.   ------------------------------  % Date: Mon, 13 Nov 2000 15:13:25 -0700r  From: Jon <jsmyth69@hotmail.com>1 Subject: Does CompaQ TCP/IP support secure shell?u2 Message-ID: <P2cQOsKHAJkafofJyjX4MvJ+bhVf@4ax.com>   Hi,i  E    Anyone know if CompaQ TCP/IP support use of a secure shell?  Also,_B if any add-ons are needed, could you give me a reference?  Thanks!   ------------------------------   Date: 13 Nov 2000 15:47:34 PSTT From: Fairfield@SLAC.Stanford.EDU (Ken Fairfield; SLAC: 650-926-2924; FAX: 926-3515)5 Subject: Re: Does CompaQ TCP/IP support secure shell? 3 Message-ID: <1AwJBebr0xcr@mccdev.slac.stanford.edu>u  2 In article <P2cQOsKHAJkafofJyjX4MvJ+bhVf@4ax.com>,'     	Jon <jsmyth69@hotmail.com> writes:   G >    Anyone know if CompaQ TCP/IP support use of a secure shell?  Also, D > if any add-ons are needed, could you give me a reference?  Thanks!  H         At present,  TCP/IP  Services  does  not  include  SSH  support,H     although  I've  heard rumors here that it will be added in a  futureE     release.  (Did anyone hear something more definite at CETS 2000?)a  H         The most recent release of  Process Software's Multinet includesH     both  an  SSH client and server (daemon), and their TCPware  productA     will also include these in the next release (soon I believe).   H         There is a freely available  SSH  Server  from Dave Jones at OSUH     (Ohio  State University) that we've been running for quite some timeH     now.  See http://er6s1.eng.ohio-state.edu/~jonesd/ssh/ .  This  willH     build  and  run  under TCP/IP Services (or the older UCX) as well asH     Multinet and TCPware.  We use  the  OSU  SSH Server in "production",H     but  on our busiest node, it "disappears" every few weeks and  needsH     to be restarted.  I haven't considered that sufficiently annoying to+     have tried to track down the problem...a  H         There is a freely  available  SSH  client  for VMS, called FISH.H     See  ftp://ftp.lp.se/vms/fishu1006.zip  (and fishu2006.patch if  youH     want to have a little more control of the enviroment  using  logicalH     names).   Also  see  http://www.free.lp.se/fish/  for  an  overview.H     Unfortunately, FISH does  _not_  do  X11  forwarding and the author,H     Richard Levitte, seems to have dropped further development, at least     at the moment.  H         There is a mailing list for  both  the  OSU SSH Server and FISH.&     Send a one-line e-mail message to:  "         MXserver@alpha.sggw.waw.pl       with the message:e       	SUBSCRIBE VMS-SSH             -Ken --M  Kenneth H. Fairfield            |  Internet: Fairfield@SLC.Slac.Stanford.Eduo:  SLAC, 2575 Sand Hill Rd, MS 46  |  Voice:    650-926-2924:  Menlo Park, CA  94025           |  FAX:      650-926-3515N  -----------------------------------------------------------------------------B  These opinions are mine, not SLAC's, Stanford's, nor the DOE's...   ------------------------------  % Date: Mon, 13 Nov 2000 17:51:00 -0600e7 From: "David J. Dachtera" <djesys.nospam@earthlink.net> # Subject: Re: f$search improvement ?i- Message-ID: <3A107E64.E9AB5A74@earthlink.net>a   Tim Llewellyn wrote: >   > "Jean-Franois Marchal" wrote: >  > > "eI > > > This is a one-off operation as f$search() has _all_ the informationaI > > > that its needs to answer the question. It does not need and so doeseJ > > > not maintain a search context after completing the action. RepeatingI > > > the command repeats the action. There is no previous context so thec > > > result is the same.  > >eI > > Could we imagine that giving a context as a second argument would let J > > f$search think it is called from within a loop, so returning "" at the" > > second call would be desired ? > >  > ( > Yes, I know where you are coming from. >  > [snip] >  > >a > > J > > Just as a background, I had written a procedure to help a user to findM > > its files in a directory (along with the shelvable attribute). EverythingoM > > worked well till a user type a name without wildcard that matched exactly C > > with a file name. The loop f$search was the key ran endless ...e > >S > K > Yes, I have had to insert a check for a wildcard in the search string and N > have two code paths, one with a loop (if "*" found in search string) and one8 > without (no "*" found) in some procedures I have made. > 3 > Unless we are both missing something obvious heres  B Well, I do something like this when a wildcard is possible but not guaranteed:o   $ PREV_FILE := $loop_label:= $ FSP = F$SEARCH( filespec_expression[, context_expression] ) + $ IF FSP .EQS. "" THEN GOTO loop_exit_label 2 $ IF FSP .EQS. PREV_FILE THEN GOTO loop_exit_label $ PREV_FILE = FSPE 	. 	. 	. $ GOTO loop_label  $loop_exit_label:c  , This has proven effective on many occasions.   -- r David J. Dachteral dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/r  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.o   ------------------------------  % Date: Mon, 13 Nov 2000 10:44:38 -0500i! From: "OHM" <ohm62@SFhotmail.com>c# Subject: Re: f$search improvement ?u* Message-ID: <8up1no$4r71@usilsuna.cai.com>  H > Just as a background, I had written a procedure to help a user to findK > its files in a directory (along with the shelvable attribute). EverythingOK > worked well till a user type a name without wildcard that matched exactly A > with a file name. The loop f$search was the key ran endless ...o  K Actually, you could scan the user specified file name for wildcard presence % and merely add one if there's none...n   ------------------------------  % Date: Mon, 13 Nov 2000 19:17:44 +0000a  From: steven.reece@quintiles.comC Subject: Re: Help... Need Software to Logoff users after 20 minutes H Message-ID: <OFA67CCD62.45E09565-ON80256996.00695021@qedi.quintiles.com>  I I go with Theo on this - I use WATCHER on an AlphaStation and haven't hadiK any problems yet.  The time period is adjustable, as is the eligibility for- getting logged out.- Steve.  ; Theo Jakobus (jakobus at iaf dot fhq dot de) wrote/quoted :  >>>Vinnie Saladino wrote:  >jB > Does anyond know the name of the company that sells software forF > Open-VMS that monitors users terminals for inactivity, and logs them > off after a time limit ? >m; >   Please e-mail responces to   vincent.saladino@gsxxi.coma	 > Thanks!h    E We use WATCHER on Alpha OpenVMS 7.2-1 systems since years without anyc problem and it's free!!!  < http://www2.wku.edu/scripts/fileserv/fileserv.com?WATCHER<<<   ------------------------------  % Date: Mon, 13 Nov 2000 17:19:35 -0400a- From: JF Mezei <jfmezei.spamnot@videotron.ca> $ Subject: Re: Malloc + debug question, Message-ID: <3A105AE0.3EE236B3@videotron.ca>   David Mathog wrote:/D > What makes you so sure that the space was not freed accidentally?   J The space was allocated very early on in the transaction, and the variableM used to store the pointer is not referenced until the end of the transaction. L But somewhere in the middle, malloc gave the same storage to another totallyI unrelated variable. I ended up with an internet address being stored in a ' located expected to hold a filename :-)g  B I the end, I fixed one bug which seemed to stop that problem. WhenM concatenating 2 strings, I read "realloc" one byte too few to account for the L space I was adding between the 2 strings. So I was writing a null beyond the" limit of what had been allocated.   J This had been "working" for about 2 years. But in all cases until now, theH fields requiring concatenation were freed as soon as they were parsed orK simply discarded. For instance, I would built the "Received:" multiline RFC?L field into a single string, and then just discard it. I would built the listK of TO: in a single string, parse it and store the TO: individually and then- free the long string.3  G But this time around, I had a Subject field that spanned 2 lines IN THEcO PREVIOUS MESSAGE, and the subject field persists until the end of that message.A    F But because I have no idea how malloc works, it is very hard for me toM understand why writing a null past your allocated space would cause malloc tosM get confused at a later stage in the program and provide an already allocatedo8 block. (of freeing the block without you asking for it).    N So, Is there any documentation on how malloc keeps track of how many bytes areI allocated to each block ? Is there a central table somewhere, or does theaM whole thing consist of a linked list of memory cells with control informatione7 stored startuing at the byte following the last block ?e   ------------------------------  + Date: Tue, 14 Nov 2000 00:34:28 +0100 (CET)n: From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>1 Subject: Re: NCP commands in the 7.2 Cluster doc?nJ Message-ID: <Pine.LNX.4.21.0011122206160.27368-100000@irys.stanpol.com.pl>  % On Wed, 8 Nov 2000, Hans Vlems wrote:a [...]t( + That said, the information you can get- +out of ncl is a lot more than NCP ever gave. H +Provided you can remember where you put the paper bits with the correct +invocations ;-)  <  Really - special effort, not less than command procedure isC expected for that simple info as NCP SHOW LINKS (or TCP SHOW DEVICEt@ and LAT SHOW LINK) if someone will find it in the tons of data !E  "The paper bit" is probably a overkill, but the % of HELP invocation > *is* much bigger than in any other VMS tools... (even, if less little than any un*x usage :]).O   +Hans Vlemst    Regards - Gotfryd   -- oE ===================================================================== F $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=MEt. $!                        GS@stanpol.zabrze.plE =====================================================================r   ------------------------------  + Date: Tue, 14 Nov 2000 01:05:15 +0100 (CET)c: From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>$ Subject: Re: Nested DCL symbol namesH Message-ID: <Pine.LNX.4.21.0011140030130.438-100000@irys.stanpol.com.pl>    Hello !9  The time is not appriopiate for long answers -;), but my3 minor word:r    On Fri, 10 Nov 2000, Fatz wrote: [...]wF +(2) CALL doesn't seem to do more than 1 level symbol substitution forE +parameters to a subroutine, whereas it does for the subroutine name.  [...example CALL vs. WRITE...], +Having said all that - any comments on (2)?  I  Yes. You can see my message in the "DCL symbol substitution hint needed"o subject.;  The "&" parsing is not *always* in case, but *only* if the- "command CLI request" is made.7  You can check, writting a simple C code where displaysa< in loop the parameters, what happens if you use "&" running  the code with:  - MCR, foreign name and DCL$PATH - a .CLD defined command7  and compare with example (b.ex. from FILE or somethinga3  freeware utility) where calls the CLI routines forr  parameter parsing.e   +  And why does & not workD +when there's a prefix?  It appears there's something wrong with DCL +because this works: +t +$ test& = 5 +$ sho sym test&2 +  TEST& = 5   Hex = 00000005  Octal = 00000000005    Ok.@  Here is much more "unsupported characters" where work, and have% hope COMPAQ will not prevent that -:)m)  In the past (VMS 4.X AFAIR) was allowed:e $ T"est"=15o9  and symbol Test was created (with *lowercase* in name), M= in other VMS version the symbol names T"est" - to be precise,a checked this moment (VMS 7.2): $ t"est"=15r $ sh sym t*r3   T"est" = 15   Hex = 0000000F  Octal = 00000000017  $ )  - but all this fiddling are unsupported.t  ?  You are *not* allowed use the special character in this place,u/ *or* you get some not wanted effect - period...-  :  And IMHO the user manual (mentioned here) warns about "&"< because the limitation - and hard way, to simple description "when & where it works" !m  I +That's probably why you can't substitute with prefixes because DCL isn'tt? +recognising the & as a special character when it appears laterf" +than position 0 in a symbol name.  =  Also. But that mean, that you can use anywhere a symbol with " the unsupported character in name: $ test&:=t.*   $ dir &test&   Directory DISK$USER:[GS]  O T.COM;9             T.COM;8             T.COM;7             T.T;1              ?  I +According to the docs, & shouldn't be a valid character in a symbol name.H +and to my mind, docs or no docs, it's a special character: it shouldn't? +be allowed in symbol, file, logical or any other kind of name.b    IMHO you are wrong !m  = - in symbol name - you right in the limited way, that you aska?  to force any unsupported character be a bug. IMHO not so good:g=  here *was* some query for non-alpha chars (my suggestion wasf:  b.ex. a "^") for some reason - and you ask to enforce the  resolution not work !  7 - in logical name *all* character are supported ! Even,D:  if the name of the part of architecture is "logical name";  it was not limited, like (in past) the file names or b.ex. :  user name. As b.ex. "process name" it is a string and can  get roughly any value(s), see:e  ; $ define "A name, where good & long=15" and value=5%DCL", -   "but the second is ' !"   WARNING:9  A name, to be *used* by the *automatic name translation*o8 in VMS kernel - really, must be "syntactically correct".6  But if a name is used for *other* reason, the name of the logical can be *any* !  , - in ODS-5 file names roughly no limitation.   +Fatz.C +All corrections, admonitions and exhortations gratefully received.f    Yes, you can count for me -:)  And the same applies !D    Regards for all
  - Gotfryd   --  E =====================================================================-F $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=MER. $!                        GS@stanpol.zabrze.plE =====================================================================w   ------------------------------  % Date: Mon, 13 Nov 2000 15:39:06 -0500i# From: Jim Agnew <agnew@hsc.vcu.edu>u4 Subject: Re: open VMS backdrop for motif or xwindows+ Message-ID: <3A10516A.A5EFD0B4@hsc.vcu.edu>P   First...   get and setup FETCH_HTTP   then   $ FETCH_HTTP -: "http://twister.sbs.ohio-state.edu/sat/GOES-8/us_ir.gif" -  inbound.jpg -B    then...   ! $ XV inbound.jpg -ROOT -MAX -QUITs  , see Hunter Goately's site for the proggies..   Enjoy!!!   Constandinos Stavrou wrote:n > M > Can anybody cprovide any interesting backdrops for modif or xwidnows?  I aml1 > stuck with the ones that came with openVMS 6.2.o >  > Thank you    ------------------------------  # Date: Mon, 13 Nov 2000 21:16:30 GMTe/ From: "Richard L. Dyson" <rick-dyson@uiowa.edu> 4 Subject: Re: open VMS backdrop for motif or xwindows) Message-ID: <3A1005CE.60FFDCE1@uiowa.edu>(   Jim Agnew wrote: > 
 > First... >  > get and setup FETCH_HTTP  = Fetch_HTTP comes with the OSU Webserver but is also available,8 as a stand-alone too. I think the last, best version is:  / 	ftp://ftp.lp.se/patches/osu-httpd/fetch_http.ce   > then >  > $ FETCH_HTTP -< > "http://twister.sbs.ohio-state.edu/sat/GOES-8/us_ir.gif" - >  inbound.jpg -Bh	 > then...d > # > $ XV inbound.jpg -ROOT -MAX -QUITk > . > see Hunter Goately's site for the proggies..  ( For a current version of XV for OpenVMS:  7 	ftp://www-pi.physics.uiowa.edu/~dyson/xv-3_10a-vms.ziph   > Constandinos Stavrou wrote:  > >aO > > Can anybody cprovide any interesting backdrops for modif or xwidnows?  I am 3 > > stuck with the ones that came with openVMS 6.2.g   rick -- tH Richard L. Dyson                                    rick-dyson@uiowa.eduH  _   _      _____                http://www-pi.physics.uiowa.edu/~dyson/H | | | |    |_   _|   Systems Analyst                     O: 319/335-1879H | | | | of   | |     The University of Iowa            FAX: 319/335-17536 | \_/ |     _| |_    Department of Physics & Astronomy-  \___/     |_____|   Iowa City, IA 52242-1479    ------------------------------   Date: 13 Nov 2000 14:40:27 PSTT From: Fairfield@SLAC.Stanford.EDU (Ken Fairfield; SLAC: 650-926-2924; FAX: 926-3515). Subject: Re: OpenVMS 7.1-2 and cluster forming3 Message-ID: <aXN71Y3w07I7@mccdev.slac.stanford.edu>m  7 In article <200011130852.JAA19452@sinet1.fom.fgan.de>, u-     	Rudolf Wingert <win@fom.fgan.de> writes:J  H         I often have a bit  of trouble understanding Rudolf's english...H     no  criticism,  just a warning of the potential for me to  misunder-'     stand the problem statement...  :-}-  H > since our upgrade of OpenVMS from 7.1-1H1 to 7.1-2 we can't reboot our > bootserver.   H         Did it _ever_ boot successfully after the upgrade?  Can you do aH     conversational boot so that you  can  at  least see the system disk?H     Is  the system disk locally attached (on a local SCSI, or an HSZ  orH     HSJ, etc.) or is it MSCP-served from another  node?   If  it  has  aH     local  system  disk,  then  you  should  be  able to boot the system:     standalone, e.g., assuming SYS0 is this system's root,           >>> b -fl 0,1p!         SYSBOOT> set vaxcluster 0o%         STSBOOT> set writesysparams 0u         SYSBOOT> continuen  H     The system will boot as a  stand-alone  system.  It won't be able toH     access  any  of the cluster-shared disks, etc., but it will let  youH     look at SYSGEN parameters and other items of interest on the  system	     disk.x  K >             In the previous version, we could reboot without any problem.r  H         Here I have a little trouble understanding your meaning.  Do youH     mean "reboot" as in "shutdown, then reboot automatically", or do youH     mean "boot", as in "starting a the >>> prompt, successfully boot theH     system"?  If  you  really  mean  _re_boot,  then  check  the consoleH     parameter  'autostart_action' to see what it is set to.  Choices areH     'halt', 'boot' and 'restart', and 'halt' would keep the system  fromH     rebooting  after  a  crash (but should be OK for a "normal" shutdown     with "reboot" requested).b  J > Sometimes a reboot of the bootserver is nesseray like dismounted/mountedH > failure, installed patches. Also should a after crash reboot function.I > IMHO, yesterday I found the case for our problem. A new forming policy.eH > After reboot I did see, that the bootserver sends a membership requestH > to one of the satellites. Then it lost the path (Closed virtuel port).K > The bootserver disconnects the network (transciever and speed LED go off)i > and nothing happens any more.   H         Would you please show  us  your  cluster  topology?  I'd like toH     know  each  system's  VOTES, what is EXPECTED_VOTES,  where  do  theH     SYSUAF and RIGHTSLIST, etc., reside, where is CLUSTER_AUTHORIZE.DAT,H     where is the system disk, is there more than  one  system  disk,  isH     there  a  quorum disk, are there both VAXes and Alphas, and anything     else you can think of.  G >                               In the previous version of OpenVMS, the G > bootserver did not send any membership request and all the satellites-6 > did send there membership request to the bootserver.  H         That's just  an  artifact.   _Any_  node  in  the  cluster _may_H     respond to a membership request.  It is _usually_ the last node thatH     joined  the  existing  cluster, although that can  be  difficult  toH     determine after many reboots of nodes...  It also depends on the the4     various cluster interconects that are available.  I >                                                      IMHO, that was theiH > right way, because it eliminates some problems, when the booting nodesM > lost the path to this member (the node hangs). Now, every member, includingeN > the bootserver, sends a membership request to the node with the highest name > (alphabetical order).t  H         That's the way  a  VMScluster  works.   Members  are members andH     _all_ participate in cluster communications.  There is nothing wrong	     here.e  H >                       In case of this, the bootserver can't reboot (it' > hangs), or the whole cluster reboots.   7         I don't understand this.  What exactly happens?a  J >                                       May be, the new policy is also theH > reason for the hang of a lot of clustermembers (8 of 30) after a whole > cluster "clueexit" crash.a  H         I don't believe  there  is  any  "new policy".  VMSclusters have'     worked like this for a long time...e  I >                           Is there any patch out, which do reenable theeJ > old behavior (that I can reboot the bootserver without any effect to theF > satellites)? I think the new behavior is a blame for OpenVMS and un-" > acceptable in a production area.  H         I think you're barking up the wrong  tree.  It may be a hardwareH     issue  with the bootserver's network device, or it may be a  clusterH     incarnation issue, or cluster ID/password problem, or a quorum  diskH     problem.  Have you looked through the list of ECOs for VMS 7.1-2 andH     applied any that rated 1 or otherwise apply to your environment?  [IH     generally  keep  that list on disk, but since upgrading to VMS 7.2-1:     last January, I haven't kept up with the 7.1-2 ECOs...  H         I ran  VMS  7.1-2  in  a  mixed-architecture, mixed-interconnectH     environment  of  35  nodes  for  a long  time  without  any  problemH     rebooting disk servers or satellites.  I fear you're looking in  theH     wrong  place for the problem.  But we need many more details to sort
     this out..           -Ken --  M  Kenneth H. Fairfield            |  Internet: Fairfield@SLC.Slac.Stanford.Edua:  SLAC, 2575 Sand Hill Rd, MS 46  |  Voice:    650-926-2924:  Menlo Park, CA  94025           |  FAX:      650-926-3515N  -----------------------------------------------------------------------------B  These opinions are mine, not SLAC's, Stanford's, nor the DOE's...   ------------------------------  % Date: Tue, 14 Nov 2000 13:27:11 +1100-/ From: "Phil Howell" <phowell@snowyhydro.com.au>4  Subject: Re: Printing blank page3 Message-ID: <Hz1Q5.31493$SF5.550532@ozemail.com.au>f  ? USERNET NEWSGROUP <MARK.MAZURETS@VISTACOM.COM> wrote in message9 news:%STP5.15$5i6.391@client...  > On@ > DIGITAL TCP/IP Services for OpenVMS Alpha Version V4.2 - ECO 1! > on a AlphaServer 1200 5/533 4MBi > running OpenVMS V7.1-1H2 >g. > I have a problem with a client, in printing.K > They print out invoices thru an interactive program in which each invoiceo isI > its own print job. Every invoice is preceded by a blank page, using andrJ > therefore wasting preprinted invoice stationery. I have read FAQ's, fromF > "Ask the Wizard" on the OpenVms.digital website, about this problem, howeversK > none of the solutions worked. I understand that TCP/IP services ver4.2 is H > the latest. What is the significance of ECO 1? Is this the latest ECO?I > I have made sure that the form and queue settings are correct, so as toe not,J > invite a page overflow problem (which would cause a blank page to follow thej > invoice, anyway).u > Here are the settings:? > Printer queue HKINV3, idle, on ACTIVE::"205.198.253.84:9100",k% > mounted form HKINV3 (stock=DEFAULT)r@ > /AUTOSTART_ON=(ACTIVE::"205.198.253.84:9100") /BASE_PRIORITY=4D > /DEFAULT=(FORM=HKINV3 (stock=DEFAULT)) /LIBRARY=HPDEVCTL Lowercase6 > /OWNER=[1,4] /PROCESSOR=UCX$TELNETSYM /NO_INITIAL_FF2 > /PROTECTION=(S:M,O:D,G:R,W:S) /SCHEDULE=(NOSIZE) >t > Form nameaF >                                        Number   ------   ------------ > HKINV3 (stock=DEFAULT)                  101 B > /LENGTH=50 /SETUP=(HKINVBOT) /STOCK=DEFAULT /TRUNCATE /WIDTH=132 >tI > I have tried defining a new logical: ucx$telnetsym_suppress_formfeeds 1 @ > I have tried inserting <esc>]vms;2<esc>\ into the reset moduleJ > I have tried bracketing the existing PCL in the setup module with <esc>P ande > <esc>\ > This is the setup: >eL <esc>P<esc>E<esc>&l2A<esc>&l5H<esc>&l1O<esc>&l51P<esc>&l1E<esc>&l60U<esc>&l51 > 0F<esc>&l7.72C<esc>(8U<esc>(s0P<esc>(s12H<esc>\c >eI > Note: I have since taken out reference to the reset module. I have also I > included /default=nofeed and /no_initial_ff in the queue settings. This G > resulted in blank pages being preceded by an invoice, but the printer  pullsu: > from the plain paper tray, rather than using stationery.L > I have also noted, that, with the current settings, if I print out a file, 3 J > times in quick succession, then only the first job is preceded by  blank > page.lI > The printer itself does not appear to have any settings associated with  the J > blank pages, and the same thing happens on another printer as well. Both aret > HP5000's.nF > The spool files themselves do not have any formfeed or other control! > characters to throw a new page.o) > I appreciate any input on this problem.w	 > Thanks.eJ This applies to ucx 5.0a and all the logical names are now tcpip$ not ucx$  H you can tell TELNETSYM to log everything that it writes to, and receives" from, the TCP/IP link by entering:  ( $ DEFINE /SYSTEM TCPIP$TELNETSYM_DEBUG 4   also   TCPIP$TELNETSYM_RAW_TCPrC Suppresses all TELNET type modifications of the print output streamn   Phil   ------------------------------  + Date: Tue, 14 Nov 2000 00:33:59 +0100 (CET) : From: "Gotfryd Smolik, VMS lists" <gotfryd@stanpol.com.pl>( Subject: RE: RECALL question/suggestion.J Message-ID: <Pine.LNX.4.21.0011122116530.27368-100000@irys.stanpol.com.pl>  ' On Thu, 9 Nov 2000, John Vottero wrote:    +> -----Original Message-----sC +> From: koehler@eisner.decus.org [mailto:koehler@eisner.decus.org]  +>  - +> In article <8trmdj$t8p$1@nnrp1.deja.com>, e$ +> richard_maher@my-deja.com writes: [...] H +> > great interest, but in your example is it a case of the sub-process3 +> > having it's own "empty" command recall buffer?s  	  Exactly.    +> Appearently not:  +>   +> $recall/all
 +>   1 sho us  +>   2 sho mem
 +>   3 dir +> $pipe recall/all=
 +>   1 sho us= +>   2 sho mem
 +>   3 dir  -  But "PIPE" is not a synonym of "suprocess" !L9  Be aware of the fact, that "pipe subprocess" is started S
 only when:
 - you use "&"=
 - you use "|"= - you use "()" pair.  C  PIPE with command separated ";" or "logical sequences" - && or ||.,9  You are pleased check (in place of SHOW PROC you may useV8 WRITE SYS$OUTPUT F$GETJPI(0,"PID") to get short output):   $ PIPE SHOW PROC ; SHOW PROC $ PIPE SHOW PROC ; SHOW PROC $ PIPE (SHOW PROC)? $ PIPE SHOW PROC | SHOW PROC	! The first command output is lost ! $ PIPE SHOW PROC | TYPE SYS$PIPE:i  C +> I've verified that recall/all writes to sys$output by redefininghA +> sys$output, so it should get connected to sys$input via pipe. e  +>  Don't know what's happening.  8  - exactly the PIPE subprocess has empty RECALL buffer !  E +PIPE seems to be smart enough to see when it's not needed.  Try this 
 +sequence: +n +$ PIPE RECALL/ALL  +$ PIPE RECALL/ALL | SHOW SYSTEM + K +Note that only one of those two commands is in your current recall buffer.t  D  That is supported & expected RECALL behaviour and has nothing to do with PIPE command !!;  *If* a command executes the RECALL command, then the wholeC' command is not saved in command buffer.m1  The "whole command" can be a complex one, check:t $ if "false" then recall/all $ if "true" then recall/all   ;  and you see what I mean: only if RECALL is *executed*, ther?  command is skipped in RECALL buffer - check, that [if "false"]t#  *is* saved, but [if "true" - not !=  *  *Also", if you create pseudocommand like: $ X="RECALL"=  the "changed in view" command - when RECALL is executes - isM
 not saved.    *The same* applies for PIPE !  #  And in your examples all is clear:1< - without "|" (or as have before say) - the RECALL executes,  and the command is not savedo8 - with "|" the RECALL is *not* executed *in the process*9  (like you do b.ex. SPAWN RECALL) - then the execution is00  out of control of the RECALL buffer control... B (och.... the RECALL controls themself, really hard to describe !).    Regards - Gotfryd   -- oE =====================================================================nF $ ON F$ERROR("LANGUAGE","ENGLISH","IN_MESSAGE").GT.F$ERROR("NORMAL") - 		THEN EXCUSE/OBJECT=MEn. $!                        GS@stanpol.zabrze.plE =====================================================================g   ------------------------------    Date: 13 Nov 2000 14:45:59 -0500/ From: jordan@lisa.gemair.com (Jordan Henderson)y; Subject: Re: Research Systems Inc. dropping support for VMS * Message-ID: <8upgdn$4jq$1@lisa.gemair.com>  ) In article <3A0FA810.3B8EC8EB@uiowa.edu>, . Richard L. Dyson <rick-dyson@uiowa.edu> wrote: >spolato@my-deja.com wrote:  >> o6 >> In article <3A016BF8.56B821D8@toyvax.Tucson.AZ.US>,7 >>   Vance Haemmerle <vance@toyvax.Tucson.AZ.US> wrote:rA >> > Just received my IDL 5.4 cover letter that says that versioniD >> > 5.4 is the last version to support OpenVMS.  That's certainly aD >> > kick in the pants to scientific computing on VMS.  RSI was just= >> > bought by Kodak, maybe that had something to do with it.a >> > >> > --s >> > Vance Haemmerle >> > vance@toyvax.Tucson.AZ.US >> > >> fH >> We have a complex scientific computing environment based on VMS - and% >> IDL is a strategic product for us. I >> That RSI cover letter announce make us considering to migrate to an OS 2 >> where IDL will be still supported (unix/linux). >> aA >> It is a sad way to say "goodby VMS" after more then 15 years !e >> "G >> I'd like to know whether Compaq has intention to accept the dropping G >> support of IDL on VMS as a fate or intend to find an accord (money!)o+ >> for the future development of IDL on VMSl >cG >We use it too, but I would never consider dropping the OS just becausei >of the loss of an application!f > G >Why don't you contact RSI/Kodak and complain about the loss of OpenVMS  >support instead!u >   I Do both!  See if you can get Compaq interested and complain to RSI/Kodak.g  J Andrew is right this.  Compaq should do more to encourage ISV support for 5 OpenVMS if they expect OpenVMS to be taken seriously.G   >Rick_ >-- I >Richard L. Dyson                                    rick-dyson@uiowa.edu I > _   _      _____                http://www-pi.physics.uiowa.edu/~dyson/=I >| | | |    |_   _|   Systems Analyst                     O: 319/335-1879 I >| | | | of   | |     The University of Iowa            FAX: 319/335-1753e7 >| \_/ |     _| |_    Department of Physics & Astronomy6. > \___/     |_____|   Iowa City, IA 52242-1479   -Jordan Hendersonh jordan@greenapple.comh   ------------------------------  % Date: Mon, 13 Nov 2000 18:10:17 -0600 7 From: "David J. Dachtera" <djesys.nospam@earthlink.net>u; Subject: Re: Research Systems Inc. dropping support for VMSd- Message-ID: <3A1082E9.70517085@earthlink.net>e   Jan Vorbrueggen wrote: > . > "Richard Brodie" <R.Brodie@rl.ac.uk> writes: > N > > Of course, this is computing. It would be confusing if an abbreviation had! > > only one meaning. DNA anyone?M > G > Why - is there any other meaning than "Digital Network Architecture"?a   Deoxyribonucleic acid?   -- s David J. Dachtera0 dba DJE Systems( http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/a  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  # Date: Mon, 13 Nov 2000 21:49:42 GMT 2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>> Subject: Re: Resurrecting an Alphastation 200/250: help needed3 Message-ID: <WlZP5.15$QV2.2457@typhoon.aracnet.com>d  0 In comp.sys.dec Roy Omond <Roy@omond.net> wrote:0 > Anyone able to suggest what I should do next ?  L This answer isn't very helpful, but this problem sounds familiar.  Have you M tried a DejaNews search at http://www.deja.com/=dnc/ ?  I think this has beeniL discussed in the last few months.  It might even be in the OpenVMS FAQ (I've* not checked, but it wouldn't surprise me).   				Zane   ------------------------------  % Date: Mon, 13 Nov 2000 17:30:30 -0400i- From: JF Mezei <jfmezei.spamnot@videotron.ca>  Subject: Re: RWAST, Message-ID: <3A105D6E.E4DCE003@videotron.ca>   Javier Henderson wrote:rI >         . Hack something to deliver the AST the process is waiting for, 4 >           then use a regular process stopping gun.  N Does RWAST mean that the process is waiting for an IO to complete while in AST mode ?  G If an AST routine does a QIO to a mailbox device that is full, does the > process become RWAST or RWMBX ? or is there a RWASTMBX state ?  A Is RWAST really a LEF state for when a wait is done in AST mode ?n   ------------------------------  % Date: Mon, 13 Nov 2000 18:01:49 -0600e7 From: "David J. Dachtera" <djesys.nospam@earthlink.net>v Subject: Re: RWAST- Message-ID: <3A1080ED.9DAC72A9@earthlink.net>    Martin Francois wrote: > " > I have a process in RWAST state. > How can I stop it ?????s  	 1. RebootdD 2. Spend some quality time with SDA (ANALYZE/SYSTEM) and try to find< what it's waiting for, then try to find a way to deliver it.   -- s David J. Dachtera  dba DJE Systemsa http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/Y  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.G   ------------------------------  % Date: Mon, 13 Nov 2000 19:31:13 +0000 8 From: John Macallister <J.Macallister1@physics.ox.ac.uk>6 Subject: Scarey use of Technology in the US Elections.L Message-ID: <35666012DF4CD411BE940090279FA240111FB6@ppnt41.physics.ox.ac.uk>  K Supposing the US Elections has been run using some of the latest technologypL rather than, in computer terms, the Stone Age tools of punched cards imagine  some of the potential scenarios.  I If the voting and counting had been conducted on Linux systems, with eachoG state free to choose its version of Linux each state would be open to aa? different set of exploits from the unscrupulous. Realising thishK inconsistency, both candidates would demand a recount. However, by the timelK they were ready to begin the recount there would be at least one new set ofsK essential patches for the OS. Because they couldn't rerun with the original J OS version owing to the possibility of an exploit and because they couldntJ conduct the recount with the patched OS as it would behave differently theL only option would be to rerun the election again ... and again ... and again ...   I If the voting and counting had been conducted on WindowsNT/2000 all would J proceed with smoothness. After all, nothing could possibly go wrong as theL procedure had been tested several times with real people. However, after theK polls closed, and the votes were read back from disk it was discovered that J all candidates had identical totals - 0 votes. It was quickly realised, ofK course, that the data had not been saved to disk during voting in order not K to impair response times and nobody had thought of doing it after the pollsnH closed. Why would they? The trials had worked fine with reading the dataK back in from memory: aren't memory and disk the same thing? There was now a B serious decision to be made: take the vote as it stood to save theH Constitution or rerun the election again. States would be free to chooseG whether or not they saved data to disk and so the election could be runo" again ... and again ... and again.  L If the voting and counting had been run on VMS there would be no problems atI the polling stations. There would be no problems with counting the votes. I As there was one single National VMS Cluster all the voting data could beeL held in one place. There would be no problems with protocols. You could haveF any protocol you wanted and each state would choose TCP/IP, DECnet IV,I DECnet/OSI, LAT, SEND/FOREIGN with mail, SMTP, PMDF, All-in-1, All-for-1,mH One-for-All, etc according to taste. Each state would choose a differentJ cluster group number and password to assert their independence. Each stateL would accurately declare its result and each state would report the nationalL total on local news. There would be great rejoicing. Someone would happen toJ hear the (different) result reported from another state and the first hintL that all was not well would emerge. Each state's node had, of course, bootedH up into its own cluster with its own copy of what appeared to be the NVDJ (National Votes Disk). The folk in DC had, of course, already declared theL winner and loser and, as far as they were concerned, the president-elect hadK already been chosen. There was a Constitutional crises and the election wasoJ rerun again ... and again ... and again ... Although all the states agreedJ to use the same cluster number no state would divulge its cluster password to another state ...     John  B Name: John B. Macallister  E-mail: j.macallister1@physics.ox.ac.ukH Post: Nuclear and Astrophysics Laboratory, Keble Road, Oxford OX1 3RH,UKA Phone: +44-1865-273388 (direct)  273333 (reception)  273418 (Fax)1   ------------------------------  % Date: Tue, 14 Nov 2000 01:07:49 -0500t2 From: rdeininger@mindspring.com (Robert Deininger): Subject: Re: Scarey use of Technology in the US Elections.L Message-ID: <rdeininger-1411000107500001@user-2ivebim.dialup.mindspring.com>   In article <35666012DF4CD411BE940090279FA240111FB6@ppnt41.physics.ox.ac.uk>, John Macallister <J.Macallister1@physics.ox.ac.uk> wrote:  M > Supposing the US Elections has been run using some of the latest technologysN > rather than, in computer terms, the Stone Age tools of punched cards imagine" > some of the potential scenarios.  Y Check back with me in a couple or days, and I'll tell you whether this is funny or not...:   :-)5   -- 5 Robert Deininger rdeininger@mindspring.com-   ------------------------------    Date: 13 Nov 2000 16:39:03 -0500* From: young_r@eisner.decus.org (Rob Young)7 Subject: Re: Severity of eBay's ecache problem detailed1+ Message-ID: <BrJB5SmP+QCh@eisner.decus.org>y  ] In article <3A0FFC6D.3DE5CA24@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:u > Rob Young wrote: >> M` >> In article <3A0C0865.497FBF17@uk.sun.com>, andrew harrison <andrew.nospam@uk.sun.com> writes:+ >> > "Brian Schenkenberger, VAXman-" wrote:- >> >> ^ >> >> In article <vXePVDXmGW4S@eisner.decus.org>, young_r@eisner.decus.org (Rob Young) writes: >> >> >{...snip...} * >> >> >       Come spin with us Andrew!!!! >> >> >n) >> >> >                               Robo >> >> >i >> >>iP >> >> Rob, really!  Andrew's contribution here is to tout Sun and FUD Compaq andO >> >> VMS technologies.  You continue to invite him.  I know that you and KerrytN >> >> and several others are trying to dispell his FUD and show what a FUDsterN >> >> he is himself but the truth is that I believe we all know what a FUDster- >> >> he is.  Can we try to stop baiting him?a >> >>o >> > >> e >> >I >> > Just occasionally someone pops out of the woodwork and says no thatso/ >> > B****H**T but only very very occasionally., >> > >> > No wonder I stayed. >> >J >> > You have mostly Rob to blame for this so you are posting to the right >> > person. >> >F >> > Incedentally despite labeling me as a FUDSTER and a spin merchantB >> > the fact is that my objections to mainly Robs hype on Galaxy,M >> > Spiralog, ISV interest in OpenVMS, the performance of the WildFire boxesM >> > etc have been correct.Z >> > >> oG >>         No ... I made a few mistakes.  I bought into Spiralog and it-I >>         is/was great technology.  I stand fast on the other things you=P >>         sling at me.  I am ahead of the curve on Wildfire and Son-of-WildfireM >>         but faster 21264 and 21364 are a long time coming (and yes, otherssN >>         aren't standing still but Compaq is doing VERY well in the high-endH >>         as they and IBM are suddently the platform of choice in HPC). >> cJ > No Rob Spiralog wasn't great technology, it was a great white paper. Had
 > it been ) > great technology it might have worked.      > 	It was great technology and it does work.  Folks are using it( 	today, it doesn't have a future though.  " It also illustrates perfectly the B > problem you seem to have, all the things you refer to, Spiralog,
 > WildFireC > etc have been great white paper material its the translation intos
 > products > thats been the problem.s >   ! 	No... Wildfire is a product too.q  G > You were probably ahead of the curve on WildFire for the same reason,  > you I > read too many white papers published about a product that didn't exist.l >    	Both products exist.   H > You and Jordan also appear to share the same rather cavalier attitiude	 > towardsvH > time lines, you I think pre-announced that WildFire was going to whip J > everyones asses nearly 2 years before it actually shipped. Ahead of the ' > Curve or Totally Jumped the Gun ?????e >   ? 	Ahead of the curve.  Jumping the gun is when Sun presumed theyt: 	would win the would win the Lawrence Livermore ASCI Q bid 	and they didn't.  T  ; http://www5.compaq.com/hpc/tsn/iss017/hptc_iss017_fa.html#1p    I > And ahead of the curve assumes that you think that WildFire is going to-H > get fixed ? What changes are Compaq going to make to WildFire's memoryN > subsystems to reduce local memory latency and reduce the ratio between local > and remote memory latency ?- > ; > What predictions are you making for Son of WildFire ?????> >   6 	Higher bandwidth, lower latency.  Much lower latency:  , http://www.alphapowered.com/ev7-and-ev8.html    G >>         But yes, you are a FUDSTER and a spin merchant as you rarelypG >>         if ever address the contents of a post and typically lie lowiF >>         until the contents have shifted ever so slightly to warrantR >>         a total change of subject.  You sure you didn't co-invent the Internet? >> >I >>         By the way, regarding the subject line, how about this anyhow:l >> >7 >> http://public.wsj.com/sn/y/SB973636942369499535.htmls >> aK >> At online auctioneer eBay Inc., a major Sun customer, the memory problemsP >> initially resulted in one to two big-system crashes every month, says MaynardP >> Webb, president of eBay Technologies. The memory problem wasn't behind eBay'sO >> most serious outage, a 22-hour failure in 1999, but has caused other service O >> outages, Mr. Webb says, although he adds that with Sun's fixes in place eBay 0 >> can now go "months" without seeing a problem. > I > How remarkable you actually included the punchline in one of your postsv > instead of cutting it out. e > K > "he adds that with Sun's fixes in place eBay can now go "months" without i > seeing a problem." o > ! > Remarkable and a first for you.y >   > 	Andrew, that's not something to be proud of.  Boxes shouldn't4 	crash every few months because of CPU cache issues.    O >>         Do something and shock us.  Admit that you have attempted to totallylI >>         mislead us for the past year on this issue.  Perhaps you thinksN >>         it will simply go away or will not longer make headlines?  In lightH >>         of the above, how do those postings in December 1999 now makeG >>         you look?  No different.  "We" essentially knew all along it,L >>         wasn't a matter of datacenter temperature or environmentals.  ButF >>         you wouldn't give it up.  Now you are forced to give up theG >>         datacenter spin but it is all out there for us to see and torH >>         trot out as we wish.  But you are oh so sure of yourself whenH >>         you write some things and the DANGER is kids that are lurkingK >>         wouldn't have a clue that what you pen is often VERY misleading.  >> uI >>         Your reputation took a hit.  Reputation?  Well.... never mind.o >> eE > Despite your claim that I havn't answered your questions Rob I haveu@ > in fact explained this to you a number of times and each time @ > you either cannot or will not understand what the issues are.  >   E 	Sure I understand the issue.  You dodge every question put to you.  =: 	Ask Jordan.  Ask Kerry.... ask anyone else in this forum.   > ; > 1.	There is a problem with commodity SRAM supplied by one)7 > 	vendor to Sun. This problem causes bits to flip whenS2 > 	they should not. These can be multi-bit errors. >     ? 	So... if as John Shoemaker relates Sun hadn't made the blunder B 	of using ECC L2 cache in their servers this would be a non-issue.C 	IBM, HP and Compaq choose to use ECC L2 in their high-end servers.     D > 2.	The bits don't just flip by themselves they require an external& > 	force, static, gamma radation etc.  > 9 	Trying to sneak static in there on us?  How about gamma.   ) > 3.	The fixes for the above problem are.. > F > A.	Reduce the incidence of flipping be reducing the incidence of theF > 	external force that causes the bits to flip, e.g make sure that theF > 	systems are not running in an environment where there is excessive  > 	heat/static.  >    	Won't eliminate the problem.     > B.	Install the cache scrubber. >    	Won't eliminate the problem.>  ; > C.	Change the cache to ecache made by a different vendor.S >   B 	Or do as your top 3 competitors have done.  Design ECC correction, 	in from the outset as Sun did for the UIII.   > D > For most customers actions A and B have resulted in either a major > reduction + > in or the elimination of ecache failures.  >   @ 	Reduction yes.  Elmination, no.  Hence as John Shoemaker states6 	mirror cache will soon be available for the UII CPUs.  J > Some customers cannot do much about A and so their best option is B and  > C. > 0 > All this information is in my previous posts.  > G > I don't know which of the three eBay applied but as they said and youeE > so kindly reproduced the course of action recommended by Sun fixed I > their problem. > E > You of course have tried to rubbish all the remedial actions which u; > is I assume why you claim I havn't answered the question.r > I > The problem origionally surfaced with Sun's 400 Mhz 4 MB ecache modules H > and at the time the SRAM supplier claimed to have rectified the issue. > H > Sun replaced 400 Mhz 4 MB cache modules with new 400 Mhz 8 MB modules B > only to discover that the problem that was supposed to have been > rectified_C > was still there in the 8 MB modules. We are now using a different 
 > supplier > of SRAM. o >   @ 	Too bad.  Testing might have revealed that.  But the real issue= 	is ECC L2.  As you pointed out in another thread, be careful4A 	what we say as Compaq was using the same SRAM.  But Compaq isn't-8 	having problems with it are they?  No.  ECC masks that.  E > For some time Sun and our first SRAM supplier did not know why the :H > SRAM was suceptible to having the bits flip, we have however known for > some r< > time what the external forces are that make the bits flip. >   @ 	Sure.  Gamma effects have been around for a while.  Hence, your> 	wiser competitors have been using ECC L2 for quite some time.  C > You of course don't want to contemplate the fact that it was youro
 > FUDDING K > Sun over eBay and as you appear to have admitted your ludicrous boosting (F > of Compaq future products which got me involved in this group in the > firsto > place. >   ? 	No... this isn't the case.  As you pointed out in another post A 	it was eBay issues that got you involved.  And of course, JordaneB 	points out you have been posting to comp.os.vms for quite a while< 	before that.  So the story changes ever so slightly to say:= 	"[Rob's] boasting got me involved."  Once again we catch youo@ 	changing your story in realtime mode.  You didn't co-invent the 	Internet , did you?   F > Now I am here you still don't seem to have learnt your lesson as the	 > Marvel/7J > Son of WildFire posts show, at one point recently you actually published > the A > expected Marvel memory latencies (from a white paper no doubt).h >   E 	Son-of-Wildfire they are in the URL above and yes they will be that   	good or better.    G 	Regarding Wildfire, what were those great latencies I published?  You i? 	see of course that most of us are willing to cut and paste to t; 	substantiate what we say.  You dredge something up with no B 	supporting evidence at all.  Recalling posts from the top of your> 	head.  But you are right , I did post latencies and there was$ 	nothing wrong with those latencies.   				Robr   ------------------------------  + Date: Mon, 13 Nov 2000 15:29:31 -0400 (EDT) ( From: Alan Simon <SIMON@SHRSYS.HSLC.ORG> Subject: SSU License Pak0 Message-ID: <01JWHVS7EXT68ZGILA@SHRSYS.HSLC.ORG>   Hello -T  G Now that Compaq has dropped SSU (Session Support Utility) from the CSLGdK program as a discontinued product, is a perpetual license PAK available fors	 download?l   Thanks for your help.w  O -------------------------------------------------------------------------------0@ Alan Simon                                      simon@hslc.org  E Associate Director                              simon@shrsys.hslc.orgkG Health Sciences Libraries Consortium            VOICE: (215) 222-1532   E 3600 Market Street, Suite 550                     FAX: (215) 222-0416t Philadelphia, PA  19104dO -------------------------------------------------------------------------------s   ------------------------------   Date: 13 Nov 2000 13:21:34 PSTT From: Fairfield@SLAC.Stanford.EDU (Ken Fairfield; SLAC: 650-926-2924; FAX: 926-3515) Subject: Re: SSU License Pak3 Message-ID: <ksoZpj5vD0bM@mccdev.slac.stanford.edu>t  1 In article <01JWHVS7EXT68ZGILA@SHRSYS.HSLC.ORG>, f/     	Alan Simon <SIMON@SHRSYS.HSLC.ORG> writes:a  I > Now that Compaq has dropped SSU (Session Support Utility) from the CSLG M > program as a discontinued product, is a perpetual license PAK available form > download?i  H         They are available, I  don't  know about downloadable.  With theH     most  recent  set  of  CSLG  PAKs we  also  received  a  file  namedH     CSLGVMS-MASTER-RETIRED.TXT, which is actually  a  command  file  andH     contains  non-expiring  PAKs for a wide variety of retired products.7     I just checked and verified that SSU is included...c           -Ken -- tM  Kenneth H. Fairfield            |  Internet: Fairfield@SLC.Slac.Stanford.Edu :  SLAC, 2575 Sand Hill Rd, MS 46  |  Voice:    650-926-2924:  Menlo Park, CA  94025           |  FAX:      650-926-3515N  -----------------------------------------------------------------------------B  These opinions are mine, not SLAC's, Stanford's, nor the DOE's...   ------------------------------  % Date: Mon, 13 Nov 2000 22:42:37 -0600h+ From: "Main, Kerry" <Kerry.Main@compaq.com>b' Subject: RE: Sun "uptime" belly laugherdN Message-ID: <910612C07BCAD1119AF40000F86AF0D8052849C5@kaoexc3.kao.cpqcorp.net>   Robert,   K >>> You are defining "user impact" too narrowly.  Some users of VMS will bee impacted, some will not.<<<   6 I expressely stated the following in a previous reply:I "While there are exceptions where some users staying logged into a systemn9 for days, most shops would say this is a security risk. "   K Does this not state clear enough that this is not a solution for everyone? l  K I appreciate that you are very likely what might be considered a power usereJ (programmer, developer ect) who would likely object to any IT group takingJ steps like this, my point is that a OpenVMS Cluster Customer can implementH this solution with no application middleware or changes to their cluster aware application. e  K If there are no multi-day sessions or batch jobs, then this is an excellentdA solution for being able to shutdown systems with ZERO application  availability impact.  : This is NOT fantasy as we have Customers doing this today.  7 >>> Don't tell them they don't know their own needs.<<<.  C I don't. They tell me their requirements and I recommend solutions.u  F Why do you think there are so many different third party products like
 "Hitman"?   F Obviously, many Customers feel (as I do) that multi-day sessions are a? security risk and are willing to pay $'s to resolve this issue.p  L You seem to feel long multi-day sessions are not a security risk. Fine. Lets agree to disagree. n  L What constitutes a security risk could lead to a discussion that could go onJ for days. Each site needs to determine what is best for their environment.   Regards,  
 Kerry Main Senior Consultante Compaq Canada Inc. Professional Servicese Voice: 613-592-4660s Fax  :  819-772-7036 Email: Kerry.Main@Compaq.com             -----Original Message-----B From: rdeininger@mindspring.com [mailto:rdeininger@mindspring.com]  Sent: November 13, 2000 11:22 AM To: Info-VAX@Mvb.Saic.Com ' Subject: Re: Sun "uptime" belly laugher     
 In articleJ <910612C07BCAD1119AF40000F86AF0D8052849BD@kaoexc3.kao.cpqcorp.net>, "Main,% Kerry" <Kerry.Main@compaq.com> wrote:e    e > and NO@ > user impact. Just re-read previous postings on how to do this.  ? Sigh.  It was ME, not Andrew, who started out objecting to yoursK over-statement.  I ignored it for months, but it finally got under my skin.a  G You are defining "user impact" too narrowly.  Some users of VMS will be I impacted, some will not.  All the other log out/don't log out change yourl4 password/lock your screen stuff is beside the point.  G Any work model that requires very long, uninterrupted "sessions" cannot+I survive the shutdown of the VMS node where the sessions are running.  You-J prefer to think in terms of "sessions" that are very short, and situationsG where new sessions can be routed to a different node on the fly.  Fine. / None of your arguements apply to long sessions.1  L It would be more comfortable if you would stop claiming more for VMS than isL real.  While there are valid reasons to avoid a "very long session" model ofI computing in many cases, Compaq should not think that all their customerssJ fit into this week's slick marketing campaign.  You can't convince me thatK VMS does something that my own experience tells me it does not.  The danger E is if you, and everyone else at Compaq, start to convince yourselves.i  J Listen to your customers.  Even if they don't agree with you.  Even if youI can't immediately remedy their concerns.  Don't tell them they don't knowr their own needs.  E Note.  Nothing in this post should be considered a defense of Andrew.t   --   Robert Deininger rdeininger@mindspring.com    ------------------------------  % Date: Tue, 14 Nov 2000 00:32:09 -0500@2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Sun "uptime" belly laugher L Message-ID: <rdeininger-1411000032090001@user-2ivebim.dialup.mindspring.com>  { In article <910612C07BCAD1119AF40000F86AF0D8052849C5@kaoexc3.kao.cpqcorp.net>, "Main, Kerry" <Kerry.Main@compaq.com> wrote:>    4 > If there are no multi-day sessions or batch jobs,    ^^Aha!  At last, the word I was looking for.  As long as you give me an "if", I don't have any big complaints with your statements.  Just remember that if, and others that might crop up occasionally.  You don't want us to think you are some kind of fanatic! :-)e     > then this is an excellentoC > solution for being able to shutdown systems with ZERO application  > availability impact.  N > You seem to feel long multi-day sessions are not a security risk. Fine. Lets > agree to disagree.   I didn't say that.  I only said that multiple-day sessions are not compatible with forced reboots.  I don't know who dragging in the security red herring.( They are two completely separate issues.  e (I did say that multi-day sessions are not a security risk, IF the system is locked when unattended.)    -- e Robert Deininger rdeininger@mindspring.comp   ------------------------------  % Date: Mon, 13 Nov 2000 17:56:36 -0600e7 From: "David J. Dachtera" <djesys.nospam@earthlink.net>o# Subject: Re: System disk filling upa- Message-ID: <3A107FB4.12BB3671@earthlink.net>    Christoph Gartmann wrote:s >  > Hi,n > I > under OpenVMS 7.1-2 I noticed today that the system disk is filling up.dJ > I have diskquota enabled and today the ERRFMT process crashed because it > had no disk quota. [snip]   F Quotas on the system disk are generally not recommended. The reason is' because of the experience you just had.e   -- e David J. Dachtera  dba DJE Systems  http://www.djesys.com/  : Unofficial Affordable OpenVMS Home Page and Message Board: http://www.djesys.com/vms/soho/w  F This *IS* an OpenVMS-related newsgroup. So, a certain bias in postings is to be expected.  @ Feel free to exercise your rights of free speech and expression.  F However, attacks against individual posters, or groups of posters, are strongly discouraged.    ------------------------------  % Date: Mon, 13 Nov 2000 19:11:24 -0500e  From: John Santos <JOHN@egh.com># Subject: Re: System disk filling upa5 Message-ID: <1001113191002.2093A-100000@Ives.egh.com>   ) On 13 Nov 2000, Christoph Gartmann wrote:h  P > In article <8uofmk$ehp$1@n.ruf.uni-freiburg.de>, I (Christoph Gartmann) wrote:J > >under OpenVMS 7.1-2 I noticed today that the system disk is filling up.K > >I have diskquota enabled and today the ERRFMT process crashed because itr2 > >had no disk quota. The actual quota setting is:* > >   UIC [SYSTEM] has 2965732 blocks used6 > >   of 2700000 authorized, 5000 permitted overdraft.N > >SYSTEM has BYPASS privilege. Page- and swapfile are on a separate disk, theM > >same is true for SYSUAF.DAT and similar files. In addition OPERATOR.LOG isi* > >purged daily, the same for ERRLOG.SYS .O > >I looked for files larger than 5000 blocks but found nothing explaining whataN > >is happening here. Next, I looked for files created since November 1, againA > >nothing uncommon. An "ANALYZE/DISK/REPAIR" didn't help either.aO > >Thus, what else can I try to find out what is filling up SYSTEM's diskquota?  > L > I solved it: I didn't think about an open file and therefore missed a huge7 > file created by some DCL procedure. Unfortunately theUL > DIR/SELECT=(SIZE=(MINIMUM... doesn't offer the possibility to look for the > allocated size.y  . $ DIR/SIZE=ALLOCATION/SELECT=(SIZE=(MINIMUM...    does look at the allocated size.  
 > Regards, >    Christoph Gartmannt   -- s John Santost Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------  % Date: Mon, 13 Nov 2000 21:43:05 -0500l, From: Howard S Shubs <hshubs@mindspring.com># Subject: Re: System disk filling up,> Message-ID: <hshubs-9528AC.21430513112000@news.mindspring.com>  J In article <8uoqns$ot7$1@nnrp1.deja.com>, Uwe Zessin <zessin@my-deja.com>  wrote:   >Famous last words, #7214:" >"Backups? We have a RAID system." >;-)   HAHAHAHAHAHAHA....  good 'un!9 -- t Howard S ShubsD "Run in circles, scream and shout!"  "I hope you have good backups!"   ------------------------------  % Date: Tue, 14 Nov 2000 16:29:21 +0010 % From: paddy.o'brien@zzz.tg.nsw.gov.aud# Subject: Re: System disk filling up 5 Message-ID: <01JWJC4HLR82006K4B@tgmail.tg.nsw.gov.au>h   David J. Dachtera wrote:   >Christoph Gartmann wrote: >> i >> Hi, >> "J >> under OpenVMS 7.1-2 I noticed today that the system disk is filling up.K >> I have diskquota enabled and today the ERRFMT process crashed because itc >> had no disk quota. [snip] >oG >Quotas on the system disk are generally not recommended. The reason is ( >because of the experience you just had.  N I don't use them myself, but you could give the SYSTEM account a quota bigger 3 than or as big as the number of blocks on the disk.d   Regards, Paddy   Paddy O'Brien, Transmission Development,?
 TransGrid, PO Box A1000, Sydney South,  NSW 2000, Australia    Tel:   +61 2 9284-3063 Fax:   +61 2 9284-3050& Email: paddy.o'brien@zzz.tg.nsw.gov.au  M Either "\'" or "\s" (to escape the apostrophe) seems to work for most people,i; but that little whizz-bang apostrophe gives me little spam.    ------------------------------  % Date: Mon, 13 Nov 2000 16:39:21 -0400s- From: JF Mezei <jfmezei.spamnot@videotron.ca>i' Subject: Re: Technology of US electionsw, Message-ID: <3A105174.A0B8E7F1@videotron.ca>   Martin Knoblauch wrote:tL > - producing correct results from the voting process. I think a nation likeL > the US could be much better by utilizing modern technologies (same is trueH > for Germany). At least the votes themselves would be "unquestionable".  < There was an interesting programme on canadian CBC yesterday4  (undercurrents at http://tv.cbc.ca/undercurrents/ )  N One of the fears of electronic voting is the lack of paper ballots that permit  manual recount.  N I guess that ideally, you answer question on a screen which captures your voteI elctronically and it then prints 2 receipts. Once which you keep, and oneaH which is then deposited in the paper ballot box which woudl allow manualL recount. You woudl then also have the ability to make sure that the computer got your vote right.   ------------------------------  % Date: Mon, 13 Nov 2000 16:52:58 -0400n- From: JF Mezei <jfmezei.spamnot@videotron.ca> ' Subject: Re: Technology of US electionsH, Message-ID: <3A1054A4.3A1E4344@videotron.ca>   Richard Brodie wrote:.Q > This is rather different: the integrity of the lottery largely rests on that ofcJ > the draw itself. Taking a copy of the entry database and simply making aH > few copies on a non-volatile medium gives a high degree of protection.  M But in those lotteries when you choose the numbers, you will typically inserttM a paper ballot into a machine at the store, and that machine reads the ballot M and sends your selected numbers to the central lottery computers and it comesi@ back with a checksum/control number to authenticate your ticket.  L When you win something, your bring your ticket to the store and they read itI into the machine which then sends the transaction to the main computer tonJ verify your ticket and how much you have won, the computer then marks thatN ticket as having been cashed in. This prevents falsified tickets, and prevents0 you from cashing the same ticket multiple times.  I It is ironic that in those lotteries, the winning numbers are selected byt> analogue means... balls falling out of a thumbling cylinder...   ------------------------------  % Date: Tue, 14 Nov 2000 00:45:32 -0500t2 From: rdeininger@mindspring.com (Robert Deininger)' Subject: Re: Technology of US ElectionseL Message-ID: <rdeininger-1411000045320001@user-2ivebim.dialup.mindspring.com>   In article <35666012DF4CD411BE940090279FA240111FB5@ppnt41.physics.ox.ac.uk>, John Macallister <J.Macallister1@physics.ox.ac.uk> wrote:  K > The "technology" such as it is appears to be concentrated in the countingsF > process with no consideration given to the fact that it can be quiteN > difficult to punch a clear hole in any piece of card especially if you don'tG > use some pressure-assisted tool. I find it incredible that people areoN > prepared to accept that counts should be accurate within error limits. We'reN > not discussing here some experimental scientific measurement but the precise' > number of votes cast in an election. s  w Ahem.  It depends on what the meaning of "cast" is.  And it appears that for some folks, common sense does NOT apply...   ' > In manual counts there will always bekK > some errors in the first count but, when the results are close, there areHN > recounts until all candidates are satisfied that the count is 100% accurate,M > not within a 2%, or whatever, error margin. In any election n votes  has to N > mean n votes and not n +/- e .   A manual recount of the whole country mightE > be the only thing which will convince people that accuracy matters.    I think you are missing a very crucial point.  Every time a bunch of ballots is handled, they are changed.  Sometimes this is deliberate, sometimes accidental.t| In practice, recounts often fail to give exactly matching answers.  Eventually, the ballots wear out, and they have to stop.  O It's also worth noting that they aren't really "recounting" ballots in Florida.i$They are examining ambiguous, spoiled, or damaged ballots, and making subjective decisions about what the voter "intended".  The machine-counted ballots aren't especially in dispute (except when a pile was missed or counted twice).  They are fighting about the ones the machines gave up on.    N > On the subject of accuracy and that little chad crittur, is it only a matterN > of time before the Gun Lobby demans the right to shoot that chad right outta > the back of the card?s   Weren't you just calling for a pressure-assisted tool a couple of paragraphs ago?  Now you seem to have second thoughts.  I will pass your suggestion along to my congresscritter, however.r   -- . Robert Deininger rdeininger@mindspring.comd   ------------------------------  % Date: Mon, 13 Nov 2000 15:18:54 -0500-! From: Dan Allen <dallen@nist.gov> L Subject: RE: Technology of US elections  OR why Palm Beach selected  Solaris: Message-ID: <NEBBIALHDHJMJINPGMOAIEHBCNAA.dallen@nist.gov>  Z   Oh Pelease!  It's just like the old keypunch verification used by businesses for decadesX   in the tab card world.  Just "read" my ballot using whatever (non)technology the localI   experts have chosen (lowest bidder!) and let me confirm my submission! -  ]   In my personal voting precinct we completed an arrow using a felt marker in multiple littlecd   screened cubicles.  The completed ballot was then fed (by me) to an optical scanner which recorded`   the result.  At the least, should my ballot have an illegal (i.e. two votes for a single seat)b   selection(s) on it, it should refuse to take it. NOW I KNOW I SCREWED UP AND CAN TAKE CORRECTIVE4   ACTION OR SIMPLY GIVE UP AND DRIVE MY CADDIE HOME.  a   This has nothing to do with how the results are ultimately recorded.  What it does is avoid theC_   situation of 20K ballots being tossed for data irregularities that could/should be identifieda   at the time of proffer!c   > -----Original Message-----; > From: David Mathog [mailto:mathog@seqaxp.bio.caltech.edu] * > Sent: Monday, November 13, 2000 11:31 AM > To: Info-VAX@Mvb.Saic.ComL) > Subject: Re: Technology of US elections  >  > \ > In article <g5vu0tgbvm8oma7astds74rofagb7n8c91@4ax.com>, LBohan@dbc.spam_less..com writes:9 > >the first remedy,  might be something to help a voter  8 > >confirm that his/her ballot is sound, and/or perhaps D > >even that the punched ballot really reflects their final choices. > M > Somewhat easier said than done.  You don't have to vote for President, but dI > if you did, and the little tab didn't fall all the way out, your ballotiL > could pass the validity testing and still not be what you intended (as it T > might read as you intended on the local machine, and another on the main machine).F > Of course, if you really care you should check your own ballot card I > by eye before handing it in, which is something you can do now.  Maybe  H > polling sites should require that you inspect it in front of the poll G > watchers before you turn it in.  That could be done 10 feet from the :N > watchers and they wouldn't be able to read it, but could check off that you  > had inspected your card. > K > This sort of validation would be much easier to handle with an electronicoI > system, where you could get a "you voted for xxx, is this correct" menulK > before the thing was actually transmitted.  Electronic systems have theiriG > own problems though, mostly in the area of preventing outright fraud,tF > especially if internet voting is allowed.  The little old ladies whoI > couldn't make sense of the Florida ballot aren't likely to do very welloD > with digital signatures, one time passwords, and the like either!  > 
 > Regards, >  > David Mathog > mathog@seqaxp.bio.caltech.edunA > Manager, sequence analysis facility, biology division, Caltech o >  >    ------------------------------  % Date: Mon, 13 Nov 2000 18:06:45 -0400u- From: JF Mezei <jfmezei.spamnot@videotron.ca> L Subject: Re: Technology of US elections  OR why Palm Beach selected  Solaris, Message-ID: <3A1065EB.CBC61030@videotron.ca>   Dan Allen wrote: > \ >   Oh Pelease!  It's just like the old keypunch verification used by businesses for decadesZ >   in the tab card world.  Just "read" my ballot using whatever (non)technology the localJ >   experts have chosen (lowest bidder!) and let me confirm my submission!  I The problem is that "business" punched cards were generated by punch cardoI machines and read by punch card readers. Each were built with establisheds tolerance levels.   H But when you ask a human to manually punch a hole in a card, how can youL really expect to have a stack of cards each punched to the same tolerances ?   ------------------------------  % Date: Mon, 13 Nov 2000 19:59:31 -0500C  From: John Santos <JOHN@egh.com>J Subject: Re: Technology of US elections OR why Palm Beach selected Solaris5 Message-ID: <1001113194532.2093B-100000@Ives.egh.com>-  $ On Mon, 13 Nov 2000, JF Mezei wrote:   > Dan Allen wrote: > > ^ > >   Oh Pelease!  It's just like the old keypunch verification used by businesses for decades\ > >   in the tab card world.  Just "read" my ballot using whatever (non)technology the localL > >   experts have chosen (lowest bidder!) and let me confirm my submission! > K > The problem is that "business" punched cards were generated by punch cardrK > machines and read by punch card readers. Each were built with establishedt > tolerance levels.t > J > But when you ask a human to manually punch a hole in a card, how can youN > really expect to have a stack of cards each punched to the same tolerances ?  F That isn't how the cards work.  The holes are pre-punched, but not allE the way around.  There are little strands of card still attaching thev little rectangle to the card.w  ! ASCII art, enourmously magnified:    	+---+ 	|   | 	|   | 	|   | 	+---+  A The spaces between the minus signs, plus signs, and vertical bars A are NOT artifacts.  They are really there, and are what keeps the = little rectangle from falling out of the card.  However, they  are easily torn!  D You insert a blank card in a device that looks a little like a metalC book with one office on each page.  The candidates names are listediF and there is a hole for each candidate.  The device has multiple pagesF of elective offices.  When you flip the pages in the book, a differentG set of holes is exposed.  They are offset slightly, so the holes in theo9 frame align with a different column of holes in the card.o  J You have a little metal plunger that you insert through one of the holes, I punching out the hole in the punch card that happens to be lined up with o the hole you selected.  E Then you flip the book to the next page and vote for the next office.dF (Often, if there are only 2 or 3 candidates for an office, or if there= is a yes/no question, there will be more than one on a page.)R  B When you finish, you pull the card out of the frame, and it is nowE (you hope) punched with a pattern of holes that reflect your choices.n  5 It doesn't require much force to punch out the holes.3  C Has anyone seen a baseball All-Star ballot?  They are very similar. ? You can easily punch out the holes with a finger nail or a pen.t  F There is no issue of alignment because the holes are at pre-determined
 locations.   -- l John Santosu Evans Griffiths & Hart, Inc. 781-861-0670 ext 539   ------------------------------   Date: 13 Nov 2000 19:12:27 GMT0 From: fdc@watsun.cc.columbia.edu (Frank da Cruz)= Subject: Re: Technology of US elections x Brazilian electionsR5 Message-ID: <8upeer$jgt$1@newsmaster.cc.columbia.edu>   L In article <OF95970324.F3BACBF4-ON03256996.00637698@ep-bc.petrobras.com.br>,-  <fabio_compaq@ep-bc.petrobras.com.br> wrote:mG : We had elections in my country this year, 100% using electronic vote,w : with no fraud and etc ...  : H : Sounds  funny to see a country of the "third world" with success in anJ : electronic election with almost 100.000.000 of votes and the great power( : of the world with a messy election.... : I In fact, Brazil began with electronic elections in 1994, based largely on1, creative use of inexpensive Kermit software:  3   http://www.columbia.edu/kermit/newsn6.html#brazill  J and existing telephone lines.  The cost was minimal and the benefits good.K I understand the more recent elections were Microsoft-based, but I have note3 seen any comparisons of the cost/benefit ratios :-)d  E Of course it is not as ironic as it might first seem that developing  I countries can outstrip "first world" countries in technology.  Countries  J like the USA have telephone networks and switches in which some componentsL can count their age in centuries, not just decades or years.  Countries likeK China, on the other hand, can begin in many areas with fiber optics withoutrH having to worry about an installed base of copper and mechanical relays.  E Similarly with elections.  In the USA, many areas invested heavily in,J punchcard-based automatic vote-tabulating equipment in the 1960s, which isH still in use.  For example, our voting booths in New York City date fromI about 1962 and there are tens of thousands of them, each probably costingl thousands of dollars.1   - Frank2   ------------------------------  % Date: Mon, 13 Nov 2000 17:58:12 -0400 - From: JF Mezei <jfmezei.spamnot@videotron.ca>.= Subject: Re: Technology of US elections x Brazilian electionst, Message-ID: <3A1063EB.5A2B903A@videotron.ca>   Frank da Cruz wrote:G > Similarly with elections.  In the USA, many areas invested heavily inoL > punchcard-based automatic vote-tabulating equipment in the 1960s, which isJ > still in use.  For example, our voting booths in New York City date fromK > about 1962 and there are tens of thousands of them, each probably costingo > thousands of dollars.f  K Using your argument, USA corporation would still be using punched cards not H only as data storage but also for programmers to type in their programs.  J Considering that the USA population has grown significantly, at what pointE will the old technology, even if paid for, be considered inadequate ?h  K For instance, Florida is supposed to certify its votes within 7 days of theqI election (even though it continues to accept incoming votes until 10 daysmM after the election). Perhaps that 7 day limit for counting votes is no longerEC adequate considering the number of votes which must now be counted.   F If you're deciding to keep the same technology, then the deadlines forN processing the growing amount of data should be increased. If you want to keepM the same deadlines, then you need to bring in technology which will allow you-5 to handle the increased workload in the time alloted.0  F There are also cost savings that could be acheived. So, your old 1960sN keypunch machines are paid for. But how much does it cost you to pay the staffM to operate those machines ? If corporations figured out it was cheaper to buyr, new technology, how come government hasn't ?   ------------------------------   Date: 13 Nov 2000 23:47:56 GMT0 From: fdc@watsun.cc.columbia.edu (Frank da Cruz)= Subject: Re: Technology of US elections x Brazilian elections-5 Message-ID: <8upujc$1u8$1@newsmaster.cc.columbia.edu>m  , In article <3A1063EB.5A2B903A@videotron.ca>,/ JF Mezei  <jfmezei.spamnot@videotron.ca> wrote:8 : Frank da Cruz wrote:I : > Similarly with elections.  In the USA, many areas invested heavily inHN : > punchcard-based automatic vote-tabulating equipment in the 1960s, which isL : > still in use.  For example, our voting booths in New York City date fromM : > about 1962 and there are tens of thousands of them, each probably costingy : > thousands of dollars.n : M : Using your argument, USA corporation would still be using punched cards notcJ : only as data storage but also for programmers to type in their programs. : K But corporations invest (e.g. in new equipment) in order to make money; theeL same is not true for local governments.  A big public investment, once made,M is difficult to repeat.  Brazil and China, 20 years from now, might be in theuN same position as the USA today, except that the "legacy, outmoded, discreditedH technology" on which they depend might be fiber optics, cell phones, and7 Internet, rather than punchcards and mechanical relays.2  L : Considering that the USA population has grown significantly, at what pointG : will the old technology, even if paid for, be considered inadequate ?n : J When the voters stop voting for parties that cut the taxes which would payI for new voting technology?  (Not to mention education, health care, etc.)o  M Actually, a more accurate answer might be... "tomorrow"?  Clearly the systemssJ used in many parts of the USA are inadequate, even when you don't considerL fraud and manipulation, which can pervert even the most advanced system.  ItL takes a close election like this one to bring the problems the forefront andB produce a clamor to fix them.  But where will the money come from?   - Frank    ------------------------------  % Date: Tue, 14 Nov 2000 01:00:02 -0500 2 From: rdeininger@mindspring.com (Robert Deininger)= Subject: Re: Technology of US elections x Brazilian electionslL Message-ID: <rdeininger-1411000100030001@user-2ivebim.dialup.mindspring.com>  [ In article <3A1063EB.5A2B903A@videotron.ca>, JF Mezei <jfmezei.spamnot@videotron.ca> wrote:o    M > For instance, Florida is supposed to certify its votes within 7 days of thedK > election (even though it continues to accept incoming votes until 10 dayst > after the election). r  p The in-person ballots have to be certified in 7 days.  There's a separate statute covering the absentee ballots.  : > Perhaps that 7 day limit for counting votes is no longerE > adequate considering the number of votes which must now be counted.   l The problem isn't counting, as I said in another post.  It's deciding how to treat the questionable cases.    tH > If you're deciding to keep the same technology, then the deadlines forP > processing the growing amount of data should be increased. If you want to keepO > the same deadlines, then you need to bring in technology which will allow youo7 > to handle the increased workload in the time alloted.c  T This idea sounds good, until you get to the cases your technology can't cope with.    H > There are also cost savings that could be acheived. So, your old 1960s" > keypunch machines are paid for.   L They're not keypunch machines.  And the cards aren't hollerith-style punchedccards.  The machines are simple -- basically a template to hold the card in place, and a sharp thing to jab it with.  The holes are pre-perforated.  And this system is only used in a portion of Florida counties.  One or two counties still use written ballots and purely hand counting.  There are other systems as well, and that's only considering Florida.a  0 > But how much does it cost you to pay the staff > to operate those machines ?   K You mean the party hacks who tamper with them?  They're a dime a dozen. :-)   3 > If corporations figured out it was cheaper to buy . > new technology, how come government hasn't ?  2 Seems like two totally different situations to me.   -- o Robert Deininger rdeininger@mindspring.comr   ------------------------------  # Date: Mon, 13 Nov 2000 18:41:06 GMTe From: richard_maher@my-deja.comrC Subject: Re: Two Phase Commit (2PC) on VMS - the nature of the fusst) Message-ID: <8upck0$94l$1@nnrp1.deja.com>w   Hi,   1 >When Richard talks about a "lack of 2PC" on VMS,o0 >he is making a gross exaggeration.  VMS has had3 >2PC for years between RMS and Rdb, across multiplef >nodes.t   That's funny I thought I wrote:h  1 >>They were considering keeping their VMS systemse3 >>in parrallel to their new Solaris systems but due,0 >>to the lack of 2PC they are getting off VMS as >>quickly as they can.  E Not that I'm not occasionaly given to exaggeration but I think you'lltG find that _you_ inserted the "on VMS", I was talking VMS<->Solaris. NownC you might argue that we do have this functionality with RTR but I'ds advise against it.  - BTW I believe that DBMS also supports DECdtm.e  1 >VMS does not implement Two phase commit via somei2 >new protocols in which Richard is interested.  Of. >course those protocols did not exist when VMS! >first supplied Two phase commit.y  3 Didn't stop ACMSxp or Tandem or MTS/DTC catching upa  1 >Even if Compaq did not support the new protocolsj2 >as Richard wants, someone else could do so if the" >DECdtm specification were public.  > This is very true and in my opinion very deliberate. But LarryG obviously has access to the riveting read that is the DECdtm FunctionaltA Specification V1.2. The way I see it working is that as long as aa@ resource manager (Rdb) specifies the ddtm$m_full_state flag whenB calling $getdtiw to find out whether or not to commit a "prepared"A transaction then all a third party has to do to call the shots isyG $join_rmw with ddtm$m_coordinator set. So the Rdb DBR process will just @ sit there protecting the database until you (or your TIP monitorE process as the case may be) do a $setdti then DECdtm will pass on the0E news to Rdb. I haven't had a chance to try it out and am still hopingc@ that I won't have to. (But if someone has the money, I'm free in	 February)B  D The _real_ downside of rolling your own is that you have to keep youB own journal as well as DECdtm's. Obviously this is not optimal letF alone sensible. But then Rich Riscoe was telling Compaq this 12 months ago!   Regards Richard Maher.    & Sent via Deja.com http://www.deja.com/ Before you buy.    ------------------------------  # Date: Mon, 13 Nov 2000 23:04:46 GMTr2 From: "Zane H. Healy" <healyzh@shell1.aracnet.com>* Subject: Re: VAXstation 4000-90 Error Code3 Message-ID: <is_P5.21$QV2.4106@typhoon.aracnet.com>s  5 John Laird <john@laird-towers.freeserve.co.uk> wrote: 3 > On Thu, 09 Nov 2000 21:28:05 GMT, "Zane H. Healy"t% > <healyzh@shell1.aracnet.com> wrote:C	 >>-DKA400R >>%VMB-F-ERR, PC = 00001340V >>%VMB-I-STS, R0 = 00000912. >>>>>h >>) >>DKA0: is the HD, DKA400: is the CD-ROM.l  I > Status 00000912 is "nosuchfile" - which usually indicates the disks areiH > not seen, not file-structured, or non bootable.  However, you say yourI > model 60 could boot from either, which suggests either you do have some[J > scsi funniness on the model 90 (do you have the sacrificial goat handy?)I > or just maybe, BFLG is set to something other than 0 which will lead toiI > the boot ROM code looking for a non-existent system root.  Try SET BFLGo > 0 or BOOT/0 DKAnnn.w  F I'm happy to report that this was the answer.  Yesterday I found a fewL minutes to once again pull 64MB out of the -60, and put it in the -90.  ThisI time I was lazy though and didn't bother to pull the HD, I just connected  the CD.u  D Doing a "B DKA400" resulted in the above error once again.  Doing a J "B/0 DKA400" resulted in the system booting into Standalone backup off the	 CD-ROM!  l  J So it looks like I've got myself a -60 with 40MB and a -90 with 64MB.  I'mJ happy :^)  Now I just need to find a permanant place to stick them and get them clustered in.  / My thanks to everyone for their help in this!!!w   				Zane   ------------------------------  # Date: Tue, 14 Nov 2000 02:04:42 GMTn From: halkuff@my-deja.coml Subject: Wildfire GS160 ) Message-ID: <8uq6jn$vod$1@nnrp1.deja.com>m  D We are looking at a GS160 with four QBBs.. Two for TRU64 and two forA OpenVMS... We intend to partition the two OpenVMS QBBs with eightd$ processors as one Galaxy instance...  : There will be two clusters: (Initial CPU Allocation shown)+         Node OESA                    2 CPUse+              OESB                    2 CPUss*              OESC                    1 CPU  +         Node WMSA                    2 CPUst*              WMSB                    1 CPU  E We're looking for people with similar configurations, the alternative A would be two four node ES40 systems for OpenVMS with eight cards.   F The per-CPU performance on the ES40s is better, due to the cache size,@ however the Wildfire offers better all round performance and the. migration path to a full GS320 is available...  @ Is anyone running a GS160 at ths time?  Impressions with Galaxy?   Kuff@Tessco.Come    & Sent via Deja.com http://www.deja.com/ Before you buy.c   ------------------------------   End of INFO-VAX 2000.636 ************************